summaryrefslogtreecommitdiff
diff options
context:
space:
mode:
authorKen Kellner <ken@kenkellner.com>2017-09-26 18:12:04 (GMT)
committerKen Kellner <ken@kenkellner.com>2017-09-26 18:12:04 (GMT)
commit9f78ffebab982615230bb6229abd1ed05ac83d17 (patch)
tree6184cb3e2ab9b84f0711ff8893e2060d531d3bbe
Initial uploadHEADmaster
-rw-r--r--Rstyle.tex17
-rw-r--r--appendices.tex19
-rw-r--r--chapter2.tex142
-rw-r--r--chapter3.tex106
-rw-r--r--chapter4.tex115
-rw-r--r--chapter5.tex245
-rw-r--r--chapter6.tex155
-rw-r--r--chicago.bst1653
-rw-r--r--conclusion.tex24
-rw-r--r--dissertation.tex49
-rw-r--r--dissertation.toc118
-rw-r--r--endnotes.sty339
-rw-r--r--figures/fig1-1.R49
-rw-r--r--figures/fig1-1.pdfbin0 -> 14056 bytes
-rw-r--r--figures/fig2-1.pdfbin0 -> 13509 bytes
-rw-r--r--figures/fig2-2.pdfbin0 -> 12446 bytes
-rw-r--r--figures/fig2-3.pdfbin0 -> 13433 bytes
-rw-r--r--figures/fig2-4.pdfbin0 -> 12911 bytes
-rw-r--r--figures/fig3-1.pdfbin0 -> 12014 bytes
-rw-r--r--figures/fig3-2.pdfbin0 -> 17661 bytes
-rw-r--r--figures/fig3-3.pdfbin0 -> 11836 bytes
-rw-r--r--figures/fig4-1.pdfbin0 -> 15551 bytes
-rw-r--r--figures/fig4-2.pdfbin0 -> 13101 bytes
-rw-r--r--figures/fig5-1.pdfbin0 -> 94073 bytes
-rw-r--r--figures/fig5-2.pdfbin0 -> 41795 bytes
-rw-r--r--figures/fig5-3.pdfbin0 -> 32544 bytes
-rw-r--r--figures/fig5-4.pdfbin0 -> 23995 bytes
-rw-r--r--figures/fig5-5.pdfbin0 -> 13709 bytes
-rw-r--r--figures/fig5-6.pdfbin0 -> 15330 bytes
-rw-r--r--figures/fig5-7.pdfbin0 -> 14419 bytes
-rw-r--r--figures/fig5-8.pdfbin0 -> 31594 bytes
-rw-r--r--figures/fig6-1.pdfbin0 -> 92063 bytes
-rw-r--r--figures/fig6-2.pdfbin0 -> 17523 bytes
-rw-r--r--figures/fig6-3.pdfbin0 -> 18875 bytes
-rw-r--r--figures/fig6-4.pdfbin0 -> 18869 bytes
-rw-r--r--front.tex24
-rw-r--r--genmodel.sh17
-rw-r--r--introduction.tex57
-rw-r--r--jabowamodel.tex560
-rw-r--r--model_browse.R94
-rw-r--r--model_growth.R78
-rw-r--r--model_surv.R80
-rw-r--r--netlogomodel.tex1113
-rw-r--r--netlogostyle.tex18
-rw-r--r--notoccite.sty43
-rw-r--r--pulongtable.sty446
-rw-r--r--puthesis.cls1605
-rw-r--r--references.bib2619
-rw-r--r--tables/tab2-1.tex23
-rw-r--r--tables/tab2-2.tex18
-rw-r--r--tables/tab3-1.tex24
-rw-r--r--tables/tab4-1.tex27
-rw-r--r--tables/tab4-2.tex27
-rw-r--r--tables/tab5-1.tex19
-rw-r--r--tables/tab5-2.tex16
-rw-r--r--tables/tab5-3.tex24
-rw-r--r--tables/tab5-4.tex27
-rw-r--r--tables/tab6-1.tex25
-rw-r--r--tables/tab6-2.tex24
-rw-r--r--tables/tab6-3.tex24
-rw-r--r--tables/tab6-4.tex16
-rw-r--r--tables/tab6-5.tex16
-rw-r--r--tables/tab6-6.tex20
-rw-r--r--vita.tex153
64 files changed, 10268 insertions, 0 deletions
diff --git a/Rstyle.tex b/Rstyle.tex
new file mode 100644
index 0000000..501abc2
--- /dev/null
+++ b/Rstyle.tex
@@ -0,0 +1,17 @@
+\lstdefinelanguage{R}{
+ morekeywords=[1]{for, in, logit, dnorm, abs, dcat, sum, pow, dunif, sort,
+ function, paste, if, else, library, NLStart, NLLoadModel,
+ makeCluster,clusterExport,invisible, parLapply, NLCommand,
+ is.null, detectCores, NLDoReport, return, sapply, clusterApply, length,
+ NLQuit, stopCluster, list, Sys.info},
+ extendedchars=true,
+ breaklines=true,
+ breakatwhitespace=true,
+ basicstyle=\linespread{0.5}\scriptsize,
+ columns=fullflexible,
+ tabsize=4,
+ comment=[l]{\#},
+ literate={~}{$\sim$ }{1},
+ keywordstyle=[1]\bfseries,
+}
+
diff --git a/appendices.tex b/appendices.tex
new file mode 100644
index 0000000..7e8cb8c
--- /dev/null
+++ b/appendices.tex
@@ -0,0 +1,19 @@
+\appendices
+
+\chapter{{CHAPTER 3 CODE}}
+\section{Mixed-Effects Proportional-Odds Logistic Regression Model}
+\lstinputlisting[language=R]{model_browse.R}
+
+\chapter{{CHAPTER 4 CODE}}
+\section{Mixed-Effects Seedling Survival Model}
+\lstinputlisting[language=R]{model_surv.R}
+\section{Mixed-Effects Seedling Growth Model}
+\lstinputlisting[language=R]{model_growth.R}
+
+\chapter{{CHAPTER 5 CODE}}
+\section{{SOEL} {N}et{L}ogo Code}
+\lstinputlisting[language=netlogo]{netlogomodel}
+\section{{JABOWA} Implementation in {N}et{L}ogo}
+\lstinputlisting[language=netlogo]{jabowamodel}
+\section{R Function to Run {N}et{L}ogo Models}
+\lstinputlisting[language=R]{../oak-lifecycle/sim_function.R} \ No newline at end of file
diff --git a/chapter2.tex b/chapter2.tex
new file mode 100644
index 0000000..b65a9bf
--- /dev/null
+++ b/chapter2.tex
@@ -0,0 +1,142 @@
+\chapter{{MIDSTORY REMOVAL ALTERS PREDATION AND DISPERSAL OF OAK (\textit{QUERCUS}) ACORNS BY SMALL MAMMALS IN THE CENTRAL HARDWOOD FOREST REGION}}
+
+\section{Introduction}
+
+Oak (\textit{Quercus}) is an important overstory tree species in hardwood forests throughout the eastern and midwestern United States \citep{Fral04, Elli05}, that provides a crucial food resource (acorns) to nearly 100 animal species \citep{Mart61}, and an economically valuable source of timber \citep{John09, Carm13}. In general, oaks have intermediate shade tolerance, requiring canopy disturbance to recruit into the forest overstory \citep{Niin06}. Oak dominance in these forests has therefore been maintained via a forest disturbance regime with multiple components including natural and human-induced fire \citep{Abra92, Abra03}, other natural disturbances like windthrows \citep{Crow88, Abra96}, and land clearing for agriculture \citep{Abra92, Abra96, Carm13}.
+
+Over the past century, an interruption of this disturbance regime highlighted by widespread fire suppression has led to an oak regeneration failure, and overstory oaks are being replaced by other, more shade-tolerant species \citep{Crow88, Abra92, Aldr05, Nowa08}. Loss of oak in the hardwood forest overstory could be catastrophic for the many wildlife species that rely on acorns as a food source \citep{Mcsh07}. Silviculture can be used to create artificial disturbance (overstory removal via harvesting) that favors oak regeneration and therefore maintenance of oak as a canopy dominant \citep{John09}. However, silvicultural methods for regenerating oak have so far met with mixed results \citep{Schl93, Morr08, Dey09, Swai13}.
+
+The disturbance caused by silviculture also affects forest species other than the tree species targeted for regeneration. Interactions between the target tree species and other plants and animals may be altered, potentially with important consequences for successful tree regeneration. A well-studied example is the response of tulip poplar (\textit{Liriodendron tulipifera}) to forest management with the ostensible goal of regenerating oak. In some forests, particularly those with fertile soils, opening the canopy via clearcutting to increase the available light for oak advance regeneration results in an increase in dominance of tulip poplar, which typically grows more quickly than oak when light is available \citep{Morr08, Swai13}.
+
+Mammalian dispersal agents also influence oak regeneration and may be affected by habitat changes due to timber harvest. In eastern deciduous forests, a suite of granivorous small mammal species rely on acorns as a primary food source over winter (e.g. the eastern chipmunk [\textit{Tamias striatus}], white-footed mouse [\textit{Peromyscus leucopus}], gray squirrel [\textit{Sciurus carolinensis}], and fox squirrel [\textit{Sciurus niger}]). In contrast to other animals (e.g. white-tailed deer [\textit{Odocoileus virginianus}] and wild turkey [\textit{Meleagris gallopavoa}]), granivorous small mammals do not always act strictly as acorn predators. Scatter-hoarders like the gray squirrel remove and store acorns in a series of small, shallow caches for later consumption \citep{Vand90, Moor07}. Larder-hoarders like the eastern chipmunk store a large number of acorns in a single large cache, often in a den or burrow \citep{Vand90}; some species may display both hoarding behaviors \citep{Clar94}. Depending on a number of factors including food availability \citep{Janz71, Jans04, Vand10, Berg11, Lich14}, and granivore population density \citep{Janz71, Ostf96, Schn02}, some caches may go unrecovered and allow the stored acorn(s) to germinate \citep{Vand01, Steel02, Haas05}.
+
+Microsite conditions for dispersed and unrecovered acorns are likely to be more favorable than those left untouched under the parent tree for two primary reasons. First, burial under a shallow layer of soil and leaves increases acorn germination probability both by increasing available moisture and reducing the chance the acorn is found and consumed by another predator \citep{Garc02, Haas05}. Second, acorns that are moved away from the source tree may suffer less density-dependent mortality due to intraspecific competition, infestation and disease, or seed predation (the Janzen-Connell escape hypothesis; \citealp{Janz70, Conn71, Howe82, Hirs12}). Overall, the effect of granivorous small mammals on oak falls somewhere on the spectrum ranging from seed predator (negative) to dispersal agent (positive), varying with forest type, overall food availability, small mammal population size, and many other factors \citep{Schu10, Xiao15}.
+
+Forest disturbance due to silviculture can have behavioral and numerical effects on small mammals, and hence the important interaction between oaks and their seed predators/dispersers. At the extreme end of the disturbance scale, clearcutting removes all forest overstory and drastically alters the abundance and composition of understory vegetation (and by extension, the food available to small mammals; \citealp{Perr04, Reyn06}). The resulting habitat is likely to be host to an altered community of granivorous small mammals \citep{Kirk90}. For example, eastern chipmunks appear to respond positively to clearcutting \citep{Kell13}. On the other hand, arboreal small mammals like the eastern gray squirrel are less likely to make use of the clearcut habitat, and are unlikely to disperse acorns into the opening \citep{Nixo80}.
+
+As a disturbance event, shelterwood harvests are less extreme than clearcutting because the disturbance (i.e., overstory removal) happens in stages over a period of years \citep{John09}. The initial phase of a shelterwood harvest typically involves removing midstory and some small overstory trees, while leaving most of the canopy intact. In the context of management for oak, the primary goal of the shelterwood approach is to generate enough competitive advance oak regeneration underneath the partial canopy so that removal of the remaining canopy results in continued dominance of oak \citep{Loft90, Schl93, John09}. In the initial stage(s) of shelterwood harvests, acorns are still being produced by remaining oaks in the overstory. Granivores continue to disperse, hoard, and consume these acorns in the partially harvested stand. Therefore, granivore activity during the time over which the shelterwood stages occur will impact the amount of oak regeneration that accumulates prior to the final harvesting stage.
+
+Habitat disturbance due to shelterwood harvesting is not as rapid as the immediate effects of a clearcut, but the changes can still be significant for small mammals. Besides the expected reduction in overstory and midstory canopy closure and tree density (Parker and Dey 2008), shelterwood harvesting may also result in increased volume of coarse woody debris (e.g., from harvesting leftovers like treetops), altered density and/or composition of understory vegetation due to increased light \citep{Quin00, Paqu06}, and altered temperature and moisture conditions \citep{Harp99}. Changes to the density and composition of forest overstory, midstory, and understory due to shelterwood harvest may also alter the total abundance and composition of food resources available to granivores. Some tree species (e.g. sugar maple [\textit{Acer saccharum}] and American beech [\textit{Fagus grandifolia}]) that produce large mast crops are typically removed in the initial shelterwood harvest, potentially lowering the overall availability of food. This could be offset by an increasing density of soft-mast producing plants in the understory in response to increased light availability \citep{Paqu06}.
+Despite these habitat changes, the abundance and community composition of granivorous small mammals does not appear to be strongly affected by the early stages of shelterwood harvests, at least in eastern deciduous forests (e.g., \citealp{Kell13}). However, granivore behavior may change in response to shifting vegetation structure and composition during the shelterwood harvesting process, with consequences for oak regeneration. Canopy cover is an important mitigator of predation risk for small mammals \citep{Bowe93, Mans98}, and small mammals adjust their behavior in response to varying levels of risk. For example, eastern chipmunks gathered and carried smaller seed loads at lower levels of canopy closure \citep{Bowe93}. In contrast, gray squirrels consumed food more quickly when under high predation risk (i.e., further from cover; \citealp{Newm88}). Gray squirrels also may trade the increased risk of caching seeds outside tree canopies for the reduced rate of pilferage that comes with this greater predation risk \citep{Steel14, Steel15}. Understory vegetation provides small mammals with protection from predators. Increasing understory density allows small mammals to devote more attention to foraging, resulting in higher seed removal rates \citep{Royo08, Pere08}.
+
+Given the potential effects of harvest-related disturbance on small mammals, understanding how acorn survival and dispersal is altered in recent shelterwoods is important for the design of silvicultural approaches that maximize the probability of sufficient oak advance regeneration. Several studies have examined how rates of seed removal by seed predators and/or dispersal agents change under various forest management regimes including shelterwood harvests \citep{Bell05, Pere08, Kell14}. In the case of acorns in hardwood forests, the general conclusion of these studies has been that removal is unaffected by harvest \citep{Bell05, Kell14}. Measuring removal is attractive because it is easy to estimate: seeds are monitored and the rate at which they disappear is noted. However, equating removal with seed survival is problematic; many seeds removed by predators may ultimately be cached and survive to germination \citep{Vand05, Moor08}, and these cached seeds may in fact be more likely to recruit successfully than seeds left under the parent tree \citep{Garc02, Haas05}. Further, solely estimating removal provides no information about dispersal distance, direction, or the characteristics of the seed’s final location. To better understand how silviculture, and shelterwood harvests in particular, affect seed fate (and by extension germination, regeneration and recruitment), it is necessary to determine seed fate following dispersal.
+
+I tested the null hypothesis that survival and dispersal distance for acorns of two oak species (black oak [\textit{Q. velutina}] and white oak [\textit{Q. alba}]) would not be impacted by disturbance associated with the first stage of a shelterwood harvest. Nested within this hypothesis, I also tested for differences in survival and dispersal of acorns between acorn species, among years and between two different groups of dispersal agents: all small mammal species vs. small mammal species other than gray squirrel, the primary scatter-hoarding agent in my study area.
+
+\section{Methods}
+
+\subsection{Study Location}
+
+The study was conducted in Morgan-Monroe and Yellowwood State Forests in south-central Indiana, U.S.A, part of the Central Hardwood Forest region \citep{Fral03}. The two forests together comprise \textgreater 19,000 ha and are composed predominantly of upland areas with steep slopes (25-35\%) and silt-loam soils \citep{Jenk98}. The overstory of the forests is dominated by oaks (\textit{Q. alba}, \textit{Q. velutina}, \textit{Q. rubra}, and \textit{Q. prinus}) and hickories (\textit{Carya} spp.) whereas the midstory and understory is composed mainly of shade tolerant species including sugar maple and American beech \citep{Saun13}. Tree density ranges from 923-1,527 trees/ha and basal area from 21.7-29.9 m\textsuperscript{2} ha\textsuperscript{-1}. The predominant acorn predators include white-tailed deer, wild turkey, gray squirrel, Eastern chipmunk, white-footed mouse, blue jay (\textit{Cyanocitta cristata}), and acorn weevil \citep{Gibs72, Mcsh02, Kell13, Rich13}. Of these, the scatter-hoarding gray squirrel and blue jay, and to a lesser extent the larder-hoarding chipmunk and white-footed mouse, also serve as important dispersal agents for oaks \citep{Sork84, Steel06, Moor06, Moor07}.
+
+The state forests are managed for multiple uses including recreation and timber production. Historically, forest management in the area consisted of group and single-tree selection \citep{Jenk98, Carm13}. In 2006, the Hardwood Ecosystem Experiment (HEE) was established to examine the long-term effects of different timber harvesting regimes (even- and uneven-aged) on the forest ecosystem \citep{Kalb13}. As part of the HEE, three-stage shelterwood harvests were conducted within three stands across the forest area. The first phase in winter 2008-2009 was the removal of non-oak trees \textless 25.4 cm dbh (essentially, the midstory) with a minimum residual basal area of 13.8 m\textsuperscript{2} ha\textsuperscript{-1}. After 7-10 more years, an establishment cut will reduce basal area to 13.8-16.1 m\textsuperscript{2} ha\textsuperscript{-1}, and after ca. 20 years the final overstory trees will be removed \citep{Kalb13}. Only the first phase of the shelterwood was complete during this study (2010 - 2013).
+
+\subsection{Experimental Design}
+
+In fall 2010, following the first phase of the shelterwood harvests, ten focal trees were selected for inclusion in the study. Focal trees were a subset of mature (dbh \textgreater 30 cm) black or white oaks (total n = 113) that were part of a larger mast monitoring study \citep{Kell14}. Half the focal trees (n = 5) were located at least 30 m inside the recent shelterwood harvests, and the other half were located in unharvested control areas.
+
+In each of the 4 years of the study (2010 - 2013) either four (2010) or six (2011 - 2013) of the ten focal trees were sampled, so that each tree was sampled in 1-2 years. One focal tree by year combination (in 2012) was censored from the final dataset because it had a very irregular dispersal pattern (most relocated seeds were found in a small area 13 m from the tree).
+
+A 40 m $\times$ 40 m grid of wire flags, divided into 16 10 m\textsuperscript{2} cells, was established and centered on the each focal tree. Two semi-permeable exclosures, each covering 0.56 m\textsuperscript{2} areas, were placed under the tree. One exclosure allowed access to squirrels and smaller animals (hereafter designated exclosure ``S'') but excluded larger animals like deer and turkey, and another allowed access by mice and chipmunks but excluded access to larger animals including squirrels (designated ``M''). Exclosure S consisted of 3.8 cm chicken wire wrapped around the sides and top of four wooden stakes, leaving a 15 cm gap at the bottom. Exclosure M consisted of a 0.56 m\textsuperscript{2} wooden frame with a height of 0.2 m, covered in 3.8 cm chicken wire.
+
+In late October or early November (depending on acorn availability) of a sampling year, each of the two exclosures at a given focal tree was provisioned with either 150 black oak acorns (in 2010 only) or 100 black and 100 white oak acorns (2011 - 2013) yielding a total of 300 provisioned acorns (2010) or 400 provisioned acorns (2011 - 2013) per focal tree per year. Acorns were collected from Indiana oaks when available, and purchased from suppliers (Sheffield's Seed Co., Inc., Locke, NY and FW Schumacher Co., Inc., Sandwich, MA) otherwise. In all cases, unusually large or small acorns, damaged acorns, and acorns that had already germinated (in the case of white oak) were not included to keep acorn characteristics consistent within species, among sites and years. A small steel nail was carefully inserted into each acorn to allow recovery using metal detectors. Small mammal seed predators readily ate around the nails, allowing me to determine the final fate of a given acorn (i.e., if the nail was found alone, the acorn had been consumed; \citealp{Sork84}). Several unique nail types were used to allow identification of the original acorn species and source exclosure: black oak acorns received a 1.27 cm nail (0.10 g), and white oak acorns received a 1.59 cm nail (0.63 g); acorns from exclosure M were colored either via paint or a brass coating, while acorns from exclosure S were not. This method has been used in many previous studies to allow tracking of acorns after dispersal (e.g. \citealp{Sork84, Moor07, Lich14}). In a pilot study, tagged acorns were readily taken within a day or two by small mammal seed predators.
+
+In mid-March of the subsequent year, the 40 m $\times$ 40 m region around the focal tree was exhaustively searched for tagged acorns and loose nails on a cell-by-cell basis using metal detectors (Fisher F2, Fisher Research Labs, El Paso, TX). Detectors were set to maximum sensitivity at all times; whenever a metal signal was detected, the surface and top 5-7 cm of soil in a 0.3 m radius circle around the signal center was searched with the help of a powerful magnet to identify the source of the signal (tagged acorn, nail, or otherwise). When a tagged acorn or nail was recovered, the fate (survived or consumed), species, source exclosure, location in the soil (buried or surface), and distance and angle from the source exclosure were recorded. In 2011-2013, microsite type (six categories: base of tree, fallen treetop, fallen log/coarse woody debris, stump, bush/dense vegetation, and none/open) was also recorded. Recovered acorns and nails were removed from the study area. Acorns and nails remaining inside the two exclosures (i.e., not dispersed and/or eaten \textit{in situ}) were counted, recorded, and removed.
+
+\subsection{Covariate Data}
+
+Information on forest structure around the focal trees was obtained from a concurrent HEE study. Permanent 0.1 ha overstory monitoring plots were established throughout the HEE in 2008 and sampled in 2013 \citep{Saun13}. A subset of 36 plots that were within 150 m of the focal trees were selected, located either in unharvested areas (n = 13) or entirely within a shelterwood harvest (n = 23). In each plot, the location, height, and dbh of all trees \textgreater 11.5 cm dbh, as well as all saplings and tree advance regeneration in a 0.025 ha subset of the plot were recorded \citep{Saun13}. All measured trees were classified into one of three canopy strata: (1) overstory, if most of the live tree crown was within or above the general forest canopy and dbh $\geq$ 3 cm; (2) midstory, if most of the tree crown was below the general canopy and dbh $\geq$ 3 cm; and (3) regeneration, if small (\textless 3 cm dbh). Density and basal area metrics were compared between harvest treatments using \textit{t}-tests in R \citep{Rcor15}.
+
+An index of ambient mast availability data for the two harvest treatments in the study area was obtained from a concurrent HEE study on acorn production and predation \citep{Kell14}. Acorn production was monitored yearly for n = 46 mature black and white oaks (23 of each species) in the forest matrix and n = 18 oaks (9 of each species) inside shelterwood harvests. Acorn production by each tree in each year was estimated by counting all acorns that fell into 0.34 m\textsuperscript{2} mast traps under the tree \citep{Kell14}. Acorn production was summarized for each of the two harvest treatments as mean acorns produced per m\textsuperscript{2} of oak canopy area.
+
+\subsection{Acorn Fate Analysis}
+
+I assumed perfect detectability of acorns and tags remaining in the exclosures (that is, acorns that were not removed/dispersed). Removal probability and probability of survival of unremoved acorns were estimated as functions of acorn species, exclosure type, and harvest treatment using a logistic regression fit in R 3.1.0 using function glm \citep{Rcor15}. The relationship between microsite type and acorn survival was also analyzed using logistic regression in R.
+
+For acorns dispersed outside the exclosures, I could not assume perfect detection. This is the most important challenge in dealing with recovery of tagged acorns - some tagged acorns will be moved to locations unreachable with a metal detector (e.g., high in a tree, or buried in a cache deep underground). Moreover, some will be dispersed outside the search area, and still others may be detectable but nonetheless missed due to observer error. Assuming perfect detection could lead to biased estimates of acorn survival and dispersal distance \citep{Lich14}. Recently, a maximum-likelihood approach that incorporates detection probability into models estimating seed survival and dispersal distance has been developed (described in detail in \citealp{Lich12} and applied in \citealp{Lich14}). The model has been shown, via simulation, to improve model fit relative to naïve models that ignore imperfect detection of dispersed seeds \citep{Lich12}. Briefly, the model seeks to estimate survival probability and dispersal distance for a combined dataset of both observed (recovered) seeds and unobserved (unrecovered) seeds. The model was fit using a modified expectation-maximization (EM) algorithm \citep{Demp77} that integrated over potential values of the vector of survival states (\textit{s}) and dispersal distances (\textit{r}) for unobserved seeds. To improve estimation of detection probability, I incorporated prior information on detection from a pilot study \citep{Moor07} that used a seed tagging approach identical to the one in this study \citep{Lich12}. Prior estimates for detection of cached (0.45) and surface (0.85) acorns from Moore et al. (2007) were consistent with my own pilot study, in which 40 acorns with known locations in a 20 m\textsuperscript{2} area were relocated by trained technicians (detection probabilities of 0.35 and 0.75, respectively). Dispersal distance was modeled as coming from an isotropic 2Dt distribution \citep{Clar99}. To conduct the survival and dispersal analysis, the dataset of removed seeds was divided into a series of discrete strata depending on acorn species, exclosure type, and harvest treatment; these strata were then further divided by study year. Detection was modeled as a function of acorn species (since tag size differed slightly between species) and acorn condition (i.e., cached or surface).
+
+The EM algorithm was run until convergence in R using the data from each individual stratum, generating unique estimates for tag detection probabilities, survival probability, and dispersal parameters along with an estimate of variability for each parameter. The 95\% confidence intervals around survival parameters could then be directly compared to determine statistically significant differences among strata. Comparing dispersal distance between factor combinations was more complicated, since dispersal distance was driven by a combination of 2Dt shape and scale parameters estimated by the model. To translate parameter estimates into comparable estimates of dispersal distance, I used a bootstrapping approach. First, a set of shape and scale parameter values was drawn from the appropriate parameter posterior distributions for each of the factor combinations. Then, for each set of parameter values, 100 distance samples were drawn from a 2Dt distribution, and a median dispersal distance value was calculated. Differences between median values for each factor combination were calculated to yield a test statistic. This process was repeated 100 times to yield 100 median values for each factor combination and a distribution of 100 difference values for each binary comparison. A quasi-\textit{p} value was then calculated for each multiple comparison as the proportion of difference values less than 0. The critical quasi-\textit{p} value was adjusted to account for multiple comparisons using a Bonferonni correction.
+
+\section{Results}
+
+\subsection{Plot Characteristics}
+
+Acorn production by black and white oaks in the vicinity of the study plots was highly variable among years, particularly for white oak (Figure 2.1), typical of masting species like the oaks \citep{Lusk07}. In general, acorn production was similar between treatments, although there was a trend toward greater production in the control treatment in the final years of the study (Figure 2.1). Summarized across all sample points (mean $\pm$ standard error), overstory stem density was 145 $\pm$ 10.8 ha\textsuperscript{-1}, midstory stem density was 179 $\pm$ 16.9 ha\textsuperscript{-1}, overstory basal area was 20.4 $\pm$ 1.2 m\textsuperscript{2} ha\textsuperscript{-1}, and midstory basal area was 3.99 $\pm$ 0.33 m\textsuperscript{2} ha\textsuperscript{-1}. There were no significant differences between harvest treatments for overstory density, overstory basal area, midstory density, or midstory basal area (all \textit{p} \textgreater 0.05), thought there was a trend of lower midstory density and basal area in the shelterwood harvests. The stem density of regeneration was significantly higher (\textit{t} = -2.72, \textit{p} = 0.01) inside shelterwood harvests (445 $\pm$ 66 ha\textsuperscript{-1}) than in the unharvested forest matrix (203 $\pm$ 20 ha\textsuperscript{-1}) but there was no difference in regeneration basal area (\textit{t} = -1.45, \textit{p} = 0.16; 1.16 $\pm$ 0.14 m\textsuperscript{2} ha\textsuperscript{-1} across all plots).
+
+\begin{figure}
+\centering
+\includegraphics[scale=1]{figures/fig2-1.pdf}
+\caption{Average acorn production (per unit crown area) for mature black (\textit{Q. velutina}) and white (\textit{Q. alba}) oak trees in the study region, by year and harvest treatment. Error bars are standard errors around the mean.}
+\label{fig:2.1}
+\end{figure}
+
+\subsection{Acorn Removal}
+
+A total of 8000 metal-tagged acorns (4600 black oak, 3400 white oak) were released at the study plots from 2010 - 2013. Across all years, the mean weight of included black oak acorns ($\pm$ SD) was 1.89 $\pm$ 0.12 g and the mean weight of white oak acorns was 3.43 $\pm$ 0.21 g. A total of 2199 acorns (27\%) were eventually recovered (tag and acorn or just tag); 1308 (16\%) were recovered from inside the original exclosure and 891 (11\%) were relocated using metal detectors. Overall, 84\% of acorns were ultimately removed from the exclosures, ranging yearly from 73\% in 2010 to 89\% in 2012. All covariates (oak species, harvest treatment, and exclosure type) had significant effects on both probability of removal from the exclosure and the probability that an acorn was not eaten if it was not removed (Table 2.1). White oak acorns were 2.9 times less likely to be removed from the exclosure and were 2.9 times less likely to remain uneaten in the exclosures than black oak acorns. In the shelterwood harvest treatment, acorns were 1.7 times more likely to be removed from the exclosures and 1.6 times more likely to remain uneaten. Acorns that were not accessible to squirrels were 2.2 times less likely to be removed, and 5 times more likely to remain uneaten in the exclosure.
+
+\input{tables/tab2-1}
+
+\subsection{Acorn Survival and Dispersal}
+
+The EM-MCEM analysis revealed differences in acorn dispersal and survival after accounting for imperfect detection. There were significant differences in detection probability between cached and buried acorns (i.e., 95\% confidence intervals did not overlap), but not between species. As expected, detection probability was higher for acorns on the surface (0.97, CI: 0.88 - 0.99 for black oak and 0.96, CI: 0.87 - 0.99 for white oak) than for buried acorns (0.55, CI: 0.49 - 0.61 for black oak and 0.48, CI: 0.42 - 0.55 for white oak).
+
+For all study years pooled together, survival of dispersed seeds differed by acorn species, harvest treatment and exclosure type (i.e., which small mammal species had access to the acorns). For black oak, acorns accessible to all small mammals had a significantly lower survival probability in the control treatment (0.27) than in the shelterwood treatment (0.47; Figure 2.2b). This trend was reversed for acorns accessible to only mice and chipmunks; survival was significantly higher in the control treatment (0.54) than the shelterwood treatment (0.19; Figure 2.2a). For white oak, survival probabilities for both exclosure types in the control treatment (0.31 for acorns accessible to all small mammals, 0.43 for acorns accessible only to mice and chipmunks) were higher than in the shelterwood treatment (0.14 and 0.11 respectively; Figure 2.2). When survival probabilities were estimated for each year separately in addition to the other factors, I observed generally the same trends in differences between exclosure type and harvest treatment (Figure 2.3). However, survival probability was highly variable among years, notably in 2012 and 2013 when survival was close to zero in the shelterwood treatment (Figure 2.3)
+
+\begin{figure}
+\centering
+\includegraphics[scale=1]{figures/fig2-2.pdf}
+\caption{Mean acorn survival probabilities based on the maximum likelihood model (with 95\% confidence intervals), by acorn species (BO = black oak, WO = white oak) and harvest treatment (control or shelterwood harvest) for acorns accessible to (a) mice and chipmunks only and (b) squirrels, mice and chipmunks.}
+\label{fig:2.2}
+\end{figure}
+
+\begin{figure}
+\centering
+\includegraphics[scale=1]{figures/fig2-3.pdf}
+\caption{Mean acorn survival probabilities based on the maximum likelihood model (with 95\% confidence intervals), by year, acorn species (BO = black oak, WO = white oak) and harvest treatment (control or shelterwood harvest) for (a) acorns accessible to mice and chipmunks only and (b) squirrels, mice and chipmunks. For bars marked with an asterisk, I could not obtain reasonable 95\% confidence intervals (since estimated survival was 0)}
+\label{fig:2.3}
+\end{figure}
+
+There was a significant effect of microsite type on acorn survival (Table 2.2). Acorns relocated at the base of trees, adjacent to fallen logs, in bushes or dense vegetation, or under fallen treetops had a lower probability of survival than acorns located in open microsites. Relocation near tree stumps was not associated with survival probability.
+
+\input{tables/tab2-2}
+
+Dispersal parameters also differed between treatments. For acorns accessible only to mice and chipmunks, bootstrapped median dispersal distance was significantly higher in shelterwood harvest plots than in control plots for both white and black oak (Figure 2.4a). There was no difference in median dispersal distance among acorn species or harvest treatment when squirrels also had access to the acorns (Figure 2.4b). Median dispersal distances when squirrels had access tended to be in between the median values for the two harvest treatments when squirrels did not have access (Figure 2.4).
+
+\begin{figure}
+\centering
+\includegraphics[scale=1]{figures/fig2-4.pdf}
+\caption{Distribution boxplots of bootstrapped median dispersal distances from 2Dt distributions for each combination of acorn species (BO = black oak, WO = white oak) and harvest treatment (control or shelterwood harvest) for (a) acorns accessible to mice and chipmunks only and (b) squirrels, mice and chipmunks. Factor combinations that do not share a common letter (A-C) had significantly different median dispersal distances.}
+\label{fig:2.4}
+\end{figure}
+
+\section{Discussion}
+
+\subsection{Acorn Removal}
+
+Acorn removal by small mammal predators observed in this study contrasted with patterns in the literature. A high proportion of seeds were ultimately removed by predators (84\%), more than the 43\% observed by \citet{Kell14} in earlier years at the same study sites and the 38\% observed by \citet{Bell05}. The four years of this study (2010 - 2013) comprised a period of generally lower overall acorn production (Figure 2.1) than in the previous 4 years in the study region (2006 - 2009; \citealp{Kell14}). With more scarce overall food resources, predation pressure likely increased on the remaining acorns and resulted in higher overall removal probability. I also observed a higher probability of removal in shelterwood harvests than in unharvested controls (Table 2.1), whereas harvest treatment had no effect in the other two studies \citep{Bell05, Kell14}. Given that data collection in this study occurred between 2 and 6 years after the initial shelterwood harvest whereas the other studies occurred immediately following harvest \citep{Bell05, Kell14}, my results suggest that it may take several years for differences in removal between harvest treatments to manifest.
+
+Unsurprisingly, when more small mammal species (specifically, gray squirrels) had access to acorns, removal probability increased (Table 2.1). A larger suite of potential predators has meant higher probability of removal in studies of several regions and small mammal communities (e.g. \citealp{Hulm99, Bell05, Jink12, Kell14}). White oak acorns were less likely to be removed and more likely to be eaten in situ in the exclosure box than black oak (Table 2.1). Small mammal predators likely account for the increased perishability of white oak acorns (which germinate in the fall, rather than in the spring like black oak) by consuming them immediately and instead caching black oak, especially when both are presented together \citep{Hadj96, Smal01, Lich14}.
+
+\subsection{Acorn Survival and Dispersal}
+
+Comparing removal data with actual seed fate data obtained from tag recoveries highlights the important difference between the two \citep{Vand05, Moor08}. While 84\% of acorns were removed, actual mortality of removed seeds was lower. Of the 891 acorns recovered outside the exclosures, 199 (22\%) survived the winter and more than half (106, 54\%) of those that survived had been cached in the soil, likely increasing their probability of germination by providing more favorable environmental conditions and reducing the chance of predation by other animals \citep{Howe82, Garc02, Haas05}.
+
+For most (3 out of a possible 4) factor combinations of exclosure type and acorn species, acorn survival was reduced in shelterwood harvests relative to controls (Figure 2.2), often substantially, evidence for rejection of my null hypothesis that survival would not differ by treatment. In addition to survival differences between treatments, I also observed differences in dispersal distance; generally median dispersal distance was greater in the shelterwoods than in the controls for acorns that squirrels did not have access to (Figure 2.4a). There are several potential explanations for these differences. First, the first stage of the shelterwood harvest was designed to remove the majority of midstory trees, changing forest structure and light availability as well as introducing other types of disturbance to the site like soil compaction and a modest increase in downed coarse woody debris \citep{Kell13}. Granivores change their foraging behavior based on environmental conditions including available vegetative cover, which provides protection from predators, with dense vegetation at the ground level facilitating more or longer foraging bouts \citep{Yahn82, Brow88, Kotl91, Bowe93, Bowe93b, Tcha01}. I did not observe significant differences in the density or total basal area of overstory and midstory trees between silvicultural treatments; however, the mean density of tree advance regeneration more than doubled. More effective foraging (and higher risk of cache pilferage; \citealp{Hirs12}) could be the reason for the increased acorn removal (Table 2.2), decreased acorn survival (Figure 2.2), and greater dispersal distances (Figure 2.4a) I observed. Compared to the other two parameters, dispersal distance is less tightly connected to successful oak recruitment; however, greater dispersal distances lead to a higher probability that acorns will be relocated outside the parent tree canopy, where they will face less parental competition for light. Other studies have also shown positive correlations between understory vegetation density and seed removal \citep{Royo08, Pere08} but have not estimated seed survival.
+
+The effects of harvest treatment were consistent between acorn species, but white oak acorns were overall less likely to be removed, less likely to survive following dispersal (Figure 2.2), and more likely to be eaten inside the exclosure from which they originated (Table 2.1). Because white oaks begin to germinate as soon as they fall and thus begin to deplete the nutrients contained in the acorn, granivores preferentially consume white oaks immediately and cache members of the black oak section for later \citep{Hadj96, Smal01}. Since some of these caches will go unrecovered, overwinter survival is higher for black oak acorns. However, gray squirrels are known to excise the embryo of white oak acorns, halting germination and preserving the acorn’s nutrient content \citep{Steel01}. Decreased survival of white oak acorns relative to black oak may also reflect a more volatile background level of white oak acorn availability in the environment, including 2 years of complete acorn crop failure (Figure 2.1). With few white oak acorns available in the environment, the supplied, tagged acorns may have been especially attractive.
+
+Harvest treatment had an interactive effect with exclosure type (i.e., the suite of small mammals that had access to a given acorn) on acorn survival and dispersal. In the control treatment, acorns in exclosures with access only to mice and chipmunks had a higher probability of survival than when gray squirrels also had access (Figure 2.2). At first glance, it makes sense that a greater number of potential seed predators with access should always mean a lower probability of survival (and indeed this was also true for probability of removal, Table 2.1). However, this pattern was reversed in the shelterwood harvest; allowing gray squirrels access actually increased the survival probability. This again emphasizes the danger of conflating removal with survival probability. Though more seeds will be removed when a greater number of predator species have access (Table 2.1), the behavior and foraging/caching decisions made by these species differ and so survival need not follow the same pattern as removal \citep{Moor08}. The large number of small, scattered caches made by gray squirrels increases the probability that any one acorn will go unrecovered and survive to germination, relative to an acorn in the larder hoard of a white-footed mouse or chipmunk \citep{Vand01}. In this study, the shelterwood harvest appeared to have disproportionate impact on the process underlying the foraging behavior exhibited by the scatterhoarder (gray squirrel) relative to the primarily larder hoarding mice and chipmunks. The changes in the density of regeneration-stage vegetation cover may have provided a disproportionately large benefit to the foraging of the primarily ground-dwelling mice and chipmunks, relative to the more arboreal gray squirrel.
+
+While there are key differences in dispersal and survival between treatments, by far the most striking differences in these parameters were among years. Notably, survival probability for acorns in the shelterwood treatment in 2012 and in both treatments in 2013 was essentially zero (Figure 2.3). Yearly variation in acorn survival is likely due to two interacting factors: food availability in the environment and granivore abundance. Since oaks (and other species at my study sites like hickories) typically exhibit ``boom and bust'' cycles of yearly seed production \citep{Sork93a, Lusk07}, the overall abundance of food available to rodents also fluctuates greatly between years. Acorn production by two common oak species at my study sites (white and black oak) varied greatly, with two complete white oak mast failures in 2011 and 2013 (Figure 2.1). Small mammal abundance at these study sites is positive correlated with mast abundance in the previous year \citep{Kell13}. The relatively strong mast crop in 2012 likely boosted the overall abundance of the small mammal granivore community. Total captures of white-footed mice and chipmunks increased by an average of 119\% on HEE study sites from summer 2012 to summer 2013 (Kellner, \textit{unpublished data}). With low acorn availability in fall 2013 (especially for white oak; Figure 2.1), granivores had to eat rather than cache the few available acorns, likely contributing to low survival for both acorn species (Figure 2.3).
+
+Interestingly, there was some evidence that acorn production within years was lower for both black and white oak species in the shelterwood harvests relative to the controls, particularly in the last two years of the study (Figure 2.1). Few differences in acorn production immediately following the first stage of a shelterwood harvest have been reported (\citealp{Bell05}, and \citealp{Kell14} at the same sites as this study), despite the expectation that the harvest would increase light availability throughout the canopies of the remaining oaks, increasing available energy for seed production \citep{Verm53, John94}. However, these studies have generally been conducted 2 - 3 years following the harvest. The final 2 years of this study were 4 - 5 years after the initial harvest; as with removal, a longer study period may be necessary to elucidate differences in mast production between harvest treatments. In either case, lower overall acorn availability at the shelterwood treatment sites would also contribute to reduced acorn survival and presumably to increased dispersal distance, since granivores would be willing to expend extra effort to cache in open areas away from the source where the seeds are less likely to be pilfered by inter- or intraspecific competitors \citep{Steel14}. One caveat to my data on mast production is that while white and black oak are among the most dominant mast producers on my study sites, there are other (unmeasured) food sources including other oaks, hickories, and soft mast producing species which also play a role in overall food availability.
+
+\subsection{Microsite Effects}
+
+Independent of treatment, I found that surviving acorns (both cached and uncached) were associated with certain microsite conditions. Acorns relocated at the base of trees, in bushes or dense vegetation, adjacent or under coarse woody debris, and under fallen treetops were negatively associated with acorn survival probability relative to open microsites (Table 2.2). One characteristic common to all of these microsites is that they provide some degree of protective cover to small mammals. Small mammals were likely removing acorns from the exclosures and then moving to these protected microsites to consume the seed, resulting in a higher probability of non-surviving acorns (i.e., bare tags) being found at these locations. In contrast, acorns found out in open microsites (where predation risk is greater) were likely brought there to be cached for future use rather than immediately consumed. Caching in more dangerous locations reduces the probability that the cache will be pilfered by other animals (the Habitat Structure Hypothesis; \citealp{Muno11, Steel14}.
+
+\subsection{Conclusion}
+
+Establishing adequate oak advance regeneration following the first stage of a shelterwood harvest is a critical goal for forest managers \citep{John09}. Seed predation by granivorous small mammals hinders that goal, while seed dispersal by the same granivores is likely to help it. I provide evidence that forest disturbance associated with shelterwood harvesting impacts the relationship between oaks and their seed predators and dispersal agents in several ways. First, acorn removal probability is higher in recent shelterwoods relative to controls; second, the survival of dispersed acorns is generally reduced in shelterwoods; and finally, acorns are generally dispersed farther in shelterwoods. Taken together, these impacts have important implications for successful acorn germination, seedling recruitment, and therefore, generation of adequate oak advance regeneration. Lower survival of dispersed acorns in shelterwoods is unquestionably a negative result, yielding fewer oak seedlings. However, if fewer acorns survive in the shelterwoods, but the survivors are dispersed farther and to better open microsites, the lower overall number of initial seedlings may be offset by a greater number of seedlings located at sites where they are able to take advantage of increased light following the harvest. This study did not track acorns all the way to germination and subsequent growth/survival of the resulting seedling; this would be a valuable area for future study.
+
+Shelterwood sites in my study exhibited a characteristic increase in the density of tree regeneration (both of target oaks and other non-target tree species) and other understory plants, likely due to increased light availability. This vegetation, absent or reduced in density in the control sites, may have facilitated increased foraging efficiency by small mammal seed predators \citep{Brow88, Kotl91, Bowe93, Bowe93b, Tcha01} and contributed to the reduced probability of acorn survival. Other shelterwood harvesting approaches have included a post-harvest prescribed burn (e.g. \citealp{Bros99}) intended to reduce competitive pressure on oak seedlings and saplings from less fire-tolerant tree species like tulip poplar. The combined shelterwood harvest/prescribed fire approach may be more successful in regenerating oak than harvesting alone \citealp{Bros99}. My results suggest that reduction in understory vegetation density via fire may serve not only to reduce competitive pressure on oak seedlings and acorn predation by rodents. Further work is needed to confirm my hypotheses about how shelterwood harvesting affects granivore activity and to determine the extent to which the combined harvest/fire approach confers these dual benefits for oak regeneration. \ No newline at end of file
diff --git a/chapter3.tex b/chapter3.tex
new file mode 100644
index 0000000..eedcfef
--- /dev/null
+++ b/chapter3.tex
@@ -0,0 +1,106 @@
+\chapter{{OAK SEEDLING HERBIVORY BY MAMMALS AND INSECTS ALONG A DISTURBANCE GRADIENT CREATED BY TIMBER HARVEST IN THE CENTRAL HARDWOODS}}
+
+\section{Introduction}
+
+Plant survival and growth is influenced directly by a multitude of environmental factors including climate, light availability, moisture, and soil type. Environmental conditions also influence plants indirectly by driving intra- and interspecific competition and herbivory. Disentangling these relationships within and between trophic levels is crucial to understand the ecology of natural systems and in many cases to conduct applied conservation and management activities. Plant-herbivore interactions are particularly complex. Herbivores come from multiple taxa, and herbivory takes a number of forms, with differing consequences for affected plants. For example, herbivorous insects frequently defoliate plants or feed on roots, whereas mammalian herbivores frequently excise stems and branches or consume bark (reviewed by \citealp{Hunt91}, \citealp{Swih01}). Browse damage is often detrimental to plant survival and/or growth, but this is not always the case \citep{Bels86}.
+
+Edges create unique gradients of environmental conditions that influence plant composition and herbivore behavior \citep{Harp05}. The quantity and configuration of edge habitat is shifting due to human-induced fragmentation of habitat \citep{Saun91, Mart08}, and disturbance regimes are changing \citep{Dale01}. Hence, edge effects on the herbivore community, and the potential implications for plant communities, is an important area of research \citep{Wirt08}. In forested areas, for instance, the abundance and species composition of tree seedlings present is a key driver of future overstory composition \citep{Oliv96}. Disturbances that lead to canopy openings alter environmental conditions across the gap and newly created edge, with direct and indirect consequences for seedlings and therefore the future forest stand. For example, increased light in openings may favor shade-intolerant ``pioneer'' tree species over other, more shade-adapted species depending on the opening size \citep{Whit89}. The diversity of new habitat created post-disturbance can also attract a different community of herbivores than is typically present in the surrounding forest \citep{Cade00, Mein02, Wirt08}. The behavior and preferences of these herbivores influence which seedlings ultimately survive and recruit into the forest overstory \citep{Davi93, Roon03, Holm13}. Interactions between disturbance, resulting edge habitat, tree seedlings, and herbivores are important for understanding forest succession, yet they are still not fully understood in many forest systems.
+
+One such system is the hardwood forests of the eastern United States. Oak (\textit{Quercus}) and hickory (\textit{Carya}) dominate the overstory in many areas, including the Central Hardwoods Region \citep{Fral03}. Oaks in particular are foundation species in the Central Hardwoods \citep{Fral04, Elli05} that provide numerous ecosystem services (notably the production of acorns, a crucial food resource for many wildlife species; \citealp{Mcsh93, Mcsh00}). Oak is still common in the overstory, but it is rare in the understory - typically outcompeted by more shade tolerant tree species like maple (\textit{Acer} spp.) and American beech (\textit{Fagus grandifolia}) \citep{Crow88, Aldr05, Nowa08}. If uninterrupted this successional trajectory will lead to a decline in oak dominance (termed ``oak regeneration failure'') with negative effects on wildlife that depend on acorns as a food resource \citep{Mcsh07}. Historically, oak was maintained in the canopy via frequent disturbances (primarily fire) that thinned the overstory and understory thus promoting the survival and growth of oak seedlings, but such disturbances have been suppressed in the past century \citep{Abra92, Bros01}. Recently, timber harvesting has been used as a means to emulate natural disturbance in order to promote oak regeneration \citep{Dey02, John09}. Harvesting alters many aspects of the forest ecosystem including the creation of edge habitat conditions (e.g. after clearcutting). Thus, harvesting is likely to affect the interaction between herbivores and the plant community, including oak seedlings, post-disturbance.
+
+Numerous herbivores browse on oak including rabbits (\textit{Sylvilagus} spp.), voles (\textit{Microtus} spp.), arthropods (orders Coleoptera, Hymenoptera, Lepidoptera, Orthoptera), and white-tailed deer (\textit{Odocoileus virginianus}) \citep{Marq94, Ostf97, Russ01, Mein02}. Under some conditions, herbivory can limit the establishment of oak and other seedlings \citep{Marq76, Kitt95, Hors03, Abra12}. Therefore, understanding how oak-herbivore interactions change post-disturbance (including within new edge habitat) is crucial for predicting and promoting successful oak regeneration in managed forests. The available literature on herbivory following harvest disturbance in eastern hardwood forests is limited and occasionally contradictory. For example, \citet{Mill09} found that browse rates by white-tailed deer on seedlings both inside and outside harvested areas declined as the proportion of the study forest that was harvested increased, presumably because more early successional vegetation became available as alternative food sources. However, \citet{Crim10} found that overall browse rates by deer on clearcut edges were higher than either the adjacent forest interior or clearcut interior. Additionally, seedlings are browsed by other vertebrates (e.g. rabbits and voles) and suffer foliar damage from insects (especially lepidopteran caterpillars) \citep{Marq94, Mein02, Stan98}, but less is known about the impacts of these herbivores on oak and less still on their interactions with harvesting disturbance and edges.
+
+I examined how herbivory on oak seedlings by two mammals (white-tailed deer and eastern cottontail rabbits \textit{S. floridanus}) and leaf-eating insects varied across forest edges created by clearcuts. I also measured how finer-scale variables like seedling height and interspecific plant competition affected seedling herbivory. Most studies have treated herbivory as a binary variable (browsed or not-browsed; e.g. \citealp{Mein02, Mill09, Crim10}). I measured and analyzed browse damage on an ordinal scale to better capture the variability in browsing events. I predicted that browse damage from mammalian sources would be highest near the forest edge, since the primary mammal herbivore in my study system (white-tailed deer) prefers edge habitat \citep{Alve88, Wall97}. In contrast, I predicted that browse damage on oak seedlings from insect sources would be highest in intact forest, where abundance and diversity of leaf-eating insects may be higher than in edge or harvest opening habitat \citep{Summ11, Summ02}.
+
+\section{Methods}
+
+\subsection{Study Location}
+
+This study was conducted in Yellowwood and Morgan-Monroe State Forests in south-central Indiana, USA. The two state forests total about 19,000 ha in size and are managed for multiple uses including timber harvest \citep{Carm13}. Overstory trees are mainly oaks and hickories (\textit{Carya} spp.) while the understory is dominated by sugar maple (\textit{Acer saccharum}) and American beech (\textit{Fagus grandifolia}) \citep{Saun13}. My research was conducted within the framework of the Hardwood Ecosystem Experiment (HEE), a 100-year landscape-scale study of the effects of timber harvest for oak regeneration on forest ecosystems located in Yellowwood and Morgan-Monroe \citep{Kalb13}. For a more detailed site description, see Chapter 2.
+
+\subsection{Experimental Design}
+
+In May 2011, 3 replicate blocks (hereafter, sites) within the HEE were selected. A 4 ha portion of each site had been clearcut between fall 2008 and spring 2009; the clearcuts removed all stems greater than 1 cm diameter \citep{Kalb13}. A transect perpendicular to and across the clearcut edge was established at each site. Transects ranged in length from 230-300 m depending on site conditions (slope, aspect). Four plot locations were designated along each transect: one plot \textgreater 50 m (range 50-175 m) inside the adjacent intact forest, one located directly on the clearcut harvest edge, and two inside the clearcut harvest (both at least 50 m from the harvest edge and 20 m apart from each other) for a total of 12 plots. Within each plot location, 4 subplots (25 m\textsuperscript{2}) were established. For the edge plots, the subplots were located in a row along the clearcut edge, with at least 10 m spacing between subplots. For the other plot locations, the subplots were arranged in a 2 $\times$ 2 grid, again at least 10 m apart. Each set of subplots received a random permutation of 4 treatments. The 4 treatment options were based on a 2 $\times$ 2 factorial design crossing herbivory (deer excluded vs. deer allowed) with competition (plant competitors controlled vs. plant competitors not controlled). Deer were excluded from half the subplots using permanent deer exclosure fences, and interspecific plant competitors were controlled at half the plots using a combination of initial herbicide (glyphosate) application followed by semiannual hand-weeding. The straight-line distance from each subplot to the nearest forest edge was obtained using GPS (Earthmate PN-40 GPS units); distance was considered positive going from the edge into the harvest and negative going from the edge into the adjacent forest.
+
+\subsection{Oak Plantings}
+
+In May 2011, following the selection of subplot locations and initial treatment application, white and black oak seedlings were planted in each subplot. Seedlings came from two sources. The first source was a cohort of greenhouse-grown seedlings. To produce these seedlings, I planted acorns collected from the HEE study area in 5.6 cm $\times$ 10.2 cm Jiffy pellets (Jiffy Products of America Inc., Lorain, OH, USA) in February-March 2011. When the resulting seedlings reached ~10 cm they were removed from the greenhouse and placed in an enclosed outdoor area for 1-2 weeks to harden them for transplanting in the study sites. The greenhouse-grown seedlings were supplemented with year-old bare-root seedlings from the Indiana Department of Natural Resources tree nursery at Vallonia. The largest and smallest 25\% of nursery seedlings were discarded to insure they were of a similar size and condition.
+
+Available seedlings were divided evenly among subplots. Since there were different numbers of available seedlings by species and source this resulted in an unbalanced design. Each subplot received 24 greenhouse-origin seedlings (8 white oak, 16 black oak) and 15 one-year-old bare-root seedlings (8 white oak, 7 black oak) for a total of 39 seedlings/subplot. Seedlings were planted in a grid using a planting bar, alternated by species, and separated from each other by \textgreater 0.30 m (greenhouse-origin) or \textgreater 0.60 m (nursery-origin). One month after planting, seedlings were revisited and survival was assessed. Mortality within the first month was attributed to transplant shock; seedlings that died in this period were excluded from further analysis.
+
+\subsection{Seedling Measurements}
+
+Seedlings were re-visited in May and October each year of the study (2011-2014, following the initial re-visit in June 2011) for a total of 8 sampling occasions. On each occasion, the survival status of the seedling was recorded, along with intensity of browse damage from mammalian and insect sources. If a seedling was recorded as dead, it was not sampled on future occasions. Browse damage from mammalian herbivores was recorded on a 4-point ordinal scale: 0 for seedlings with no visible browse damage; 1 if 1-15\% of shoots on the seedling were browsed; 2 if 15-50\% of shoots were browsed; and 3 if 50\% or greater of shoots were browsed or the apical meristem was browsed (modified from \citealp{Boul09}). Browse damage from insects was recorded on a separate ordinal scale, identical except that percent leaf area damaged was substituted for percent shoots browsed. Mammalian browse damage was further narrowed to a deer or rabbit source based on the stem excision characteristics \citep{Verc05}. The same observer (KFK) recorded browse damage throughout the study to maintain consistency. Following each growing season (October sampling occasions), the height (to the tallest point) of the seedling was recorded.
+
+Since I was not successful in completely excluding interspecific competitors from the competition-controlled subplots, I recorded several vegetation characteristics at each subplot during the May sampling occasion in each year to generate a continuous index of competition. Percent woody and herbaceous cover at the ground level were estimated visually, and the density of non-oak stems in several height categories (\textless 50 cm, 50-100 cm, \textgreater 100 cm) within the subplot was recorded. These vegetation variables were combined into a single index value for competition in a given subplot and year using principal components analysis (PCA).
+
+\subsection{Pellet Count Transects}
+
+To estimate the intensity of habitat use by deer and rabbits near my study plots, I established 75 m pellet count transects at each of the 12 plots in March 2012, 2013, and 2014. Transects were centered on a deer exclosure area and divided into three 25 m sections: one headed downslope from the deer exclosure, one inside the deer exclosure, and one upslope from the deer exclosure. At 5 m intervals along the transect, a 1 $\times$ 1 m area was exhaustively searched for pellet groups. Pellet groups were identified to source species (deer or rabbit) and counted \citep{Neff68}.
+
+\subsection{Analysis}
+
+I conducted separate but identical analyses of browse damage from three different herbivore sources (deer, rabbits, and insects). The dependent variable was browse damage (0-3 scale, hereafter browse intensity) on a given oak seedling for a given sampling occasion. For the deer browse analysis, only seedlings that deer had access to were included. Since browse intensity was measured on an ordinal scale, I used a mixed-effects proportional-odds logistic regression for each analysis \citep{Jack09}. The random effects were seedlings nested within subplots nested within plots. Fixed effects included in the analysis were seedling species (white or black oak), competition index, distance to forest edge, and seedling height. I initially included second-order terms for distance to edge and height to test for quadratic relationships, but the second order distance terms were not important predictors in any of the models so they were discarded. I also tested interactions of the fixed effects, but none were important and thus were not retained in the final models. The regressions were fit in a Bayesian framework using JAGS (Plummer 2003) called from within R \citep{Rcor15} using package jagsUI \citep{Kell15}. Model code can be found in Appendix A.
+
+I tested for differences in pellet counts among disturbance treatments (forest, edge, harvest) using permutation tests in R. Deer and rabbit pellet data were analyzed separately. Briefly, for each herbivore, an \textit{F}-statistic and post-hoc differences between pellet count treatment means were calculated for my actual dataset using ANOVA and a Tukey HSD test. Then, the treatment assignments were randomly permuted and the same statistical tests conducted again. This process was repeated 1000 times for each herbivore to obtain distributions of \textit{F}-statistics and post-hoc differences between means. The \textit{F}-statistic and post-hoc treatment means obtained from my actual pellet count data were then compared to these distributions to determine significant differences.
+
+\section{Results}
+
+Over the four years of the study, 1155 seedlings were sampled at least once (range 1-8 samples); the mean was 4.19 samples/seedling prior to seedling mortality. Of those 1155 seedlings, 5.0\% were browsed at least once by rabbits, 26.2\% by deer, and 55\% by insects. For non-zero rabbit browse events, 44\% were the next lowest ordinal value (1), 40\% received a value of 2, and 16\% received a maximum value of 3. For deer, the breakdown was 53\% (1), 37\% (2), and 10\% (3), and for insects it was 73\% (1), 23\% (2), and 4\% (3).
+
+The magnitude of variability in browse intensity explained by the random effects was higher overall for rabbits than for deer and insects (Table 1). For both rabbits and deer there were no important differences between plot, subplot, and seedling variation (i.e., 95\% credible intervals on estimated standard deviations overlapped). For insects, in contrast, variation among plots was higher than among subplots nested within plot or seedlings nested within subplot (Table 3.1).
+
+\input{tables/tab3-1}
+
+Distance into the harvest opening was negatively related to intensity of insect herbivory (96\% of the parameter posterior distribution was the same sign as the mean - the closer this value is to 100\% the more confidence the parameter is positive or negative; hereafter I refer to this value as \textit{f}) and positively related to intensity of rabbit browsing (\textit{f} = 0.93; Table 3.1). Deer browse intensity was also positively related to distance into the harvest from the forest edge, but the relationship was weaker (\textit{f} = 0.81). When distance to edge is converted from a continuous to a categorical variable (i.e., plots are classified as either forest, edge, or harvest interior), these patterns are reinforced (Figure 3.1).
+
+\begin{figure}
+\centering
+\includegraphics[scale=0.9]{figures/fig3-1.pdf}
+\caption{Mean proportion of seedlings browsed at least once during the study, at three locations along a transect (\textgreater 50 m inside forest, forest edge, and \textgreater 20 m inside clearcut harvest) and for three different herbivores: rabbits (\textit{Sylvilagus floridanus}), deer (\textit{Odocoileus virginianus}) and insects (primarily orders Coleoptera, Hymenoptera, Lepidoptera, Orthoptera). Error bars represent 95\% confidence intervals around the means.}
+\label{fig:3.1}
+\end{figure}
+
+Plant competition was negatively associated with both deer and rabbit browse (\textit{f} = 1.00 for both), while competition had essentially no effect on insect herbivory (\textit{f} = 0.60; Table 3.1). All three herbivores browsed more intensively on white oak relative to black oak (\textit{f} = 0.98 for deer, 0.93 for rabbits, and 0.92 for insects). Seedling height was also an important predictor of browse intensity for all three herbivores (Table 3.1). For deer and rabbits, the first- and second-order height effects had opposite signs and high values of \textit{f} (\textgreater 0.99), indicating a quadratic relationship between height and browse intensity with intensity peaking at an intermediate value. As expected this peak was at a greater height for deer than for rabbits (Figure 3.2). For insects, in contrast, the relationship between seedling height and browse intensity was linear since only the first-order height effect had a high value of \textit{f} (Table 3.1, Figure 3.2).
+
+\begin{figure}
+\centering
+\includegraphics[scale=0.9]{figures/fig3-2.pdf}
+\caption{Scaled probability of browse damage in a given time step to a given oak seedling based on its height (cm), for three different herbivores: rabbits (\textit{Sylvilagus floridanus}), deer (\textit{Odocoileus virginianus}) and insects (primarily orders Coleoptera, Hymenoptera, Lepidoptera, Orthoptera). Predictions are based on estimated parameter values from a proportional-odds logistic regression model (Table 3.1). Probability of browse damage for each herbivore is scaled so that the maximum predicted probability is 1, allowing for comparison on a common scale; see Figure 3.1 for unscaled browse probabilities for each species.}
+\label{fig:3.2}
+\end{figure}
+
+For rabbits, there were significant differences in pellet counts among the disturbance treatments (\textit{p} = 0.019). Post-hoc tests revealed that pellet counts were significantly lower in the intact forest than on the edge or inside the harvest, but there was no difference between the edge and harvest treatments (Figure 3.3). There were no significant differences in pellet counts among the disturbance treatments for deer, although pellet counts were highest on the forest edge (\textit{p} = 0.16, Figure 3.3)
+
+\begin{figure}
+\centering
+\includegraphics[scale=0.9]{figures/fig3-3.pdf}
+\caption{Mean pellet count density for rabbits and deer over three disturbance treatments: intact forest, forest edge, and clearcut harvest interior. Within each herbivore, treatment means with different letters were significantly different from each other based on permutation tests.}
+\label{fig:3.3}
+\end{figure}
+
+\section{Discussion}
+
+Herbivores play a key role in driving change in plant communities \citep{Russ01}. In forested areas, patterns of herbivory on tree seedlings can influence the ultimate forest stand composition decades into the future \citep{Holm13}. The intensity of herbivory experienced by a given tree seedling is a function of many interacting factors, including species, seedling size, and location. I demonstrated that browse damage on oak seedlings varies horizontally across a forest-edge-harvest opening gradient and vertically with seedling height; and further, this variation differed among herbivore species.
+
+\subsection{Edge and Harvest Effects}
+
+I attribute the differences in browse intensity across the forest edge primarily to changes in habitat structure following harvest, which in turn affected the distribution of herbivore species. Two major consequences of overstory removal as part of the clearcut harvest were (1) increasing density and diversity of understory vegetation (including, but not limited to, tree seedlings) thanks to increased light availability, and (2) an influx of coarse woody debris (primarily treetops and other unmerchantable timber) left behind after the harvest \citep{Kalb13}. This represents ideal habitat for cottontail rabbits, who prefer dense early-successional habitat that provides them with ample food and protection from aerial predators \citep{Alth97, Mank99, Bond02}. Pellet count surveys confirmed that cottontail rabbits made greater use of the edge and harvest interior habitat relative to the forest interior (Figure 3.3), and where they were more abundant, they were more likely to browse on oak seedlings (Figure 3.1).
+
+Browse damage by insects exhibited a trend opposite that of cottontail rabbits; intensity of browse by insects decreased moving from the forest interior across the edge and into the clearcut harvest (Figure 3.1). Although large number of insect orders can feed on the leaves of oak seedlings (e.g. Coleoptera, Hymenoptera, Lepidoptera, Orthoptera; \citealp{Lini86, Galf91}), the primary source of browse damage to the oak seedlings in this study was likely lepidopteran caterpillars \citep{Lini86, Marq94}. Concurrent research at my study sites found 30-40\% lower lepidopteran richness and shifts in dominant taxa in harvested areas relative to control forest \citep{Summ11}. In a similar forest, an oak-specialist moth \textit{Herculia olinalis} declined in abundance in clearcuts \citep{Summ02}, and \citet{Jeff06} found that abundance and diversity of arthropods that browse oak was lower in young (recently harvested) forests in the Missouri Ozarks.
+
+As with cottontail rabbits, changes in habitat conditions are likely to be a primary factor driving observed trends in insect herbivory. Habitat conditions including altered microclimate and herbaceous ground cover post-harvest may have contributed to this difference \citep{Zhen00, Bass01, Berg04}. Additionally, for oak specialists, loss of large overstory oaks and associated leaf biomass following harvest represents a major decrease in food availability that would lead to population decline \citep{Summ02}. In any case, a decline in abundance of lepidopterans (especially oak specialists) is one likely explanation for the trend in reduced foliar damage due to insects that I observed in the harvest interior and on the edge. Plant defense may also play a role; \citet{Adam03} found elevated tannin levels in the leaves of white oak seedlings on recently burned sites, perhaps due to increased light penetration. Greater light penetration in the harvest opening and on the edge, relative to the forest interior, may have made the leaves of the corresponding seedlings less palatable to herbivorous insects in my study.
+
+I predicted browse damage from deer would be highest on the forest edge since deer abundance is typically maximized in disturbed/edge habitat \citep{Alve88, Wall97, Buck98}, but browse damage from deer on oak seedlings showed the weakest spatial trend of the three herbivore types (Figure 3.1). Clearcut edges are ideal habitat for deer because they provide easy access to complementary habitat - abundant vegetation in the clearcut during the growing season, and cover/food during the winter in the adjacent forest \citep{John95}, but the interior of clearcut harvests become difficult for deer to traverse when vegetation becomes too dense (\textgreater 5 years following harvest; \citealp{Blym77}). Habitat use by deer showed a weak trend of being highest on the harvest edge (Figure 3.3), but that did not translate to differences in browse intensity (Figure 3.1). It is possible that the entire transect length I sampled (230-300 m across the edge) represented edge habitat for deer relative to their home-range size (45-233 ha; \citealp{Tier85, Beie90, Verc98}), in which case I did not have the resolution to detect differences induced by habitat selection. Another factor was moderate deer densities throughout my study area (~7.5 deer per km\textsuperscript{2}, \citealp{Haul09}). A greater deer density may be necessary to elucidate differences in browsing between treatments \citep{Cast00, Ross05}.
+
+Examining the effects of distance to edge is complicated by the effect of surrounding vegetation (i.e., oak's competitors) on browse intensity. Harvesting increased light available inside the opening and (to a lesser degree) on the edge, resulting in the development of a dense layer of early successional vegetation including tree seedlings. While this new habitat structure and abundance of food may attract herbivores (e.g. cottontail rabbits, as discussed previously), it could also reduce the probability that any given oak seedling is browsed simply because there is so much vegetation available. This type of plant defense, in which a seedling is protected by being hard for herbivores to find or access, has been called the Plant Apparency Hypothesis \citep{Feen76}. Browse damage from cottontail rabbit and deer was in fact negatively correlated with the amount of competition oak seedlings faced from other plant species (Table 3.1). A similar negative correlation between competition and herbivory has been observed for American chestnut (\textit{Castanea dentata}) seedlings grown in canopy openings \citep{Dalg15}. \citet{Mill09} found that browsing rates on several tree species, including red oak, declined as the proportion of early successional habitat (with greater understory density) increased, and dense multiflora rose (\textit{Rosa multiflora}) might offer some protection to oak seedlings from herbivores \citep{Mein02}. The impact of interspecific competitors may also explain why \citet{Mein02} observed lower browse damage by rabbits in edge habitat relative to adjacent forests and old fields (in contrast to my results): they hypothesized that a dense patch of \textit{Solidago canadensis} on the edge prevented rabbit access to tree seedlings. Of course, seedlings only benefit from this “protection” if the reduction in browse intensity outweighs the other costs of competition.
+
+\subsection{Seedling Height}
+
+Browse intensity also varied with seedling height in a manner specific to the source of herbivory (Table 3.1). For cottontail rabbits and deer, browse intensity increased with seedling height, peaked at intermediate values, and then declined (Figure 3.2). Extremely short seedlings may provide fewer nutrients per unit browsing effort (due to their small size) and so are not worth consuming, or they may simply be missed below taller vegetation. Deer typically ignore short seedlings, even when deer density is high \citep{Kitt95, Dobs15}. Taller seedlings were likely browsed less intensively simply because they escaped the reach of herbivores. Predictably, browsing intensity by the smaller herbivore (cottontail rabbit) peaked at a lower seedling height (71.6 cm) than the larger deer (93.2 cm) based on the model (Figure 3.2). Note that these seedling heights do not represent the height at which browse damage actually occurred; browse damage could occur at any height on the seedling (thus a peak at 71.6 cm for rabbits, which under most circumstances could not reach that height). In contrast to the quadratic relationship I observed for the mammalian herbivores, browse intensity had a positive, linear relationship with seedling height for insects, (Figure 3.2), which matches the findings of \citet{Mein02} that browse probability on red oak seedlings was positively associated with seedling height. Taller seedlings may simply be more obvious targets for insect herbivory (more leaves, less hidden by other vegetation; another example of the apparency hypothesis). Alternatively, taller seedlings may be associated with higher plant quality (e.g. more foliar nitrogen) that came about either via spatial variation in available soil nutrients across the planting sites, or due to the nursery origin of some seedlings. Increased foliar nitrogen and lower concentrations of foliar phenolics has been shown to increase insect herbivory on oak saplings \citep{Fork00}.
+
+\subsection{Implications for Regeneration}
+
+I present evidence that the activity of herbivores is altered by harvesting; intensity of insect browse damage was higher in intact forest, and cottontail rabbit browse intensity was higher inside the adjacent harvest. Ultimately, however, these shifts are unlikely to be important drivers of oak seedling growth and survival in this system relative to the impacts of deer. Cottontail rabbits browsed a very low proportion of oak seedlings relative to the other two herbivores even in the harvest interior (5\% of all seedlings, Figure 3.1). Additionally, the height at which a seedling can functionally escape rabbit browsing is shorter, and therefore can be reached more quickly, than the equivalent for deer (Figure 3.2). Though few seedlings were browsed by rabbits throughout the study, when browsing did occur it tended to be more severe (16\% of seedlings browsed by rabbits showed the highest damage category compared to 10\% for deer) with seedlings frequently clipped close to the ground; this type of damage may be harder for the seedling overcome \citep{Cade00}. A greater proportion of seedlings suffered damage at least once from herbivorous insects (55\%) than the two mammalian herbivores combined (31.2\%), but the intensity of the damage tended to be low. There is little evidence in the literature that foliar damage by insects has a major impact on the oak regeneration process (\citealp{Lini86, Adam01, Mein02}, \textit{but see} \citealp{Mcph93, Marq94}).
+
+Relative to the other two herbivore types, deer browse intensity was much more consistent across the gradient (Figure 3.1) despite a trend of more deer activity on the harvest edge (Figure 3.3). Therefore, it appears that the relatively small harvest openings (4 ha) created as part of the HEE are not having a major impact, either positively or negatively on rates of deer browsing on oak seedlings. The data presented in this study do not permit me to determine the overall impacts of deer herbivory on oak seedling growth and survival (and therefore, regeneration success) by itself. However, these results coupled with the moderate deer densities observed in my study area \citep{Haul09} suggest that deer herbivory is likely not a crucial limiting factor on oak regeneration in these forests. Chapter 4 addresses this question more directly by examining the interacting effects of competition, herbivory, and timber harvesting on oak seedling survival and growth.
diff --git a/chapter4.tex b/chapter4.tex
new file mode 100644
index 0000000..d9425e3
--- /dev/null
+++ b/chapter4.tex
@@ -0,0 +1,115 @@
+\chapter{{TIMBER HARVEST AND DROUGHT INTERACT TO IMPACT OAK SEEDLING GROWTH AND SURVIVAL IN THE CENTRAL HARDWOOD FOREST}}
+
+\section{Introduction}
+
+Oaks (\textit{Quercus}) are a common, and frequently dominant, member of hardwood forest canopies in eastern North America \citep{John09}. The oak species group has been identified as ‘keystone’ \citep{Fral04} and ``foundational'' \citep{Elli05} species in these forests because they play a number of critical ecological roles; for example, the acorns they produce are a crucial food resource for many species \citep{Mcsh00, Mcsh93}, particularly since the decline of American chestnut (\textit{Castanea dentata}) in the early 1900s \citep{Dalg12}. They also provide substantial economic value to humans as a timber-producing species \citep{John09}. In recent decades, forest managers and ecologists have noted differences in the tree species composition of the understory and midstory in forests where oaks make up a large part of the overstory canopy. Many oak-dominated forests have little to no oak recruiting underneath the canopy; instead, these forest strata are dominated by seedlings and saplings of more shade-tolerant, mesophytic species like maple (\textit{Acer} spp.) and beech (\textit{Fagus grandifolia}; \citealp{Aldr05, Crow88, Nowa08}). If this pattern persists, oak will eventually be lost from the overstory of the stand in the future, along with the ecological benefits it provides \citep{Mcsh07}.
+
+The current failure of oak to successfully regenerate (i.e., to recruit new stems into the population) is a complex problem with many potential causes and no simple solution. The growth, survival, and ultimately the successful recruitment of oak seedlings in eastern deciduous forests hinges on multiple, and potentially interacting, environmental factors including light and moisture availability, interspecific competition, and herbivores. Oaks generally have intermediate shade tolerance although there is variability within the genus \citep{Niin06}, with maximum growth at 50-70\% full sunlight \citep{Dey08}. Initially, oak seedlings devote substantial energy to increasing root biomass at the expense of shoot growth \citep{John09}. This strategy, together with resprouting capability \citep{Crow88, John09} and resistance to fire \citep{Abra92, John09}, increases the probability an oak seedling will be able to survive until adequate light is available for growth.
+
+Abiotic factors that affect oak, particularly light, are also tightly tied to the abundance and composition of competing tree seedlings (and other vegetation). Oak seedlings face competitive pressure across the spectrum of possible abiotic conditions. At highly productive sites and/or when light availability is high, pioneer species like tulip poplar (\textit{Liriodendron tulipifera}), sassafras (\textit{Sassafras albidum}) and black cherry (\textit{Prunus serotina}) are able to more quickly convert abundant resources into shoot growth and so can outcompete oak for light \citep{Beck03, Jenk98}. In contrast, when available light is low (e.g., under a dense canopy), shade tolerant species like sugar maple or American beech outcompete oak seedlings and eventually replace oak in the canopy (``mesophication''; \citealp{Nowa08}). This two-pronged competitive pressure on oaks in high-disturbance habitats (e.g., canopy openings) and low-disturbance habitats (e.g. intact canopy) suggests the importance of intermediate disturbance regimes like low-intensity fires and land clearing for agriculture in driving past patterns of oak dominance \citep{Abra92, Abra96, Abra05, Nowa08}. Over the past century, patterns of fire suppression have reduced the frequency of disturbance \citep{Bros01} and therefore contributed to oak regeneration failure.
+
+Even under ideal abiotic conditions and a favorable disturbance regime, successful oak recruitment is still challenged by browsing pressure from herbivores. Oak seedlings are browsed by a variety of species in eastern hardwood forests, including mammals like white-tailed deer (\textit{Odocoileus virginianus}), rabbits (\textit{Sylvilagus} spp.), voles (\textit{Microtus} spp.) \citep{Mein02, Ostf97, Russ01}, and arthropods from several orders (e.g. Coleoptera, Hymenoptera, Lepidoptera, Orthoptera; \citealp{Galf91, Lini86}). Deer are a particularly important herbivore, especially in areas where they have become unusually abundant due to habitat changes and declines in predator populations \citep{Wall97}. The effects of deer herbivory on oak regeneration are variable, but usually range from neutral \citep{Adam01, Cast00} to negative \citep{Abra12, Heal97, Kitt95, Roon03}.
+
+Disturbance caused by timber harvesting can influence seedling competitive interactions and browse pressure, and therefore the ultimate fate of oak seedlings. In addition to providing wood products \citep{John09}, timber harvesting can thus be used by forest managers to promote certain desired tree species including oak \citep{Dey02}. A number of silvicultural approaches with the goal of regenerating oak have been applied, with varied success across different regions and site qualities \citep{Dey09, Morr08, Schl93, Swai13}. Part of the reason for this varied success is the impact of different timber harvesting approaches on competitive interactions between oak and other tree seedlings. While traditional clearcuts and single-tree selection harvesting are the most widely applied timber harvesting methods, harvesting approaches that introduce intermediate levels of disturbance (e.g. shelterwood harvesting, often combined with low-intensity controlled burns) have been successful in encouraging oak regeneration \citep{Bros99, Dey02, Dey08}, perhaps because they emulate the intermediate disturbance regime that led to historical oak dominance. Larger, more intensive disturbances (e.g. clearcutting) may provide a competitive edge to pioneer species like tulip poplar or sassafras over oak \citep{Dey08}, while smaller disturbances (single-tree selection) do not create enough light for oak seedlings to thrive and outcompete shade-tolerant species \citep{Dey08}.
+
+In addition to its effect on seedling competitive interactions, timber harvest also alters the habitat, and therefore the behavior, of mammal and insect herbivores with implications for the browse damage done to oak seedlings. Harvesting creates edge habitat, which is preferred by key browse species including white-tailed deer \citep{John95, Will85}, while harvest interiors (particularly in larger openings) have a greater abundance and diversity of woody and herbaceous vegetation, attracting forager attention. The dense layer of understory vegetation that quickly develops after harvest, along with a (potentially) large increase in the volume of coarse woody debris left over from harvest activity also represents good habitat for other herbivores like the eastern cottontail (\textit{S. floridanus}) that prefers early successional habitat \citep{Mank99}. Conversely, the elevated density of plant stems and variety of plant species present post-disturbance \citep{John95} means that a given oak seedling may experience a lower probability of browse damage than a seedling in intact forest. Dense vegetation post-disturbance and piles of coarse woody debris (e.g. treetops) left behind after harvest can also offer protection for seedlings if they prevent larger herbivores like deer from accessing some parts of the disturbed area \citep{Blym77}. This type of defense against herbivores has been called the “apparency” defense \citep{Feen76}.
+
+Interspecific competition, herbivory, and disturbance (e.g., timber harvest) all play crucial roles in determining if oak regeneration will be successful in a given stand and under a given management scenario. There is considerable literature on the response of oak seedlings to these factors individually, but less is known about their interactions and concurrent effects. To quantify the effects of competition, herbivory, and harvest treatment on oak seedling survival and growth, I experimentally planted oak seedlings of multiple species (black oak \textit{Q. velutina} and white oak \textit{Q. alba}) and ages (new seedlings and year-old nursery seedlings). Treatments were applied in a location (inside forest, on harvest edge, inside harvest opening) $\times$ herbivory (protected or not) $\times$ competition (competitors removed or not) factorial design. I hypothesized that seedlings protected from both interspecific competitors and herbivory, and located within the harvest opening where sunlight is maximized, would have the highest levels of survival and growth.
+
+\section{Methods}
+
+\subsection{Study Site}
+
+This study took place in Morgan Monroe and Yellowwood State Forests in south-central Indiana, United States, part of the Central Hardwood Forest region \citep{Fral03}. The state forests together comprise \textgreater 19,000 ha in three Indiana counties and are managed for multiple uses including recreation, conservation, and timber production \citep{Carm13}. Past timber harvest in the state parks has been mainly single-tree and group selection \citep{Carm13, Jenk98}. The forest overstory is dominated by oak and hickories (\textit{Carya}), while the understory and midstory strata are primarily shade tolerant species including sugar maple and American beech \citep{Saun13}. Prior to the timber harvests implemented as part of this study, basal area in the study area ranged from 21.7-29.9 m\textsuperscript{2}/ha, and total tree density ranged from 923-1,527 trees/ha \citep{Saun13}. The region is characterized by steep slopes (23-35\%) and silt-loam soils \citep{Indi84, Jenk98}. Annual mean temperature is 13-16$^{\circ}$ C and mean precipitation is 112-137 cm \citep{Mcna94}. The most common mammalian herbivores in the study area are white-tailed deer (~7.5 deer per km\textsuperscript{2}; \citealp{Haul09} and the eastern cottontail rabbit.
+
+Research was conducted within the framework of the Hardwood Ecosystem Experiment (HEE), a landscape-scale, long-term study of forest ecosystem response to timber harvest \citep{Kalb13}. As part of the HEE, ten research units (each 78-110 ha in size) were established in 2006, each receiving one of three silvicultural treatments: even-aged harvest (rotations of 4 ha clearcuts and three-stage shelterwood harvests), uneven-aged harvest (0.4-2 ha patch cuts and single-tree selection), or unharvested control. The three-stage shelterwood consisted of (1) an initial removal of non-oaks \textless 25.4 cm dbh and a minimum residual basal area of 13.8 m\textsuperscript{2}/ha, followed by (2) an establishment cut 5-10 years later reducing basal area to 13.8-16.1 m\textsuperscript{2}/ha, and finally (3) removal of remaining overstory trees 5-10 years after stage 2 \citep{Kalb13}. The initial round of harvests, including clearcuts and the first shelterwood stage, occurred in fall 2008-winter 2009. Only the first shelterwood stage had been completed when this study was conducted.
+
+\subsection{Experimental Plots}
+
+In spring 2010, plots were established in the three HEE research units selected for even-aged harvest. Within each unit, a transect of four plots spanned the edge of a 4 ha clearcut (two plots inside the harvested area, one directly on the harvest edge, and one \textgreater 100 m into the surrounding forest matrix). An additional plot was established inside a 4 ha shelterwood harvest in each unit, for a total of 15 plots across the three units.
+
+At each of the four clearcut plot locations, I established four 5 $\times$ 5 m subplots. In the harvest edge plot, the subplots were arranged randomly along the edge, at least 10 m apart from each other. In the other plots, subplots were arranged randomly in a 2 $\times$ 2 grid, again at least 10 m apart from each other. The distance from each subplot to the harvest boundary (in the clearcut areas) was obtained using a handheld GPS unit and ArcMap 10 geographic information systems (GIS) software (ESRI, Redlands, CA). To examine the effects of deer herbivory and interspecific competition, I applied a 2 $\times$ 2 factorial design to the four subplots. Specifically, two subplots were located within permanent deer exclusion fences (no herbivory treatment), and two outside. One subplot inside the fence and one outside were also treated in spring 2010 with glyphosate herbicide and then manually weeded to remove all existing vegetation (no competition treatment). The no-competition treatment subplots were then manually weeded twice annually at the beginning and end of the growing season, with all non-oak vegetation removed. At the shelterwood plots, two subplots were established: one inside and one outside a deer exclusion fence.
+
+\subsection{Oak Plantings}
+
+A pilot study in spring 2010 in which subplots were directly planted with \textgreater 5,000 total acorns intended to generate oak seedlings revealed that granivore pressure was too high to generate an acceptable number of seedlings (\textgreater 95\% of planted acorns pilfered). Therefore, in May 2011 I planted oak seedlings instead of acorns in the subplots. Oak seedlings came from two sources. The first source was 1-0 (year-old) nursery-grown, bare-root seedlings from the Indiana Department of Natural Resources tree nursery at Vallonia, IN. Upon receipt of 1-0 seedlings, the largest and smallest 25\% were discarded to insure all seedlings were of similar size and condition. The second group of seedlings were greenhouse-grown beginning in March 2011 (hereafter referred to as 0-0 seedlings). Acorns collected on the HEE sites and at the Vallonia state nursery were planted in 5.6 cm $\times$ 10.2 cm Jiffy pellets (Jiffy Products of America Inc., Lorain, OH, USA). When the 0-0 seedlings reached ~ 10 cm in height, they were moved from the greenhouse to a partially shaded, enclosed, outdoor area for 1-2 weeks to harden them for transplanting. In May 2011, all seedlings were planted in the field using a tree planting bar. Each subplot in the clearcut transects received 24 0-0 seedlings (8 white oak, 16 black oak) and 15 1-0 seedlings (8 white oak, 7 black oak) for a total of 39 seedlings/subplot. Number of seedlings/species were not equal due to differences in germination success in the greenhouse. Seedlings were planted in a grid with all 0-0 seedlings separated by \textgreater 0.30 m and 1-0 seedlings separated by \textgreater 0.60 m.
+
+\subsection{Seedling Measurements}
+
+One month after planting, seedlings were assessed for initial mortality. Seedling mortality during this period was assumed to be primarily due to transplant shock, and seedlings that were dead were excluded from future analysis. The initial height of surviving seedlings (from base to tallest part of the plant) was recorded along with the root collar diameter. In October 2011, at the end of the growing season, plots were re-visited to check seedling mortality and re-measure height and root collar diameter. In addition, browse damage by herbivores was assessed on a 4 point ordinal scale (1 = no browse damage, 2 = minor browse damage, 3 = medium browse damage, 4 = heavy browse damage) and attributed to either deer or smaller herbivores based on the excision characteristics \citep{Verc05}. Any resprouting by the seedlings was noted. In early May of the following year (2012) plots were again visited immediately following initial leaf-out to assess mortality, herbivore browse and leaf damage, and resprouting. This pattern of May and October visits continued in each year of the study through October 2014.
+
+\subsection{Competition Measurement}
+
+Keeping the ``no-competition'' subplot treatment completely free of all interspecific competitors at all times was essentially impossible, particularly in the recent clearcuts. Thus, I measured vegetation characteristics within each subplot in early summer of each study year to create a quantitative metric of competition. Visual estimates of percent woody and herbaceous cover were recorded, along with a count of all non-oak stems falling into three height classes: \textless 0.5 m, 0.5-1.0 m, and \textgreater 1.0 m. I created a light competition index for each individual oak seedling in each year by calculating the density of stems in the sample plot that could potentially compete with the seedling for light (i.e., competing stems equal to or greater than the height of the oak seedling).
+
+\subsection{Analysis}
+
+I selected two dependent variables for analysis: (1) seasonal oak seedling survival, and (2) yearly seedling height growth. For both variables, oak species were modeled separately using a similar hierarchical generalized linear mixed modeling approach. Random effects were organized into a nested structure: individual seedlings were nested within subplots, nested within plots. Fixed effects included competition (I ran models for a binary competition/no competition covariate as well as the continuous metric of light competition described above), presence or absence of browse damage, harvest treatment (control, clearcut harvest edge, clearcut harvest interior, and shelterwood harvest), site aspect, sprout status (i.e., if the seedling has been previously top-killed and was resprouting), and time. I also tested interactions of these fixed effects, but none were statistically important and so they were not included in the final model. Seedling survival over a given sampling occasion/season (n=7 sampling occasions; October 2011- October 2014) was modeled using using a logit link function (i.e., logistic regression). Individual seedling random effects were omitted from the survival analysis due to convergence issues. Seedling root collar diameter and season (growing season or winter) were additionally included as fixed effects on survival. Seedling growth during a given growing season (n=4; 2011-2014) was modeled using an identity link function (i.e., normal linear regression).
+
+All analyses were fit in a Bayesian framework in program JAGS \citep{Plum03}, called from within R 3.1.0 \citep{Rcor15} using package jagsUI \citep{Kell15}. Each analysis included 3 parallel Markov chain Monte Carlo (MCMC) chains of at least 10,000 iterations each. Chains were run until adequate convergence was reached (Rhat values \textless 1.1, \citealp{Broo98}). Model fit was assessed by plotting posterior predictive checks and calculating a Bayesian \textit{p}-value for each analysis \citep{Kery10}. Model parameters were considered significant (i.e., significantly different from 0) if their 95\% credible intervals (based on the posterior distribution) did not overlap 0. Model code can be found in Appendix B.
+
+\section{Results}
+
+\subsection{Survival}
+
+In a combined multi-species model, white oak had a higher overall survival probability than black oak (Figure 4.1); hereafter, oak species were analyzed separately. There were few significant predictors on per-interval survival at the spatial scales of the plot and subplot. Black oak had significantly lower survival inside recent clearcuts and on slopes with southwestern aspects (Table 4.1, Figure 4.1) but survival on harvest edges and inside first-stage shelterwoods was similar to unharvested forest. For white oak, neither aspect nor harvest treatment had a significant effect.
+
+\begin{figure}
+\centering
+\includegraphics[scale=0.9]{figures/fig4-1.pdf}
+\caption{Proportion of all seedlings of two species (white oak \textit{Q. alba} and black oak \textit{Q. velutina}) surviving over a 4 year period in 4 silvicultural treatments. Error bars are standard errors around the mean. Seedling survival was assessed 8 times total during the 4 year period, once in the spring/summer (designated S on the x-axis) and once in the fall (designated F) of each year (2011 - 2014). Statistically important differences between harvest treatment means (based on the hierarchical logistic regression) are represented by different letters.}
+\label{fig:4.1}
+\end{figure}
+
+\input{tables/tab4-1}
+
+At the spatial scale of individual seedlings, increasing competition led to significantly lower probability of survival for both species, but browse damage had no effect (Table 4.1). Independent of other predictors, increasing time since the seedling was planted and seedling root collar diameter (as a surrogate for seedling root biomass / overall size) had significant positive effects on survival for both species (Table 4.1). Survival probability was lower during growing seasons than in the winter (Figure 4.1, Table 4.1). In all intervals following a dieback event that resulted in resprouting, affected seedlings of both species had a significantly lower probability of survival (Table 4.1).
+
+\subsection{Growth}
+
+Overall yearly height growth was similar between black and white oak seedlings (Figure 4.2). Yearly growth was significantly higher inside clearcut harvest openings for white oak, and marginally so for black oak; growth on clearcut edges and inside first-stage shelterwood harvests was not significantly different from unharvested forest (Table 4.2, Figure 4.2). Yearly growth was also marginally higher on southwest-facing slopes for both species (Table 4.2).
+
+\begin{figure}
+\centering
+\includegraphics[scale=0.9]{figures/fig4-2.pdf}
+\caption{Mean yearly seedling growth (cm) by oak species (black oak \textit{Q. velutina} and white oak \textit{Q. alba}) and harvest treatment (forest matrix, first-phase shelterwood harvest, clearcut edge, and clearcut interior). Error bars represent standard error around the mean. Within-species means that do not share a common letter are significantly different based on Tukey HSD tests.}
+\label{fig:4.2}
+\end{figure}
+
+\input{tables/tab4-2}
+
+For individual seedlings, the presence of browse damage in the previous year had a significant negative impact on growth for both species, and interspecific competition had a marginal negative effect (Table 4.2). As with survival, time since planting had a positive effect on yearly growth, but height growth following dieback and resprouting was significantly lower (Table 4.2).
+
+\section{Discussion}
+
+Oak seedlings can be outcompeted under the most intensive timber harvest disturbances (e.g., clearcuts), particularly on productive and mesic sites (\citealp{Beck86, Jenk98, John09, Swai13}, \textit{but see} \citealp{Morr08}). The ostensible explanation is that competing shade-intolerant tree species (tulip poplar, sassafras, etc.) are better able to quickly grow and monopolize the newly available light in a clearcut while oaks are diverting more resources into root biomass \citep{John09}. However, other biotic and abiotic factors may play a role in determining if oak successfully regenerates at these sites. My study did not examine oak seedling competitive success relative to other species per se, but I was able to quantify the relative impacts of several environmental factors including competition, browse intensity, and post-harvest environment on oak seedling survival and growth (which are key metrics of competitive success).
+
+\subsection{Site Conditions}
+
+Previous work has shown that both growth and survival of red oak (\textit{Q. rubra}) seedlings increased with decreasing canopy cover (e.g. \citealp{Beck03, Crow92}, \textit{but see} \citealp{Buck98}). I therefore expected oak seedling survival and growth in my study to be maximized in the interior and on the edge of clearcuts, where available sunlight is maximized. However, for black oak, seedling survival probability was lowest inside clearcut harvests, particularly those on southwest-facing slopes where sunlight is the most direct; Table 4.1), and harvest treatment had no effect on growth (Table 4.2). The negative effect of harvests and southwestern slopes is counterintuitive - planting sites in these conditions should be more xeric and maximize the available sunlight (both competitive advantages for the water-stress tolerant, but shade-intolerant oaks; \citealp{John09}).
+
+Part of the explanation for this pattern lies in the severity of the weather conditions for the first two growing seasons in the study (2011 and 2012). These two years had unusually hot and dry summers for southern Indiana. The study area reached “severe drought” status (category D2) and above for a portion of the growing season in both years \citep{Nati15}. Mean July temperatures in 2011 and 2012 were 25.3$^{\circ}$ and 26.6$^{\circ}$ C, respectively, compared to the 30 year monthly average of 23.1°. Moreover, total July precipitation was 1.63 and 1.42 cm, respectively, compared to a 30 year average of 10.9 cm (National Climatic Data Center 2014). This likely also contributed to the overall lower survival of black oaks in the growing season (Table 4.1), particularly obvious in the first two years of the study where the steepest declines in survivorship occurred in the first two growing seasons (Figure 1). Black oak seedlings located in the clearcuts and on southwestern-facing slopes were exposed to the hottest and driest conditions in these two summers, perhaps exceeding their tolerance for drought \citep{Andr14, Mart87, Mcca94, Mein02}. The following two summers (2013 and 2014) were milder; contributing to the positive effect of time overall on survival (Table 4.1). The drought coincided with the first two growing seasons after the seedlings were transplanted; if it came instead in the final two years of the study, after the seedlings had more time to develop root biomass, the negative effect of being inside the harvest would probably not have been as strong. Still, many of the seedlings were year-old and nursery-grown and already had substantial root systems in place when they were transplanted.
+
+For white oak, in contrast, there was no significant effect of harvest treatment or aspect on survival (Table 4.1), and as expected there was a significant positive effect of harvest on white oak growth (Table 4.2, Figure 4.2). Further, when the two species were pooled together in the models, white oak had higher overall survival probability and growth. There were some environmental effects, as survival probability was lower for white oak during growing seasons and survival probability (and growth) increased over time as it did with black oak (Table 4.1, Figure 4.1), but overall white oak seedlings appeared to be more resilient than black oak to the drought conditions in the first two years of the study, at least in the driest areas. That I observed these differences between \textit{Quercus} species is itself not surprising; congeneric physiological differences appear most frequently when seedlings are under high temperature and moisture stress \citep{Abra90, Bart96, Abra94, Kubi94}. However, my finding that white oak was more resilient than black oak to the extremely dry conditions inside harvest openings and on southwestern-facing slopes contrasts with evidence in the literature that black oak saplings have a higher photosynthetic rate and water use efficiency under drought conditions than white oak \citep{Abra90}, though both species appear to be better adapted to dry sites than red oak, also common in my study area \citep{Abra90, Baha85}. The reason for the differences in survival between the two species in the harvested sites may simply be that white oaks grew more quickly in the harvested sites (Table 4.2) and larger seedlings had a higher probability of survival (root collar diameter was positively related with survival probability for both species; Table 4.1).
+
+Regardless of the underlying cause, quantifying the survival and growth response of tree seedlings generally and typically drought-tolerant oaks specifically to extreme drought should continue to be a research focus. Climate change is predicted to bring warmer temperatures and more frequent drought events to eastern U.S. forests \citep{Mish10}. These patterns may lead to an overall increase in tree mortality \citep{Alle10} and are likely already contributing to increased oak mortality in the midwestern and southeastern U.S. \citep{Clin93, Voel08}. While black oak, and to a lesser extent white oak, were negatively impacted by drought in this study, more frequent and/or more severe drought events may ultimately be a net positive for oak regeneration success since many competing tree species are less tolerant to drought conditions. This is especially true on fertile sites where oak is normally out-competed by fast-growing pioneer species like tulip poplar. \citet{Morr08} found that several consecutive drought years resulted in high tulip poplar mortality on clearcut sites in southern Indiana; this reduced competition with oak seedlings and sprouts and thus contributed to successful oak regeneration. Simulation studies also predict changes in forest composition with changing regimes of drought brought on by climate change, with white oak appearing to benefit from longer drought events \citep{Gust13}. \citet{Gust13} also highlighted the relative paucity of empirical studies examining tree response to drought in the mixed hardwood forests of the eastern United States; my study provides relevant data for two oak species.
+
+\subsection{Competition and Herbivory}
+
+Interspecific competition had mixed, but generally negative, effects on seedling survival and growth (Tables 4.1-4.2). Competition reduced survival for both oak species, but the magnitude of the effect was relatively small compared to other predictors like harvest type, season, and aspect (Table 4.1); growth was unaffected by understory competition (Table 4.2). The competition metric is most aptly described as understory competition, because it included only those stems in the same forest strata as the oak seedlings (e.g., seedlings of other tree species and other woody vegetation) and not competing midstory and overstory trees. Overstory and midstory competition was altered by the harvest treatments, and these effects were part of the effect of harvest treatment (Tables 4.1 and 4.2). Thus, the total impact of all levels of competition on oak seedlings is likely larger, but it is difficult to separate the changes in competition as a result of harvest from their other effects.
+
+In a study of similar length to mine (5 years), \citet{Lori94} found that the presence of competing understory stems reduced red oak seedling survival and growth. However, all the seedlings in that study were in mature forest stands whereas the majority of seedlings in my study were in harvested areas. The reduced competition with mature trees for light in the harvested stands may have masked any competitive effects of the understory in this study. \citet{Beck70} observed a similar pattern; removal of understory vegetation had no effects on red oak seedling growth but removal of the overstory (or both overstory and understory) greatly increased growth.
+
+I found no significant effect of mammalian browse damage on the survival of either oak species within a given sampling interval (Table 4.1). Oak seedlings have some short-term tolerance to browsing \citep{Harm99, Bide15}. However, repeated browsing (especially on slower-growing tree species like oak) will eventually deplete carbon reserves and impact seedling survival \citep{Cote04, Russ01}. I did not analyze the cumulative effect of multiple browsing events on seedling survival, but less than half (40\%) of oak seedlings browsed in this study were browsed more than once. This despite the fact that the majority of the seedlings (98\%) in my study remained relatively short (\textless 1 m tall) throughout the duration of the study, meaning their tallest shoot was still accessible at least to deer if not also rabbits and other herbivores.
+
+Though browsing did not affect survival in the short term, there was a relatively large negative effect of herbivory on seedling growth (Table 4.2). Since growth was measured as the change in seedling height (from the ground to the apical meristem) over the growing season, this implies that herbivores were often browsing the tallest shoots on a given seedling (and the apical meristem), thus negating some or all of the growth in a growing season. This type of browse damage is especially problematic for oaks as it reduces the ability of the seedling to compete for light with surrounding vegetation. In recent harvests deer browsing preferences exacerbate the competitive disadvantages of oak versus faster-growing pioneer species like tulip poplar that are less preferred \citep{Wake09}. Oak is at a disadvantage in mature forests as well, where shade tolerant seedlings like maple and beech are not only better able to tolerate low-light conditions, but are also less preferred by deer \citep{Mill09}. So, even if browsing does not strongly affect oak seedling survival in the short term, browsed seedlings are unlikely to ever reach a height where they can escape herbivory and compete for space in the canopy \citep{Marq81}.
+
+Prior studies have reported a range of impacts of mammal herbivory on oak seedlings; in some situations herbivory has had a minimal effect on oak regeneration \citep{Cast00, Adam01} while in others it has had a negative impact \citep{Abra12, Heal97, Kitt95, Mein02, Roon03, Ross05}. In this study, only 17\% of seedlings were ever browsed and only 6.7\% faced repeated browsing. The most important reason for this range of impacts is likely variation in herbivore density. When herbivore (particularly white-tailed deer) density was low, herbivory was not an issue for oak regeneration (e.g. \citealp{Cast00}). Deer are actively hunted in all my study areas, keeping populations at moderate levels \citep{Haul09, Hoov13}.
+
+There was no evidence of interactions between harvest treatment, herbivory, and/or understory competition. Lack of an herbivory-harvest interaction may be temporary. Browsing by white-tailed deer (by far the most important herbivore in this system) was similar in forest interior, edge, and harvest habitat (Chapter 3). Though I did not find an interaction between the effects of competition and herbivory, the two processes are related - seedlings exposed to a higher level of competition had a lower probability of being browsed (Chapter 3). Increased competition may reduce browse damage simply because a given oak seedling is less likely to be targeted by herbivores when surrounded by many other stems (the Plant Apparency Hypothesis; \citealp{Feen76}). Additionally, as the clearcut harvests age, they become less accessible to herbivores (especially deer) because of the dense vegetation \citep{Blym77}.
+
+\subsection{Conclusions}
+
+Environmental conditions, harvesting, competition, herbivory, and related processes like seed predation, dispersal, and germination, act as biotic filters driving tree recruitment, and thus ultimate forest composition \citep{Zamo14}. Identifying how the relative importance of these processes varies, both spatially and temporally, is crucial to understand the ecology of the system and choose appropriate management strategies. For oak-focused forest management, accumulation of sufficient oak advance regeneration may be primarily limited by competitors in some cases, and herbivory in others \citep{Ross05, Swai13}.
+
+For black and white oak seedlings in a southern Indiana forest, we found that abiotic factors and environmental conditions (site characteristics, climate) played a more important role in driving oak seedling survival and growth than seedling interactions with understory competitors and herbivores. Most importantly, severe drought conditions coincided with high mortality and minimal growth in the first two years after seedlings were planted, and it appeared that the effects of the drought on black oak seedlings were strongest in harvested areas (which ostensibly are better for shade-intolerant oak). Shelterwood harvest systems are well-suited to promote accumulation of oak advance regeneration \citep{Loft90, Schl93}, and it appears that they also buffered against the drought-related mortality observed in the first two years of the study (Figure 4.1). \ No newline at end of file
diff --git a/chapter5.tex b/chapter5.tex
new file mode 100644
index 0000000..1b620a8
--- /dev/null
+++ b/chapter5.tex
@@ -0,0 +1,245 @@
+\chapter{{SOEL: AN INDIVIDUAL-BASED FOREST GAP MODEL FOCUSED ON PROCESSES DRIVING DEMOGRAPHY OF EARLY OAK (\textit{QUERCUS}) LIFE STAGES}}
+
+\section{Introduction}
+
+Understanding the development and function of forests is a significant research challenge. Processes that drive tree reproduction (e.g. seed predation, dispersal, seedling herbivory, competition) occur on a short time scale relative to the potential lifespan of a tree (hundreds to thousands of years). Thus, the ultimate consequences of small perturbations to a forest ecosystem may take centuries to manifest. Empirical forest research typically has been limited to much shorter time intervals, which makes it difficult to ascertain what factors caused a forest to reach its current state and (perhaps more importantly) what the forest will look like in the future. Some forest research projects have collected or are working to collect long-term datasets (e.g. \citealp{Swan88, Fran90, Bier92, Sher02, Kalb13}), but these projects are arguably the exception and not the rule. Historical patterns can be reconstructed using natural records like tree rings (Lafon and Speer 2002, Speer 2010), fire scar chronologies \citep{Arno77, Guye06}, pollen records \citep{Pitk99}, or even woodrat middens \citep{Beta90, Thom00}. These approaches can help us understand trends in tree growth, climate, and vegetation composition over time but are less useful for understanding interactions between trees and other organisms or predicting the effects of new management strategies. To address the latter issues and many others, forest researchers have often turned to simulation modeling \citep{Liu95, Mlad99, Bugm01a}. While modeling allows for great freedom in answering research questions, the complexity of the questions and the time interval in which they are addressed are limited by computational power and by the modeling framework (i.e., structure and parameterization). Thus, a wide variety of modeling frameworks have been developed, each with strengths and weaknesses, to address different types of forest research questions.
+
+These forest modeling frameworks can be organized conceptually by the degree to which a model incorporates details about individual organisms. At one end of this conceptual scale are forest ecosystem (or forest landscape) models \citep{Mlad99}. Modeling detailed dynamics of individual trees (and/or individual tree species) is not a goal of these models; instead they focus on overall stand density, composition, and biomass. This approach sacrifices individual detail for tractability in estimating large-scale, long-term patterns at a landscape scale (i.e., effects of climate change or carbon fixation). Probably the best example is the LANDIS II model developed by Eric Gustafson, R. J. Scheller, and colleagues \citep{Mlad04}.
+
+Matrix projection models track counts of individual organisms in discrete age, size, or life stage classes. Within a given class, all individuals are considered identical. Matrices of transition probabilities (“vital rates”) are parameterized to describe how individuals move between the categories. Useful information about population dynamics is straightforward to obtain from matrix models; for example, the population’s asymptotic growth rate λ and the stable age/size distribution can be calculated as the dominant eigenvalue and right dominant eigenvector, respectively, of the transition matrix \citep{Casw01}. These parameters are useful for assessing the trajectory of the population and how it is affected by different vital rates. Thus, they have been applied to a wide variety of plants and animals since they were introduced \citep{Lesl45} including trees \citep{Huen87, Pric94, Lian05, Lian13}.
+
+For long-lived, relatively slow-growing organisms like trees, matrix projection models do have some disadvantages. First, trees in matrix models are usually classified based on size (e.g., height or diameter at breast height), even though size actually varies continuously \citep{Zuid10}. Thus, heterogeneous individuals in a given size class will be treated as identical, which introduces error. Increasing the number of discrete classes can reduce this error, at the cost of reduced sample size and therefore reduced precision in transition probability estimation \citep{Elln06}. Changing the number of discrete classes in the model can also influence elasticity and sensitivity measurements \citep{Enri95}. Second, matrix models have only limited ability to incorporate variability in growth rates among individuals \citep{Zuid10}. For long-lived species especially, individual variability in growth may strongly influence population growth rate \citep{Zuid09}.
+
+Integral projection models (IPMs) are a logical extension to matrix models that address these disadvantages. The key difference between IPMs and matrix models is that while vital rates in matrix models are class-specific, in IPMs vital rates can be functions of the underlying continuous trait (e.g., size) \citep{Coul12}. This characteristic allows IPMs to incorporate individual variability in vital rates and relax the requirements for discrete classes \citep{Zuid10}. IPMs address these limitations of matrix models while still yielding similar information about population dynamics including population growth rate and stable size distribution \citep{Elln06}. While IPMs are flexible overall, some ecological processes are currently more challenging to incorporate into the IPM framework, including spatially explicit variability and dispersal (\citealp{Jong08, Adle10}, \textit{but see} \citealp{Jong11}).
+
+Individual-based models (IBMs; also called agent-based models) use a simulation approach to examine how an overall system (e.g. a population or community of organisms) affects, and is affected by, its autonomous component agents \citep{Grim05}. In ecology, they share a general structure with analytical approaches like matrix and integral projection models in that they are driven by an underlying set of ecological processes (e.g. survival and growth). Individual-based models allow for straightforward inclusion of spatially explicit processes, interaction between individuals, and adaptive behavior of individual agents, which are challenging to implement in analytical modeling approaches \citep{Grim06}. However, because they are based on simulation they give up some of the analytic results that can be obtained from matrix and integral projection modeling \citep{Jong11}. Additionally, they typically require a large number of parameters and substantial computing power \citep{Grim06}.
+
+IBMs have a long history in forest modeling and have become increasingly popular in ecology in recent years owing to new software tailored to IBM analysis and increases in computing power \citep{Grim05}. Forest IBMs can take many forms but typically consist of a schedule of submodels or rules that govern the vital rates, movement, and interaction of individuals with each other and the environment over a series of time steps \citep{Grim05, Bugm01a}. Among the first and most influential forest IBMs was JABOWA, which inspired a class of IBMs called “gap models” \citep{Botk93}. The basic structure of JABOWA consists of a number of independent 10 $\times$ 10 m cells, each containing a set of trees that regenerate and grow according to species-specific rules and compete for light with other trees in the same cell \citep{Botk93}. JABOWA was widely implemented, in part because of the ease with which it is parameterized (only 5-10 parameters per tree species are required), and also because the independent cells can be simulated sequentially to cut down on computing power required \citep{Bugm01a}.
+
+Of course, the simplicity of the model detracts from realism; JABOWA makes many assumptions about the shape of tree crowns, trees in adjacent cells do not interact in any way, and ecological processes connecting adult and sapling trees (seed production, dispersal, etc.) are ignored. Other gap models built upon the foundation of JABOWA while adding additional realism, including FORET \citep{Shug77} and ZELIG \citep{Urba90}. Among other additions, ZELIG added more detail to the simulation of tree crown shapes and allowed for large trees to shade adjacent cells \citep{Urba92}. ZELIG has been further modified to examine the effects of herbivory on tree regeneration \citep{Seag01, Holm13}. Even more complex forest IBMs like SORTIE have taken advantage of modern increases in computing power to offer a highly realistic representation of forest structure and plug-ins to simulate numerous ecological processes that drive forest dynamics, including seed production, dispersal, and herbivory \citep{Paca93, Paca96}. Of course, in addition to their computational requirements, these more detailed modeling frameworks require many more parameters to be estimated (either from the literature or via experiments) than the original types of gap models and must be carefully calibrated to a specific region and forest type; thus, they may not be the most logical choice for simpler studies of forest dynamics.
+
+Forest simulation models have been employed to answer a wide variety of research questions across many study systems, regions, and forest types \citep{Bugm01a, Bugm01b}, but fine-scale ecological processes driving dynamics of early-tree life stages (e.g. seed predation, seed dispersal, and herbivory) have not been a major modeling focus (\textit{but see} e.g. \citealp{Seag01, Holm13}). One system that would benefit from a better understanding of how these processes ultimately influence regeneration and forest succession is the central hardwood forest region of the eastern United States \citep{Fral03}. Oak (\textit{Quercus}) is a dominant canopy species in this forest region providing key value to wildlife as a food resource \citep{Mcsh00, Mcsh93} and to humans as a timber species \citep{John09}. While dominant in the canopy, oak is typically absent from the forest midstory and understory in the central hardwoods; over time this is expected to result in oak being replaced in the forest canopy by more shade-tolerant tree species (``oak regeneration failure'', \citealp{Aldr05, Nowa08}).
+
+A primary culprit in oak regeneration failure is shifts in disturbance regimes, namely suppression of wildfires that favored regeneration of fire-adapted oak (Abrams 1992). Timber harvesting is one way of reintroducing disturbance to promote oak regeneration (Dey 2002), and a number of different silvicultural techniques have been developed to do so, with mixed success \citep{Schl93, Morr08, Dey09, Swai13}. However, oak regeneration is also challenged by a multitude of other factors, including a large suite of seed predators and herbivores that rely on oak acorns and oak seedlings for food \citep{Mcsh00, Mcsh93, Mcsh07}. Acorn-seed predator and oak seedling-herbivore interactions may themselves be affected by timber harvesting disturbance (e.g. \citealp{Bell05, Lomb08, Mill09, Crim10, Kell14}). Understanding and predicting the outcome of these trophic interactions is challenging, but would be valuable for designing silvicultural prescriptions that promote oak regeneration.
+
+The ecological processes occurring in the early life stages of oak (i.e., acorn to seedling) occur over a short interval of time (a few years) relative to the oak life expectancy (hundreds of years). Observed and experimental data related to the early life history of oak (e.g. Chapters 2-4) is often similarly collected over short periods of time. This temporal and logistical limitation of oak studies makes it difficult to project the ultimate consequences of factors affecting early oak life stages on the final population or forest stand. I developed a model to bridge that gap; i.e., a simulation-based approach to determine how changes in parameters of oak early life stages ultimately affect the persistence of oak as a canopy dominant species in managed forest stands hundred(s) of years later.
+
+\section{Selection of a Modeling Framework}
+
+The model structure was designed with several important considerations in mind. First, empirical data for parameterizing early oak life history processes were collected at the scale of individual acorns and seedlings (\citealp{Kell14}, Chapters 2-4), meaning the forest ecosystem modeling approach (e.g. LANDIS), which occurs at the stand scale and does not track individual trees, was not appropriate. Second, early oak life history involves several spatially-explicit processes including acorn dispersal and interaction between oak seedlings and nearby competitors (primarily competition for light). Given these considerations, I judged an individual-based simulation model to be the most straightforward approach to incorporating my field data into a spatially-explicit modeling framework.
+
+I therefore developed a spatially-explicit, individually-based model called SOEL: Simulation of Oak Early Life history. SOEL is composed of two primary submodels: the early life history submodel, and the contextual forest submodel simulating the forest in which the early life history submodel occurs (Figure 5.1). The early oak life history submodel tracks oaks from acorn to sapling and is governed by a set of key parameters parameterized using field data (Chapters 2-4, \citealp{Kell14}). The contextual forest submodel simulates growth and mortality of sapling-size and larger trees. I did not have field data to parameterize the contextual forest submodel; thus, I adopted an existing modeling framework to do so, JABOWA (Botkin 1993). JABOWA was selected due to its widespread use yet relatively simple parameterization; it was modified in several ways to match the spatially-explicit structure of the early oak life history submodel. An overview of the model is given in section 5.3, followed by a detailed description of the early life history submodel in 5.4 and the contextual forest submodel in 5.5.
+
+\begin{figure}
+\centering
+\includegraphics[scale=0.7]{figures/fig5-1.pdf}
+\caption{Overview of SOEL structure including (a) the contextual forest submodel derived from JABOWA and (b) the early life history submodel parameterized with field data. Key parameters in (b) are in gray ovals. For probability parameters in (b), black arrows indicate the result of a ``success'' and gray arrows the result of a ``failure''.}
+\label{fig:5.1}
+\end{figure}
+
+\section{Model Overview}
+
+SOEL is summarized below using the ODD (Overview, Design concepts, Details) protocol \citep{Grim10} and was implemented in the NetLogo 5.2.0 programming language \citep{Wile99}. Model code for SOEL can be found in Appendix C.
+
+\subsection{Purpose}
+
+The purpose of SOEL is to determine how processes involved in the early life history of oak (e.g., seed production, predation, dispersal, competition and herbivory) interact with timber harvest to drive oak regeneration, and ultimately forest succession, following harvest in a multi-species stand.
+
+\subsection{Entities, State Variables, and Scales}
+
+There are three types of entities included in SOEL: patches, trees, and seeds (specifically, acorns). Patches are 1 $\times$ 1 m in size and have one primary state variable: the amount of canopy cover in that patch. Canopy cover values range from 0 to 1, where 1 represents complete shade from tree canopies (no light passing through). These cover values are calculated based on the height and canopy radius of the trees in and around each patch and are used to calculate the light available to each tree. The simulated area consists of a “core” square area of patches surrounded by a buffer, both of which can be variable in size. The default is a 100 $\times$ 100 “core” grid of patches (1 ha) surrounded by a 20 patch (20 m) buffer on all sides for a total simulated area of 1.96 ha.
+
+Trees in the model are treated differently based on species, of which there are four: white oak, black oak, tulip poplar, and sugar maple. Tulip poplar was included in the model to represent oak competitors in recently harvested areas \citep{Jenk98}, while maple was included to represent oak competitors under mature forest canopies \citep{Nowa08}. Each tree species has its own set of parameters governing growth, survival, and reproduction under different environmental conditions. Oaks, the main focus of the model, are simulated in more detail than the other species. This difference manifests in two ways. First, oaks have three life stage categories: seeds (acorns), seedlings (\textless 1.4 m tall), and sapling/adult trees ($\geq$ 1.4 m tall). For sugar maple and tulip poplar, only sapling/adult trees are simulated. All life-stage categories have a state variable associated with spatial location within the simulated stand. Oak acorns have one additional primary binary state variable to indicate weevil infestation status. Oak seedlings are characterized by a single additional state variable, height. Sapling/adult trees have several additional state variables: (1) height; (2) diameter at breast height, dbh; (3) canopy radius, since canopies are assumed to be circular and flat; and (4) age. Oaks additionally have (5) fecundity in terms of average number of acorns produced per unit canopy area.
+
+\subsection{Process Overview and Scheduling}
+
+Each model time step represents one year. Broadly, four processes run in each time step in the following order: (1) tree growth (2) tree survival (3) tree reproduction (4) timber harvest, if applicable (Figure 5.1). The tree growth and survival process for sapling/adult trees (all species) is separate from the process for seedlings (oaks only; Figure 5.1); both depend on site characteristics, light availability, and species-specific parameters. Reproduction for oaks involves several sub-processes including acorn production, predation, dispersal, and germination (Figure 5.1). For the other tree species, reproduction is simpler and uses the JABOWA approach - saplings are spawned on patches directly, skipping the earlier life stages. When a timber harvest is conducted, trees are removed (equivalent to mortality) from the simulation based on a set of rules specific to the timber harvesting type; the timber harvesting options are clearcut (all trees removed every 100 years), shelterwood (all trees removed in three phases every 100 years), and single-tree selection (a subset of individual merchantable trees removed each year). The individual submodels are described in greater detail in sections 5.4-5.5.
+
+\subsection{Design Concepts}
+
+In this IBM, light availability is the most important factor driving tree survival and growth. Trees implicitly “sense” light availability in the environment based on their height, and their growth and survival is scaled accordingly. The height and spatial configuration of trees leads to a key interaction influencing light availability; trees reduce the light available to surrounding shorter competitors (i.e., shade), and in turn are shaded by taller competitors. When stem density is high, individual trees will have little light available and so growth will slow and density-dependent mortality will be high. However, there are species-specific differences in the ability to tolerate shade so some tree species (e.g., tulip poplar) are more negatively affected by shade than others (e.g. sugar maple). Thus, via changes in survival and growth, abundance and distribution of each tree species is an emergent property of its ability to grow and survive under light conditions which vary across the landscape, with time, and with timber harvesting treatment.
+
+\subsection{Initialization}
+
+The IBM is initialized by creating a mature (80-100 year old) forest stand. The basal area, dbh distribution, and species distribution of the tree species are based on pre-harvest forest survey data from the Hardwood Ecosystem Experiment (HEE), in the forest system on which the IBM is based (Saunders and Arseneault 2013). The HEE tree distribution data (which included many more tree species than I am including in SOEL) was divided into three categories: oaks, shade-tolerant species, and shade-intolerant species (Table 5.1). An equal number of black and white oak trees of appropriate size were initialized to match the HEE data for oaks. Tulip poplars and sugar maples of appropriate size were initialized to match the basal area and size distribution of the pooled shade-intolerant and shade-tolerant species groups, respectively. Overstory trees of all species are assigned locations within the simulation in a random order; the assigned location is also random, subject to the constraint that the location is \textgreater 7 m from any other mature trees (to avoid unrealistic clusters of overstory trees). Midstory and understory oaks and tulip poplars are assigned random locations subject to the constraint that the location has at least 60\% full sunlight. Maple midstory and understory trees are assigned completely random locations throughout the simulation. Before any experiments are conducted on the simulated forest, a 30 time-step (30-year) burn-in is conducted to minimize the effects of the initial conditions.
+
+\input{tables/tab5-1}
+
+\subsection{Input}
+
+All parameter values and initial conditions are controlled within the NetLogo modeling interface; no external data is read into the model. There are several primary variables set within NetLogo before each model run. First, the size of the simulated forest stand and the buffer around the stand are specified. Second, site conditions are specified (mean temperature, moisture, and soil fertility). Finally, the timber harvest treatment (clearcut, shelterwood, single-tree selection, or no harvest) is selected.
+
+\section{Early Life History Submodel}
+
+\subsection{Acorn Production}
+
+Oaks are a masting species, characterized by a boom-and-bust cycle of acorn production \citep{Lusk07}. To simulate that process, each time step in the model is assigned a different value for mean acorn production per m\textsuperscript{2} of canopy area, parameter \textit{meanAcorn} (Figure 5.1). The value of \textit{meanAcorn} is based on field data \citep{Kell14} and can be fixed or follow a set pattern (for example, a sequence of average, good, and poor mast crop years). Table 5.2 gives a set of example values for \textit{meanAcorn}. Within each year, every oak is assigned a random value for acorns produced per m\textsuperscript{2} of canopy area from an exponential distribution with mean meanAcorn. The tree then drops a number of acorns underneath its canopy equal to this random value multiplied by its canopy area. In some oak species (including white oak), acorn production declines after a peak size is reached \citep{Down44}. This decline was not incorporated into the model because I did not have field data on acorn production from very large, very old oaks and I was more interested in production by trees around the age typically observed in a managed forest (~80-120 years) when harvesting takes place.
+
+\input{tables/tab5-2}
+
+\subsection{Acorn Predation, Dispersal, and Germination}
+
+After acorns are produced, they can be predated by acorn weevils \citep{Gibs72, Gibs82} and small mammals, and may also be dispersed by small mammals \citep{Vand90, Moor07}. In the basic version of SOEL described here, these probabilities of predation and dispersal are fixed, based on empirical data from the Hardwood Ecosystem Experiment (\citealp{Kell14}, Chapter 2); however they can be varied as functions of other covariates depending on the research goal (see Chapter 6). Infestation occurs first with probability \textit{pWeevil}; weevil infested acorns are not dispersed but may be consumed by predators \citep{Steel96}. Uninfested acorns may be either left where they fell, or dispersed with probability \textit{pDispersal}. The dispersal kernel is assumed to be isotropic; distance is a random value from a Weibull distribution with scale parameter \textit{dispDist} and shape parameter fixed at 1.4. Dispersed acorns are cached with probability \textit{pCache} \citep{Steel06}. Dispersal probability and distance is based on HEE data (Chapter 2). Acorns that were not dispersed are consumed by granivores with probability \textit{pUndispEaten} and acorns that were dispersed are consumed with probability \textit{pDispEaten}.
+
+All acorns that escape predation (either dispersed or not) have a probability \textit{pGerm} of germinating into a seedling. Germination probability was obtained from the literature and was higher for acorns that have been cached and lower for acorns that are infested by weevils \citep{Haas05, Lomb09}. Acorns that do not germinate in the year in which they are produced are removed from the simulation (i.e., there is no year-to-year seed bank for acorns).
+
+\subsection{Oak Seedling Growth and Survival}
+
+The expected growth of a given seedling meanGr depends on oak species, available light (see section 5.5.1), and the presence of browse damage in a given year which occurs with fixed probability \textit{pBrowse} (Chapter 3; Figure 5.1). A predictive growth model was fit using empirical data (Chapter 4) as a function of oak species \textit{sp} (binary variable with white oak = 1), percentage of shade cover \textit{sh} (see section 5.5.1 for calculation), presence/absence of browse damage \textit{br}, a random seedling effect $rs \sim N(0, 0.25)$ and residual error $\epsilon \sim N(0, 1.19)$. I transformed the empirical growth data using the neg-log transformation \citep{Whit05}, thus yearly seedling growth was calculated as:
+\begin{equation}
+ \text{Growth} =
+ \begin{cases}
+ 1 - \exp{(-1 \times \mu)} & \text{if } \mu \leq 1 \\
+ \exp{(\mu - 1)} & \text{if } \mu > 0
+ \end{cases}
+\end{equation}
+where
+\begin{equation}
+\mu = 1.24 - 0.48 \times sh + 0.12 \times sp - 0.94 \times br + rs + \epsilon
+\end{equation}
+
+The probability of oak seedling survival meanSurv was also based on empirical data (Chapter 4). Survival probability $\psi$ for a given year is a function of oak species \textit{sp} (binary variable with white oak = 1), seedling age, and shade \textit{sh}:
+\begin{equation}
+\psi = \frac{1}{1 + e^{-\mu}}
+\end{equation}
+where
+\begin{equation}
+\mu = -0.60 + 0.10 \times sp + 0.37 \times sh + 0.58 \times age
+\end{equation}
+
+\section{Contextual Forest Submodel}
+
+\subsection{Calculate Environmental Conditions}
+
+Growth, survival, and regeneration of mature trees in the JABOWA-derived contextual forest submodel depend on two key environmental conditions: site quality and light availability. During model initialization, a species-specific metric of site quality \textit{qE} is calculated according to \citet{Botk93}. Site quality is based on four environmental characteristics, assumed constant across space and time in the simulation: (1) site climate (temperature), measured as the number of degree-days at the site; (2) soil fertility, measured as kg/ha nitrogen; (3) soil moisture (or wilt potential), measured based on a comparison between potential and actual evapotranspiration; and (4) soil water saturation measured as depth to the water table (m). The effect of each environmental variable is obtained from response functions bounded on [0,1] where 1 represents ideal conditions for tree growth and survival (e.g., Figure 5.2). Each response function is parameterized with species-specific values (Table 5.3), so ideal conditions may differ between species. The site quality \textit{fQ} for each species is calculated as the product of the values from the four response functions and is therefore also bounded on [0,1] with 1 representing ideal site quality for the species.
+
+\begin{figure}
+\centering
+\includegraphics[scale=1]{figures/fig5-2.pdf}
+\caption{White oak (\textit{Quercus alba}) growth response function curves for five environmental characteristics (1) degree-days at the site; (2) soil fertility, measured as kg/ha nitrogen; (3) wilt potential, measured based on a comparison between potential and actual evapotranspiration; (4) soil water saturation measured as depth to the water table (m); and (5) proportion of full light available. The product of the first four response functions is the overall site quality index \textit{fQ} for white oak.}
+\label{fig:5.2}
+\end{figure}
+
+\input{tables/tab5-3}
+
+Available light (ranges from 0 to 1) is calculated for each individual seedling, sapling, and adult tree. The baseline light availability above the canopy ($\phi$) is set to 1 in every patch at the beginning of each time step. Simulated trees have canopies in which all of the leaf area is compressed into a flat disk of known radius located at the top of the tree (similar to \citet{Botk93}). Beginning with the tallest tree in the simulation, available light is calculated as the mean light available in patches that overlap the tree’s canopy; by definition, for the tallest tree, this will be 1. The shade cast by the first tree is then determined. The tree canopy shades individual patches underneath it according to the Beer-Lambert law (\citep{Nowa96}). The amount of light that passes through the canopy (\textit{AL}) is a function of the tree’s leaf area index (\textit{LAI}, itself a function of dbh) and a light extinction coefficient (\textit{k} = 1/4000; \citealp{Botk93}):
+\begin{equation}
+AL = \phi \times e^{-k \times LAI}
+\end{equation}
+Shade is then calculated as 1 - \textit{AL}.
+
+If the sun’s position were fixed and directly overhead, the tree canopy would shade a circular region of patches directly underneath it at all times; this was the approach of JABOWA-3 (although the shaded area in JABOWA was the entire square cell instead of a circular region based on tree size) \citep{Botk93}. In reality, as the time of day, date, and latitude change, the angle of the canopy relative to the sun changes and therefore the exact area shaded relative to the tree shifts. This distinction is particularly important when simulating timber harvests (or other disturbance events), since a perfectly vertical shade profile underestimates the shade cast by a tree onto adjacent canopy openings. Therefore, I modified the shade profile in this simulation. Given the yearly time step, daily and seasonal variation in shading by a given tree were simplified to a single “snapshot” that approximates the pooled yearly shading influence of that tree on the surrounding area. I used a series of concentric ovals of shade on the northern side of a tree’s spatial location, the size of which depended on the tree’s canopy radius, as an approximation (Figure 5.3). The pattern of shade using this approach was similar to the pattern generated by simulations run in the software package ShadeMotion \citep{Soma13}, which calculates the shape and spatial position of shading across an entire growing season based on inputs including tree height, canopy radius, and latitude (Figure 5.3).
+
+\begin{figure}
+\centering
+\includegraphics[scale=1]{figures/fig5-3.pdf}
+\caption{Comparison of simulated shade pattern across an entire growing season in the individual-based model (a) with shade pattern generated by a tree with similar height and canopy radius in the software package ShadeMotion (b). Tree location is marked in both (a) and (b) with a black square and the darker the gray, the stronger the shading effects (i.e., the area was shaded by the tree for a higher proportion of time in the growing season). The concentric elliptical approach in (a) was chosen to maximize similarity to the ShadeMotion results while minimizing complexity to facilitate faster simulations.}
+\label{fig:5.3}
+\end{figure}
+
+After the shaded region is calculated for the tallest tree, the process repeats for the next-tallest tree, which may be partially shaded by the first tree (if it is nearby) and thus have mean available light \textless 1. Available light and thus shade in patches under the second tree is then calculated according to equation 5.5. If the tree shades an already-shaded patch (or patches), i.e., multiple tree canopies overlap above the patch, then total shade in the patch is calculated as
+\begin{equation}
+S = S_0 + (1 - S_0) \times S_N
+\end{equation}
+where $S_0$ is the existing shade in the patch and $S_N$ is the shade cast by the focal tree. This process continues for all remaining sapling and mature trees in descending order of height. Oak seedlings (\textless 1.4 m in height) do not shade each other in the model and do not have a canopy, so light available to seedlings is simply $1 - S$ where $S$ is the total shade in the patch where the seedling is located.
+
+As with the site quality, light availability is translated to an effect on growth using a response function \textit{fAL} (Figure 5.2; note it is not bounded on [0,1]). In SOEL, there are three response functions, one for each shade tolerance category \citep{Bona92}. In general, more shade tolerant species do well at low light levels, but cannot translate higher light availabilities to growth as effectively as shade-intolerant species.
+
+\subsection{Growth}
+
+In the contextual forest submodel, growth is defined as an increase in dbh and growth is calculated similar to JABOWA. Briefly, a maximum or ideal growth rate for each tree is calculated and then adjusted downward based on various factors. Maximum potential growth, $\delta D_a$, in a time step is a function of the maximum height and dbh attainable for the tree species, adjusted by a factor \textit{G} determining how early in its life the tree achieves most of its growth; increasing values of \textit{G} mean that the tree will reach one-half its maximum dbh more quickly \citep{Botk93}. Actual growth $\delta D_a$ is calculated as the maximum growth adjusted downward by site quality \textit{fQ}, available light \textit{fAL} (see section 5.5.1), and stem count \textit{ST} in a 3.5 m radius:
+\begin{equation}
+\delta D_a = \delta D_i \times fQ \times fAL \div ST
+\end{equation}
+The stem count parameter ST is not in the original JABOWA model. It was added to this IBM to simulate the negative effects of competition by saplings for finite resources on growth, thereby preventing unrealistic stem densities and basal areas (particularly following harvesting). In this way it is analogous to the basal area limit parameter included in the original JABOWA model \citep{Botk93}. Several different values for the radius of the stem count (the stem density parameter d) were tested and a value of 3.5 m was chosen because it resulted in the most realistic trajectory of basal area change following forest harvest (Figure 5.4; \citealp{Oliv81}). Low values (\textit{d} = 1,2) resulted in unrealistically high basal areas during the stem exclusion / understory reinitiation phases of forest succession while at higher values (d = 4,5), total stand basal area recovered to pre-harvest levels too slowly (Figure 5.4; \citep{Oliv81}). Though varying \textit{d} had a large effect on final stand basal area it did not have a strong effect on final stand breakdown by tree species (with the exception of \textit{d} = 1) meaning that the choice of d is unlikely to have any impact on inference about final stand composition (Figure 5.4).
+
+\begin{figure}
+\centering
+\includegraphics[scale=1]{figures/fig5-4.pdf}
+\caption{Comparison of (a) changes in basal area over time and (b) final basal areas for each tree species, for different values of the stem density control parameter \textit{d}. The value of \textit{d} used in the final model (\textit{d} = 3.5) is highlighted in bold. The dashed lines represent pre-harvest basal area in the simulated forest stand.}
+\label{fig:5.4}
+\end{figure}
+
+Once $\delta D_a$ is calculated for a given sapling or mature tree and added to the existing dbh of the tree, the other size and shape attributes of the tree (height, basal area and leaf area index LAI) are calculated from the new dbh using allometric equations as in JABOWA \citep{Botk93}. The one difference in this IBM relative to JABOWA is calculation of canopy radius. In JABOWA, the canopy size of each tree was fixed as the area of the simulated cell (typically 10 $\times$ 10 m). In SOEL, the canopy is circular and changes in size as the tree grows. Canopy radius is calculated as
+\begin{equation}
+\text{Canopy Radius} = 0.385 \times \text{tree height} \div 2
+\end{equation}
+\citet{Kene99} report a range of 0.376 - 0.393 for the first value in the allometric equation. I selected 0.385 because it resulted in simulated oaks with canopy radii that most closely matched empirical data \citep{Kell14}.
+
+\subsection{Survival}
+
+Survival is directly dependent on growth, therefore reducing the number of parameters necessary to fit the model. Survival of sapling and mature trees has two components: (1) intrinsic or ``background'' mortality, and (2) mortality related to growth \citep{Botk93}. The model assumes that 2\% of trees that reach a height of 1.4 m will eventually reach the maximum age for the species, given acceptable growing conditions. Therefore, there is a species-specific intrinsic background mortality of 4 / maximum age. The second component of mortality is dependent on growth. The model assumes that a tree has only a 1\% chance of surviving 10 consecutive years in which no diameter growth occurs \citep{Botk93}. Thus, the additional probability of mortality in a given year in which diameter growth is very small or zero (\textless 0.01 cm) can be calculated as:
+\begin{equation}
+1 - \sqrt[10]{0.01} = 0.369
+\end{equation}
+In this way, survival is tied to all the same factors (light, environmental conditions, stem density) as growth.
+
+\subsection{Reproduction}
+
+Non-oak trees (i.e., sugar maple and tulip poplar) regenerate directly as saplings. In each time step, saplings enter the simulation in a given cell with a certain probability based on site conditions and light availability \citep{Botk93}. The process is identical to regeneration in JABOWA except that SOEL has a smaller cell size (1 $\times$ 1 m instead of 10 $\times$ 10 m) and the sapling generation probability is adjusted accordingly. For shade-tolerant species (i.e., sugar maple), the probability $\zeta$ that a 1 $\times$ 1 m cell spawns a sugar maple sapling in a time step is:
+\begin{equation}
+\zeta = fQ \times fAL \times S_i \times 0.01
+\end{equation}
+In equation 5.10, \textit{fQ} and \textit{fAL} are values from the species-specific response functions for site quality and light in the patch, respectively, $S_i$ is a species-specific maximum number of saplings that can be added in a 10 $\times$ 10 m area in one time step (Table 5.3), and 0.01 scales the probability from a 10 $\times$ 10 m area to a 1 $\times$ 1m area. For shade-intolerant species (tulip poplar), the probability that a sapling spawns is identical to equation 5.10, except that when \textit{AL} \textless 0.99, $\zeta$ is fixed at 0 (so tulip poplar saplings can only spawn in high-light patches). Saplings enter the simulation at a random height between 1.4 and 1.67 m \citep{Botk93}.
+
+In JABOWA, mortality events are final and sprouting is not modeled. However, sprouting plays a key role in the successional process in eastern deciduous forests and is a key source of oak advance regeneration in particular \citep{John09}. Stump sprouting was therefore modeled explicitly in SOEL. Trees between 5 and 80 cm dbh that suffer mortality have a chance to generate a stump sprout. Probability of stump sprouting is a species-specific function of dbh and age based on logistic regression equations obtained from the literature (Figure 5.5; \citealp{True53, Wend75, Beck81, Macd83, Weig02}). Stump sprouts enter the simulation at a random height between 1.4 and 1.67 m.
+
+\begin{figure}
+\centering
+\includegraphics[scale=1]{figures/fig5-5.pdf}
+\caption{Probability of stump sprouting following harvest or mortality based on diameter at breast height (dbh) for the four tree species in the model. Probabilities were based on fitted regression models from the literature \citep{True53, Wend75, Beck81, Macd83, Weig02}. }
+\label{fig:5.5}
+\end{figure}
+
+\subsection{Timber Harvest}
+
+The individual-based structure of SOEL facilitates a straightforward incorporation of timber harvesting, including more complicated multiple-stage harvests. Harvests occur in simulation years defined by a rotation time. A set of rules specific to each harvesting scenario governs which trees will be harvested; the selected trees then die (and may resprout). The three types of timber harvest included in the model are (1) clearcutting, (2) shelterwood harvesting, and (3) single-tree selection. Harvesting rules are defined based on the harvest approach used as part of the Hardwood Ecosystem Experiment \citep{Kalb13}. Only the core simulation area (i.e., not the buffer) is harvested.
+
+Clearcutting is the simplest of the three scenarios. At the beginning of each rotation period (the default is a rotation of 100 years), all trees with a dbh greater than 1 cm are harvested and removed from the simulation (Figure 5.6). Under the shelterwood alternative, trees are harvested in three phases with a 100-year rotation. In the first phase, non-oaks in the midstory (dbh $\leq$ 25.4 cm) are removed and not allowed to resprout. In the second phase (7 years later), overstory trees are removed down to a minimum residual basal area of 16.1 m\textsuperscript{2}/ha with a preference for removing smaller, non-oak species. In the final phase (8 years after phase 2), the remaining overstory trees are harvested (Figure 5.6). The final harvesting approach is single-tree selection; every 20 years, trees of any species (dbh \textgreater 10 cm) are removed to maintain a basal area of 25 m\textsuperscript{2}/ha in the stand with a preference for larger, more merchantable trees (Figure 5.6).
+
+\begin{figure}
+\centering
+\includegraphics[scale=1]{figures/fig5-6.pdf}
+\caption{Changes in total stand basal area over time under three timber harvesting scenarios: clearcutting, 3-phase shelterwood harvest, and single-tree selection.}
+\label{fig:5.6}
+\end{figure}
+
+\section{Model Validation}
+
+Given the scope and complexity of the model described above, I wanted to ensure that (given default values) it produced ecologically realistic output. The model has two primary components: the early oak life history component that is driven by HEE data (from acorns to saplings), and the JABOWA-driven component (saplings to adult trees). I focused on the former, given that the focus of the model (and of this dissertation) is on early oak life history. Several model outputs, including acorn production and oak regeneration (i.e., seedling) density, were compared to empirical estimates from the HEE sites as well as the literature.
+
+\subsection{Acorn Production}
+
+Yearly acorn production has been reported in the literature several ways, including total acorn mass year\textsuperscript{-1} hectare\textsuperscript{-1}, total number of acorns produced year\textsuperscript{-1} hectare\textsuperscript{-1}, and mean acorns year\textsuperscript{-1} tree\textsuperscript{-1}. I compared model-derived estimates (from a ten-year 1 ha forest simulation using standard initial values; see section 5.3.5) of the latter two metrics with acorn production datasets and reported values from the literature. Mean acorns year\textsuperscript{-1} tree\textsuperscript{-1} for mature trees (dbh \textgreater 15.2 cm) from the model was 922 with a standard error of 209, which fell in the middle of the range reported in the literature (382-2147, Table 5.4). At the stand scale, a mean of 90,757 acorns year\textsuperscript{-1} hectare\textsuperscript{-1} were produced by the simulation model. This value was similar to values reported by other studies in oak-dominated forests (Table 5.4), with the important qualification that acorn production is dependent on the basal area and size distribution of oak within the stand. The main purpose of this component of the validation exercise was simply to show that the number of acorns produced in the simulation was comparable in magnitude to empirical estimates.
+
+\input{tables/tab5-4}
+
+\subsection{Oak Seedling and Sapling Density}
+
+The total number and size distribution of young oak stems is a predictor of the ability of oak to successfully regenerate following a disturbance event or timber harvest \citep{Sand72}. Stump sprouts are a key source of oak regeneration (\citealp{John09}, section 5.4.4) but seed-origin regeneration also plays a role. Given the importance of stem density for predicting oak regeneration success, I determined if the model accurately predicted the density of seedling (\textless 1.4 m) and sapling oaks ($\geq$ 1.4 m height and \textless 1.5 cm dbh; \citealp{Saun13}) following several types of timber harvest. I conducted 30 replicate simulations in SOEL and using a standard JABOWA model \citep{Botk93} and estimated the mean seedling and sapling oak density five years after three harvesting treatments (clearcut, shelterwood, and no harvest). I compared these model-derived estimates with corresponding empirical data and oak seedling and sapling densities collected from Hardwood Ecosystem Experiment sites five years following harvest (Kellner, \textit{unpublished data}).
+
+Generally, SOEL accurately predicted densities of oak seedlings and saplings (Figure 5.7). JABOWA predicted much higher oak sapling densities (particularly for the no harvest and shelterwood treatments) than were observed (Figure 5.7). The contrast in sapling density between SOEL (similar to the empirical data) and JABOWA suggests that the simplistic regeneration process used in the latter ignores the challenges facing oak early in its life history (weevil infestation, predation, herbivory) thus overestimating the number of oaks that reach the sapling stage.
+
+\begin{figure}
+\centering
+\includegraphics[scale=1]{figures/fig5-7.pdf}
+\caption{Predicted densities of oak seedlings (\textless 1.4 m height) from SOEL and saplings ($\geq$ 1.4 m and \textless 1.5 cm DBH) from SOEL and JABOWA under three harvesting scenarios, compared to actual mean densities from the Hardwood Ecosystem Experiment (HEE). Model results are means and 95\% confidence intervals from 30 replicated simulations (measured 5 years post-harvest to match HEE data). The dashed line represents the mean from the HEE field data and the shaded area is the 95\% confidence interval around the mean.}
+\label{fig:5.7}
+\end{figure}
+
+\subsection{Mature Forest Structure}
+
+As previously noted, the focus of SOEL is on early oak life history; predicting overall forest structure was a secondary concern. However, I still wanted to test the ability of the model to simulate forests (following a stand initiation / disturbance event) that approximated the structure of the forest stands in the Hardwood Ecosystem Experiment. For both SOEL and JABOWA, I conducted 30 replicate simulations that began with a stand initiation event (a clearcut harvest). The trajectory over time of several key metrics of forest structure (basal area, stem density, and quadratic mean diameter) are plotted in Figure 5.8 and compared with the range of forest structure data collected at the HEE.
+
+\begin{figure}
+\centering
+\includegraphics[scale=1]{figures/fig5-8.pdf}
+\caption{Comparison of three metrics of forest structure (basal area, stem density, and quadratic mean diameter) between forest stands simulated JABOWA and SOEL (means of 30 replicates) and empirical data gathered at the HEE sites. Comparisons are for all trees (dbh \textgreater 1.5 cm) and overstory trees only (dbh \textgreater 30 cm). HEE data represents the range of values reported in 9 forest stands approximately 80-100 years old. Gray lines represent one standard deviation above and below the simulation mean.}
+\label{fig:5.8}
+\end{figure}
+
+SOEL simulated forests 80-100 years post-disturbance generally had similar structure to the HEE stands, particularly when all stems \textgreater 1.5 cm dbh were considered. When only overstory trees (dbh \textgreater 30 cm) were measured, the model slightly overestimated density with a corresponding underestimate of quadratic mean diameter (Figure 5.8). Still, SOEL more accurately approximated forest structure than did JABOWA, particularly for stand basal area. A complicating factor is that the HEE forest stands used for this comparison have been subjected to some single-tree selection harvesting since stand initiation \citep{Carm13}, likely reducing overall stand basal area, while the simulated stands were not harvested following the initial disturbance event.
+
+\section{Conclusion}
+
+The pathway to establishment for an oak seedling is complex, involving highly variable acorn crops \citep{Lusk07, Kell14}, predation by insects and mammals \citep{Bell05, Lomb08, Kell14}, dispersal by small mammals (\citealp{Moor07}, Chapter 2), herbivory (Chapter 3), and competition and other environmental effects (Chapter 4). Oak regeneration success depends on the interaction of these and other factors (e.g., stump sprouting probability, site quality; \citealp{John09}). No modeling structure I am aware of has brought together all of these components into a single detailed, comprehensive model of early oak life history. To tackle this complex modeling challenge, I harnessed the power of individual-based models \citep{Grim05}, which allowed me to track oaks (from acorns to mature trees) through time, space, and varying environmental conditions. As with all models, I had to make numerous simplifying assumptions. However, based on preliminary validation tests, SOEL was able to realistically simulate key metrics including acorn production, seedling and sapling densities, and forest structure with a close correspondence to empirical data collected at the Hardwood Ecosystem Experiment sites (Table 5.4, Figures 5.7 - 5.8). The simulations performed in this chapter were simplistic, holding many key parameters constant at mean values (e.g. dispersal kernels, weevil infestation probability) with the goal of testing general model behavior. In the next chapter, I will manipulate these and other parameters to determine how they affect key metrics of oak regeneration success. \ No newline at end of file
diff --git a/chapter6.tex b/chapter6.tex
new file mode 100644
index 0000000..0b1cec4
--- /dev/null
+++ b/chapter6.tex
@@ -0,0 +1,155 @@
+\chapter{{INTERACTIONS OF TIMBER HARVEST WITH EARLY OAK LIFE HISTORY: IMPLICATIONS FOR OAK REGENERATION VIA APPLICATION OF A NOVEL INDIVIDUAL-BASED MODEL, SOEL}}
+
+\section{Introduction}
+
+Early tree life history and demography is driven by a combination of intrinsic- and extrinsically-controlled interactions with the environment. Trees exert a degree of direct control over intrinsic interactions by virtue of ``decisions'' on e.g., resource allocation, whereas for extrinsic processes the outcome is primarily controlled by other biotic or abiotic components of the environment. The relative importance of intrinsic or extrinsic factors falls on a continuum varying with time, space, and the ecological process in question. For example, seed production is primarily an intrinsic process in which trees choose when to devote resources to reproduction, but weather can also influence seed crops \citep{Koen00}. In contrast, seed predation is primarily an extrinsic, biotic interaction driven by predators, but seed traits (as determined by the parent tree) can play a role in the predation and dispersal process \citep{Janz71}. Seedling growth and survival are also typified by substantial impacts of extrinsic, biotic factors including herbivory \citep{Hanl98, Bigg99} but also abiotic environmental conditions including light and drought \citep{Fote01, Vall08, Suar08}. Together, these early life history processes drive the abundance and composition of tree seedlings in the forest understory and influence the future successional path of the forest \citep{Oliv96}. Tree early life history is therefore important for understanding forest ecology and informing forest management to shape successional trajectories. Thus, understanding the interactions between management and the intrinsic and extrinsic factors that influence early tree life history is an important research goal \citep{Guar98}.
+
+Oaks (\textit{Quercus}) are a frequent subject of both forest ecology research and forest management efforts \citep{Lars98, John09}. They are a dominant component of the canopy in many eastern hardwood forests, a key food source for numerous species of mammals and birds, and a valuable timber species \citep{Mcsh07, John09}. The decline of oak throughout eastern forests over the past century, likely a consequence of changing patterns of forest disturbance \citep{Bros01}, has prompted numerous studies on oak ecology and management for oak regeneration \citep{Lars98, John09}. Successful oak management requires the presence of advance regeneration; that is, young oak stems already in the understory, originating from seed or sprout, ready to capture the increased light and newly created growing space \citep{Lars98}.
+
+Seed-origin regeneration in early oak life history involves (1) acorn production, (2) acorn predation and dispersal, and (3) seedling growth and survival. To the detriment of managers, these processes are frequently unpredictable, correlated with each other, and/or altered by disturbances associated with forest management. Acorn production, for example, is highly variable year-to-year and highly spatially autocorrelated within a given oak species \citep{Koen02}. The overall effect of harvest on acorn predation is unclear, but any effect is likely to be small compared to the yearly variation \citep{Heal99, Bell05, Lomb08, Kell14, Olso15b}. The cycle of acorn production induces correlated cyclical patterns in acorn predators including acorn weevils (\textit{Curculio} and \textit{Conotrachelus}) and granivorous small mammals; rates of infestation and predation are typically maximized when acorns are scarce \citep{Craw95, Wolf96, Kell14}. Direct management of acorn predators at a large scale is not typically realistic. However, forest management can have indirect impacts on predator populations - in one study, midstory removal as part of a shelterwood harvest decreased acorn infestation by weevils (\citealp{Kell14}, \textit{but see} \citealp{Bell05, Lomb08}), and many studies have reported changes in the abundance and composition of granivorous small mammal populations following timber harvest (e.g. \citealp{Kirk90, Kell13}). In addition to numerical effects, changes in habitat structure following harvest may change the behavior of acorn dispersers with consequences for ultimate acorn fate (Chapter 2).
+
+Seedling growth and survival are driven by their own suite of extrinsic factors. Available light - usually a direct effect of forest management - is among the most important predictors of seedling growth given oak’s relative intolerance to shade \citep{Dey08}. However, an increase in available light following harvest can also promote faster-growing oak competitors (e.g. tulip poplar, \textit{Liriodendron tulipifera}) and thus have an indirect negative impact on oak \citep{Jenk98, Beck03}. Additionally, under drought conditions, seedlings in high-light environments may suffer greater physiological consequences than seedlings in shade \citep{vanh97, Quer06}. Biotic factors, namely herbivory by insects and mammals on oak seedlings, can limit establishment \citep{Kitt95, Hors03, Abra12}. Forest management can alter the abundance and habitat use of herbivores and thus impact the level of herbivory on oak seedlings \citep{Mein02, Mill09, Crim10}.
+
+Variable success in management for oak may be partly driven by early life history processes \citep{Marq76, Lars98}, motivating substantial body of research on early oak life history and its interaction with management. Previous field studies have generally focused on a single process (e.g., seed production) or a single oak life history stage (e.g., seedling) (e.g. \citealp{John84, Buck98, Bell05, Lomb08, Kell14, Olso15b}). A narrow focus is helpful in understanding the process being studied, but it precludes direct assessment of the relative importance of different component processes (e.g. seed production, predation, herbivory) on advance oak reproduction under different management regimes. Demographic modeling permits simultaneous examination of multiple processes and life stages. Approaches for trees include matrix models \citep{Lian13}, integral projection models \citep{Zuid10}, and individual-based models \citep{Bugm01a}; several oak-specific models including ACORn \citep{Dey91} and SIMSEED \citep{Roge98} have also been developed.
+
+To simultaneously examine the intrinsic and extrinsic components of early oak life history, I developed an individual-based demographic model (hereafter SOEL, Simulation of Oak Early Life history) for black (\textit{Q. velutina}) and white (\textit{Q. alba}) oak that included acorn production, acorn predation and dispersal, and seedling growth and survival components. Importantly, oaks in the model were placed in the context of a spatially explicit simulated forest, including competitor species and spatially variable light availability. The modeling framework allowed me to run experiments via simulation that would be difficult or impossible to conduct in the field and track the impact of changes in individual life history parameter values on the density and size distribution of oak regeneration. I performed global sensitivity analysis on the model to determine the relative importance of early life history parameters (Table 6.1) on oak demography. I then applied SOEL to three case studies involving the interaction of early oak life history and timber harvest, based on findings from previous field work (Chapters 2-4). First, I examined how variable acorn production interacts with timing of timber harvest. Second, I examined the effect of shelterwood harvest-induced differences in seed predation. Finally, I examined the consequences of interactions between drought, seedling growth and survival, and timber harvest.
+
+\input{tables/tab6-1}
+
+\section{Methods}
+
+\subsection{Study Location}
+
+Empirical data to parameterize SOEL was collected at Morgan-Monroe and Yellowwood State Forests in southern Indiana, USA. Together the state forests cover \textgreater 18,000 ha of mainly upland areas with steep slopes \citep{Jenk98}. The forest canopy is dominated by hardwoods, primarily oaks (\textit{Q. alba}, \textit{Q. velutina}, \textit{Q. prinus} and \textit{Q. rubra}) and hickories (\textit{Carya}) and the understory is mainly shade tolerant maple (\textit{Acer} spp.) and beech (\textit{Fagus grandifolia}) \citep{Saun13}. The primary acorn predators are small mammals, including white footed mice (\textit{Peromyscus leucopus}), eastern chipmunks (\textit{Tamias striatus}) and gray squirrels (\textit{Sciurus carolinensis}), and the primary herbivores are white-tailed deer (\textit{Odocoileus virginianus}) and eastern cottontail rabbits (\textit{Sylvilagus floridanus}). Data collection was centered on long-term monitoring sites established as part of the Hardwood Ecosystem Experiment (HEE), a long-term study of the effects of silviculture on forest ecosystems \citep{Kalb13}. The HEE was established in 2006 and comprises numerous studies monitoring forest structure, acorn production and removal, and mammal, bird, reptile, and amphibian populations \citep{Kalb13}.
+
+\subsection{Demographic Data}
+
+Data on acorn production, acorn predation and dispersal, and seedling survival and growth for black and white oak were obtained from multiple studies conducted at the HEE between 2006 and 2014. All studies were replicated across multiple sites and timber harvest treatments (no harvest, clearcut harvest, and the first, midstory-removal phase of a 3-phase shelterwood harvest) which were implemented in 2008-2009 \citep{Kalb13}. Hereafter, the initial shelterwood harvest will be referred to as ‘midstory removal’.
+
+Yearly acorn production data (in acorns per m\textsuperscript{2} of canopy) was collected from 56 white and 57 black oaks over the 9-year period (64 in unharvested areas, 31 on the edge of harvest openings, and 18 in midstory removal harvests). Collected acorns were X-rayed to determine weevil infestation status and calculate a yearly infestation probability \citep{Kell14}. Under each study tree, fallen acorns were marked in the fall and monitored for removal by seed predators to calculate a dispersal probability \citep{Kell14}. For a subset of 4 years (2011-2014) acorns at 10 trees (5 each in unharvested and midstory removal harvest sites) were tagged and then relocated using metal detectors the following spring to determine acorn caching and survival probabilities (Chapter 2).
+
+To obtain data on oak seedling survival and growth, 39 black and white oak seedlings (24 first-year seedlings and 15 year-old seedlings) were experimentally planted in each of 27 plots located within unharvested control (n=6), clearcut (n=18), and midstory removal harvest sites (n=3) in May 2011 (Chapters 3-4). Seedlings were re-visited twice yearly through 2014 to assess survival, measure growth and light availability, and identify herbivore damage and insect defoliation.
+
+\subsection{Modeling Approach}
+
+I modeled oak demography with SOEL using a multiple-step process \citep{Caug15}. First, I developed individual regression models for early oak life history parameters using empirical data and generalized linear mixed modeling techniques, and organized these parameters into an early life history submodel tracking oak from acorn production to seedlings (defined as \textless 1.4 m height). Second, I developed a contextual forest submodel adapted from the established JABOWA forest modeling framework to simulate vital rates of saplings ($\geq$ 1.4 m height, \textless 1.5 cm d.b.h.) and mature trees ($\geq$ 1.5 cm d.b.h) \citep{Botk93}. Finally, I linked the two submodels to form comprehensive forest model focused on oak.
+
+The early oak life history submodel was governed by 11 key oak early life history parameters, of which 8 were related to the acorn life stage and the remaining 3 to the seedling life stage (Table 6.1, Figure 6.1). For the 10 parameters for which I had empirical data, I fit unique generalized linear mixed models in R \citep{Rcor15}. For each regression, I defined a saturated model of candidate predictor variables based on prior analyses of the empirical data (Chapters 2-4) and the research goals of the study. For acorn-related parameters, oak species, midstory removal harvest effect, yearly ambient mast availability, and yearly random effects were included in the candidate set at a minimum; for seedling-level parameters, species, age, height and effect of light availability were included. Ultimately, only significant predictor variables were retained in the final model for each early life history parameter (Table 6.2).
+
+\begin{figure}
+\centering
+\includegraphics[scale=0.8]{figures/fig6-1.pdf}
+\caption{Flowchart of the early life history submodel of SOEL, including key oak life stages (white squares), parameters (gray ovals), and mortality events (black squares). For probability parameters, black arrows indicate the result of a “success” and gray arrows the result of a ``failure''. Dashed arrows indicate the presence of parameter effects on other parameters (e.g., weevil infestation affects germination probability; Table 6.2). Parameter descriptions are found in Table 6.1.}
+\label{fig:6.1}
+\end{figure}
+
+\input{tables/tab6-2}
+
+The contextual forest submodel simulated the spatially explicit forest structure contextualizing the early oak life history submodel, comprising trees $\geq$ 1.4 m in height. Four tree species were included: (1) black and white oaks (which have passed through the early life history submodel and reached $\geq$ 1.4 m height); (2) a shade-tolerant oak competitor (sugar maple, \textit{Acer saccharum}); and (3) a shade-intolerant oak competitor (tulip poplar, \textit{Liriodendron tulipifera}). Since I did not have empirical data for saplings or mature trees, I used an established modeling framework (JABOWA) to simulate their yearly growth and survival due to its ease of parameterization from the literature \citep{Botk93, Holm13}. Briefly, JABOWA simulates tree diameter growth as a species-specific function of current tree dbh, height, light availability, a site quality index and stand basal area \citep{Botk93}. Trees shade nearby shorter trees, and light availability at any given spatial location and height (a necessary input for the early life history submodel) is calculated based on the depth of the canopy above using the Beer-Lambert absorbance law \citep{Botk93}. Yearly tree survival is a function of tree age relative to maximum age, with a penalty to survival probability if diameter growth is below a threshold level of 0.01 cm \citep{Botk93}. For the two non-oak species, JABOWA was also used to simulate regeneration; in JABOWA new saplings of a given species are added to the simulation with a probability based on site quality and light availability \citep{Botk93}.
+
+JABOWA is only partially spatially explicit. Individual trees are located indistinctly within 10x10 m cells, and the canopies of all trees in a cell are assumed to fill the entire cell area. To match the more spatially explicit early life history submodel, I modified several aspects of JABOWA. First, all trees were given exact spatial coordinates. Second, rather than all trees having 10 $\times$ 10 m square canopies, canopy dimensions (and thus patterns of shade cast) were circular and related to tree height based on allometric equations. Third, the negative effect of stand basal area on individual tree diameter growth was replaced with a negative effect of nearby stem density (in a 3.5 m radius) which facilitated a more realistic forest structure and accumulation of basal area. Finally, trees were allowed to sprout directly into new saplings following mortality events based on their diameter, using species-specific predictive sprouting models from the literature \citep{True53, Wend75, Beck81, Macd83, Weig02}. For details and justification for the modifications made to JABOWA and the parameter values used, see Chapter 5.
+
+The two component submodels were linked together in NetLogo \citep{Wile99} to form the complete forest model. NetLogo was called from within R using RNetLogo \citep{Thie14}. Complete code for SOEL and the R script used to control simulations can be found in Appendix C. Simulated forests were 140 $\times$ 140 m in size (a 2 ha area with 20 m buffers) and initialized with saplings and mature trees to match the size and species distribution of forest structure data collected at the HEE \citep{Saun13}. In each annual model time step, acorn production, acorn dispersal and survival, seedling growth and survival, and sapling/mature tree growth and survival were simulated stochastically using parameter values generated by the set of predictive regression equations (Table 6.2) and the modified JABOWA equations. Multiple output parameters related to oak regeneration were recorded following each time step: the percent of produced acorns that ultimately emerged as seedlings, the total number of new seedlings that entered the simulation in the time step and the overall density of oak seedlings (\textless 1.4 m height), and finally the density of oak saplings (defined as $\geq$ 1.4 m height and \textless 1.5 cm d.b.h.).
+
+\subsection{Global Sensitivity Analysis}
+
+I conducted a global sensitivity analysis for the subset of early oak life history parameters in SOEL for which I had empirical data. In a global sensitivity analysis, all parameters of interest are simultaneously perturbed within a given distribution, and the impact of this variability on model outputs is determined \citep{Caug15}. The input distribution for each parameter was normal, with a mean equal to the estimate of the regression model intercept for the parameter; the standard deviation was equal to the standard deviation around the estimate, representing the level of uncertainty in the parameter mean. The analysis was performed using a Monte Carlo approach: 500 sets of input parameters were drawn from the appropriate distributions in a Latin hypercube design \citep{Megr01, Xu08}. For each set of input parameter values, the forest simulation was run for 37 years: a 30 year burn-in period followed by a 7 year period in which output metrics were recorded. Output metrics were the mean percent acorn emergence over the final 7 years of the simulation, the total new seedlings entering the simulation over the final 7 years, and the density of saplings in year 37. Sensitivities were calculated for each output metric described by separating the uncorrelated component of sensitivity for a parameter with the component correlated with the other input parameters \citep{Xu08}.
+
+\subsection{Interaction of Timing of Harvest with Acorn Production}
+
+To assess the impact of timing of harvesting events to coincide with high acorn production, I defined two variables: the definition of ``high'' acorn production (either 1 or 2 standard deviations above the mean value, as defined by the fitted predictive equation for mast production from 6.2.3) and the number of years of ``high'' acorn production immediately prior to harvest implementation (either 1 or 2). Crossing these two variables led to four total pre-harvest acorn production scenarios. In an additional ``average'' scenario, and in all years in the preceding four scenarios in which acorn production was not defined as high, acorn production was allowed to vary randomly. I then crossed the five total scenarios with three harvest treatments applied to the simulated forest: no harvest, a midstory-removal harvest, and a clearcut harvest. The midstory removal harvest removed non-oaks \textless 25.4 cm d.b.h. down to a minimum residual basal area of 13.8 m\textsuperscript{2}/ha and the clearcut removed all stems \textgreater 1 cm dbh in a single harvest event \citep{Kalb13}. Each scenario $\times$ harvest treatment combination was replicated in a simulated forest (initialized as described in section 6.2.3) 36 times, with each replicate consisting of a 30-year burn-in period, followed by application of harvest (if any) and associated increases in acorn production, followed by a 7-year data collection period. The 7-year data collection period represents the period after the midstory removal but before the second shelterwood phase (typically 5-10 years later, \citealp{Kalb13}). We did not have empirical data for shelterwood harvest effects after the second harvest so we limited the data collection period to 7 years.
+
+\subsection{Interaction of Seed Predation and Midstory Removal}
+
+By building predictive equations for a suite of parameters impacting the acorn life stage (Figure 6.1) and combining them into a comprehensive individual-based model, I was able to quantify the combined impacts of multiple parameters on metrics of oak regeneration. For parameters affecting the acorn life stage (Figure 6.1), I focused on two important sources of variability: harvest effects and yearly-varying effects. More specifically, the former was the impact of midstory removal on acorn-level parameters (given that the acorn was located inside a harvested area), and the latter included both impacts of variable acorn abundance and other random yearly effects. I manipulated the entire set of predictive equations for acorn-related life history parameters to include a harvest effect (given that the acorn was inside a harvest; hereafter treatment scenario, TE), yearly-varying effects (hereafter yearly scenario, YE), or both (TE + YE) in a given simulation (Table 6.2). These three scenarios were then crossed with two harvest treatments: no harvest and a midstory removal harvest (as described in section 6.2.5). Yearly effects (YE) on oak parameters were present in both harvest treatments, but treatment effects (TE) on parameters could only apply if the acorn was actually located in a simulated post-harvest forest. Each predation scenario $\times$ harvest treatment combination was replicated in 36 simulated forests as in section 6.2.5.
+
+\subsection{Effects of Drought on Seedling Growth and Survival}
+
+The empirical data used to build predictive equations for seedling growth and survival (section 6.2.3) included two drought years and two non-drought years. The two drought years were defined by very low precipitation during the growing season (total July precipitation was 1.63 and 1.42 cm, respectively, compared to a 30 year average of 10.9 cm) \citep{Nati14}. The entire study area reached at least ``severe drought'' (category D2) status for at least part of the growing season in both years, according to the U.S Drought Monitor \citep{Nati15}. To simulate the impact of a drought year on seedling growth and survival, I fit separate regression models for seedling growth and survival to the data for the drought and non-drought period. The models were otherwise identical to the full 4-year model. Then, I defined a probability parameter within the simulation that represented the chance that a given simulation year was a drought year or not, applying the appropriate predictive equations for seedling growth and survival. I defined six drought scenarios, one for each of 6 values for drought probability (ranging from 0 to 1 in increments of 0.2). The six scenarios were crossed with three harvest scenarios: no harvest, midstory removal harvest, a clearcut harvest (both harvests as described in section 6.2.5). As with previous case studies, each drought scenario x harvest treatment combination was replicated 36 times, with each replicate including a 30-year burn-in period, followed by the harvest (if any), followed by a 7-year data collection period.
+
+\subsection{Analysis of Model Output}
+
+For the acorn production and seed predation case studies described above, I examined the interactive effects of harvest treatment and scenario on each of the three metrics of oak regeneration using analysis of variance in R \citep{Rcor15}. Post-hoc differences between means were assessed with Tukey’s honest significance test. For the drought case study, the variable scenario was continuous (drought probability). Thus, I examined the interaction of harvest and drought probability on oak regeneration metrics using simple linear regression. For all analyses, alpha was set at 0.05.
+
+\section{Results}
+
+\subsection{Sensitivity Analysis}
+
+The global sensitivity analysis identified several parameters that had the large influence on oak regeneration metrics. Proportion of emerged acorns and total number of new seedlings over the 7-year data collection period were most sensitive to variability in acorn caching probability (\textit{pCache}) (Table 6.3). Sapling density was sensitive to caching probability, mean seedling growth (\textit{meanGr}), and to a lesser extent, probability of seedling survival (\textit{pSurv}) (Table 6.3). Both seedling and sapling densities were sensitive to acorn production (\textit{meanAcorn}), but only a small amount of this sensitivity was unique and uncorrelated with other parameters. A high level of correlation existed between acorn production and other acorn-level parameters, a relationship that was also reflected in the fitted models for several acorn-level parameters (Table 6.2).
+
+\input{tables/tab6-3}
+
+\subsection{Interaction of Timing of Harvest with Acorn Production}
+
+Seedling density prior to harvest was affected significantly by acorn production scenario but not harvest type (Table 6.4). For sapling density 7 years post-harvest, harvest and scenario main effects and their interaction were significant (Table 6.4). Seedling and sapling density increased with number of good mast years prior to harvest and with the strength of a good year, and the trend was stronger for seedling density (Figure 6.2). The most extreme scenario (2 good acorn production years in the stand prior to harvest, each 2 standard deviations above the mean; 2YR2SD) significantly increased final sapling density in clearcut harvests by an average of 25.3\% and midstory removal by an average of 19.0\% relative to the average scenario where acorn production varied randomly.
+
+\input{tables/tab6-4}
+
+\begin{figure}
+\centering
+\includegraphics[scale=1]{figures/fig6-2.pdf}
+\caption{Boxplots of model-simulated output values for oak seedling (a, \textless 1.4 m height) and sapling (b, $\geq$ 1.4 m height and \textless 1.5 cm d.b.h.) densities under five different scenarios for mast production immediatley prior to harvest and three different timber harvest treatments. Mast production scenarios were compared to an “average” scenario (Avg; random production prior to harvest) and were defined by crossing 2 variables: the magnitude of “good” acorn production (1 or 2 standard deviations above the average crop; 1SD or 2SD respectively) and the number of years prior to harvest of ``good'' acorn production (1 year or 2 years; 1YR or 2YR respectively). Within each subfigure (a) – (b), different letters represent significantly different means based on post-hoc Tukey HSD tests. }
+\label{fig:6.2}
+\end{figure}
+
+\subsection{Interaction of Seed Predation and Midstory Removal}
+
+Independent of seed predation scenario, there were several trends in simulation output. Total acorn production over the 7-year post harvest period was highly variable but similar across scenarios (Table 6.5, Figure 6.3a). The altered light conditions following midstory removal had minimal impact on acorn production, percent emergence, and total new seedlings in the 7 year period post-harvest (relative to the no-harvest treatment), but increased sapling density by an average of 23.5\% (Table 6.5, Figures 6.3b-d).
+
+\input{tables/tab6-5}
+
+\begin{figure}
+\centering
+\includegraphics[scale=1]{figures/fig6-3.pdf}
+\caption{Boxplots of model-simulated output values for four metrics (a) – (d) related to oak regeneration (see text for description), under two harvesting treatments (no harvest and midstory removal harvest) and four seed predation model scenarios: (1) a constant control C; (2) harvest treatment effects on seed predation TE; (3) variable yearly effects on seed predation YE; and (4) treatment and yearly effects together TE + YE. Within each subfigure (a) – (d), different letters represent significantly different means based on post-hoc Tukey HSD tests.}
+\label{fig:6.3}
+\end{figure}
+
+Inclusion of yearly-varying effects only (scenario YE; comprising both mast availability effects and random yearly variation) on seed predation parameters in the simulation model increased mean percent emergence by 28.6\% (\textit{p} \textless 0.001), the 7-year sum of new seedlings by 30.4\% (\textit{p} \textless 0.001), and the total density of saplings 7 years after the initial harvest by 33.2\% (\textit{p} \textless 0.001, Figure 6.3b-d) in the midstory removal simulations relative to the control scenario (C). Along with an increase in the means of these metrics there was also an increase in year-to-year variability (Figure 6.3b-d). The unharvested control simulations showed similar relative increases in regeneration metrics and associated variability when yearly effects were included (Figure 6.3b-d). Inclusion of treatment effects on seed predators given that the yearly effects were included (i.e., comparing YE to TE + YE) did not significantly decrease percent emergence (\textit{p} = 0.14) or total new seedlings (\textit{p} = 0.38) and as before had no effect on sapling density (\textit{p} = 0.87; Figure 6.3b-d).
+
+\subsection{Effects of Drought on Seedling Growth and Survival}
+
+There were significant main effects of both harvest treatment and yearly drought probability on the 7-year sum of new seedlings and the final density of saplings 7 years post-harvest (Table 6.6, Figure 6.4). There were significantly fewer seedlings present post-harvest in the clearcut simulations relative to the other two treatments, and both the midstory removal and clearcut treatments had significantly higher numbers of oak saplings 7 years post-harvest. Notably, the effect size of the clearcut harvest on sapling density was an order of magnitude higher than the effect of midstory removal (Table 6.6, Figure 6.4b). As yearly drought probability increased, both the seedling and sapling metrics decreased significantly; each increase of 0.1 in the probability of drought decreased the expected 7-year sum of new seedlings by 895 per ha (a reduction of 4.4\%) and the expected final density of saplings by 65 per ha (11.5\%; Table 6.6). This negative relationship was not as strong for the seedling metric in clearcuts (significant, positive interaction term between drought and clearcut harvest; Table 6.6) relative to the other two harvest types. However, increasing drought probability had a larger negative effect on sapling density in both clearcut and midstory removal simulations relative to no harvest (Table 6.6).
+
+\input{tables/tab6-6}
+
+\begin{figure}
+\centering
+\includegraphics[scale=1]{figures/fig6-4.pdf}
+\caption{Boxplots of model-simulated output values for two metrics under a range of yearly drought probabilities and three different timber harvest treatments. Total new seedlings (a) is the sum total of all newly recruited seedlings (height \textless 1.4 m) per ha over a 7-year period; and sapling density (b) is total seed-origin saplings (height $\geq$ 1.4 m and dbh \textless 1.5 cm) at the conclusion of the 7 year period. Seedling growth and survival were governed by different regression models in drought and non-drought years. The ``Avg'' treatment on the far right of each subfigure represents output for models combined across drought and non-drought years.}
+\label{fig:6.4}
+\end{figure}
+
+\section{Discussion}
+
+The process of forest succession depends on the presence and composition of advance regeneration in the forest stand, which in turn depends on biotic and abiotic filters influencing the pathway from seed to seedling \citep{Beck03, Zamo14}. I have synthesized field data into SOEL, a comprehensive and flexible individual-based model of the early oak life cycle (Figure 6.1) to quantify how key filters including seed predation, seed dispersal, and drought interact with silvicultural disturbance to influence the regeneration of oak, a keystone species group in eastern U.S. deciduous forests \citep{Fral04}. Three individual case studies demonstrate the ability of the model to connect intrinsic and extrinsic processes driving early oak life history with their ultimate consequences for oak regeneration.
+
+\subsection{Interaction of Harvest Timing with Acorn Production}
+
+Oak regeneration metrics were sensitive to variability in the total number of acorns produced year-to-year (Table 6.3). High variability among years in seed production is characteristic of the oaks \citep{Koen02}, and years of high acorn production are often followed by a high density of oak seedlings \citep{Loft90, Craw95}. Previous work on the efficacy of forest management practices for oak has demonstrated benefits from timing harvest to coincide with these periods of high seedling density, to maximize the number of stems that have a chance to compete following release \citep{Bros11}. This is a difficult task, given the challenge of predicting acorn production \citep{Sork93a} and logistical constraints on when harvests are implemented. SOEL output supported the conclusion that timing harvest to follow large acorn crops can measurably increase the density of oak regeneration in the short term following harvest, at least under some conditions (Figure 6.2). The impact of this acorn production ``pulse'' was strongly visible in the seedling density metric just before harvest for all scenarios; for the most extreme masting scenario seedling density was 132\% higher than the control (Figure 6.2a). The difference was not nearly as strong for sapling-sized oaks 7 years post-harvest, with the most extreme masting scenario yielding an average increase in sapling density of just 19\% across all harvest treatments (Figure 6.2b). The difference is attributable to the increasing influence of other factors on oak seedlings over time (e.g. light availability, herbivory, and eventually density-dependent mortality among saplings).
+
+Targeting forest management for oak to take advantage of multiple consecutive large acorn crops may not always be realistic. In the 9-year mast production dataset that parameterized SOEL, mean white and black oak acorn production (per m\textsuperscript{2} canopy) exceeded one standard deviation above the mean just twice, and only approached two standard deviations above the mean once. Additionally, in the model, there were only two oak species, and they produced synchronous large acorn crops. In reality, different oak species (especially oaks in different sections) do not necessarily have matching year-to-year masting schedules \citep{Koen00, Gree02, Lusk07}. Managers may not have the flexibility to wait multiple years for maximal acorn production to implement harvest, especially given the fairly modest increases in sapling density even in a perfect scenario (Figure 6.2). Avoiding bad mast years may be a more realistic criteria for timing harvest.
+
+\subsection{Interaction of Seed Predation and Midstory Removal}
+
+At the acorn life stage, several parameters were altered following midstory removal harvest (Table 6.2), though the effect (in terms of consequences for acorns) was not in a consistent direction; for example, midstory removal decreased probability of weevil infestation but increased probability of consumption by granivores (Chapter 2, \citealp{Kell14}). An exploration of the reasons for these harvest effects on individual acorn-related parameters is beyond the scope of this study. The more relevant, and arguably more important question (from a management perspective) is whether or not these harvest effects, pooled across the entire early oak life history, have a measurable effect.
+
+The answer depends on the focal oak life stage. In the absence of yearly variability, the cumulative impact of the midstory removal, across all acorn parameters, was a small but significant reduction in the probability a given acorn established into a seedling and the total number of new seedlings that accumulated across the 7-year post-harvest period (Figure 6.3b-c). The difference became negligible at the sapling life stage (Figure 6.3d). Thus, my results suggest that acorn predator-midstory removal harvest interactions have a measurable effect on the oak regeneration process initially, but these effects decline over time and thus likely do not drive the dominance of oak in the future canopy. The relatively weak overall impact of harvest is not surprising given the modest level of disturbance to the forest stand, primarily just a partial removal of the midstory \citep{Kalb13}. This research only examined the effects of harvest on predation through the initial phase of a three-phase shelterwood. Additional work is needed to determine if the more intensive disturbance in the later phases of the shelterwood harvest (i.e., greater amounts of basal area removal) have stronger effects on acorn-level life stage parameters of oak.
+
+The impact of including year-to-year variation in acorn-level parameters was an increase in the mean value for all three metrics of oak regeneration by roughly a third, with a commensurate increase in variability. The magnitude of these changes, relative to scenarios in which year-to-year variation was absent, were such that they masked the much smaller effects of harvest on oak regeneration (Figure 6.3). A primary driver of year-to-year variability in parameter values was the mean acorn input into the model system, which had an overall negative correlation with several metrics of acorn predation (Table 6.2). While predator populations were not modeled explicitly, this negative relationship implicitly includes the impact of predator satiation \citep{Janz71, Craw95, Kell14}. In years with large acorn crops, overall seed predation was reduced and more acorns escaped to germinate into seedlings (Figure 6.3b).
+
+\subsection{Effects of Drought on Seedling Growth and Survival}
+
+Previous field work (Chapter 4) demonstrated differences in seedling growth and survival in drought years. Specifically, drought changed the predictive relationship between shade and seedling survival and growth. In drought years, seedlings in high-light environments had lower survival and growth, presumably because they were under greater water stress \citep{Mart87, Mcca94, Mein02}. It would be difficult, if not impossible, to implement a study that experimentally varied the frequency of drought at a large spatial scale to determine the consequences for oak regeneration; SOEL provides a means of doing so via simulation. Given the decrease in survival and growth in drought years for a given level of light, the ultimate consequence of increasing yearly drought probability was a large reduction in both the seedling and sapling regeneration metrics to the point where essentially no seedlings reached the sapling size category (\textgreater 1.4 m height) when there was a 100\% probability of drought (Figure 6.4). The impact of drought was greater for black oak relative to white oak (which reflects the field data on which the model is based); when drought year probability was 0, 42\% the final density of oak saplings were black oaks whereas when drought probability was 0.8, only 27\% were black oak. Thus when drought is frequent, high-light environments represent a tough trade-off for oak seedlings between growth and survival, and there are species-specific differences in overall response even with the Quercus genus.
+
+While there are clearly negative consequences of drought for oak regeneration, the impact of drought is likely to be even greater for competing tree species that do not have oak’s adaptations to xeric environments \citep{John09}). Drought could therefore facilitate competitive success of oak, particularly in harvest openings where competition from pioneer tree species is fierce \citep{Morr08}. In the drought case study, drought impacted only oak seedlings and saplings and not the competing maple and tulip poplar saplings. A logical extension of this simulation would be to use field data to parameterize the impacts of drought on competing saplings as well, thus yielding a more realistic composition of regeneration in different drought scenarios.
+
+More frequent droughts (along with higher temperatures) are predicted to come to eastern U.S. forests due to climate change \citep{Mish10}. The impacts of climate change on forests have been simulated at large spatial scales (e.g. \citealp{Gust13}), but there is less focus on finer-scale inference. This case study, examining the impact of drought, represents a first step towards an approach to simulating the impacts of climate change on oak regeneration. Tying the frequency of drought years in the simulation to predictions from climate models would give managers useful information on expected oak regeneration outcomes from different harvesting regimes in the future. Additional field data that allowed for incorporation of continuous climate data in predictive models of oak seedling growth and survival (in contrast to the binary drought/non-drought predictor used in this case study) would allow for more accurate predictions.
+
+\subsection{Synthesis and Outlook}
+
+Individual-based modeling provides a useful, spatially explicit framework for integrating multiple datasets across life stages to examine questions about the tree regeneration process. SOEL quantified the importance of multiple parameters impacting early oak life history (Figure 6.1), and confirmed via three case studies that both intrinsic and extrinsic processes influenced early oak life history and thus the oak regeneration process. At the acorn life stage, acorn production was positively related to the system both directly (Table 6.2) and indirectly by altering variability in the probability of acorn infestation and predation (Figure 6.3). Additionally, the interaction of acorn production with harvest (i.e., timing harvest to coincide with large acorn crops) continued to have measurable effects under the most extreme scenarios on sapling density multiple years later (Figure 6.2). At the seedling life stage, the interactive effects of light, drought, and harvest drove growth and survival.
+
+The case studies presented here represent only a few straightforward examples of what could be accomplished within the framework of SOEL. Further studies could put more focus on different early life history parameters (Figure 6.1), or connect the parameter values to different predictor variables based on field data. For example, the case studies in this study did not focus on caching probability despite the fact that variability in caching had a strong effect on oak regeneration parameters (Table 6.3). With appropriate field data, the model could be extended to assign utility to individual acorns based on seed characteristics (species, size, nutrient composition) and model caching probability as a function of these attributes \citep{Sund15}. Likewise, the three cases studies had a minimal emphasis on the impact of browse damage on seedling growth and survival, despite the demonstrated ability of herbivores (particularly the white-tailed deer) to greatly impact oak regeneration \citep{Marq76, Roon03, Abra12}. Deer densities, and overall rates of herbivory, were low across my study area (Chapter 3, \citealp{Haul09}). In areas where deer are more numerous and exert more control over the regeneration process, SOEL could be used to examine the effects of browse damage that varies spatially across harvest opening edges \citep{Mein02}. Given field data from multiple sites with different deer densities, the effect of herbivore density on browse probability could be incorporated into the model to predict the impact of different deer management strategies on oak regeneration.
+
+In addition to further study of processes impacting oak early life history and the immediate consequences for oak regeneration, model inference could be extended to examine longer-term consequences of changes in early life history parameters, e.g. in terms of ultimate stand structure decades later. With a longer-term outlook, SOEL becomes increasingly sensitive to parameters impacting adult tree growth and mortality in the JABOWA-derived contextual forest submodel and thus selection of proper parameter values is critical. By manipulating metrics of site quality (e.g. available nitrogen and depth to the water table), harvest regime, and/or including a different composition of tree species, the model can be generalized to simulate the successional process in different regions, where oak regeneration is limited by different factors \citep{Dey09}. Ideally, field data on adult tree growth and survival and site conditions would be used to parameterize the model; where such data is unavailable, however, the literature can be a source of parameter values (e.g. \citealp{Botk93, Holm13}). \ No newline at end of file
diff --git a/chicago.bst b/chicago.bst
new file mode 100644
index 0000000..925f9a2
--- /dev/null
+++ b/chicago.bst
@@ -0,0 +1,1653 @@
+%%% ====================================================================
+%%% @BibTeX-style-file{
+%%% author = "Glenn Paulley",
+%%% version = "4",
+%%% date = "28 August 1992",
+%%% time = "10:23:39 199",
+%%% filename = "chicago.bst",
+%%% address = "Data Structuring Group
+%%% Department of Computer Science
+%%% University of Waterloo
+%%% Waterloo, Ontario, Canada
+%%% N2L 3G1",
+%%% telephone = "(519) 885-1211",
+%%% FAX = "(519) 885-1208",
+%%% checksum = "26323 1654 5143 37417",
+%%% email = "gnpaulle@bluebox.uwaterloo.ca",
+%%% codetable = "ISO/ASCII",
+%%% keywords = "",
+%%% supported = "yes",
+%%% abstract = "A BibTeX bibliography style that follows the
+%%% `B' reference style of the 13th Edition of
+%%% the Chicago Manual of Style. A detailed
+%%% feature list is given below.",
+%%% docstring = "The checksum field above contains a CRC-16
+%%% checksum as the first value, followed by the
+%%% equivalent of the standard UNIX wc (word
+%%% count) utility output of lines, words, and
+%%% characters. This is produced by Robert
+%%% Solovay's checksum utility.",
+%%% }
+%%% ====================================================================
+%
+% "Chicago" BibTeX style, chicago.bst
+% ===================================
+%
+% BibTeX `chicago' style file for BibTeX version 0.99c, LaTeX version 2.09
+% Place it in a file called chicago.bst in the BibTeX search path.
+% You need to include chicago.sty as a \documentstyle option.
+% (Placing it in the same directory as the LaTeX document should also work.)
+% This "chicago" style is based on newapa.bst (American Psych. Assoc.)
+% found at ymir.claremont.edu.
+%
+% Citation format: (author-last-name year)
+% (author-last-name and author-last-name year)
+% (author-last-name, author-last-name, and author-last-name year)
+% (author-last-name et al. year)
+% (author-last-name)
+% author-last-name (year)
+% (author-last-name and author-last-name)
+% (author-last-name et al.)
+% (year) or (year,year)
+% year or year,year
+%
+% Reference list ordering: alphabetical by author or whatever passes
+% for author in the absence of one.
+%
+% This BibTeX style has support for abbreviated author lists and for
+% year-only citations. This is done by having the citations
+% actually look like
+%
+% \citeauthoryear{full-author-info}{abbrev-author-info}{year}
+%
+% The LaTeX style has to have the following (or similar)
+%
+% \let\@internalcite\cite
+% \def\fullcite{\def\citeauthoryear##1##2##3{##1, ##3}\@internalcite}
+% \def\fullciteA{\def\citeauthoryear##1##2##3{##1}\@internalcite}
+% \def\shortcite{\def\citeauthoryear##1##2##3{##2, ##3}\@internalcite}
+% \def\shortciteA{\def\citeauthoryear##1##2##3{##2}\@internalcite}
+% \def\citeyear{\def\citeauthoryear##1##2##3{##3}\@internalcite}
+%
+% These TeX macro definitions are found in chicago.sty. Additional
+% commands to manipulate different components of a citation can be defined
+% so that, for example, you can list author's names without parentheses
+% if using a citation as a noun or object in a sentence.
+%
+% This file was originally copied from newapa.bst at ymir.claremont.edu.
+%
+% Features of chicago.bst:
+% =======================
+%
+% - full names used in citations, but abbreviated citations are available
+% (see above)
+% - if an entry has a "month", then the month and year are also printed
+% as part of that bibitem.
+% - all conjunctions use "and" instead of "\&"
+% - major modification from Chicago Manual of Style (13th ed.) is that
+% only the first author in a reference appears last name first-
+% additional authors appear as J. Q. Public.
+% - pages are listed as "pp. xx-xx" in all entry types except
+% article entries.
+% - book, inbook, and manual use "location: publisher" (or organization)
+% for address and publisher. All other types list publishers separately.
+% - "pp." are used to identify page numbers for all entry types except
+% articles.
+% - organization is used as a citation label if neither author nor editor
+% is present (for manuals).
+% - "et al." is used for long author and editor lists, or when "others"
+% is used.
+%
+% Modifications and bug fixes from newapa.bst:
+% ===========================================
+%
+% - added month, year to bib entries if month is present
+% - fixed bug with In proceedings, added necessary comma after title
+% - all conjunctions changed to "and" from "\&"
+% - fixed bug with author labels in my.full.label: "et al." now is
+% generated when "others" is an author name
+% - major modification from Chicago Manual of Style (13th ed.) is that
+% only the first author in a reference appears last name first-
+% additional authors appear as J. Q. Public.
+% - pages are listed as "pp. xx-xx" in all entry types except
+% article entries. Unnecessary (IMHO) "()" around page numbers
+% were removed, and page numbers now don't end with a period.
+% - created chicago.sty for use with this bibstyle (required).
+% - fixed bugs in FUNCTION {format.vol.num.pages} for missing volume,
+% number, and /or pages. Renamed to format.jour.vol.
+% - fixed bug in formatting booktitles: additional period an error if
+% book has a volume.
+% - fixed bug: editors usually given redundant period before next clause
+% (format.editors.dot) removed.
+% - added label support for organizations, if both author and editor
+% are missing (from alpha.bst). If organization is too long, then
+% the key field is used for abbreviated citations.
+% - In proceedings or books of several volumes, no comma was written
+% between the "Volume x" and the page numbers (this was intentional
+% in newapa.bst). Fixed.
+% - Some journals may not have volumes/numbers, only month/year (eg.
+% IEEE Computer). Fixed bug in article style that assumed volume/number
+% was always present.
+%
+% Original documentation for newapa.sty:
+% =====================================
+%
+% This version was made by modifying the master file made by
+% Oren Patashnik (PATASHNIK@SCORE.STANFORD.EDU), and the 'named' BibTeX
+% style of Peter F. Patel-Schneider.
+%
+% Copyright (C) 1985, all rights reserved.
+% Copying of this file is authorized only if either
+% (1) you make absolutely no changes to your copy, including name, or
+% (2) if you do make changes, you name it something other than 'newapa.bst'.
+% There are undoubtably bugs in this style. If you make bug fixes,
+% improvements, etc. please let me know. My e-mail address is:
+% spencer@cgrg.ohio.state.edu or 71160.3141@compuserve.com
+%
+% This style was made from 'plain.bst', 'named.bst', and 'apalike.bst',
+% with lots of tweaking to make it look like APA style, along with tips
+% from Young Ryu and Brian Reiser's modifications of 'apalike.bst'.
+
+ENTRY
+ { address
+ author
+ booktitle
+ chapter
+ edition
+ editor
+ howpublished
+ institution
+ journal
+ key
+ month
+ note
+ number
+ organization
+ pages
+ publisher
+ school
+ series
+ title
+ type
+ volume
+ year
+ }
+ {}
+ { label.year extra.label sort.year sort.label }
+
+INTEGERS { output.state before.all mid.sentence after.sentence after.block }
+
+FUNCTION {init.state.consts}
+{ #0 'before.all :=
+ #1 'mid.sentence :=
+ #2 'after.sentence :=
+ #3 'after.block :=
+}
+
+STRINGS { s t u }
+
+FUNCTION {output.nonnull}
+{ 's :=
+ output.state mid.sentence =
+ { ", " * write$ }
+ { output.state after.block =
+ { add.period$ write$
+ newline$
+ "\newblock " write$
+ }
+ { output.state before.all =
+ 'write$
+ { add.period$ " " * write$ }
+ if$
+ }
+ if$
+ mid.sentence 'output.state :=
+ }
+ if$
+ s
+}
+
+% Use a colon to separate output. Used only for address/publisher
+% combination in book/inbook types, address/institution for manuals,
+% and organization:publisher for proceedings (inproceedings).
+%
+FUNCTION {output.nonnull.colon}
+{ 's :=
+ output.state mid.sentence =
+ { ": " * write$ }
+ { output.state after.block =
+ { add.period$ write$
+ newline$
+ "\newblock " write$
+ }
+ { output.state before.all =
+ 'write$
+ { add.period$ " " * write$ }
+ if$
+ }
+ if$
+ mid.sentence 'output.state :=
+ }
+ if$
+ s
+}
+
+FUNCTION {output}
+{ duplicate$ empty$
+ 'pop$
+ 'output.nonnull
+ if$
+}
+
+FUNCTION {output.colon}
+{ duplicate$ empty$
+ 'pop$
+ 'output.nonnull.colon
+ if$
+}
+
+FUNCTION {output.check}
+{ 't :=
+ duplicate$ empty$
+ { pop$ "empty " t * " in " * cite$ * warning$ }
+ 'output.nonnull
+ if$
+}
+
+FUNCTION {output.check.colon}
+{ 't :=
+ duplicate$ empty$
+ { pop$ "empty " t * " in " * cite$ * warning$ }
+ 'output.nonnull.colon
+ if$
+}
+
+FUNCTION {output.year.check}
+{ year empty$
+ { "empty year in " cite$ * warning$ }
+ { write$
+ " (" year * extra.label *
+ month empty$
+ { ")" * }
+ { ", " * month * ")" * }
+ if$
+ mid.sentence 'output.state :=
+ }
+ if$
+}
+
+
+FUNCTION {fin.entry}
+{ add.period$
+ write$
+ newline$
+}
+
+FUNCTION {new.block}
+{ output.state before.all =
+ 'skip$
+ { after.block 'output.state := }
+ if$
+}
+
+FUNCTION {new.sentence}
+{ output.state after.block =
+ 'skip$
+ { output.state before.all =
+ 'skip$
+ { after.sentence 'output.state := }
+ if$
+ }
+ if$
+}
+
+FUNCTION {not}
+{ { #0 }
+ { #1 }
+ if$
+}
+
+FUNCTION {and}
+{ 'skip$
+ { pop$ #0 }
+ if$
+}
+
+FUNCTION {or}
+{ { pop$ #1 }
+ 'skip$
+ if$
+}
+
+FUNCTION {new.block.checka}
+{ empty$
+ 'skip$
+ 'new.block
+ if$
+}
+
+FUNCTION {new.block.checkb}
+{ empty$
+ swap$ empty$
+ and
+ 'skip$
+ 'new.block
+ if$
+}
+
+FUNCTION {new.sentence.checka}
+{ empty$
+ 'skip$
+ 'new.sentence
+ if$
+}
+
+FUNCTION {new.sentence.checkb}
+{ empty$
+ swap$ empty$
+ and
+ 'skip$
+ 'new.sentence
+ if$
+}
+
+FUNCTION {field.or.null}
+{ duplicate$ empty$
+ { pop$ "" }
+ 'skip$
+ if$
+}
+
+%
+% Emphasize the top string on the stack.
+%
+FUNCTION {emphasize}
+{ duplicate$ empty$
+ { pop$ "" }
+ { "{\em " swap$ * "}" * }
+ if$
+}
+
+%
+% Emphasize the top string on the stack, but add a trailing space.
+%
+FUNCTION {emphasize.space}
+{ duplicate$ empty$
+ { pop$ "" }
+ { "{\em " swap$ * "\/}" * }
+ if$
+}
+
+INTEGERS { nameptr namesleft numnames }
+%
+% Format bibliographical entries with the first author last name first,
+% and subsequent authors with initials followed by last name.
+% All names are formatted in this routine.
+%
+FUNCTION {format.names}
+{ 's :=
+ #1 'nameptr := % nameptr = 1;
+ s num.names$ 'numnames := % numnames = num.name$(s);
+ numnames 'namesleft :=
+ { namesleft #0 > }
+
+ { nameptr #1 =
+ {s nameptr "{vv~}{ll}{, jj}{, f.}" format.name$ 't := }
+ {s nameptr "{f.~}{vv~}{ll}{, jj}" format.name$ 't := }
+ if$
+ nameptr #1 >
+ { namesleft #1 >
+ { ", " * t * }
+ { numnames #2 >
+ { "," * }
+ 'skip$
+ if$
+ t "others" =
+ { " et~al." * }
+ { " and " * t * } % from Chicago Manual of Style
+ if$
+ }
+ if$
+ }
+ 't
+ if$
+ nameptr #1 + 'nameptr := % nameptr += 1;
+ namesleft #1 - 'namesleft := % namesleft =- 1;
+ }
+ while$
+}
+
+FUNCTION {my.full.label}
+{ 's :=
+ #1 'nameptr := % nameptr = 1;
+ s num.names$ 'numnames := % numnames = num.name$(s);
+ numnames 'namesleft :=
+ { namesleft #0 > }
+
+ { s nameptr "{vv~}{ll}" format.name$ 't := % get the next name
+ nameptr #1 >
+ { namesleft #1 >
+ { ", " * t * }
+ { numnames #2 >
+ { "," * }
+ 'skip$
+ if$
+ t "others" =
+ { " et~al." * }
+ { " and " * t * } % from Chicago Manual of Style
+ if$
+ }
+ if$
+ }
+ 't
+ if$
+ nameptr #1 + 'nameptr := % nameptr += 1;
+ namesleft #1 - 'namesleft := % namesleft =- 1;
+ }
+ while$
+
+}
+
+FUNCTION {format.names.fml}
+%
+% Format names in "familiar" format, with first initial followed by
+% last name. Like format.names, ALL names are formatted.
+%
+{ 's :=
+ #1 'nameptr := % nameptr = 1;
+ s num.names$ 'numnames := % numnames = num.name$(s);
+ numnames 'namesleft :=
+ { namesleft #0 > }
+
+ { s nameptr "{f.~}{vv~}{ll}{, jj}" format.name$ 't :=
+
+ nameptr #1 >
+ { namesleft #1 >
+ { ", " * t * }
+ { numnames #2 >
+ { "," * }
+ 'skip$
+ if$
+ t "others" =
+ { " et~al." * }
+ { " and " * t * }
+% { " \& " * t * }
+ if$
+ }
+ if$
+ }
+ 't
+ if$
+ nameptr #1 + 'nameptr := % nameptr += 1;
+ namesleft #1 - 'namesleft := % namesleft =- 1;
+ }
+ while$
+}
+
+FUNCTION {format.authors}
+{ author empty$
+ { "" }
+ { author format.names }
+ if$
+}
+
+FUNCTION {format.key}
+{ empty$
+ { key field.or.null }
+ { "" }
+ if$
+}
+
+%
+% Format editor names for use in the "in" types: inbook, incollection,
+% inproceedings: first initial, then last names. When editors are the
+% LABEL for an entry, then format.editor is used which lists editors
+% by last name first.
+%
+FUNCTION {format.editors.fml}
+{ editor empty$
+ { "" }
+ { editor format.names.fml
+ editor num.names$ #1 >
+ { " (Eds.)" * }
+ { " (Ed.)" * }
+ if$
+ }
+ if$
+}
+
+%
+% Format editor names for use in labels, last names first.
+%
+FUNCTION {format.editors}
+{ editor empty$
+ { "" }
+ { editor format.names
+ editor num.names$ #1 >
+ { " (Eds.)" * }
+ { " (Ed.)" * }
+ if$
+ }
+ if$
+}
+
+FUNCTION {format.title}
+{ title empty$
+ { "" }
+ { title "t" change.case$ }
+ if$
+}
+
+% Note that the APA style requres case changes
+% in article titles. The following does not
+% change cases. If you perfer it, uncomment the
+% following and comment out the above.
+
+%FUNCTION {format.title}
+%{ title empty$
+% { "" }
+% { title }
+% if$
+%}
+
+FUNCTION {n.dashify}
+{ 't :=
+ ""
+ { t empty$ not }
+ { t #1 #1 substring$ "-" =
+ { t #1 #2 substring$ "--" = not
+ { "--" *
+ t #2 global.max$ substring$ 't :=
+ }
+ { { t #1 #1 substring$ "-" = }
+ { "-" *
+ t #2 global.max$ substring$ 't :=
+ }
+ while$
+ }
+ if$
+ }
+ { t #1 #1 substring$ *
+ t #2 global.max$ substring$ 't :=
+ }
+ if$
+ }
+ while$
+}
+
+FUNCTION {format.btitle}
+{ edition empty$
+ { title emphasize }
+ { title empty$
+ { title emphasize }
+ { volume empty$ % gnp - check for volume, then don't need period
+ { "{\em " title * "\/} (" * edition * " ed.)" * "." * }
+ { "{\em " title * "\/} (" * edition * " ed.)" * }
+ if$
+ }
+ if$
+ }
+ if$
+}
+
+FUNCTION {format.emphasize.booktitle}
+{ edition empty$
+ { booktitle emphasize }
+ { booktitle empty$
+ { booktitle emphasize }
+ { volume empty$ % gnp - extra period an error if book has a volume
+ { "{\em " booktitle * "\/} (" * edition * " ed.)" * "." *}
+ { "{\em " booktitle * "\/} (" * edition * " ed.)" * }
+ if$
+ }
+ if$
+ }
+ if$
+ }
+
+
+FUNCTION {tie.or.space.connect}
+{ duplicate$ text.length$ #3 <
+ { "~" }
+ { " " }
+ if$
+ swap$ * *
+}
+
+FUNCTION {either.or.check}
+{ empty$
+ 'pop$
+ { "can't use both " swap$ * " fields in " * cite$ * warning$ }
+ if$
+}
+
+FUNCTION {format.bvolume}
+{ volume empty$
+ { "" }
+ { "Volume" volume tie.or.space.connect % gnp - changed to mixed case
+ series empty$
+ 'skip$
+ { " of " * series emphasize * }
+ if$
+ "volume and number" number either.or.check
+ }
+ if$
+}
+
+FUNCTION {format.number.series}
+{ volume empty$
+ { number empty$
+ { series field.or.null }
+ { output.state mid.sentence =
+ { "Number" } % gnp - changed to mixed case always
+ { "Number" }
+ if$
+ number tie.or.space.connect
+ series empty$
+ { "there's a number but no series in " cite$ * warning$ }
+ { " in " * series * }
+ if$
+ }
+ if$
+ }
+ { "" }
+ if$
+}
+
+INTEGERS { multiresult }
+
+FUNCTION {multi.page.check}
+{ 't :=
+ #0 'multiresult :=
+ { multiresult not
+ t empty$ not
+ and
+ }
+ { t #1 #1 substring$
+ duplicate$ "-" =
+ swap$ duplicate$ "," =
+ swap$ "+" =
+ or or
+ { #1 'multiresult := }
+ { t #2 global.max$ substring$ 't := }
+ if$
+ }
+ while$
+ multiresult
+}
+
+FUNCTION {format.pages}
+{ pages empty$
+ { "" }
+ { pages multi.page.check
+ { "pp.\ " pages n.dashify tie.or.space.connect } % gnp - removed ()
+ { "pp.\ " pages tie.or.space.connect }
+ if$
+ }
+ if$
+}
+
+% By Young (and Spencer)
+% GNP - fixed bugs with missing volume, number, and/or pages
+%
+% Format journal, volume, number, pages for article types.
+%
+FUNCTION {format.jour.vol}
+{ journal empty$
+ { "no journal in " cite$ * warning$
+ "" }
+ { journal emphasize.space }
+ if$
+ number empty$
+ { volume empty$
+ { "no number and no volume in " cite$ * warning$
+ "" * }
+ { "~{\em " * Volume * "}" * }
+ if$
+ }
+ { volume empty$
+ {"no volume for " cite$ * warning$
+ "~(" * number * ")" * }
+ { "~" *
+ volume emphasize.space
+ "(" * number * ")" * * }
+ if$
+ }
+ if$
+ pages empty$
+ {"page numbers missing in " cite$ * warning$
+ "" * } % gnp - place a null string on the stack for output
+ { duplicate$ empty$
+ { pop$ format.pages }
+ { ", " * pages n.dashify * } % gnp - removed pp. for articles
+ if$
+ }
+ if$
+}
+
+FUNCTION {format.chapter.pages}
+{ chapter empty$
+ 'format.pages
+ { type empty$
+ { "Chapter" } % gnp - changed to mixed case
+ { type "t" change.case$ }
+ if$
+ chapter tie.or.space.connect
+ pages empty$
+ {"page numbers missing in " cite$ * warning$} % gnp - added check
+ { ", " * format.pages * }
+ if$
+ }
+ if$
+}
+
+FUNCTION {format.in.ed.booktitle}
+{ booktitle empty$
+ { "" }
+ { editor empty$
+ { "In " format.emphasize.booktitle * }
+ { "In " format.editors.fml * ", " * format.emphasize.booktitle * }
+ if$
+ }
+ if$
+}
+
+FUNCTION {format.thesis.type}
+{ type empty$
+ 'skip$
+ { pop$
+ type "t" change.case$
+ }
+ if$
+}
+
+FUNCTION {format.tr.number}
+{ type empty$
+ { "Technical Report" }
+ 'type
+ if$
+ number empty$
+ { "t" change.case$ }
+ { number tie.or.space.connect }
+ if$
+}
+
+FUNCTION {format.article.crossref}
+{ "See"
+ "\citeN{" * crossref * "}" *
+}
+
+FUNCTION {format.crossref.editor}
+{ editor #1 "{vv~}{ll}" format.name$
+ editor num.names$ duplicate$
+ #2 >
+ { pop$ " et~al." * }
+ { #2 <
+ 'skip$
+ { editor #2 "{ff }{vv }{ll}{ jj}" format.name$ "others" =
+ { " et~al." * }
+ { " and " * editor #2 "{vv~}{ll}" format.name$ * }
+ if$
+ }
+ if$
+ }
+ if$
+}
+
+FUNCTION {format.book.crossref}
+{ volume empty$
+ { "empty volume in " cite$ * "'s crossref of " * crossref * warning$
+ "In "
+ }
+ { "Volume" volume tie.or.space.connect % gnp - changed to mixed case
+ " of " *
+ }
+ if$
+ editor empty$
+ editor field.or.null author field.or.null =
+ or
+ { key empty$
+ { series empty$
+ { "need editor, key, or series for " cite$ * " to crossref " *
+ crossref * warning$
+ "" *
+ }
+ { "{\em " * series * "\/}" * }
+ if$
+ }
+ { key * }
+ if$
+ }
+ { format.crossref.editor * }
+ if$
+ " \citeN{" * crossref * "}" *
+}
+
+FUNCTION {format.incoll.inproc.crossref}
+{ "See"
+ " \citeN{" * crossref * "}" *
+}
+
+% format.lab.names:
+%
+% determines "short" names for the abbreviated author information.
+% "Long" labels are created in calc.label, using the routine my.full.label
+% to format author and editor fields.
+%
+% There are 4 cases for labels. (n=3 in the example)
+% a) one author Foo
+% b) one to n Foo, Bar and Baz
+% c) use of "and others" Foo, Bar et al.
+% d) more than n Foo et al.
+%
+FUNCTION {format.lab.names}
+{ 's :=
+ s num.names$ 'numnames :=
+ numnames #2 > % change number to number of others allowed before
+ % forcing "et al".
+ { s #1 "{vv~}{ll}" format.name$ " et~al." * }
+ {
+ numnames #1 - 'namesleft :=
+ #2 'nameptr :=
+ s #1 "{vv~}{ll}" format.name$
+ { namesleft #0 > }
+ { nameptr numnames =
+ { s nameptr "{ff }{vv }{ll}{ jj}" format.name$ "others" =
+ { " et~al." * }
+ { " and " * s nameptr "{vv~}{ll}" format.name$ * }
+ if$
+ }
+ { ", " * s nameptr "{vv~}{ll}" format.name$ * }
+ if$
+ nameptr #1 + 'nameptr :=
+ namesleft #1 - 'namesleft :=
+ }
+ while$
+ }
+ if$
+}
+
+FUNCTION {author.key.label}
+{ author empty$
+ { key empty$
+ { "no key, author in " cite$ * warning$
+ cite$ #1 #3 substring$ }
+ 'key
+ if$
+ }
+ { author format.lab.names }
+ if$
+}
+
+FUNCTION {editor.key.label}
+{ editor empty$
+ { key empty$
+ { "no key, editor in " cite$ * warning$
+ cite$ #1 #3 substring$ }
+ 'key
+ if$
+ }
+ { editor format.lab.names }
+ if$
+}
+
+FUNCTION {author.key.organization.label}
+%
+% added - gnp. Provide label formatting by organization if author is null.
+%
+{ author empty$
+ { organization empty$
+ { key empty$
+ { "no key, author or organization in " cite$ * warning$
+ cite$ #1 #3 substring$ }
+ 'key
+ if$
+ }
+ { organization }
+ if$
+ }
+ { author format.lab.names }
+ if$
+}
+
+FUNCTION {editor.key.organization.label}
+%
+% added - gnp. Provide label formatting by organization if editor is null.
+%
+{ editor empty$
+ { organization empty$
+ { key empty$
+ { "no key, editor or organization in " cite$ * warning$
+ cite$ #1 #3 substring$ }
+ 'key
+ if$
+ }
+ { organization }
+ if$
+ }
+ { editor format.lab.names }
+ if$
+}
+
+FUNCTION {author.editor.key.label}
+{ author empty$
+ { editor empty$
+ { key empty$
+ { "no key, author, or editor in " cite$ * warning$
+ cite$ #1 #3 substring$ }
+ 'key
+ if$
+ }
+ { editor format.lab.names }
+ if$
+ }
+ { author format.lab.names }
+ if$
+}
+
+FUNCTION {calc.label}
+%
+% Changed - GNP. See also author.organization.sort, editor.organization.sort
+% Form label for BibTeX entry. The classification of which fields are used
+% for which type of entry (book, inbook, etc.) are taken from alpha.bst.
+% The change here from newapa is to also include organization as a
+% citation label if author or editor is missing.
+%
+{ type$ "book" =
+ type$ "inbook" =
+ or
+ 'author.editor.key.label
+ { type$ "proceedings" =
+ 'editor.key.organization.label
+ { type$ "manual" =
+ 'author.key.organization.label
+ 'author.key.label
+ if$
+ }
+ if$
+ }
+ if$
+
+ author empty$ % generate the full label citation information.
+ { editor empty$
+ { organization empty$
+ { "no author, editor, or organization in " cite$ * warning$
+ "??" }
+ { organization }
+ if$
+ }
+ { editor my.full.label }
+ if$
+ }
+ { author my.full.label }
+ if$
+
+% leave label on the stack, to be popped when required.
+
+ "}{" * swap$ * "}{" *
+% year field.or.null purify$ #-1 #4 substring$ *
+%
+% save the year for sort processing afterwards (adding a, b, c, etc.)
+%
+ year field.or.null purify$ #-1 #4 substring$
+ 'label.year :=
+}
+
+FUNCTION {output.bibitem}
+{ newline$
+
+ "\bibitem[\protect\citeauthoryear{" write$
+ calc.label write$
+ sort.year write$
+ "}]{" write$
+
+ cite$ write$
+ "}" write$
+ newline$
+ ""
+ before.all 'output.state :=
+}
+
+FUNCTION {article}
+{ output.bibitem
+ format.authors
+ "author" output.check
+ author format.key output % added
+ output.year.check % added
+ new.block
+ format.title
+ "title" output.check
+ new.block
+ crossref missing$
+ { format.jour.vol output
+ }
+ { format.article.crossref output.nonnull
+ format.pages output
+ }
+ if$
+ new.block
+ note output
+ fin.entry
+}
+
+FUNCTION {book}
+{ output.bibitem
+ author empty$
+ { format.editors
+ "author and editor" output.check }
+ { format.authors
+ output.nonnull
+ crossref missing$
+ { "author and editor" editor either.or.check }
+ 'skip$
+ if$
+ }
+ if$
+ output.year.check % added
+ new.block
+ format.btitle
+ "title" output.check
+ crossref missing$
+ { format.bvolume output
+ new.block
+ format.number.series output
+ new.sentence
+ address output
+ publisher "publisher" output.check.colon
+ }
+ { new.block
+ format.book.crossref output.nonnull
+ }
+ if$
+ new.block
+ note output
+ fin.entry
+}
+
+FUNCTION {booklet}
+{ output.bibitem
+ format.authors output
+ author format.key output % added
+ output.year.check % added
+ new.block
+ format.title
+ "title" output.check
+ new.block
+ howpublished output
+ address output
+ new.block
+ note output
+ fin.entry
+}
+
+FUNCTION {inbook}
+{ output.bibitem
+ author empty$
+ { format.editors
+ "author and editor" output.check
+ }
+ { format.authors output.nonnull
+ crossref missing$
+ { "author and editor" editor either.or.check }
+ 'skip$
+ if$
+ }
+ if$
+ output.year.check % added
+ new.block
+ format.btitle
+ "title" output.check
+ crossref missing$
+ { format.bvolume output
+ format.chapter.pages
+ "chapter and pages" output.check
+ new.block
+ format.number.series output
+ new.sentence
+ address output
+ publisher
+ "publisher" output.check.colon
+ }
+ { format.chapter.pages "chapter and pages" output.check
+ new.block
+ format.book.crossref output.nonnull
+ }
+ if$
+ new.block
+ note output
+ fin.entry
+}
+
+FUNCTION {incollection}
+{ output.bibitem
+ format.authors
+ "author" output.check
+ author format.key output % added
+ output.year.check % added
+ new.block
+ format.title
+ "title" output.check
+ new.block
+ crossref missing$
+ { format.in.ed.booktitle
+ "booktitle" output.check
+ format.bvolume output
+ format.number.series output
+ format.chapter.pages output % gnp - was special.output.nonnull
+% left out comma before page numbers
+ new.sentence
+ address output
+ publisher "publisher" output.check.colon
+ }
+ { format.incoll.inproc.crossref
+ output.nonnull
+ format.chapter.pages output
+ }
+ if$
+ new.block
+ note output
+ fin.entry
+}
+
+FUNCTION {inproceedings}
+{ output.bibitem
+ format.authors
+ "author" output.check
+ author format.key output % added
+ output.year.check % added
+ new.block
+ format.title
+ "title" output.check
+ new.block
+ crossref missing$
+ { format.in.ed.booktitle
+ "booktitle" output.check
+ format.bvolume output
+ format.number.series output
+ address output
+ format.pages output
+ new.sentence
+ organization output
+ publisher output.colon
+ }
+ { format.incoll.inproc.crossref output.nonnull
+ format.pages output
+ }
+ if$
+ new.block
+ note output
+ fin.entry
+}
+
+FUNCTION {conference} { inproceedings }
+
+FUNCTION {manual}
+{ output.bibitem
+ author empty$
+ { editor empty$
+ { organization "organization" output.check
+ organization format.key output } % if all else fails, use key
+ { format.editors "author and editor" output.check }
+ if$
+ }
+ { format.authors output.nonnull }
+ if$
+ output.year.check % added
+ new.block
+ format.btitle
+ "title" output.check
+ organization address new.block.checkb
+% Reversed the order of "address" and "organization", added the ":".
+ address output
+ organization "organization" output.check.colon
+% address output
+% ":" output
+% organization output
+ new.block
+ note output
+ fin.entry
+}
+
+FUNCTION {mastersthesis}
+{ output.bibitem
+ format.authors
+ "author" output.check
+ author format.key output % added
+ output.year.check % added
+ new.block
+ format.title
+ "title" output.check
+ new.block
+ "Master's thesis" format.thesis.type output.nonnull
+ school "school" output.check
+ address output
+ new.block
+ note output
+ fin.entry
+}
+
+FUNCTION {misc}
+{ output.bibitem
+ format.authors output
+ author format.key output % added
+ output.year.check % added
+ title howpublished new.block.checkb
+ format.title output
+ new.block
+ howpublished output
+ new.block
+ note output
+ fin.entry
+}
+
+FUNCTION {phdthesis}
+{ output.bibitem
+ format.authors
+ "author" output.check
+ author format.key output % added
+ output.year.check % added
+ new.block
+ format.btitle
+ "title" output.check
+ new.block
+ "Ph.\ D. thesis" format.thesis.type output.nonnull
+ school "school" output.check
+ address output
+ new.block
+ note output
+ fin.entry
+}
+
+FUNCTION {proceedings}
+{ output.bibitem
+ editor empty$
+ { organization output
+ organization format.key output } % gnp - changed from author format.key
+ { format.editors output.nonnull }
+ if$
+% author format.key output % gnp - removed (should be either
+% editor or organization
+ output.year.check % added (newapa)
+ new.block
+ format.btitle
+ "title" output.check
+ format.bvolume output
+ format.number.series output
+ address output
+ new.sentence
+ organization output
+ publisher output.colon
+ new.block
+ note output
+ fin.entry
+}
+
+FUNCTION {techreport}
+{ output.bibitem
+ format.authors
+ "author" output.check
+ author format.key output % added
+ output.year.check % added
+ new.block
+ format.title
+ "title" output.check
+ new.block
+ format.tr.number output.nonnull
+ institution
+ "institution" output.check
+ address output
+ new.block
+ note output
+ fin.entry
+}
+
+FUNCTION {unpublished}
+{ output.bibitem
+ format.authors
+ "author" output.check
+ author format.key output % added
+ output.year.check % added
+ new.block
+ format.title
+ "title" output.check
+ new.block
+ note "note" output.check
+ fin.entry
+}
+
+FUNCTION {default.type} { misc }
+
+MACRO {jan} {"January"}
+
+MACRO {feb} {"February"}
+
+MACRO {mar} {"March"}
+
+MACRO {apr} {"April"}
+
+MACRO {may} {"May"}
+
+MACRO {jun} {"June"}
+
+MACRO {jul} {"July"}
+
+MACRO {aug} {"August"}
+
+MACRO {sep} {"September"}
+
+MACRO {oct} {"October"}
+
+MACRO {nov} {"November"}
+
+MACRO {dec} {"December"}
+
+MACRO {acmcs} {"ACM Computing Surveys"}
+
+MACRO {acta} {"Acta Informatica"}
+
+MACRO {ai} {"Artificial Intelligence"}
+
+MACRO {cacm} {"Communications of the ACM"}
+
+MACRO {ibmjrd} {"IBM Journal of Research and Development"}
+
+MACRO {ibmsj} {"IBM Systems Journal"}
+
+MACRO {ieeese} {"IEEE Transactions on Software Engineering"}
+
+MACRO {ieeetc} {"IEEE Transactions on Computers"}
+
+MACRO {ieeetcad}
+ {"IEEE Transactions on Computer-Aided Design of Integrated Circuits"}
+
+MACRO {ipl} {"Information Processing Letters"}
+
+MACRO {jacm} {"Journal of the ACM"}
+
+MACRO {jcss} {"Journal of Computer and System Sciences"}
+
+MACRO {scp} {"Science of Computer Programming"}
+
+MACRO {sicomp} {"SIAM Journal on Computing"}
+
+MACRO {tocs} {"ACM Transactions on Computer Systems"}
+
+MACRO {tods} {"ACM Transactions on Database Systems"}
+
+MACRO {tog} {"ACM Transactions on Graphics"}
+
+MACRO {toms} {"ACM Transactions on Mathematical Software"}
+
+MACRO {toois} {"ACM Transactions on Office Information Systems"}
+
+MACRO {toplas} {"ACM Transactions on Programming Languages and Systems"}
+
+MACRO {tcs} {"Theoretical Computer Science"}
+
+READ
+
+FUNCTION {sortify}
+{ purify$
+ "l" change.case$
+}
+
+INTEGERS { len }
+
+FUNCTION {chop.word}
+{ 's :=
+ 'len :=
+ s #1 len substring$ =
+ { s len #1 + global.max$ substring$ }
+ 's
+ if$
+}
+
+
+
+FUNCTION {sort.format.names}
+{ 's :=
+ #1 'nameptr :=
+ ""
+ s num.names$ 'numnames :=
+ numnames 'namesleft :=
+ { namesleft #0 > }
+ { nameptr #1 >
+ { " " * }
+ 'skip$
+ if$
+ s nameptr "{vv{ } }{ll{ }}{ f{ }}{ jj{ }}" format.name$ 't :=
+ nameptr numnames = t "others" = and
+ { " et~al" * }
+ { t sortify * }
+ if$
+ nameptr #1 + 'nameptr :=
+ namesleft #1 - 'namesleft :=
+ }
+ while$
+}
+
+FUNCTION {sort.format.title}
+{ 't :=
+ "A " #2
+ "An " #3
+ "The " #4 t chop.word
+ chop.word
+ chop.word
+ sortify
+ #1 global.max$ substring$
+}
+
+FUNCTION {author.sort}
+{ author empty$
+ { key empty$
+ { "to sort, need author or key in " cite$ * warning$
+ "" }
+ { key sortify }
+ if$
+ }
+ { author sort.format.names }
+ if$
+}
+
+FUNCTION {editor.sort}
+{ editor empty$
+ { key empty$
+ { "to sort, need editor or key in " cite$ * warning$
+ ""
+ }
+ { key sortify }
+ if$
+ }
+ { editor sort.format.names }
+ if$
+}
+
+FUNCTION {author.editor.sort}
+{ author empty$
+ { "missing author in " cite$ * warning$
+ editor empty$
+ { key empty$
+ { "to sort, need author, editor, or key in " cite$ * warning$
+ ""
+ }
+ { key sortify }
+ if$
+ }
+ { editor sort.format.names }
+ if$
+ }
+ { author sort.format.names }
+ if$
+}
+
+FUNCTION {author.organization.sort}
+%
+% added - GNP. Stack author or organization for sorting (from alpha.bst).
+% Unlike alpha.bst, we need entire names, not abbreviations
+%
+{ author empty$
+ { organization empty$
+ { key empty$
+ { "to sort, need author, organization, or key in " cite$ * warning$
+ ""
+ }
+ { key sortify }
+ if$
+ }
+ { organization sortify }
+ if$
+ }
+ { author sort.format.names }
+ if$
+}
+
+FUNCTION {editor.organization.sort}
+%
+% added - GNP. Stack editor or organization for sorting (from alpha.bst).
+% Unlike alpha.bst, we need entire names, not abbreviations
+%
+{ editor empty$
+ { organization empty$
+ { key empty$
+ { "to sort, need editor, organization, or key in " cite$ * warning$
+ ""
+ }
+ { key sortify }
+ if$
+ }
+ { organization sortify }
+ if$
+ }
+ { editor sort.format.names }
+ if$
+}
+
+FUNCTION {presort}
+%
+% Presort creates the bibentry's label via a call to calc.label, and then
+% sorts the entries based on entry type. Chicago.bst adds support for
+% including organizations as the sort key; the following is stolen from
+% alpha.bst.
+%
+{ calc.label sortify % recalculate bibitem label
+ year field.or.null purify$ #-1 #4 substring$ * % add year
+ " "
+ *
+ type$ "book" =
+ type$ "inbook" =
+ or
+ 'author.editor.sort
+ { type$ "proceedings" =
+ 'editor.organization.sort
+ { type$ "manual" =
+ 'author.organization.sort
+ 'author.sort
+ if$
+ }
+ if$
+ }
+ if$
+ #1 entry.max$ substring$ % added for newapa
+ 'sort.label := % added for newapa
+ sort.label % added for newapa
+ *
+ " "
+ *
+ title field.or.null
+ sort.format.title
+ *
+ #1 entry.max$ substring$
+ 'sort.key$ :=
+}
+
+ITERATE {presort}
+
+SORT % by label, year, author/editor, title
+
+STRINGS { last.label next.extra }
+
+INTEGERS { last.extra.num }
+
+FUNCTION {initialize.extra.label.stuff}
+{ #0 int.to.chr$ 'last.label :=
+ "" 'next.extra :=
+ #0 'last.extra.num :=
+}
+
+FUNCTION {forward.pass}
+%
+% Pass through all entries, comparing current entry to last one.
+% Need to concatenate year to the stack (done by calc.label) to determine
+% if two entries are the same (see presort)
+%
+{ last.label
+ calc.label year field.or.null purify$ #-1 #4 substring$ * % add year
+ #1 entry.max$ substring$ = % are they equal?
+ { last.extra.num #1 + 'last.extra.num :=
+ last.extra.num int.to.chr$ 'extra.label :=
+ }
+ { "a" chr.to.int$ 'last.extra.num :=
+ "" 'extra.label :=
+ calc.label year field.or.null purify$ #-1 #4 substring$ * % add year
+ #1 entry.max$ substring$ 'last.label := % assign to last.label
+ }
+ if$
+}
+
+FUNCTION {reverse.pass}
+{ next.extra "b" =
+ { "a" 'extra.label := }
+ 'skip$
+ if$
+ label.year extra.label * 'sort.year :=
+ extra.label 'next.extra :=
+}
+
+EXECUTE {initialize.extra.label.stuff}
+
+ITERATE {forward.pass}
+
+REVERSE {reverse.pass}
+
+FUNCTION {bib.sort.order}
+{ sort.label
+ " "
+ *
+ year field.or.null sortify
+ *
+ " "
+ *
+ title field.or.null
+ sort.format.title
+ *
+ #1 entry.max$ substring$
+ 'sort.key$ :=
+}
+
+ITERATE {bib.sort.order}
+
+SORT % by sort.label, year, title --- giving final bib. order.
+
+FUNCTION {begin.bib}
+
+{ preamble$ empty$
+ 'skip$
+ { preamble$ write$ newline$ }
+ if$
+ "\begin{thebibliography}{}" write$ newline$
+}
+
+
+EXECUTE {begin.bib}
+
+EXECUTE {init.state.consts}
+
+ITERATE {call.type$}
+
+FUNCTION {end.bib}
+{ newline$
+ "\end{thebibliography}" write$ newline$
+}
+
+EXECUTE {end.bib}
+
diff --git a/conclusion.tex b/conclusion.tex
new file mode 100644
index 0000000..c0a2ae9
--- /dev/null
+++ b/conclusion.tex
@@ -0,0 +1,24 @@
+\chapter{{CONCLUSION}}
+
+\section{Overview}
+
+The pathway from acorn to oak (\textit{Quercus}) sapling comprises multiple biotic and abiotic filters capable of impacting oak regeneration (Chapter 1). Several studies have quantified these early oak life history processes; most have been focused on a subset of parameters or a single life stage \citep{John84, Buck98, Bell05, Lomb08, Kell14, Olso15a}. In this dissertation, I drew on a rich collection of datasets collected in a long-term study of forest ecosystem responses to management (the Hardwood Ecosystem Experiment; \citealp{Kalb13}), and augmented them to obtain a detailed picture of the black (\textit{Q. velutina}) and white oak (\textit{Q. alba}) regeneration process under several forest management scenarios. My approach to quantifying early oak life history combined empirical research on trophic interactions between oak, seed predators (Chapter 2), and herbivores (Chapters 3-4) as well as the development of SOEL, a comprehensive simulation model of the oak life cycle (Chapters 5-6).
+
+\section{Conclusions From Empirical Research}
+
+Timber harvest with the goal of regenerating oak is intended to create environmental conditions that favor the competitive success of oak advance regeneration
+\citep{John09}. However, the disturbance associated with harvesting can also alter the composition and/or behavior of the suite of plant and animal species that interact with oak. Thus, timber harvest can have unintended effects on the oak regeneration process. In Chapter 2, I examined the impact of a midstory removal harvest (implemented as part of a 3-stage shelterwood harvest; \citealp{Kalb13}) on the interaction of oak with small mammals that act as seed predators and dispersal agents. Despite the relatively minor disturbance created by the midstory removal, I identified an increase in acorn removal, a reduction in survival and in some cases, an increase in dispersal distance following midstory removal harvests (Chapter 2). Increased understory vegetation may have facilitated more efficient foraging by small mammals, resulting in these shifts in predation and dispersal parameters. The consequences for oak regeneration are unclear; despite increased removal and reduced survival, if in total more acorns are dispersed to favorable microsites regeneration could be positively impacted. I also identified large year-to-year differences in acorn survival parameters, likely tied to shifts in overall acorn abundance and small mammal population fluctuations \citep{Kell13, Kell14}.
+
+Survival and growth of experimentally planted oak seedlings was more affected by abiotic conditions altered by harvest than by biotic factors (Chapters 3-4). Growth and survival had a particularly complex relationship with light. As expected, seedling growth of oak, a relatively shade-intolerant tree, was generally positively associated with increased light in harvest openings (Chapter 4). However, the study spanned multiple years with drought during the growing season. In harvest openings, seedlings were more exposed and thus at a greater risk of mortality due to water stress during the drought than seedlings under the forest canopy. Thus, increased light can be a trade-off for oak seedlings in drought years. This is a concern for forest management for oak given that the frequency of drought is expected to increase in the Midwestern U.S. with climate change \citep{Mish10}.
+
+\section{Conclusions from Simulation Modeling}
+
+SOEL (Simulation of Oak Early Life history) combines a detailed individual-based model of early oak life history (acorn through seedling) with a simplified model of sapling and mature tree growth and survival (Chapter 5). The individual-based framework of SOEL is extremely flexible and adaptable to a variety of empirical input data, allowing for the positions and other individual characteristics of acorns and seedlings to be tracked over time. As a demonstration of the capabilities of SOEL, I implemented three individual case studies examining different parts of early oak life history and their ultimate impacts on oak regeneration: (1) the impact of good mast years immediately prior to timber harvest; (2) the overall effect of changes in acorn predation and dispersal parameters post-midstory removal harvest; and (3) impact of changes in the frequency of drought (Chapter 6).
+
+Multiple above-average mast production years immediately following harvest had a positive impact on the accumulation of oak stems, but the effect declined with time. The same was true of the impacts of changes in oak predation and dispersal parameters following midstory removal; there was a small negative effect on accumulation of new seedlings, but the difference was negligible for density of oak saplings (a more important indicator of oak regeneration success). Yearly variability in acorn-level parameters had a much stronger (positive) impact on oak regeneration, lending support for the predator satiation hypothesis \citep{Janz71, Craw95, Kell14}. Of the three case studies examined, the impacts of changing the frequency of drought were most striking. Increasing probability of a drought year greatly decreased the density of oak saplings, to the point that if growth and survival were impacted every single year by drought the oak regeneration process would be brought to a complete halt.
+
+\section{Future Directions}
+
+Opportunities exist to build on both the empirical and simulation-based research described in this dissertation. The long-term data collection efforts at the HEE (specifically, acorn production, removal, and infestation by weevils) will continue to strengthen our understanding of yearly variability (and yearly correlation) in acorn-level parameters (which SOEL confirmed play an important role in driving accumulation of oak regeneration). The study of acorn fate and dispersal was limited given that only a midstory removal harvest had occurred prior to data collection. A continuation of this study following the second and third phase of the shelterwood harvest (which should have much larger impacts on forest structure; \citealp{Kalb13}) would help to confirm if changes in habitat structure are affecting the foraging behavior of small mammals.
+
+The flexibility of SOEL enables a diverse set of hypotheses to be tested using a broad range of field data. For example, using only the datasets currently available from the HEE, SOEL could be used for a more in-depth examination of the predator satiation hypothesis, by explicitly modeling weevil and small mammal populations and tying them to acorn survival parameters. Using other datasets, SOEL could incorporate variability in acorn characteristics (mass, nutrient content) that impact seed predation, dispersal, and caching decisions by small mammals \citep{Sund15}, or examine the impacts of spatial variability in browsing by deer in areas were herbivore pressure is greater than it was in the HEE system. \ No newline at end of file
diff --git a/dissertation.tex b/dissertation.tex
new file mode 100644
index 0000000..37e5521
--- /dev/null
+++ b/dissertation.tex
@@ -0,0 +1,49 @@
+%
+% Dissertation: Kenneth F. Kellner, December 2015
+%
+\documentclass[ece,dissertation,chicago]{puthesis}
+\usepackage{amsmath}
+\usepackage{booktabs}
+\usepackage{lmodern}
+\usepackage{listings}
+
+\include{netlogostyle}
+\include{Rstyle}
+%Note: in RStudio, need to allow escape shell commands for this to work; must be done manually in Windows
+%\immediate\write18{./genmodel.sh}
+
+\title{Interactive Effects of Animals and Silviculture\\ on the Early Life History of Oak}
+
+\author{Kenneth F. Kellner}{Kellner, Kenneth F.}
+\majorprof{Robert K. Swihart}
+\pudegree{Doctor of Philosophy}{Ph.D.}{December}{2015}
+\campus{West Lafayette}
+
+
+\begin{document}
+
+\volume
+
+\include{front}
+
+\include{introduction}
+
+\include{chapter2}
+
+\include{chapter3}
+
+\include{chapter4}
+
+\include{chapter5}
+
+\include{chapter6}
+
+\include{conclusion}
+
+\bibliography{references}
+
+\include{appendices}
+
+\include{vita}
+
+\end{document} \ No newline at end of file
diff --git a/dissertation.toc b/dissertation.toc
new file mode 100644
index 0000000..41f1fa7
--- /dev/null
+++ b/dissertation.toc
@@ -0,0 +1,118 @@
+\contentsline {chapter}{LIST OF TABLES}{ix}
+\contentsline {chapter}{LIST OF FIGURES}{xii}
+\contentsline {chapter}{ABSTRACT}{xvi}
+\contentsline {chapter}{\numberline {1}{INTRODUCTION}}{1}
+\contentsline {section}{\numberline {1.1}Oak Forests in the Eastern United States}{1}
+\contentsline {section}{\numberline {1.2}Role of Oak as a Foundational Species}{2}
+\contentsline {section}{\numberline {1.3}Oak Regeneration Failure}{3}
+\contentsline {section}{\numberline {1.4}Silviculture as a Means to Promote Oak Regeneration}{6}
+\contentsline {section}{\numberline {1.5}Interaction of Silviculture with Seed Predators and Herbivores}{7}
+\contentsline {section}{\numberline {1.6}Goals of this Dissertation}{8}
+\contentsline {chapter}{\numberline {2}{MIDSTORY REMOVAL ALTERS PREDATION AND DISPERSAL OF OAK (\textit {QUERCUS}) ACORNS BY SMALL MAMMALS IN THE CENTRAL HARDWOOD FOREST REGION}}{9}
+\contentsline {section}{\numberline {2.1}Introduction}{9}
+\contentsline {section}{\numberline {2.2}Methods}{14}
+\contentsline {subsection}{\numberline {2.2.1}Study Location}{14}
+\contentsline {subsection}{\numberline {2.2.2}Experimental Design}{15}
+\contentsline {subsection}{\numberline {2.2.3}Covariate Data}{17}
+\contentsline {subsection}{\numberline {2.2.4}Acorn Fate Analysis}{18}
+\contentsline {section}{\numberline {2.3}Results}{20}
+\contentsline {subsection}{\numberline {2.3.1}Plot Characteristics}{20}
+\contentsline {subsection}{\numberline {2.3.2}Acorn Removal}{20}
+\contentsline {subsection}{\numberline {2.3.3}Acorn Survival and Dispersal}{23}
+\contentsline {section}{\numberline {2.4}Discussion}{27}
+\contentsline {subsection}{\numberline {2.4.1}Acorn Removal}{27}
+\contentsline {subsection}{\numberline {2.4.2}Acorn Survival and Dispersal}{28}
+\contentsline {subsection}{\numberline {2.4.3}Microsite Effects}{31}
+\contentsline {subsection}{\numberline {2.4.4}Conclusion}{32}
+\contentsline {chapter}{\numberline {3}{OAK SEEDLING HERBIVORY BY MAMMALS AND INSECTS ALONG A DISTURBANCE GRADIENT CREATED BY TIMBER HARVEST IN THE CENTRAL HARDWOODS}}{34}
+\contentsline {section}{\numberline {3.1}Introduction}{34}
+\contentsline {section}{\numberline {3.2}Methods}{37}
+\contentsline {subsection}{\numberline {3.2.1}Study Location}{37}
+\contentsline {subsection}{\numberline {3.2.2}Experimental Design}{37}
+\contentsline {subsection}{\numberline {3.2.3}Oak Plantings}{38}
+\contentsline {subsection}{\numberline {3.2.4}Seedling Measurements}{39}
+\contentsline {subsection}{\numberline {3.2.5}Pellet Count Transects}{40}
+\contentsline {subsection}{\numberline {3.2.6}Analysis}{40}
+\contentsline {section}{\numberline {3.3}Results}{41}
+\contentsline {section}{\numberline {3.4}Discussion}{46}
+\contentsline {subsection}{\numberline {3.4.1}Edge and Harvest Effects}{47}
+\contentsline {subsection}{\numberline {3.4.2}Seedling Height}{50}
+\contentsline {subsection}{\numberline {3.4.3}Implications for Regeneration}{51}
+\contentsline {chapter}{\numberline {4}{TIMBER HARVEST AND DROUGHT INTERACT TO IMPACT OAK SEEDLING GROWTH AND SURVIVAL IN THE CENTRAL HARDWOOD FOREST}}{53}
+\contentsline {section}{\numberline {4.1}Introduction}{53}
+\contentsline {section}{\numberline {4.2}Methods}{57}
+\contentsline {subsection}{\numberline {4.2.1}Study Site}{57}
+\contentsline {subsection}{\numberline {4.2.2}Experimental Plots}{58}
+\contentsline {subsection}{\numberline {4.2.3}Oak Plantings}{59}
+\contentsline {subsection}{\numberline {4.2.4}Seedling Measurements}{59}
+\contentsline {subsection}{\numberline {4.2.5}Competition Measurement}{60}
+\contentsline {subsection}{\numberline {4.2.6}Analysis}{60}
+\contentsline {section}{\numberline {4.3}Results}{62}
+\contentsline {subsection}{\numberline {4.3.1}Survival}{62}
+\contentsline {subsection}{\numberline {4.3.2}Growth}{62}
+\contentsline {section}{\numberline {4.4}Discussion}{67}
+\contentsline {subsection}{\numberline {4.4.1}Site Conditions}{67}
+\contentsline {subsection}{\numberline {4.4.2}Competition and Herbivory}{70}
+\contentsline {subsection}{\numberline {4.4.3}Conclusions}{73}
+\contentsline {chapter}{\numberline {5}{SOEL: AN INDIVIDUAL-BASED FOREST GAP MODEL FOCUSED ON PROCESSES DRIVING DEMOGRAPHY OF EARLY OAK (\textit {QUERCUS}) LIFE STAGES}}{74}
+\contentsline {section}{\numberline {5.1}Introduction}{74}
+\contentsline {section}{\numberline {5.2}Selection of a Modeling Framework}{79}
+\contentsline {section}{\numberline {5.3}Model Overview}{80}
+\contentsline {subsection}{\numberline {5.3.1}Purpose}{80}
+\contentsline {subsection}{\numberline {5.3.2}Entities, State Variables, and Scales}{82}
+\contentsline {subsection}{\numberline {5.3.3}Process Overview and Scheduling}{83}
+\contentsline {subsection}{\numberline {5.3.4}Design Concepts}{83}
+\contentsline {subsection}{\numberline {5.3.5}Initialization}{84}
+\contentsline {subsection}{\numberline {5.3.6}Input}{84}
+\contentsline {section}{\numberline {5.4}Early Life History Submodel}{86}
+\contentsline {subsection}{\numberline {5.4.1}Acorn Production}{86}
+\contentsline {subsection}{\numberline {5.4.2}Acorn Predation, Dispersal, and Germination}{86}
+\contentsline {subsection}{\numberline {5.4.3}Oak Seedling Growth and Survival}{88}
+\contentsline {section}{\numberline {5.5}Contextual Forest Submodel}{89}
+\contentsline {subsection}{\numberline {5.5.1}Calculate Environmental Conditions}{89}
+\contentsline {subsection}{\numberline {5.5.2}Growth}{93}
+\contentsline {subsection}{\numberline {5.5.3}Survival}{97}
+\contentsline {subsection}{\numberline {5.5.4}Reproduction}{97}
+\contentsline {subsection}{\numberline {5.5.5}Timber Harvest}{98}
+\contentsline {section}{\numberline {5.6}Model Validation}{100}
+\contentsline {subsection}{\numberline {5.6.1}Acorn Production}{102}
+\contentsline {subsection}{\numberline {5.6.2}Oak Seedling and Sapling Density}{102}
+\contentsline {subsection}{\numberline {5.6.3}Mature Forest Structure}{104}
+\contentsline {section}{\numberline {5.7}Conclusion}{107}
+\contentsline {chapter}{\numberline {6}{INTERACTIONS OF TIMBER HARVEST WITH EARLY OAK LIFE HISTORY: IMPLICATIONS FOR OAK REGENERATION VIA APPLICATION OF A NOVEL INDIVIDUAL-BASED MODEL, SOEL}}{108}
+\contentsline {section}{\numberline {6.1}Introduction}{108}
+\contentsline {section}{\numberline {6.2}Methods}{111}
+\contentsline {subsection}{\numberline {6.2.1}Study Location}{111}
+\contentsline {subsection}{\numberline {6.2.2}Demographic Data}{113}
+\contentsline {subsection}{\numberline {6.2.3}Modeling Approach}{114}
+\contentsline {subsection}{\numberline {6.2.4}Global Sensitivity Analysis}{118}
+\contentsline {subsection}{\numberline {6.2.5}Interaction of Timing of Harvest with Acorn Production}{119}
+\contentsline {subsection}{\numberline {6.2.6}Interaction of Seed Predation and Midstory Removal}{120}
+\contentsline {subsection}{\numberline {6.2.7}Effects of Drought on Seedling Growth and Survival}{120}
+\contentsline {subsection}{\numberline {6.2.8}Analysis of Model Output}{121}
+\contentsline {section}{\numberline {6.3}Results}{122}
+\contentsline {subsection}{\numberline {6.3.1}Sensitivity Analysis}{122}
+\contentsline {subsection}{\numberline {6.3.2}Interaction of Timing of Harvest with Acorn Production}{122}
+\contentsline {subsection}{\numberline {6.3.3}Interaction of Seed Predation and Midstory Removal}{126}
+\contentsline {subsection}{\numberline {6.3.4}Effects of Drought on Seedling Growth and Survival}{126}
+\contentsline {section}{\numberline {6.4}Discussion}{131}
+\contentsline {subsection}{\numberline {6.4.1}Interaction of Harvest Timing with Acorn Production}{131}
+\contentsline {subsection}{\numberline {6.4.2}Interaction of Seed Predation and Midstory Removal}{132}
+\contentsline {subsection}{\numberline {6.4.3}Effects of Drought on Seedling Growth and Survival}{134}
+\contentsline {subsection}{\numberline {6.4.4}Synthesis and Outlook}{135}
+\contentsline {chapter}{\numberline {7}{CONCLUSION}}{138}
+\contentsline {section}{\numberline {7.1}Overview}{138}
+\contentsline {section}{\numberline {7.2}Conclusions From Empirical Research}{138}
+\contentsline {section}{\numberline {7.3}Conclusions from Simulation Modeling}{139}
+\contentsline {section}{\numberline {7.4}Future Directions}{140}
+\contentsline {chapter}{REFERENCES}{142}
+\contentsline {chapter}{\numberline {A}{CHAPTER 3 CODE}}{161}
+\contentsline {section}{\numberline {A.1}Mixed-Effects Proportional-Odds Logistic Regression Model}{161}
+\contentsline {chapter}{\numberline {B}{CHAPTER 4 CODE}}{163}
+\contentsline {section}{\numberline {B.1}Mixed-Effects Seedling Survival Model}{163}
+\contentsline {section}{\numberline {B.2}Mixed-Effects Seedling Growth Model}{164}
+\contentsline {chapter}{\numberline {C}{CHAPTER 5 CODE}}{166}
+\contentsline {section}{\numberline {C.1}{SOEL} {N}et{L}ogo Code}{166}
+\contentsline {section}{\numberline {C.2}{JABOWA} Implementation in {N}et{L}ogo}{181}
+\contentsline {section}{\numberline {C.3}R Function to Run {N}et{L}ogo Models}{188}
+\contentsline {chapter}{VITA}{193}
diff --git a/endnotes.sty b/endnotes.sty
new file mode 100644
index 0000000..ccd5cdd
--- /dev/null
+++ b/endnotes.sty
@@ -0,0 +1,339 @@
+%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
+% Save file as: ENDNOTES.STY
+% modified by --bg (B.Gaulle) 09/14/94 for:
+% 1) replace (why a 8bit char here?) by ^ as a default.
+% 2) force \catcode of > to be 12 (implied by \@doanenot).
+% by --bg again 03/22/95 for:
+% 3) reseting appropriate catcode of > in case it were
+% used as an active char before \@endanenote (was
+% pointed by Ch. Pallier).
+%%%% free to distribute by John_Lavagnino@Brown.edu on Thu, 23 Mar 1995
+%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
+%
+% ****************************************
+% * ENDNOTES *
+% ****************************************
+%
+% Date of this version: 24 September 1991.
+%
+% Based on the FOOTNOTES section of
+% LATEX.TEX (VERSION 2.09 - RELEASE OF 19 April 1986), with
+% "footnote" changed to "endnote" and "fn" changed to "en" (where
+% appropriate), with all the minipage stuff pulled out, and with
+% some small changes for the different operation of endnotes.
+%
+% Uses an extra external file, with .ENT extension, to hold the
+% text of the endnotes. This may be deleted after the run; a new
+% version is generated each time.
+%
+% This code does not obey \nofiles. Perhaps it should.
+%
+% John Lavagnino (lav@brandeis.bitnet), 9/23/88
+% Department of English and American Literature,
+% Brandeis University
+%
+% To turn all the footnotes in your documents into endnotes, say
+%
+% \let\footnote=\endnote
+%
+% in your preamble, and then add something like
+%
+% \newpage
+% \begingroup
+% \parindent 0pt
+% \parskip 2ex
+% \def\enotesize{\normalsize}
+% \theendnotes
+% \endgroup
+%
+% as the last thing in your document.
+%
+% ****************************************
+% * CHANGE LOG *
+% ****************************************
+%
+% JL Modified to include \addtoendnotes. JL, 10/22/89.
+%
+% JK Modification by J"org Knappen 25. 2. 1991:
+% JK
+% JK Introduced \notesname in the spirit of international \LaTeX.
+% JK \notesname is set per default to be {Notes}, but can easily
+% JK be redifined, e.g. for german language
+% JK \renewcommand{\notesname}{Anmerkungen}
+%
+% DW Modification by Dominik Wujastyk, London, 19 September 1991:
+% DW
+% DW Moved the line
+% DW \edef\@currentlabel{\csname p@endnote\endcsname\@theenmark}
+% DW out of the definition of \@endnotetext and into the definition
+% DW of \@doanenote so that \label and \ref commands work correctly in
+% DW endnotes. Otherwise, the \label just pointed to the last section
+% DW heading (or whatever) preceding the \theendnotes command.
+%
+% JL Revised documentation and macros. 24 Sept 1991.
+%
+% ****************************************
+% * ENDNOTE COMMANDS *
+% ****************************************
+%
+%
+% \endnote{NOTE} : User command to insert a endnote.
+%
+% \endnote[NUM]{NOTE} : User command to insert a endnote numbered
+% NUM, where NUM is a number -- 1, 2,
+% etc. For example, if endnotes are numbered
+% *, **, etc. within pages, then \endnote[2]{...}
+% produces endnote '**'. This command does not
+% step the endnote counter.
+%
+% \endnotemark[NUM] : Command to produce just the endnote mark in
+% the text, but no endnote. With no argument,
+% it steps the endnote counter before generating
+% the mark.
+%
+% \endnotetext[NUM]{TEXT} : Command to produce the endnote but no
+% mark. \endnote is equivalent to
+% \endnotemark \endnotetext .
+%
+% \addtoendnotes{TEXT} : Command to add text or commands to current
+% endnotes file: for inserting headings,
+% pagebreaks, and the like into endnotes
+% sections. TEXT a moving argument:
+% \protect required for fragile commands.
+%
+% ****************************************
+% * ENDNOTE USER COMMANDS *
+% ****************************************
+%
+% Endnotes use the following parameters, similar to those relating
+% to footnotes:
+%
+% \enotesize : Size-changing command for endnotes.
+%
+% \theendnote : In usual LaTeX style, produces the endnote number.
+%
+% \@theenmark : Holds the current endnote's mark--e.g., \dag or '1' or 'a'.
+%
+% \@makeenmark : A macro to generate the endnote marker from \@theenmark
+% The default definition is \hbox{$^\@theenmark$}.
+%
+% \@makeentext{NOTE} :
+% Must produce the actual endnote, using \@theenmark as the mark
+% of the endnote and NOTE as the text. It is called when effectively
+% inside a \parbox, with \hsize = \columnwidth. For example, it might
+% be as simple as
+% $^{\@theenmark}$ NOTE
+%
+%
+% ****************************************
+% * ENDNOTE PSEUDOCODE *
+% ****************************************
+%
+% \endnote{NOTE} ==
+% BEGIN
+% \stepcounter{endnote}
+% \@theenmark :=G eval (\theendnote)
+% \@endnotemark
+% \@endnotetext{NOTE}
+% END
+%
+% \endnote[NUM]{NOTE} ==
+% BEGIN
+% begingroup
+% counter endnote :=L NUM
+% \@theenmark :=G eval (\theendnote)
+% endgroup
+% \@endnotemark
+% \@endnotetext{NOTE}
+% END
+%
+% \@endnotetext{NOTE} ==
+% BEGIN
+% write to \@enotes file: "\@doanenote{ENDNOTE MARK}"
+% begingroup
+% \next := NOTE
+% set \newlinechar for \write to \space
+% write to \@enotes file: \meaning\next
+% (that is, "macro:->NOTE)
+% endgroup
+% END
+%
+% \addtoendnotes{TEXT} ==
+% BEGIN
+% open endnotes file if not already open
+% begingroup
+% let \protect to \string
+% set \newlinechar for \write to \space
+% write TEXT to \@enotes file
+% endgroup
+% END
+%
+% \endnotemark ==
+% BEGIN \stepcounter{endnote}
+% \@theenmark :=G eval(\theendnote)
+% \@endnotemark
+% END
+%
+% \endnotemark[NUM] ==
+% BEGIN
+% begingroup
+% endnote counter :=L NUM
+% \@theenmark :=G eval(\theendnote)
+% endgroup
+% \@endnotemark
+% END
+%
+% \@endnotemark ==
+% BEGIN
+% \leavevmode
+% IF hmode THEN \@x@sf := \the\spacefactor FI
+% \@makeenmark % put number in main text
+% IF hmode THEN \spacefactor := \@x@sf FI
+% END
+%
+% \endnotetext ==
+% BEGIN \@theenmark :=G eval (\theendnote)
+% \@endnotetext
+% END
+%
+% \endnotetext[NUM] ==
+% BEGIN begingroup counter endnote :=L NUM
+% \@theenmark :=G eval (\theendnote)
+% endgroup
+% \@endnotetext
+% END
+%
+% ****************************************
+% * ENDNOTE MACROS *
+% ****************************************
+%
+
+\@definecounter{endnote}
+\def\theendnote{\arabic{endnote}}
+
+% Default definition
+\def\@makeenmark{\hbox{$^{\@theenmark}$}}
+
+\newdimen\endnotesep
+
+\def\endnote{\@ifnextchar[{\@xendnote}{\stepcounter
+ {endnote}\xdef\@theenmark{\theendnote}\@endnotemark\@endnotetext}}
+
+\def\@xendnote[#1]{\begingroup \c@endnote=#1\relax
+ \xdef\@theenmark{\theendnote}\endgroup
+ \@endnotemark\@endnotetext}
+
+% Here begins endnote code that's really different from the footnote
+% code of LaTeX.
+
+\let\@doanenote=0
+\let\@endanenote=0
+
+\newwrite\@enotes
+\newif\if@enotesopen \global\@enotesopenfalse
+
+\def\@openenotes{\immediate\openout\@enotes=\jobname.ent\relax
+ \global\@enotesopentrue}
+
+% The stuff with \next and \meaning is a trick from the TeXbook, 382,
+% there intended for setting verbatim text, but here used to avoid
+% macro expansion when the footnote text is written. \next will have
+% the entire text of the footnote as one long line, which might well
+% overflow limits on output line length; the business with \newlinechar
+% makes every space become a newline in the \@enotes file, so that all
+% of the lines wind up being quite short.
+
+\long\def\@endnotetext#1{%
+ \if@enotesopen \else \@openenotes \fi
+ \immediate\write\@enotes{\@doanenote{\@theenmark}}%
+ \begingroup
+ \def\next{#1}%
+ \newlinechar='40
+ \immediate\write\@enotes{\meaning\next}%
+ \endgroup
+ \immediate\write\@enotes{\@endanenote}}
+
+% \addtoendnotes works the way the other endnote macros probably should
+% have, requiring the use of \protect for fragile commands.
+
+\long\def\addtoendnotes#1{%
+ \if@enotesopen \else \@openenotes \fi
+ \begingroup
+ \newlinechar='40
+ \let\protect\string
+ \immediate\write\@enotes{#1}%
+ \endgroup}
+
+% End of unique endnote code
+
+\def\endnotemark{\@ifnextchar[{\@xendnotemark
+ }{\stepcounter{endnote}\xdef\@theenmark{\theendnote}\@endnotemark}}
+
+\def\@xendnotemark[#1]{\begingroup \c@endnote #1\relax
+ \xdef\@theenmark{\theendnote}\endgroup \@endnotemark}
+
+\def\@endnotemark{\leavevmode\ifhmode
+ \edef\@x@sf{\the\spacefactor}\fi \@makeenmark
+ \ifhmode\spacefactor\@x@sf\fi\relax}
+
+\def\endnotetext{\@ifnextchar
+ [{\@xendnotenext}{\xdef\@theenmark{\theendnote}\@endnotetext}}
+
+\def\@xendnotenext[#1]{\begingroup \c@endnote=#1\relax
+ \xdef\@theenmark{\theendnote}\endgroup \@endnotetext}
+
+
+% \theendnotes actually prints out the endnotes.
+
+% The user may want separate endnotes for each chapter, or a big
+% block of them at the end of the whole document. As it stands,
+% either will work; you just say \theendnotes wherever you want the
+% endnotes so far to be inserted. However, you must add
+% \setcounter{endnote}{0} after that if you want subsequent endnotes
+% to start numbering at 1 again.
+
+% \enoteformat is provided so user can specify some special formatting
+% for the endnotes. It needs to set up the paragraph parameters, start
+% the paragraph, and print the label. The \leavemode stuff is to make
+% and undo a dummy paragraph, to get around the games \section*
+% plays with paragraph indenting.
+
+\def\notesname{Notes}% <------ JK
+\def\enoteheading{\section*{\notesname
+ \@mkboth{\uppercase{\notesname}}{\uppercase{\notesname}}}%
+ \leavevmode\par\vskip-\baselineskip}
+
+\def\enoteformat{\rightskip\z@ \leftskip\z@ \parindent=1.8em
+ \leavevmode\llap{\hbox{$^{\@theenmark}$}}}
+
+\def\enotesize{\footnotesize}
+
+% The definition of \ETC. is needed only for versions of TeX prior
+% to 2.992. Those versions limited \meaning expansions to 1000
+% characters; in 2.992 and beyond there is no limit. At Brandeis the
+% BIGLATEX program changed the code in the token_show procedure of
+% TeX to eliminate this problem, but most ``big'' versions of TeX
+% will not solve this problem.
+
+\def\theendnotes{\immediate\closeout\@enotes \global\@enotesopenfalse
+ \begingroup
+ \makeatletter
+% the following is to save catcode of ``>'' and restore it in \@endanenote
+\edef\@tempa{`\string >}%
+\ifnum\catcode\@tempa=11\let\@ResetGT\relax% accepts also that > were active
+ \else\edef\@ResetGT{\noexpand\catcode\@tempa=\the\catcode\@tempa}%
+\fi%
+ \catcode`>=12% char > will be read as char so force it to \catcode 12 --bg\edef\GOfrench{`\string @}% temp def futher correctly defined
+ \def\@doanenote##1##2>{\def\@theenmark{##1}\par\begingroup
+ \@ResetGT%\catcode`>=13
+ \edef\@currentlabel{\csname p@endnote\endcsname\@theenmark} %DW
+ \enoteformat}
+ \def\@endanenote{\par\endgroup}%
+ \def\ETC.{\errmessage{Some long endnotes will be truncated; %
+ use BIGLATEX to avoid this}%
+ \def\ETC.{\relax}}
+ \enoteheading
+ \enotesize
+ \input{\jobname.ent}%
+ \endgroup}
+%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
diff --git a/figures/fig1-1.R b/figures/fig1-1.R
new file mode 100644
index 0000000..a7eb6eb
--- /dev/null
+++ b/figures/fig1-1.R
@@ -0,0 +1,49 @@
+#############################################################################
+##Figure showing oak dbh distribution at HEE sites
+#############################################################################
+#Read in data
+data.bo = as.numeric(na.omit(read.csv('data/hee_oakdbh.csv',header=FALSE)[,1]))*2.54
+data.sm = as.numeric(na.omit(read.csv('data/hee_oakdbh.csv',header=FALSE)[,2]))*2.54
+
+#Plot histograms of dbh for oak vs. sugar maple
+par(mfrow=c(2,1),mar=c(4,4,3,2))
+hist(data.bo, breaks=15,xlim=c(0,40),main="Black Oak", col=rgb(red=75,green=142,blue=26, maxColorValue=255), xlab="DBH")
+abline(v=median(data.bo,na.rm=TRUE),lwd=3)
+hist(data.sm, breaks=15,xlim=c(0,40), main="Sugar Maple", col=rgb(red=244,green=125,blue=66, maxColorValue=255), xlab="DBH")
+abline(v=median(data.sm),lwd=3)
+
+#Black and white version
+par(mfrow=c(2,1),mar=c(4,4,3,2))
+hist(data.bo, breaks=15,xlim=c(0,100),main="Black Oak", col="gray38", xlab="d.b.h. (cm)", freq=FALSE)
+abline(v=median(data.bo,na.rm=TRUE),lwd=3)
+hist(data.sm, breaks=15,xlim=c(0,100), main="Sugar Maple", col="gray75", xlab="d.b.h. (cm)", freq=FALSE)
+abline(v=median(data.sm),lwd=3)
+
+#Dissertation Version using Computer Modern Font
+#install.packages('extrafont')
+library(extrafont)
+font_install('fontcm')
+loadfonts()
+
+b = hist(data.bo,xlim=c(0,100),main="Black Oak", col="gray38", xlab="dbh")
+b$density = b$counts/sum(b$counts)
+
+s = hist(data.sm, breaks=15,xlim=c(0,100), main="Sugar Maple", col="gray75", xlab="dbh", freq=FALSE)
+s$density = s$counts/sum(s$counts)
+
+#library(extrafont)
+#font_install('fontcm')
+#loadfonts()
+pdf(file="../dissertation/figures/fig1-1.pdf",width=5,height=5,family="CM Roman",pointsize=10)
+par(mfrow=c(2,1),mar=c(4,4,3,2))
+#hist(b,xlim=c(0,100),main="Black Oak", col="gray38", xlab="dbh", freq=F)
+plot(b,xlim=c(0,100),main="Black Oak", col="gray38", xlab="dbh (cm)", freq=F,
+ ylab="Proportion of Stems",ylim=c(0,.3))
+abline(v=median(data.bo,na.rm=TRUE),lwd=3)
+#hist(s, breaks=15,xlim=c(0,100), main="Sugar Maple", col="gray75", xlab="dbh", freq=FALSE)
+plot(s,xlim=c(0,100), main="Sugar Maple", col="gray75", xlab="dbh (cm)", freq=FALSE,
+ ylab="Proportion of Stems",ylim=c(0,0.3))
+abline(v=median(data.sm),lwd=3)
+dev.off()
+Sys.setenv(R_GSCMD = "C:/Program Files/gs/gs9.18/bin/gswin64c.exe")
+embed_fonts("../dissertation/figures/fig1-1.pdf")
diff --git a/figures/fig1-1.pdf b/figures/fig1-1.pdf
new file mode 100644
index 0000000..d44a6a9
--- /dev/null
+++ b/figures/fig1-1.pdf
Binary files differ
diff --git a/figures/fig2-1.pdf b/figures/fig2-1.pdf
new file mode 100644
index 0000000..2b56402
--- /dev/null
+++ b/figures/fig2-1.pdf
Binary files differ
diff --git a/figures/fig2-2.pdf b/figures/fig2-2.pdf
new file mode 100644
index 0000000..fd748e7
--- /dev/null
+++ b/figures/fig2-2.pdf
Binary files differ
diff --git a/figures/fig2-3.pdf b/figures/fig2-3.pdf
new file mode 100644
index 0000000..d03665e
--- /dev/null
+++ b/figures/fig2-3.pdf
Binary files differ
diff --git a/figures/fig2-4.pdf b/figures/fig2-4.pdf
new file mode 100644
index 0000000..2761362
--- /dev/null
+++ b/figures/fig2-4.pdf
Binary files differ
diff --git a/figures/fig3-1.pdf b/figures/fig3-1.pdf
new file mode 100644
index 0000000..c74a65a
--- /dev/null
+++ b/figures/fig3-1.pdf
Binary files differ
diff --git a/figures/fig3-2.pdf b/figures/fig3-2.pdf
new file mode 100644
index 0000000..1da4f89
--- /dev/null
+++ b/figures/fig3-2.pdf
Binary files differ
diff --git a/figures/fig3-3.pdf b/figures/fig3-3.pdf
new file mode 100644
index 0000000..a78ed0e
--- /dev/null
+++ b/figures/fig3-3.pdf
Binary files differ
diff --git a/figures/fig4-1.pdf b/figures/fig4-1.pdf
new file mode 100644
index 0000000..a022616
--- /dev/null
+++ b/figures/fig4-1.pdf
Binary files differ
diff --git a/figures/fig4-2.pdf b/figures/fig4-2.pdf
new file mode 100644
index 0000000..9e5f2e9
--- /dev/null
+++ b/figures/fig4-2.pdf
Binary files differ
diff --git a/figures/fig5-1.pdf b/figures/fig5-1.pdf
new file mode 100644
index 0000000..9bfa445
--- /dev/null
+++ b/figures/fig5-1.pdf
Binary files differ
diff --git a/figures/fig5-2.pdf b/figures/fig5-2.pdf
new file mode 100644
index 0000000..aa75fcb
--- /dev/null
+++ b/figures/fig5-2.pdf
Binary files differ
diff --git a/figures/fig5-3.pdf b/figures/fig5-3.pdf
new file mode 100644
index 0000000..95a0f75
--- /dev/null
+++ b/figures/fig5-3.pdf
Binary files differ
diff --git a/figures/fig5-4.pdf b/figures/fig5-4.pdf
new file mode 100644
index 0000000..ee1d307
--- /dev/null
+++ b/figures/fig5-4.pdf
Binary files differ
diff --git a/figures/fig5-5.pdf b/figures/fig5-5.pdf
new file mode 100644
index 0000000..b9cef82
--- /dev/null
+++ b/figures/fig5-5.pdf
Binary files differ
diff --git a/figures/fig5-6.pdf b/figures/fig5-6.pdf
new file mode 100644
index 0000000..93ed8ba
--- /dev/null
+++ b/figures/fig5-6.pdf
Binary files differ
diff --git a/figures/fig5-7.pdf b/figures/fig5-7.pdf
new file mode 100644
index 0000000..7075f5f
--- /dev/null
+++ b/figures/fig5-7.pdf
Binary files differ
diff --git a/figures/fig5-8.pdf b/figures/fig5-8.pdf
new file mode 100644
index 0000000..907f03f
--- /dev/null
+++ b/figures/fig5-8.pdf
Binary files differ
diff --git a/figures/fig6-1.pdf b/figures/fig6-1.pdf
new file mode 100644
index 0000000..45c6bdd
--- /dev/null
+++ b/figures/fig6-1.pdf
Binary files differ
diff --git a/figures/fig6-2.pdf b/figures/fig6-2.pdf
new file mode 100644
index 0000000..dda32ec
--- /dev/null
+++ b/figures/fig6-2.pdf
Binary files differ
diff --git a/figures/fig6-3.pdf b/figures/fig6-3.pdf
new file mode 100644
index 0000000..69aae87
--- /dev/null
+++ b/figures/fig6-3.pdf
Binary files differ
diff --git a/figures/fig6-4.pdf b/figures/fig6-4.pdf
new file mode 100644
index 0000000..c2ccbdc
--- /dev/null
+++ b/figures/fig6-4.pdf
Binary files differ
diff --git a/front.tex b/front.tex
new file mode 100644
index 0000000..0e4def9
--- /dev/null
+++ b/front.tex
@@ -0,0 +1,24 @@
+\begin{dedication}
+For my family.
+\end{dedication}
+
+\begin{acknowledgments}
+My greatest thanks go to my advisor, Robert Swihart, for his wise and patient counsel. My committee members, Harmony Dalgleish, Michael Saunders, Michael Steele, and Patrick Zollner, provided invaluable advice and numerous improvements to this dissertation. Also, I would like to thank my lab group, Nathanael Lichti, Natasha Urban, Byju Govindan, Mekala Sundaram, Rita Blythe, Timothy Smyser, Jacob Berl, and Dana Nelson for their support. I am grateful to Jeff Riegel, Andy Meier, Rebecca Kalb, and Patrick Ma for their role in data collection. Numerous field technicians also contributed to my research.
+
+This dissertation is a contribution of the Hardwood Ecosystem Experiment, a
+partnership of the Indiana Department of Natural Resources, Purdue University,
+Ball State University, Indiana State University, Drake University, Indiana
+University-Pennsylvania, and the Nature Conservancy. Funding for the project
+was provided by the Indiana Division of Forestry, the Department of Forestry
+and Natural Resources at Purdue University, and the Purdue University Graduate School.
+\end{acknowledgments}
+
+\tableofcontents
+\listoftables
+\listoffigures
+
+\begin{abstract}
+
+Oak (\textit{Quercus}) is a dominant component of the forest canopy in many deciduous forests of the eastern United States. In the past century, a pattern of anthropogenic fire suppression throughout the eastern United States has created conditions that favor more shade-tolerant tree species at the expense of oak regeneration. Loss of oak as dominant canopy component would have a powerful impact on forest ecosystems; oak acorns are a keystone food resource for greater than 50 species of mammals and birds in these forests, due to their ubiquity and nutritional content. In response, forest managers have developed and applied silvicultural techniques, including timber harvesting, to emulate forest disturbance and regenerate oak. Timber harvesting can impact key trophic interactions with animals during early oak life history, thus potentially impacting the oak regeneration process indirectly; however, these impacts are not well understood. To address this research gap, I examined the effects of clearcut and shelterwood harvesting approaches on key early life history parameters involved in acorn predation, dispersal, seedling survival, and seedling growth for black oak (\textit{Q. velutina}) and white oak (\textit{Q. alba}) in southern Indiana. I found that the initial phase of shelterwood harvesting (midstory removal) resulted in increased acorn dispersal and reduced survival but that these changes were dependent on dispersal agent and acorn species. At the seedling life stage, growth was highest but survival was lowest (for black oak) in clearcut openings. The lower survival in clearcuts was likely due to drought during two consecutive growing seasons. Herbivory by white-tailed deer (\textit{Odocoileus virginianus}) impacted seedling growth but not survival; rates of browsing were low throughout the study. With this field data, I developed an individual-based, spatially-explicit simulation model of early oak life history connecting early life history parameters with consequences for oak regeneration. Simulation results indicated that the observed differences in acorn predation and dispersal parameters in shelterwoods had a small, but significant, negative impact on accumulation of oak seedlings (\textless 1.4 m height) in the period following harvest. However, this effect was not carried through to density of oak saplings ($\geq$1.4 m). Greater disturbance in the second and third phases of the shelterwood harvest may result in stronger impacts on seed predation and dispersal. Increasing the frequency of drought years in the simulation model had a powerful negative effect on both oak seedlings and saplings, and the effect increased with intensity of harvest disturbance. The impact of drought is concerning given predictions that climate change will bring more frequent drought to the Midwestern United States; the ultimate impacts on oak regeneration success are unclear since drought may affect oak competitors even more strongly. The simulation model I developed is a flexible tool that could be used to examine a wide range of animal impacts on early oak life history and project their impacts on oak regeneration success.
+
+\end{abstract}
diff --git a/genmodel.sh b/genmodel.sh
new file mode 100644
index 0000000..64f657d
--- /dev/null
+++ b/genmodel.sh
@@ -0,0 +1,17 @@
+#!/bin/sh
+
+##########################
+# Extract source code from NetLogo model and write it to a file
+# Adapted from https://github.com/mado89/netlogo-latex
+
+temp=`grep -n "\@\#\$\#\@\#\$\#\@" ../oak-lifecycle/oak_ibm.nlogo | head -n 1`
+LN=`echo $temp | awk -F ':' '{ print $1 }'`
+#LN=`grep -n "\@\#\$\#\@\#\$\#\@" /home/kkellner/analysis/oak-lifecycle/oak_ibm.nlogo | head -n 1 | awk -F ':' '{print $1}'`
+LN="$((LN - 1))"
+head -n $LN ../oak-lifecycle/oak_ibm.nlogo > netlogomodel.tex
+
+temp=`grep -n "\@\#\$\#\@\#\$\#\@" ../oak-lifecycle/oak_ibm_jabowa.nlogo | head -n 1`
+LN=`echo $temp | awk -F ':' '{ print $1 }'`
+#LN=`grep -n "\@\#\$\#\@\#\$\#\@" /home/kkellner/analysis/oak-lifecycle/oak_ibm.nlogo | head -n 1 | awk -F ':' '{print $1}'`
+LN="$((LN - 1))"
+head -n $LN ../oak-lifecycle/oak_ibm_jabowa.nlogo > jabowamodel.tex \ No newline at end of file
diff --git a/introduction.tex b/introduction.tex
new file mode 100644
index 0000000..ac177fc
--- /dev/null
+++ b/introduction.tex
@@ -0,0 +1,57 @@
+\chapter{{INTRODUCTION}}
+\section{Oak Forests in the Eastern United States}
+
+Oak (\textit{Quercus}) is a dominant component of the forest canopy in many eastern deciduous forests \citep{John09}. For example, in the hardwood forests of southern Indiana, oak makes up 44-76\% of the total tree basal area \citep{Saun13}. Understanding the circumstances under which oak became so dominant has been a focus of research in the past several decades - the answer involves an interaction between the physiological characteristics of oaks and the past history of disturbance events in the eastern United States. While there is significant variation within the \textit{Quercus} genus, oaks generally have low-to-intermediate tolerance to shade \citep{Niin06}. This characteristic positions oak as an early to mid-successional species group, requiring relatively large canopy openings in order to regenerate successfully \citep{John09}.
+
+Historically, fire was a primary mechanism by which large canopy openings were created in the forests of the eastern U.S., whether naturally occurring (e.g., via lightning strikes) or deliberately set by Native Americans (and later by European settlers) to clear land or to encourage beneficial vegetation \citep{Bros01, Mcew11}. Oaks are well-adapted to fire, with thick bark and the ability to re-sprout prolifically from the root collar after being top-killed \citep{Crow88, Abra96, John09}. Therefore, fire creates conditions well-suited for oak recruitment, and a historic disturbance regime of repeated fire maintained oak as a dominant component of forest canopies \citep{Abra92, Bros01}.
+
+\section{Role of Oak as a Foundational Species}
+
+Simply by virtue of its large contribution to total forest biomass, oaks play a critical role in forest ecosystems where they are abundant, and are therefore considered a foundation species \citep{Elli05}. They play a key role in habitat structure; relative to forests dominated by more shade-tolerant tree species, oak forests allow more light to reach the forest floor, promoting diverse understory vegetation \citep{Fral04}. Oak seedlings are a preferred food resource for herbivores like the white-tailed deer \textit{Odocoileus virginianus} and eastern cottontail rabbit \textit{Sylvilagus floridanus}, and oak leaves are consumed by a wide variety of herbivorous insects in the orders Coleoptera, Hymenoptera, Lepidoptera, and Orthoptera.
+
+More importantly, the seeds produced by oaks (acorns) are a crucial food resource for forest wildlife and insects. Acorns are a prized food source because they are large (relative to other seeds), containing significant amounts of nutrients and lipids, and are less perishable than other food types due to a hard outer shell and a delayed germination schedule (for some oak species) \citep{Wood05, Steel06}. As a result, acorns are a significant portion of the diet of at least 44 wildlife species in eastern deciduous forests, ranging from the white-tailed deer and black bear (\textit{Ursus americanus}), to wild turkey (\textit{Meleagris gallopavo}) and blue jay (\textit{Cyanocitta cristata}) \citep{Mcsh07}. Rodents (e.g. genera \textit{Sciurus}, \textit{Peromyscus}, \textit{Neotoma}, and \textit{Tamias}) are particularly dependent on acorns, and store large numbers of acorns to aid in overwinter survival \citep{Gohe03, Cast02}. Additionally, there is a genus of weevils (\textit{Curculio}; the acorn weevils) that require acorns as hosts for their eggs and larvae \citep{Gibs72, Gibs82}. The importance of acorns as a food resource for wildlife likely increased greatly following the functional extirpation of American chestnut (\textit{Castanea dentata}), another prolific seed producer, from the eastern U.S. in the early 1900s \citep{Dalg12}.
+
+Acorn production is closely synchronized between individuals of the same oak species (and sometimes between different species) at a large spatial scale (multiple kilometers). Additionally, the size of acorn crops can vary greatly from year to year, from barely any acorns produced one year to a bumper crop the next \citep{Koen02, Lusk07, Kell14}. Oaks therefore create resource pulses \citep{Dalg12}, the effects of which can cascade through multiple trophic levels. For example, the abundance of granivorous small mammals is tightly linked with acorn production, fluctuating with the acorn crop after a lag \citep{Wolf96, Mcsh00, Kell13}. Mast-induced peaks in small mammal populations can wreak havoc on songbird nests (as predators) and affect the abundance, survival, and behavior of raptors and rattlesnakes (as prey; \citealp{Mcsh00, Clot07, Schm08, Olso15a}). Even more complex trophic interactions were described by \citet{Ostf96}, who found that following abundant mast crops, deer tick (\textit{Ixodes scapularis}) populations increased together with their small mammal intermediate hosts, increasing the risk of Lyme disease transmission to humans. Further, high small mammal abundance following large acorn crops prevented gypsy moth population explosions (since small mammals eat gypsy moth pupae; \citealp{Ostf96}).
+
+\section{Oak Regeneration Failure}
+
+Over the past several decades, oak has begun to lose its position as a dominant canopy species in eastern decidious forests \citep{Abra03, Aldr05}. This trend is visible in tree diameter distribution data from southern Indiana forests: there are many large, dominant, overstory oaks in the canopy, but almost no oak saplings in the understory. Instead, the understory is dominated by shade tolerant tree species, such as maple (\textit{Acer} spp.) and American beech (\textit{Fagus grandifolia}) (Figure 1.1). Under these conditions, when an oak in the overstory is lost to mortality, it is very likely to be replaced by one of these shade tolerant species instead of another oak (i.e., oak regeneration failure). Since shade-tolerant tree species in the canopy typically allow less light to reach the forest floor and create more mesophytic soil conditions, a positive feedback loop is created that further stifles regeneration and growth of shade-intolerant, xeric oaks \citep{Nowa08}. Unchecked, this process (also called 'mesophication' since shade-tolerant, mesophytic tree species are benefitting) will lead to oaks no longer being an important part of the forest canopy \citep{Nowa08}. Loss of oak as a dominant canopy species could have serious consequences for forest ecosystems, given the importance of acorns as food resource (described in section 1.2; \citealp{Mcsh07}). For this and other reasons (e.g., the importance of oak as a timber species), understanding and addressing oak regeneration failure has become a priority for forest researchers, managers, and silviculturists \citep{Mcew11}.
+
+\begin{figure}
+\centering
+\includegraphics[scale=0.9]{figures/fig1-1.pdf}
+\caption{Tree diameter at breast height (dbh) distributions for black oak (\textit{Quercus velutina}) and sugar maple (\textit{Acer saccharum}) at the Hardwood Ecosystem Experiment study site in southern Indiana. Black vertical lines represent the median of each distribution.}
+\label{fig:1.1}
+\end{figure}
+
+The primary reason for oak regeneration failure can be traced back to the disturbance regimes that promoted its dominance in the first place. In the past, a regime of frequent, intermediate-intensity disturbances created canopy openings suitable for oak to successfully regenerate (as described in section 1.1; \citealp{Bros01}). In the past century, however, this disturbance regime has been interrupted, particularly via fire suppression throughout the eastern United States \citep{Nowa08}. Without these frequent disturbances, conditions have been consistently ideal for shade-tolerant, fire-intolerant tree species to flourish in the understory at the expense of oak. This process has been accelerated by silvicultural practices that have focused on single-tree selection harvests - the canopy openings created by removal of a small number of trees provide too little light for oaks to regenerate and instead promote shade tolerant species \citep{John09}.
+
+However, light availability and related environmental conditions are not the only factors that play a role in the oak regeneration process. Oaks are under intense pressure from herbivores and seed predators using the tree and its acorns as a food resource (described in section 1.2). These key trophic interactions occur entirely within the early life history of oak that is crucial in determining if oak regeneration will be competitive in the future. Seed predators are capable of consuming nearly the entire acorn crop in a given year, but some predators also act as dispersal agents and so might have a net positive effect on oak \citep{Bell05, Vand05, Kell14}. Likewise, herbivory (particularly via white-tailed deer when they are abundant) can suppress oak seedlings and saplings and keep them from being competitive \citep{Marq76, Roon03, Cote04}.
+
+\section{Silviculture as a Means to Promote Oak Regeneration}
+
+Harvesting timber (and thus creating canopy openings) is one way to emulate the natural disturbances that once maintained oak as a canopy dominant tree. However, not all silvicultural techniques are well-suited to regenerate oak successfully, and some may actually favor other, less desirable tree species at the expense of oak \citep{John09}. The large canopy openings necessary for adequate light has prompted even-aged silviculture as the approach when regenerating oak. In an even-aged silvicultural system a large, multiple-hectare area is harvested, to create a future stand in which all canopy trees are of a similar age and size \citep{John09}. Many uneven-aged selection systems do not not create large enough openings for oak to regenerate successfully \citep{Dey02}.
+
+The most often used even-aged silvicultural technique is clearcutting, where all trees in the area to be harvested are taken at once and the entire area is reverted to an early successional state. Depending on site conditions and other factors, oaks may be outcompeted in clearcuts by faster-growing shade-intolerants like tulip poplar (\textit{Liriodendron tulipifera}) and sassafras (\textit{Sassafras albidum}) \citep{Jenk98, Morr08, Swai13}. An alternative to the clearcut is the shelterwood harvest, which has the same target even-aged stand but gets there via a series of smaller partial removals of the overstory canopy instead of a single large one \citep{Loft90}. In the initial partial harvests, the stand is thinned to increase light at the forest floor for oak regeneration while maintaining large oaks as seed sources. This approach serves to promote oak over both its shade-tolerant and shade-intolerant competitors, and has met with some success \citep{Dey08}.
+
+\section{Interaction of Silviculture with Seed Predators and Herbivores}
+
+Regardless of silvicultural method, the most important concern is making sure there is adequate oak regeneration present prior to harvest \citep{John09}. Acorn predators and herbivores play a crucial role in determining the abundance and distribution of non-sprout origin oak regeneration (section 1.3). The contribution of these animals to the success (or failure) of silviculture for oak regeneration is not always an important consideration when designing a silvicultural approach, despite their potentially powerful effect on early oak life history. It is a complex interaction: the disturbance of timber harvesting not only affects the environment for oaks, but also can change the abundance and behavior of seed predators and herbivores, and therefore the nature of their trophic interactions with oak. For example, a clearcut harvest conducted soon after a year or years in which acorns were heavily predated could mean too few oak seedlings to take advantage of the new light conditions. Alternatively, the changes in the canopy following an initial shelterwood harvest might change the behavior of acorn predators who also act as dispersal agents, or promote an influx of deer to the area due to an increase in understory vegetation.
+
+Unfortunately, few studies have explored how timber harvesting might affect trophic interactions involving oak, and what the subsequent consequences for regeneration might be. There is some short-term evidence that acorn weevil infestation rates may change following timber harvest (\citealp{Kell14}, \textit{but see} \citealp{Bell05, Lomb08}), but there were no changes in acorn removal by small mammal predators \citep{Bell05, Kell14}. More research is necessary to fully understand the effects of timber harvesting on early oak life history and related trophic interactions.
+
+\section{Goals of this Dissertation}
+
+My dissertation aims to address the gap in the literature described above, by examining the effects of timber harvesting on the trophic interactions between oak and its predators and herbivores. Specifically, I will address the following questions:
+
+\begin{enumerate}
+ \item What are the effects of even-aged silviculture for oak on:
+ \begin{enumerate}
+ \item Acorn predation and dispersal (Chapter 2)
+ \item Herbivory on oak seedlings by deer (Chapter 3)
+ \item Seedling survival and growth (Chapter 4)
+ \end{enumerate}
+ \item What are the projected impacts of the trophic interactions in (1) on long-term maintenance of oak as a dominant canopy species in eastern deciduous forests? (Chapters 5 and 6).
+\end{enumerate}
+
+Studying and quantifying the early life history of oak, not to mention its entire life history, are not trivial tasks. To facilitate this research, I worked within the framework of an established experimental system, the Hardwood Ecosystem Experiment (HEE) \citep{Kalb13}. The HEE is a long-term, landscape-scale study of the effects of silviculture on hardwood forest ecosystems, set in the forests of south-central Indiana, USA. Data collection began in 2006 and the first round of timber harvests were conducted in 2008-2009. The HEE is a collaboration of several universities and many PIs and grad students, allowing for the collection of detailed datasets spanning nearly ten years and multiple taxonomic groups. This allowed me to bring a much stronger group of datasets to bear than I possibly could have working on my own. The framework and long-term datasets of the HEE, together with my own fieldwork and a diverse collection of quantitative approaches (including generalized linear mixed models, EM-MCMC, and simulation modeling) enabled me to address the difficult research questions in the following chapters. \ No newline at end of file
diff --git a/jabowamodel.tex b/jabowamodel.tex
new file mode 100644
index 0000000..44e7fdd
--- /dev/null
+++ b/jabowamodel.tex
@@ -0,0 +1,560 @@
+extensions [profiler]
+globals [
+ basal-area basal-area-ft basal-area-ovs
+ qdbh qdbh-in qdbh-ovs
+ dens dens-ac dens-ovs
+
+ ba-oak ba-oak-ovs
+ qdbh-oak qdbh-oak-ovs
+ dens-oak dens-oak-ovs
+
+ ba-map ba-map-ovs
+ qdbh-map qdbh-map-ovs
+ dens-map dens-map-ovs
+
+ ba-pop ba-pop-ovs
+ qdbh-pop qdbh-pop-ovs
+ dens-pop dens-pop-ovs
+
+ prop-oak prop-tol prop-intol
+ sitequal-boak sitequal-woak sitequal-maple sitequal-poplar
+ harvest-year shelter-phase
+ xmin xmax ymin ymax
+ seedlings-class4
+ ]
+
+turtles-own [in-core light age dbh height actual-growth ba lai fAL
+ Dmax Hmax Amax b2 b3 C G light-tol intrinsic-mortality min-increment growth-mortality]
+
+breed [oaks oak]
+oaks-own [species]
+
+breed [maples maple]
+breed [poplars poplar]
+
+patches-own [plight]
+
+
+;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;
+;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;
+
+
+to setup
+ clear-all
+ reset-ticks
+
+ set xmin buffer
+ set xmax buffer + x-core - 1
+ set ymin buffer
+ set ymax buffer + y-core - 1
+
+ let adjust (x-core + buffer * 2) * (y-core + buffer * 2) / 100
+
+ resize-world 0 (x-core + 2 * buffer - 1) 0 (y-core + 2 * buffer - 1)
+ ;resize-world (-1 * (xcutoff + buffer - 1)) (xcutoff + buffer) (-1 * (ycutoff + buffer - 1)) (ycutoff + buffer)
+
+ calc-site-quality
+
+ if HEE-mean = TRUE [
+ init-stand adjust TRUE 89 11 9 95 499 163
+ calc-global-vars
+ setup-plots
+ ]
+
+ ask patches [color-patches]
+ set harvest-year burnin
+ set shelter-phase 1
+
+end
+
+;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;
+;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;
+
+to go
+
+ ask patches [
+ regenerate
+ color-patches
+ ]
+
+ ask turtles [
+ calc-light
+ ]
+
+ ask turtles [
+ grow
+ check-survival
+ ]
+
+ conduct-harvest
+
+ calc-global-vars
+
+ tick
+end
+
+
+;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;
+;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;
+
+
+;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;
+;;;;;;;; Basic JABOWA Procedures ;;;;;;;;
+;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;
+
+
+to allometerize [dbh-input]
+ ;Based on JABOWA, Botkin 1993
+ ;b2 = 2*(Hmax - 137) / Dmax and b3 = (Hmax - 137) / Dmax ^2
+ set height (137 + b2 * (dbh-input * 100) - b3 * ((dbh-input * 100) ^ 2)) / 100
+ ;leaf weight based on dbh with C species-specific
+ ;Exponent can be anywhere from 1.5 to 3 (Botkin 1993)
+ set lai (C * (dbh-input * 100) ^ (2))
+ set ba (pi * (dbh / 2) ^ 2) ;Basal area
+end
+
+
+to-report max-growth-increment [dbh-input height-input] ;Based on JABOWA, Botkin 1993
+ let num (G * dbh-input * ((1 - (dbh-input * (137 + (b2 * dbh-input) - (b3 * dbh-input ^ 2))) / (Dmax * Hmax))))
+ let denom (274 + (3 * b2 * dbh-input) - (4 * b3 * dbh-input ^ 2))
+ report (num / denom)
+end
+
+
+to-report light-growth-index [light-input tol-input]
+ ;Based on Botkin 1993 and Bonan 1990
+ if tol-input = "low" [
+ report (2.24 * (1 - exp(-1.136 * (light-input - 0.08))))
+ ]
+ if tol-input = "intermediate" [
+ report (1.371 * (1 - exp(-2.227 * (light-input - 0.05))))
+ ]
+ if tol-input = "high" [
+ report (1 - exp(-4.64 * (light-input - 0.05)))
+ ]
+end
+
+
+to-report degree-days-index [degdays-input degd-min-input degd-max-input]
+ let tdegd (4 * (degdays-input - degd-min-input) * (degd-max-input - degdays-input) / ((degd-max-input - degd-min-input) ^ 2))
+ report max list 0 tdegd
+end
+
+
+to-report saturation-index [wt-input wt-dist-min-input]
+ let wefi (1 - (wt-dist-min-input / wt-input))
+ report max list 0 wefi
+end
+
+
+to-report nitrogen-index [N-input N-tol-input]
+ if N-tol-input = "intolerant" [
+ let a1 2.99 let a2 0.00175 let a3 207.43 let a4 -5 let a5 2.9 let a6 3.671
+ let lambdaN (a1 * (1 - 10 ^ (-1 * a2 * (N-input + a3))))
+ report (a4 + a5 * lambdaN) / a6
+ ]
+ if N-tol-input = "intermediate" [
+ let a1 2.94 let a2 0.00234 let a3 117.52 let a4 -1.2 let a5 1.3 let a6 2.622
+ let lambdaN (a1 * (1 - 10 ^ (-1 * a2 * (N-input + a3))))
+ report (a4 + a5 * lambdaN) / a6
+ ]
+ if N-tol-input = "tolerant" [
+ let a1 2.79 let a2 0.00179 let a3 219.77 let a4 -0.6 let a5 1.0 let a6 2.190
+ let lambdaN (a1 * (1 - 10 ^ (-1 * a2 * (N-input + a3))))
+ report (a4 + a5 * lambdaN) / a6
+ ]
+end
+
+
+to-report wilt-index [wilt-input wilt-max]
+ ;wilt is difference between potential and actual evapotranspiration divided by potential evapotranspiration
+ let wifi (1 - (wilt-input / wilt-max) ^ 2)
+ report max list 0 wifi
+end
+
+
+to grow
+
+ let max-growth (max-growth-increment (dbh * 100) (height * 100))
+
+ set fAL light-growth-index light light-tol
+ let sitequal 1
+ ifelse breed = oaks [ifelse species = "WO" [set sitequal sitequal-woak] [set sitequal sitequal-boak]]
+ [ifelse breed = maples [set sitequal sitequal-maple]
+ [set sitequal sitequal-poplar]]
+
+ set actual-growth (max-growth * fAL * sitequal)
+ set dbh (dbh * 100 + actual-growth) / 100
+ allometerize dbh
+
+ if dbh >= 0.1 and shape = "square" [
+ if breed = oaks [set size 1 set shape "tree"]
+ if breed = maples [set size 1 set shape "tree"]
+ if breed = poplars [set size 1 set shape "tree"]
+ ]
+
+end
+
+
+to check-survival
+
+ if random-float 1 < intrinsic-mortality [die]
+ if actual-growth < min-increment and random-float 1 < growth-mortality [die]
+ set age age + 1
+
+end
+
+to-report patch-light
+
+ let k light-extinct
+ let canopy turtles-here
+ report Exp (-1 * sum [lai * k] of canopy);]
+
+end
+
+to regenerate
+
+ set plight patch-light
+
+ let AL light-growth-index plight "high"
+ let Al-pop light-growth-index plight "low"
+ let Al-oak light-growth-index plight "intermediate"
+ let max-maple-saplings 3 let max-poplar-saplings 10 let max-oak-saplings 10
+
+ ;Shade intermediate-tolerant species
+ if plight >= 0.5 and plight < 0.99 [
+ let rw random-float 1
+ let rb random-float 1
+
+ if rw < (Al-oak * sitequal-woak)[
+ let woak-spawn (max-oak-saplings * rw)
+ sprout-oaks (round woak-spawn)[
+ set dbh random-float 0.0041
+ set species "WO"
+ init-params
+ ]]
+
+ if rb < (Al-oak * sitequal-boak)[
+ let boak-spawn (max-oak-saplings * rb)
+ sprout-oaks (round boak-spawn)[
+ set dbh random-float 0.0041
+ set species "BO"
+ init-params
+ ]]
+ ]
+
+ ;Shade tolerant species
+ let rm random-float 1
+ if rm < (Al * sitequal-maple)[
+ sprout-maples round (max-maple-saplings * rm)[
+ set dbh random-float 0.00468
+ init-params
+ ]]
+
+ ;Shade intolerant species
+ if plight >= 0.99 and sitequal-poplar > 0 [
+ let pop-spawn (random-float 1 * max-poplar-saplings * Al-pop * sitequal-poplar)
+ sprout-poplars (round pop-spawn) [
+ set dbh random-float 0.0039
+ init-params
+ ]]
+
+end
+
+
+to calc-light
+ let currh height
+ ifelse currh >= max [height] of turtles-here [set hidden? FALSE][set hidden? TRUE]
+ let k light-extinct
+ let canopy turtles-here with [height > currh]
+ set light Exp (-1 * sum [lai * k] of canopy);]
+end
+
+
+
+;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;
+;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;
+
+
+
+;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;
+;;;;;;;; Utility Procedures ;;;;;;;;
+;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;
+
+
+to calc-site-quality
+ ifelse manual-site-qual = TRUE [
+
+ set sitequal-woak sqwoak
+ set sitequal-boak sqboak
+ set sitequal-maple sqmap
+ set sitequal-poplar sqpop
+
+ ][
+ ;based bon Botkin 1993 and Holm 2012 with some guesses
+ ;white oak
+ let fT degree-days-index DegDays 1977 5894
+ let fWT saturation-index wt-dist 0.933
+ let fN nitrogen-index available-N "intermediate"
+ let fWL wilt-index wilt 0.45
+ set sitequal-woak fT * fWT * fN * fWL
+ ;black oak
+ set fT degree-days-index DegDays 2068 5421
+ set fWT saturation-index wt-dist 0.933
+ set fN nitrogen-index available-N "tolerant"
+ set fWL wilt-index wilt 0.45
+ set sitequal-boak fT * fWT * fN * fWL
+ ;maple
+ set fT degree-days-index DegDays 2000 6300
+ set fWT saturation-index wt-dist 0.567
+ set fN nitrogen-index available-N "intermediate"
+ set fWL wilt-index wilt 0.35
+ set sitequal-maple fT * fWT * fN * fWL
+ ;poplar
+ set fT degree-days-index DegDays 2171 6363
+ set fWT saturation-index wt-dist 0.544 ;;based on white spruce/red maple (similar moisture tolerance) ??
+ set fN nitrogen-index available-N "intolerant"
+ set fWL wilt-index wilt 0.245 ;;based on white spruce/red maple (similar moisture tolerance) ??
+ set sitequal-poplar fT * fWT * fN * fWL
+ ]
+end
+
+
+to init-stand [dens-adjust space-adjust n-oak n-maple n-poplar n-sap-oak n-sap-maple n-sap-poplar]
+
+ ;;Create adult trees
+ create-oaks dens-adjust * n-oak [ ;dens-adjust adds more trees when there is a buffer
+ ifelse random-float 1 > 0.5 [set species "WO"][set species "BO"]
+ set dbh (random-normal 45.75 5) / 100
+ init-params
+ set age round (Amax * (dbh * 100) / Dmax)
+ ifelse space-adjust = TRUE [
+ loop [
+ setxy random-pxcor random-pycor
+ if count turtles-here < 2 [stop]]]
+ [setxy random-pxcor random-pycor]
+ ]
+ create-maples dens-adjust * n-maple [
+ set dbh (random-normal 40.8 5) / 100
+ init-params
+ set age round (Amax * (dbh * 100) / Dmax)
+ ifelse space-adjust = TRUE [
+ loop [
+ setxy random-pxcor random-pycor
+ if count turtles-here < 2 [stop]]]
+ [setxy random-pxcor random-pycor]
+ ]
+ create-poplars dens-adjust * n-poplar [
+ set dbh (random-normal 45.07 5) / 100
+ init-params
+ set age round (Amax * (dbh * 100) / Dmax)
+ ifelse space-adjust = TRUE [
+ loop [
+ setxy random-pxcor random-pycor
+ if count turtles-here < 3 [stop]]]
+ [setxy random-pxcor random-pycor]
+ ]
+
+ ask turtles [calc-light]
+ ask patches [set plight patch-light]
+
+ ;;Create saplings
+ create-oaks dens-adjust * n-sap-oak [
+ ifelse random-float 1 > 0.5 [set species "WO"][set species "BO"]
+ set dbh max list 0.015 ((random-normal 14.9 5) / 100)
+ init-params
+ set age round (Amax * (dbh * 100) / Dmax)
+ loop [
+ setxy random-pxcor random-pycor
+ if [plight] of patch-here > 0.6 [stop]]
+ ]
+ create-maples dens-adjust * n-sap-maple [
+ set dbh max list 0.015 ((random-normal 10.3 5) / 100)
+ init-params
+ set age round (Amax * (dbh * 100) / Dmax)
+ setxy random-pxcor random-pycor
+ ]
+ create-poplars dens-adjust * n-sap-poplar [
+ set dbh max list 0.015 ((random-normal 14.9 5) / 100)
+ init-params
+ set age round (Amax * (dbh * 100) / Dmax)
+ loop [
+ setxy random-pxcor random-pycor
+ if [plight] of patch-here > 0.6 [stop]]
+ ]
+
+ ask turtles [calc-light check-in-core]
+ ask patches [set plight patch-light]
+
+end
+
+
+
+to init-params
+ ifelse dbh < 0.1 [set shape "square"] [set shape "tree"]
+ set hidden? FALSE
+ set actual-growth 0.02
+ set min-increment 0.01
+ set growth-mortality 0.369
+ set age 1
+ if breed = oaks [
+ ifelse species = "WO" [;White oak based on Botkin 1993 and Holm 2012
+ set color white
+ set Dmax 100 set Hmax 3800 ;based on HEE data
+ set Amax 400 set intrinsic-mortality 4.0 / Amax ;Based on 2% reaching max age; common to all species
+ set b2 2 * (Hmax - 137) / (Dmax) set b3 (Hmax - 137) / (Dmax ^ 2)
+ set C 1.75
+ set G 104
+ set light-tol "intermediate"
+ ][set color black ;Black oak based on Botkin 1993 and Holm 2012
+ set Dmax 100 set Hmax 3800 ;based on HEE data
+ set Amax 300 set intrinsic-mortality 4.0 / Amax ;Based on 2% reaching max age; common to all species
+ set b2 2 * (Hmax - 137) / (Dmax) set b3 (Hmax - 137) / (Dmax ^ 2)
+ set C 1.75
+ set G 122
+ set light-tol "intermediate"
+ ]]
+ if breed = maples [ ;Sugar maple based bon Botkin 1993 and Holm 2012
+ set color sky
+ set Dmax 100 set Hmax 3350 ;set Dmax 170
+ set Amax 400 set intrinsic-mortality 4.0 / Amax
+ set b2 2 * (Hmax - 137) / (Dmax) set b3 (Hmax - 137) / (Dmax ^ 2)
+ set C 1.57
+ set G 118.7
+ set light-tol "high"
+ ]
+ if breed = poplars [ ;Tulip poplar based on Holm 2012 with some guesses
+ set color red
+ set Dmax 100 set Hmax 4000
+ set Amax 300 set intrinsic-mortality 4.0 / Amax
+ set b2 2 * (Hmax - 137) / (Dmax) set b3 (Hmax - 137) / (Dmax ^ 2)
+ set C 1.75 ;assumed to be similar to oak
+ set G 140
+ set light-tol "low"
+ ]
+ allometerize dbh
+ check-in-core
+
+end
+
+to check-in-core
+ ifelse xcor <= xmax and xcor >= xmin and ycor <= ymax and ycor >= ymin [set in-core TRUE][set in-core FALSE]
+end
+
+to calc-global-vars ;;Calculate global reporter values
+
+ let adjust (x-core * y-core) / 100
+
+ set basal-area (sum [ba] of turtles with [dbh >= 0.015 and in-core = TRUE]) / adjust
+ set basal-area-ovs (sum [ba] of turtles with [dbh >= 0.3 and in-core = TRUE]) / adjust
+ set basal-area-ft basal-area * 4.356
+ set ba-oak (sum [ba] of oaks with [dbh >= 0.015 and in-core = TRUE]) / adjust
+ set ba-oak-ovs (sum [ba] of oaks with [dbh >= 0.3 and in-core = TRUE]) / adjust
+ set ba-map (sum [ba] of maples with [dbh >= 0.015 and in-core = TRUE]) / adjust
+ set ba-map-ovs (sum [ba] of maples with [dbh >= 0.3 and in-core = TRUE]) / adjust
+ set ba-pop (sum [ba] of poplars with [dbh >= 0.015 and in-core = TRUE]) / adjust
+ set ba-pop-ovs (sum [ba] of poplars with [dbh >= 0.3 and in-core = TRUE]) / adjust
+ set dens (count turtles with [dbh >= 0.015 and in-core = TRUE]) / adjust
+ set dens-ovs (count turtles with [dbh >= 0.3 and in-core = TRUE]) / adjust
+ set dens-ac dens * 0.40477
+ set dens-oak (count oaks with [dbh >= 0.015 and in-core = TRUE]) / adjust
+ set dens-oak-ovs (count oaks with [dbh >= 0.3 and in-core = TRUE]) / adjust
+ set dens-map (count maples with [dbh >= 0.015 and in-core = TRUE]) / adjust
+ set dens-map-ovs (count maples with [dbh >= 0.3 and in-core = TRUE]) / adjust
+ set dens-pop (count poplars with [dbh >= 0.015 and in-core = TRUE]) / adjust
+ set dens-pop-ovs (count poplars with [dbh >= 0.3 and in-core = TRUE]) / adjust
+ ifelse basal-area > 0 [
+ set qdbh sqrt(basal-area * adjust / (0.0000785 * count turtles with [dbh >= 0.015 and in-core = TRUE]))
+ set qdbh-in 2 * (sqrt(mean [ba] of turtles with [dbh >= 0.015 and in-core = TRUE] / pi)) * 39.37
+ set prop-oak (ba-oak / basal-area)
+ set prop-tol (ba-map / basal-area)
+ set prop-intol (ba-pop / basal-area)
+ ][set qdbh 0 set qdbh-in 0 set prop-oak 0 set prop-tol 0 set prop-intol 0]
+ set qdbh-ovs sqrt(basal-area-ovs * adjust / (0.0000785 * (max list 1 count turtles with [dbh >= 0.3 and in-core = TRUE])))
+ set qdbh-oak sqrt(ba-oak * adjust / (0.0000785 * (max list 1 count oaks with [dbh >= 0.015 and in-core = TRUE])))
+ set qdbh-oak-ovs sqrt(ba-oak-ovs * adjust / (0.0000785 * (max list 1 count oaks with [dbh >= 0.3 and in-core = TRUE])))
+ set qdbh-map sqrt(ba-map * adjust / (0.0000785 * (max list 1 count maples with [dbh >= 0.015 and in-core = TRUE])))
+ set qdbh-map-ovs sqrt(ba-map-ovs * adjust / (0.0000785 * (max list 1 count maples with [dbh >= 0.3 and in-core = TRUE])))
+ set qdbh-pop sqrt(ba-pop * adjust / (0.0000785 * (max list 1 count poplars with [dbh >= 0.015 and in-core = TRUE])))
+ set qdbh-pop-ovs sqrt(ba-pop-ovs * adjust / (0.0000785 * (max list 1 count poplars with [dbh >= 0.3 and in-core = TRUE])))
+ set seedlings-class4 (count oaks with [height > 1.4 and dbh < 0.015 and in-core = TRUE]) / adjust
+end
+
+to color-patches
+ if (1 - plight) > 0 [set pcolor lime + 1]
+ if (1 - plight) > 0.1 [set pcolor lime]
+ if (1 - plight) > 0.2 [set pcolor lime - 1]
+ if (1 - plight) > 0.3 [set pcolor green + 2]
+ if (1 - plight) > 0.5 [set pcolor green + 1]
+ if (1 - plight) > 0.7 [set pcolor green]
+ if (1 - plight) > 0.8 [set pcolor green - 1]
+ if (1 - plight) > 0.9 [set pcolor green - 2]
+ if (1 - plight) > 0.95 [set pcolor green - 3]
+end
+
+
+;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;
+;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;
+
+
+
+;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;
+;;;;;;;; Harvesting Procedures ;;;;;;;;
+;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;
+
+to conduct-harvest
+
+ if harvest-type = "none" [stop]
+
+ if (ticks + 1) != harvest-year [stop]
+
+ let adjust (x-core * y-core) / 100
+
+ if harvest-type = "clearcut" [
+ ask turtles with [dbh > 0.01 and in-core = TRUE] [die]
+ ;set harvest-year harvest-year + 100
+ ]
+
+ if harvest-type = "singletree" [
+ set basal-area (sum [ba] of turtles with [dbh >= 0.01 and in-core = TRUE]) / adjust
+ set harvest-year harvest-year + 20
+ if basal-area >= 25 [
+ loop [
+ let potential one-of turtles with [dbh >= 0.10 and in-core = TRUE]
+ ask potential [die]
+ set basal-area (sum [ba] of turtles with [dbh >= 0.01 and in-core = TRUE]) / adjust
+ if basal-area <= 25 [stop]
+ ]
+ ]
+ ]
+
+ if harvest-type = "shelterwood" [
+ ifelse shelter-phase = 1 [
+ set shelter-phase 2
+ ask turtles with [breed != oaks and dbh <= 0.254 and in-core = TRUE] [die] ;treated with herbicide
+ set harvest-year (harvest-year + 7)
+ ]
+ [
+ ifelse shelter-phase = 2 [
+ set shelter-phase 3
+ set harvest-year (harvest-year + 8)
+ if basal-area >= 16.1 [
+ loop [
+ ;this code is broken
+ ;let potential min-one-of turtles with [breed != oaks and dbh >= 0.10 and xcor < 50 and xcor > -50 and ycor < 50 and ycor > -50] [dbh]
+ let potential min-one-of turtles with [age > 20 and in-core = TRUE] [light]
+ ask potential [ifelse breed = oaks [die][die]]
+ set basal-area (sum [ba] of turtles with [dbh >= 0.01 and in-core = TRUE]) / adjust
+ if basal-area <= 16.1 [stop]
+ ]
+ ]
+ ]
+ [
+ ask turtles with [age > 20 and in-core = TRUE] [die]
+ set shelter-phase 1
+ ;set harvest-year (harvest-year + 100)
+ ]
+
+ ]]
+
+end
diff --git a/model_browse.R b/model_browse.R
new file mode 100644
index 0000000..cbd9674
--- /dev/null
+++ b/model_browse.R
@@ -0,0 +1,94 @@
+model {
+
+ #Likelihood
+
+ #Site-level mean
+ for (i in 1:nsites){
+
+ site.mean[i] ~ dnorm(grand.mean,site.tau)
+
+ }
+
+ #Plot-level mean
+ for (i in 1:nplots){
+
+ plot.mean[i] ~ dnorm(plot.pred[i], plot.tau)
+ plot.pred[i] <- site.mean[plot.sitecode[i]] + b.distance*distance[i]
+
+ }
+
+ #Seedling-level mean
+ for (i in 1:nseedlings){
+
+ seed.mean[i] ~ dnorm(seed.pred[i], seed.tau)
+ seed.pred[i] <- plot.mean[seed.plotcode[i]] + b.species*species[i]
+
+ #Observation level
+ for (j in 1:nsamples[i]){
+
+ #Linear predictor
+ mu[i,j] <- seed.mean[i]
+ + b.comp*comp[seed.plotcode[i],j]
+ + b.ht*ht[i,j] + b.ht2*ht2[i,j]
+
+ #Calculate probabilities for first category
+ logit(Q[i,j,1]) <- tau[1,j] - mu[i,j]
+ p[i,j,1] <- Q[i,j,1]
+
+ #Calculate probabilities for categories 2-3
+ for (k in 2:3){
+ logit(Q[i,j,k]) <- tau[k,j] - mu[i,j]
+ p[i,j,k] <- Q[i,j,k] - Q[i,j,k-1]
+ }
+
+ #Calculate probability for final category
+ p[i,j,4] <- 1 - Q[i,j,3]
+
+ #Connect to observed data
+ browse[i,j] ~ dcat(p[i,j,1:4])
+
+ #Expected value of categorical variable
+ ev[i,j] <- 1*p[i,j,1] + 2*p[i,j,2] + 3*p[i,j,3] + 4*p[i,j,4]
+
+ #Absolute residual for real datapoint
+ res[cucount[i,j]] <- abs(browse[i,j] - ev[i,j])
+
+ #Simulate new datapoint
+ browse.new[i,j] ~ dcat(p[i,j,1:4])
+
+ #Absolute residual for simulated datapoint
+ res.new[cucount[i,j]] <- abs(browse.new[i,j] - ev[i,j])
+
+ }}
+
+ #Derived quantities for posterior predictive check
+ fit <- sum(res[])
+ fit.new <- sum(res.new[])
+
+ #Priors
+
+ #Category thresholds
+ for (j in 1:8){
+ for(k in 1:3){
+ tau0[k,j] ~ dnorm(0,.01)
+ }
+ tau[1:3,j] <- sort(tau0[1:3,j])
+ }
+
+ #Random effects
+ grand.mean ~ dunif(-100,100)
+ site.tau <- pow(site.sd,-2)
+ site.sd ~ dunif(0,100)
+ plot.tau <- pow(plot.sd,-2)
+ plot.sd ~ dunif(0,100)
+ seed.tau <- pow(seed.sd,-2)
+ seed.sd ~ dunif(0,100)
+
+ #Fixed effects
+ b.distance ~ dnorm(0,0.01)
+ b.species ~ dnorm(0,0.01)
+ b.ht ~ dnorm(0,0.01)
+ b.ht2 ~ dnorm(0,0.01)
+ b.comp ~ dnorm(0,0.01)
+
+} \ No newline at end of file
diff --git a/model_growth.R b/model_growth.R
new file mode 100644
index 0000000..75c5a28
--- /dev/null
+++ b/model_growth.R
@@ -0,0 +1,78 @@
+model {
+
+ #Likelihood
+
+ #Site-level mean
+ for (i in 1:nsites){
+
+ site.mean[i] ~ dnorm(grand.mean,site.tau)
+
+ }
+
+ #Plot-level mean
+ for (i in 1:nplots){
+
+ plot.mean[i] ~ dnorm(plot.pred[i], plot.tau)
+ plot.pred[i] <- site.mean[plot.sitecode[i]]
+ + b.aspect*aspect[i]
+ + b.edge*edge[i] + b.harvest*harvest[i] + b.shelter*shelter[i]
+
+ }
+
+ #Seedling-level mean
+ for (i in 1:nseedlings){
+
+ seed.mean[i] ~ dnorm(seed.pred[i], seed.tau)
+ seed.pred[i] <- plot.mean[seed.plotcode[i]]
+
+ #Observation level
+ for (j in 1:nsamples[i]){
+
+ #Connect to observed data
+ growth[i,j] ~ dnorm(mu[i,j],obs.tau)
+
+ #Linear predictor
+ mu[i,j] <- seed.mean[i]
+ + b.browse*browse[i,j]
+ + b.comp*stem.comp[i,j]
+ + b.sprout*is.sprout[i,j]
+ + b.elapsed*elapsed[j]
+
+ #For posterior predictive check
+ res[cucount[i,j]] <- growth[i,j] - mu[i,j]
+ sqres[cucount[i,j]] <- pow(res[cucount[i,j]],2)
+ growth.new[i,j] ~ dnorm(mu[i,j],obs.tau)
+ res.new[cucount[i,j]] <- growth.new[i,j] - mu[i,j]
+ sqres.new[cucount[i,j]] <- pow(res.new[cucount[i,j]],2)
+
+ }
+ }
+
+ #Derived quantities
+ fit <- sum(sqres[])
+ fit.new <- sum(sqres.new[])
+
+ #Priors
+
+ #Random effects
+ grand.mean ~ dunif(-100,100)
+ site.tau <- pow(site.sd,-2)
+ site.sd ~ dunif(0,100)
+ plot.tau <- pow(plot.sd,-2)
+ plot.sd ~ dunif(0,100)
+ seed.tau <- pow(seed.sd,-2)
+ seed.sd ~ dunif(0,100)
+ obs.tau <- pow(obs.sd,-2)
+ obs.sd ~ dunif(0,100)
+
+ #Fixed effects
+ b.comp ~ dnorm(0,0.01)
+ b.aspect ~ dnorm(0,0.01)
+ b.edge ~ dnorm(0,0.01)
+ b.harvest ~ dnorm(0,0.01)
+ b.shelter ~ dnorm(0,0.01)
+ b.browse ~ dnorm(0,0.01)
+ b.sprout ~ dnorm(0,0.01)
+ b.elapsed ~ dnorm(0,0.01)
+
+} \ No newline at end of file
diff --git a/model_surv.R b/model_surv.R
new file mode 100644
index 0000000..ec8e8f4
--- /dev/null
+++ b/model_surv.R
@@ -0,0 +1,80 @@
+model {
+
+ #Likelihood
+
+ #Site-level mean
+ for (i in 1:nsites){
+
+ site.mean[i] ~ dnorm(grand.mean,site.tau)
+
+ }
+
+ #Plot-level mean
+ for (i in 1:nplots){
+
+ plot.mean[i] ~ dnorm(plot.pred[i], plot.tau)
+ plot.pred[i] <- site.mean[plot.sitecode[i]]
+ + b.edge*edge[i]
+ + b.harvest*harvest[i]
+ + b.shelter*shelter[i]
+ + b.aspect*aspect[i]
+
+ }
+
+ for (i in 1:nseedlings){
+
+ #Observation level
+ for (j in 2:nsamples[i]){
+
+ #Connect to observed data
+ surv[i,j] ~ dbern(psi[i,j])
+
+ psi[i,j] <- mu[i,j]*surv[i,j-1]
+
+ #Linear predictor
+ logit(mu[i,j]) <- plot.mean[seed.plotcode[i]]
+ + b.rcd*rcd[i,j-1]
+ + b.sprout*is.sprout[i,j-1]
+ + b.browse*browse[i,j-1]
+ + b.season*season[j]
+ + b.comp*stem.comp[i,j]
+ + b.elapsed*elapsed[j]
+
+
+ #For posterior predictive check
+ res[cucount[i,j]] <- abs(surv[i,j] - psi[i,j])
+ surv.new[i,j] ~ dbern(psi[i,j])
+ res.new[cucount[i,j]] <- abs(surv.new[i,j] - psi[i,j])
+
+ }
+ }
+
+ #Derived quantities
+ fit <- sum(res[])
+ fit.new <- sum(res.new[])
+ b.edge_harvest <- b.edge - b.harvest
+ b.edge_shelter <- b.edge - b.shelter
+ b.shelter_harvest <- b.shelter - b.harvest
+
+ #Priors
+
+ #Random effects
+ grand.mean ~ dunif(-100,100)
+ site.tau <- pow(site.sd,-2)
+ site.sd ~ dunif(0,100)
+ plot.tau <- pow(plot.sd,-2)
+ plot.sd ~ dunif(0,100)
+
+ #Fixed effects
+ b.edge ~ dnorm(0,0.01)
+ b.harvest ~ dnorm(0,0.01)
+ b.shelter ~ dnorm(0,0.01)
+ b.comp ~ dnorm(0,0.01)
+ b.aspect ~ dnorm(0,0.01)
+ b.elapsed ~ dnorm(0,0.01)
+ b.browse ~ dnorm(0,0.01)
+ b.rcd ~ dnorm(0,0.01)
+ b.season ~ dnorm(0,0.01)
+ b.sprout ~ dnorm(0,0.01)
+
+} \ No newline at end of file
diff --git a/netlogomodel.tex b/netlogomodel.tex
new file mode 100644
index 0000000..535535a
--- /dev/null
+++ b/netlogomodel.tex
@@ -0,0 +1,1113 @@
+extensions [profiler matrix]
+globals [
+
+ ;Setup Parameters
+ xcutoff ycutoff
+ sitequal-boak sitequal-woak sitequal-maple sitequal-poplar
+ harvest-year shelter-phase
+
+ ;Input Parameters
+ wo-mast-list bo-mast-list mast-year-index
+ mast-mean-bo mast-mean-wo mast-mean-comb
+ core-acorn-params-bo buffer-acorn-params-bo
+ core-acorn-params-wo buffer-acorn-params-wo
+ core-weev-prob-bo buffer-weev-prob-bo
+ core-weev-prob-wo buffer-weev-prob-wo
+ surv-params growth-params br-year-ef
+ sens-params
+ prioryears mastsd
+
+ ;Overstory Output Parameters
+ basal-area basal-area-ft basal-area-ovs
+ qdbh qdbh-in qdbh-ovs
+ dens dens-ac dens-ovs
+ ba-oak ba-oak-ovs
+ qdbh-oak qdbh-oak-ovs
+ dens-oak dens-oak-ovs
+ ba-map ba-map-ovs
+ qdbh-map qdbh-map-ovs
+ dens-map dens-map-ovs
+ ba-pop ba-pop-ovs
+ qdbh-pop qdbh-pop-ovs
+ dens-pop dens-pop-ovs
+ prop-oak prop-tol prop-intol
+
+ ;Regeneration Output Parameters
+ acorn-count total-acorns total-seedlings new-seedlings pct-germ
+ regen-dens regen-stump-dens
+ seedlings-class1 seedlings-class2 seedlings-class3 seedlings-class123 seedlings-class4 acorns-pertree
+ seedbo-class123 seedwo-class123 seedbo-class4 seedwo-class4
+ bo-acorn-count wo-acorn-count
+ new-bo-seedlings new-wo-seedlings
+ total-bo-acorns total-wo-acorns
+ pct-bo-germ pct-wo-germ
+
+ ]
+
+turtles-own [in-core]
+
+breed [oaks oak]
+oaks-own [species light age dbh height canopy-radius canopy-density actual-growth ba acorn-mean seedling sprout? fAL
+ Dmax Hmax Amax b2 b3 C G light-tol intrinsic-mortality min-increment growth-mortality density browsed indeffect
+ ]
+
+breed [maples maple]
+maples-own [light age dbh height canopy-radius canopy-density actual-growth ba seedling sprout? fAL
+ Dmax Hmax Amax b2 b3 C G light-tol intrinsic-mortality min-increment growth-mortality density
+ ]
+
+breed [poplars poplar]
+poplars-own [light age dbh height canopy-radius canopy-density actual-growth ba seedling sprout? fAL
+ Dmax Hmax Amax b2 b3 C G light-tol intrinsic-mortality min-increment growth-mortality density
+ ]
+
+breed [acorns acorn]
+acorns-own [species weevil cached germ disp-dist]
+
+patches-own [stem-dens shade in-layer1 in-layer2]
+
+
+;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;
+;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;
+
+
+to setup
+ clear-all
+ reset-ticks
+
+ set xcutoff (x-core / 2)
+ set ycutoff (y-core / 2)
+
+ let adjust (x-core + buffer * 2) * (y-core + buffer * 2) / 10000
+
+ resize-world (-1 * (xcutoff + buffer)) (xcutoff + buffer) (-1 * (ycutoff + buffer)) (ycutoff + buffer)
+
+ calc-site-quality
+
+ ;HEE mast rotation
+ set bo-mast-list [0.0827 0.0489 2.975 0.8809 0.0757 0.1153 0.0661 0.1415 0.05667]
+ set wo-mast-list [0.1355 0.0665 0.1092 2.720 0.0367 0.136 0.1070 0.238 0.0406]
+ set mast-year-index random 9
+
+ ;HEE-based initial forest
+ ifelse HEE-mean = TRUE [
+ init-stand adjust TRUE 89 11 9 95 499 163] [ ;Initial stand values based on Saunders and Arseneault 2013
+ init-stand adjust FALSE mature-oak mature-maple mature-poplar sapling-oak sapling-maple sapling-poplar] ;User defined
+
+ ask patches [color-patches]
+ calc-global-vars
+ set harvest-year burnin
+ set shelter-phase 1
+ setup-plots
+
+end
+
+;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;
+;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;
+
+to go
+
+ calc-light
+
+ set-scenario
+ set acorn-count 0 set bo-acorn-count 0 set wo-acorn-count 0
+
+ ask turtles with [seedling = FALSE] [
+ grow
+ check-survival
+ if breed = oaks and dbh >= 0.20 and seedlings != "none" [produce-acorns] ;based on Downs and McQuilkin 1944
+ ]
+
+ ask acorns [
+ disperse-mast
+ germinate-mast
+ ]
+
+ if seedlings = "hee" [
+ ask oaks with [seedling = TRUE] [
+ grow-seedling
+ check-seedling-survival
+ ]
+ ]
+
+ ask patches [
+ regenerate
+ color-patches
+ ]
+
+ conduct-harvest
+
+ calc-global-vars
+
+ tick
+end
+
+
+;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;
+;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;
+
+
+;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;
+;;;;;;;; Basic JABOWA Procedures ;;;;;;;;
+;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;
+
+
+to allometerize [dbh-input]
+ ;Based on JABOWA, Botkin 1993
+ ;b2 = 2*(Hmax - 137) / Dmax and b3 = (Hmax - 137) / Dmax ^2
+ set height (137 + b2 * (dbh-input * 100) - b3 * ((dbh-input * 100) ^ 2)) / 100
+
+ ;Based on Kenefic and Nyland 1999 and others
+ ;range reported is 0.376 - 0.393; going with .5 for now
+ set canopy-radius max list 1 (0.385 * height / 2)
+
+ ;leaf weight based on dbh with C species-specific
+ ;Exponent can be anywhere from 1.5 to 3 (Botkin 1993)
+ ;let SLA (C * (dbh-input * 100) ^ (2.4))
+ let SLA (C * (dbh-input * 100) ^ (2))
+
+ ;Beer-Lambert law where light above canopy = 1
+ let PHI 1
+ ;let k 0.0000756 ;Based on QuabbinPlot206, seems far too small
+ ;let k 1 / 6000 ;Based on Botkin 1993, allows too much light under big trees
+ ;let k 1 / 3000 ;Happy medium
+ let k light-extinct
+ set canopy-density min list 0.98 (1 - (PHI * exp(-1 * k * SLA)))
+
+ set ba (pi * (dbh / 2) ^ 2) ;Basal area
+end
+
+
+to-report max-growth-increment [dbh-input height-input] ;Based on JABOWA, Botkin 1993
+ let num (G * dbh-input * ((1 - (dbh-input * (137 + (b2 * dbh-input) - (b3 * dbh-input ^ 2))) / (Dmax * Hmax))))
+ let denom (274 + (3 * b2 * dbh-input) - (4 * b3 * dbh-input ^ 2))
+ report (num / denom)
+end
+
+
+to-report light-growth-index [light-input tol-input]
+ ;Based on Botkin 1993 and Bonan 1990
+ if tol-input = "low" [
+ report (2.24 * (1 - exp(-1.136 * (light-input - 0.08))))
+ ]
+ if tol-input = "intermediate" [
+ report (1.371 * (1 - exp(-2.227 * (light-input - 0.05))))
+ ]
+ if tol-input = "high" [
+ report (1 - exp(-4.64 * (light-input - 0.05)))
+ ]
+end
+
+
+to-report degree-days-index [degdays-input degd-min-input degd-max-input]
+ let tdegd (4 * (degdays-input - degd-min-input) * (degd-max-input - degdays-input) / ((degd-max-input - degd-min-input) ^ 2))
+ report max list 0 tdegd
+end
+
+
+to-report saturation-index [wt-input wt-dist-min-input]
+ let wefi (1 - (wt-dist-min-input / wt-input))
+ report max list 0 wefi
+end
+
+
+to-report nitrogen-index [N-input N-tol-input]
+ if N-tol-input = "intolerant" [
+ let a1 2.99 let a2 0.00175 let a3 207.43 let a4 -5 let a5 2.9 let a6 3.671
+ let lambdaN (a1 * (1 - 10 ^ (-1 * a2 * (N-input + a3))))
+ report (a4 + a5 * lambdaN) / a6
+ ]
+ if N-tol-input = "intermediate" [
+ let a1 2.94 let a2 0.00234 let a3 117.52 let a4 -1.2 let a5 1.3 let a6 2.622
+ let lambdaN (a1 * (1 - 10 ^ (-1 * a2 * (N-input + a3))))
+ report (a4 + a5 * lambdaN) / a6
+ ]
+ if N-tol-input = "tolerant" [
+ let a1 2.79 let a2 0.00179 let a3 219.77 let a4 -0.6 let a5 1.0 let a6 2.190
+ let lambdaN (a1 * (1 - 10 ^ (-1 * a2 * (N-input + a3))))
+ report (a4 + a5 * lambdaN) / a6
+ ]
+end
+
+
+to-report wilt-index [wilt-input wilt-max]
+ ;wilt is difference between potential and actual evapotranspiration divided by potential evapotranspiration
+ let wifi (1 - (wilt-input / wilt-max) ^ 2)
+ report max list 0 wifi
+end
+
+
+to grow
+
+ let max-growth (max-growth-increment (dbh * 100) (height * 100))
+
+ set fAL light-growth-index light light-tol
+ let sitequal 1
+ ifelse breed = oaks [ifelse species = "WO" [set sitequal sitequal-woak] [set sitequal sitequal-boak]]
+ [ifelse breed = maples [set sitequal sitequal-maple]
+ [set sitequal sitequal-poplar]]
+
+ set actual-growth (max-growth * fAL * sitequal / max list 1 density) ;Penalize growth based on density of larger trees
+ set dbh (dbh * 100 + actual-growth) / 100
+ allometerize dbh
+
+ if dbh >= 0.1 and shape = "square" [
+ if breed = oaks [set size 2 set shape "tree"]
+ if breed = maples [set size 2 set shape "tree"]
+ if breed = poplars [set size 2 set shape "tree"]
+ ]
+
+end
+
+
+to check-survival
+
+ ifelse dbh >= 0.05 [
+ if random-float 1 < intrinsic-mortality [create-sprout]
+ if actual-growth < min-increment and random-float 1 < growth-mortality [create-sprout]
+ ] [
+ if random-float 1 < intrinsic-mortality [die]
+ if actual-growth < min-increment and random-float 1 < growth-mortality [die]
+ ]
+ set age age + 1
+
+end
+
+
+to create-sprout
+ ifelse sprouting = FALSE [die][
+ let prob 0
+ if dbh < 0.8 and dbh > 0.05 [
+ ;if breed = oaks [set prob (1 + exp(-1 * (5.9991 - 0.2413 * (dbh * 39.3701) - 0.0475 * age))) ^ (-1)] ;From Dey 2002
+ if breed = oaks [ ;From Weigel and Peng 2002
+ if species = "WO" [
+ set prob (1 + exp(-1 * (-53.6225 - 1.7003 * ln(dbh * 100) - 0.00534 * age * ln(dbh * 100) + 25.7155 * ln(22) - 0.2913 * 22 * ln(22)))) ^ (-1)]
+ if species = "BO" [
+ set prob (1 + exp(-1 * (-8.1468 - 0.00055 * age * (dbh * 100) + 3.1678 * ln(22)))) ^ (-1)]
+ ]
+ if breed = maples [set prob -0.314 * ln(dbh) + 0.0877] ;Based on Powell 1983
+ if breed = poplars [set prob 1.1832 - 1.3638 * dbh] ;based on Wendel 1975, True 1953, Beck & Della-Bianca 1981
+ if prob > 1 [set prob 1]
+ ]
+ ifelse random-float 1 < prob [
+ set sprout? TRUE
+ set dbh 0.01
+ init-params
+ set size 2
+ ]
+ [die]
+ ]
+end
+
+
+to regenerate
+ if count turtles-here with [seedling = FALSE] > 0 [stop]
+ let AL light-growth-index ((1 - shade)) "high"
+ let Al-pop light-growth-index ((1 - shade)) "low"
+ let max-maple-saplings 3 let max-poplar-saplings 10
+
+ if random-float 1 < (0.01 * max-maple-saplings * AL * sitequal-maple)[
+ sprout-maples 1 [
+ check-in-core
+ set dbh random-float 0.00468
+ init-params
+ ]]
+
+ if shade <= 0.01 and random-float 1 < (0.01 * max-poplar-saplings * AL-pop * sitequal-poplar)[
+ sprout-poplars 1 [
+ check-in-core
+ set dbh random-float 0.0039
+ init-params
+ ]]
+
+ if seedlings = "none" [
+ let Al-oak light-growth-index ((1 - shade) / max list 1 stem-dens ) "intermediate"
+ let max-oak-saplings 10
+ let sitequal-oak 1
+ ifelse random-float 1 < 0.5 [set sitequal-oak sitequal-woak][set sitequal-oak sitequal-boak]
+
+ ;arbitrary value here
+ if shade < 0.50 and random-float 1 < (0.01 * max-oak-saplings * AL-oak * sitequal-oak)[
+ sprout-oaks 1 [
+ check-in-core
+ set dbh random-float 0.0041
+ ifelse random-float 1 > 0.5 [set species "WO"][set species "BO"]
+ init-params
+ ]]
+
+ ]
+
+end
+
+to-report gen-acorn-params [remint distint deint ueint treat yearly yearef spec]
+
+ let rem-mean remint + (treat * 0.129) + (yearly * yearef) + (spec * -0.162)
+ let rem (1 + exp(-1 * rem-mean)) ^ (-1)
+ let remw-mean rem-mean - 2.00
+ let rem-weev (1 + exp(-1 * remw-mean)) ^ (-1)
+ let dist exp(distint + (treat * -0.115) + (yearly * 0.114 * mast-mean-comb))
+ let de-mean deint + (treat * 0.5817) + (yearly * 1.1414 * mast-mean-comb)
+ let de (1 + exp(-1 * de-mean)) ^ (-1)
+ let de-cached-mean de-mean - 5.4178
+ let de-cached (1 + exp(-1 * de-cached-mean)) ^ (-1)
+ let ue-mean ueint + (treat * 1.609) + (spec * 1.4714) + (yearly * 1.2051 * mast-mean-comb)
+ let ue (1 + exp(-1 * ue-mean)) ^ (-1)
+ report (list rem rem-weev dist de de-cached ue)
+ ;pRemoval, pRemoval given weeviled, weibSh, pDispEaten, pDispEatencache, pUndispEaten
+ ;caching probability is calculated separately
+end
+
+to-report gen-weev-prob [intercept treat yearly yearef spec]
+
+ let weevmean intercept + (treat * -0.4438) + (yearly * (0.7324 * mast-mean-comb + yearef)) + (spec * -0.8249)
+ let prob (1 + exp(-1 * weevmean)) ^ (-1)
+ report prob
+
+end
+
+;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;
+;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;
+
+
+;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;
+;;;;;;;; Early Oak Lifecycle Procedures ;;;;;;;;
+;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;
+
+to set-scenario ;Set up parameter values for experiments
+
+ let random-index random 9
+
+ ;Acorn production
+ if mast-scenario = "custom" [
+ set mast-mean-wo mast-val
+ set mast-mean-bo mast-val
+ ]
+ if mast-scenario = "fixedaverage" [
+ set mast-mean-wo 11.740
+ set mast-mean-bo 11.071
+ ]
+ if mast-scenario = "fixedbad" [
+ set mast-mean-wo 8.579
+ set mast-mean-bo 3.749
+ ]
+ if mast-scenario = "fixedgood" [
+ set mast-mean-wo 22.346
+ set mast-mean-bo 16.584
+ ]
+ if mast-scenario = "random" [
+ set mast-mean-wo min (list exp(2.001 + random-normal 0 1.126) 50)
+ set mast-mean-bo min (list exp(2.001 + random-normal 0 1.126) 50)
+ ]
+ if mast-scenario = "hee" [
+ set mast-mean-wo 1 / (item mast-year-index wo-mast-list)
+ set mast-mean-bo 1 / (item mast-year-index bo-mast-list)
+ ]
+
+ if mast-scenario = "priordifference" [
+ if prioryears = 2 [
+ ifelse ticks = (burnin - 2) OR ticks = (burnin - 1) [
+ set mast-mean-wo exp(2.001 + mastsd * 1.126)
+ set mast-mean-bo exp(2.001 + mastsd * 1.126)
+ ][
+ set mast-mean-wo min (list exp(2.001 + random-normal 0 1.126) 70)
+ set mast-mean-bo min (list exp(2.001 + random-normal 0 1.126) 70)
+ ]
+ ]
+ if prioryears = 1 [
+ ifelse ticks = (burnin - 1) [
+ set mast-mean-wo exp(2.001 + mastsd * 1.126)
+ set mast-mean-bo exp(2.001 + mastsd * 1.126)
+ ][
+ set mast-mean-wo min (list exp(2.001 + random-normal 0 1.126) 70)
+ set mast-mean-bo min (list exp(2.001 + random-normal 0 1.126) 70)
+ ]
+ ]
+ ]
+
+ if mast-scenario = "sensitivity" [
+ set mast-mean-wo min (list exp(item 0 sens-params) 50)
+ set mast-mean-bo min (list exp(item 0 sens-params) 50)
+ ]
+
+ ;Models are in rate scale rather than lambda
+ set mast-mean-comb 1 / (mean (list mast-mean-wo mast-mean-bo))
+
+ ;Weevil scenarios
+ if weevil-scenario = "custom" [ ]
+ if weevil-scenario = "fixedaverage" [
+ set buffer-weev-prob-bo (gen-weev-prob -1.08 0 0 0 0)
+ set buffer-weev-prob-wo (gen-weev-prob -1.08 0 0 0 1)
+ set core-weev-prob-bo buffer-weev-prob-bo
+ set core-weev-prob-wo buffer-weev-prob-wo
+ ]
+ if weevil-scenario = "yearly-diff" [
+ let randN random-normal 0 1.01
+ set buffer-weev-prob-bo (gen-weev-prob -1.08 0 1 randN 0)
+ set buffer-weev-prob-wo (gen-weev-prob -1.08 0 1 randN 1)
+ set core-weev-prob-bo buffer-weev-prob-bo
+ set core-weev-prob-wo buffer-weev-prob-wo
+ ]
+ if weevil-scenario = "treat-diff" [
+ set buffer-weev-prob-bo (gen-weev-prob -1.08 0 0 0 0)
+ set buffer-weev-prob-wo (gen-weev-prob -1.08 0 0 0 1)
+ set core-weev-prob-bo (gen-weev-prob -1.08 1 0 0 0)
+ set core-weev-prob-wo (gen-weev-prob -1.08 1 0 0 1)
+ ]
+ if weevil-scenario = "yearly-treat-diff" [
+ let randN random-normal 0 1.01
+ set buffer-weev-prob-bo (gen-weev-prob -1.08 0 1 randN 0)
+ set buffer-weev-prob-wo (gen-weev-prob -1.08 0 1 randN 1)
+ set core-weev-prob-bo (gen-weev-prob -1.08 1 1 randN 0)
+ set core-weev-prob-wo (gen-weev-prob -1.08 1 1 randN 1)
+ ]
+ if weevil-scenario = "sensitivity" [
+ set buffer-weev-prob-bo (gen-weev-prob (item 1 sens-params) 0 0 0 0)
+ set buffer-weev-prob-wo (gen-weev-prob (item 1 sens-params) 0 0 0 1)
+ set core-weev-prob-bo buffer-weev-prob-bo
+ set core-weev-prob-wo buffer-weev-prob-wo
+ ]
+
+ ;Dispersal scenarios
+ if dispersal-scenario = "custom" [
+ set core-acorn-params-bo (list disperse-prob disperse-prob weibSc disperse-eaten-prob disperse-eaten-prob undisp-eaten-prob)
+ set core-acorn-params-wo core-acorn-params-bo
+ set buffer-acorn-params-bo core-acorn-params-bo
+ set buffer-acorn-params-wo core-acorn-params-bo
+ ]
+ if dispersal-scenario = "fixedaverage" [
+ set buffer-acorn-params-bo (gen-acorn-params 0.519 2.071 1.496 -0.179 0 0 0 0)
+ set buffer-acorn-params-wo (gen-acorn-params 0.519 2.071 1.496 -0.179 0 0 0 1)
+ set core-acorn-params-bo buffer-acorn-params-bo
+ set core-acorn-params-wo buffer-acorn-params-wo
+ ]
+ if dispersal-scenario = "yearly-diff" [
+ let randN random-normal 0 0.6322
+ set buffer-acorn-params-bo (gen-acorn-params 0.519 2.071 1.496 -0.179 0 1 randN 0)
+ set buffer-acorn-params-wo (gen-acorn-params 0.519 2.071 1.496 -0.179 0 1 randN 1)
+ set core-acorn-params-bo buffer-acorn-params-bo
+ set core-acorn-params-wo buffer-acorn-params-wo
+ ]
+ if dispersal-scenario = "treat-diff" [
+ set buffer-acorn-params-bo (gen-acorn-params 0.519 2.071 1.496 -0.179 0 0 0 0)
+ set buffer-acorn-params-wo (gen-acorn-params 0.519 2.071 1.496 -0.179 0 0 0 1)
+ set core-acorn-params-bo (gen-acorn-params 0.519 2.071 1.496 -0.179 1 0 0 0)
+ set core-acorn-params-wo (gen-acorn-params 0.519 2.071 1.496 -0.179 1 0 0 1)
+ ]
+ if dispersal-scenario = "yearly-treat-diff" [
+ let randN random-normal 0 0.6322
+ set buffer-acorn-params-bo (gen-acorn-params 0.519 2.071 1.496 -0.179 0 1 randN 0)
+ set buffer-acorn-params-wo (gen-acorn-params 0.519 2.071 1.496 -0.179 0 1 randN 1)
+ set core-acorn-params-bo (gen-acorn-params 0.519 2.071 1.496 -0.179 1 1 randN 0)
+ set core-acorn-params-wo (gen-acorn-params 0.519 2.071 1.496 -0.179 1 1 randN 1)
+ ]
+ if dispersal-scenario = "sensitivity" [
+ set buffer-acorn-params-bo (gen-acorn-params (item 2 sens-params) (item 3 sens-params) (item 5 sens-params) (item 6 sens-params) 0 0 0 0)
+ set buffer-acorn-params-wo (gen-acorn-params (item 2 sens-params) (item 3 sens-params) (item 5 sens-params) (item 6 sens-params) 0 0 0 1)
+ set core-acorn-params-bo buffer-acorn-params-bo
+ set core-acorn-params-wo buffer-acorn-params-wo
+ ]
+
+ ;Browsing
+ if browse-scenario = "fixedaverage" [set prob-browsed 0.1058]
+ if browse-scenario = "hee" [
+ set br-year-ef random-normal 0 0.59
+ ]
+ if browse-scenario = "custom" [ ]
+
+ ;Seedling survival and growth
+ if seedling-scenario = "fixedaverage" [
+ ; intercept species canopy age
+ set surv-params (list -0.6 0.101 0.366 0.576)
+ ; intercept seedSD browse canopy species obsSD
+ set growth-params (list 1.24 0.248 -0.94 -0.488 0.116 1.188)
+ ]
+ if seedling-scenario = "randomdrought" [
+ ;drought conditions
+ ifelse random-float 1 < drought-prob [
+ set surv-params (list -0.531 0.197 0.528 0.44)
+ set growth-params (list 0.911 0.157 -0.871 -0.238 0.085 1.016)]
+ [;non-drought
+ set surv-params (list 2.596 0 -0.733 0)
+ set growth-params (list 2.091 0.199 -0.976 -1.733 0.468 1.468)]
+ ]
+ if seedling-scenario = "sensitivity" [
+ set growth-params (list (item 8 sens-params) 0.248 -0.94 -0.488 0.116 1.188)
+ set surv-params (list (item 9 sens-params) 0.101 0.366 0.576)
+ ]
+
+ set mast-year-index (mast-year-index + 1)
+ if mast-year-index = 9 [set mast-year-index 0]
+
+end
+
+to produce-acorns
+
+ let acorns-produced 0 let mast-mean mast-mean-bo let weevil-probability bo-weevil-prob
+
+ if species = "WO" [
+ set mast-mean mast-mean-wo set weevil-probability wo-weevil-prob
+ ]
+
+ set acorns-produced (pi * canopy-radius ^ 2 * random-exponential mast-mean) ;per meter squared
+
+ if in-core = TRUE [
+ set acorn-count (acorn-count + acorns-produced)
+ ifelse species = "BO" [set bo-acorn-count (bo-acorn-count + acorns-produced)]
+ [set wo-acorn-count (wo-acorn-count + acorns-produced)]
+ ]
+
+ if weevil-scenario != "custom" and weevil-scenario != "sensitivity" [
+
+ ifelse species = "BO" [set weevil-probability buffer-weev-prob-bo][set weevil-probability buffer-weev-prob-wo]
+ if in-core = TRUE and harvest-type = "shelterwood" and ticks > burnin [
+ ifelse species = "BO" [set weevil-probability core-weev-prob-bo][set weevil-probability core-weev-prob-wo]
+ ]
+ ]
+
+ let tree-radius canopy-radius let temp species let coretemp in-core
+
+ hatch-acorns acorns-produced [
+ ;;drop produced acorns under canopy randomly
+ set species temp
+ set in-core coretemp
+ set hidden? FALSE
+ right random 360
+ forward (random-float (tree-radius - 0.5)) + 0.5
+
+ ifelse random-float 1 < weevil-probability [set weevil TRUE] [set weevil FALSE] ;;check to see if weeviled
+ ]
+
+end
+
+
+to disperse-mast
+ ;parameter list
+ ;pRemoval, pRemoval given weeviled, weibSh, pDispEaten, pDispEatencache, pUndispEaten
+
+ ;Set up parameter sets depending on species
+ let core-acorn-params core-acorn-params-bo
+ let buffer-acorn-params buffer-acorn-params-bo
+ if species = "WO" [
+ set core-acorn-params core-acorn-params-wo
+ set buffer-acorn-params buffer-acorn-params-wo
+ ]
+
+ let acorn-params buffer-acorn-params
+ if harvest-type = "shelterwood" and ticks > burnin and in-core = TRUE [set acorn-params core-acorn-params]
+
+ ;;move mast via "dispersers"
+ ;;removal probability - HEE dispersal data for WO
+ let rem-prob 0
+ ifelse weevil = FALSE [set rem-prob item 0 acorn-params][set rem-prob item 1 acorn-params]
+
+ ifelse random-float 1 < rem-prob [
+
+ right random 360
+ ;;Weibull dispersal distance based on HEE data
+ set disp-dist random-weibull 1.400 (item 2 acorn-params)
+ forward disp-dist
+
+ ;;check to see if still in core area after being dispersed
+ check-in-core
+ ifelse harvest-type = "shelterwood" and ticks > burnin and in-core = TRUE [set acorn-params core-acorn-params]
+ [set acorn-params buffer-acorn-params]
+
+ ;Calculate probability of caching
+ let pcache 0
+ if dispersal-scenario = "yearly-diff" or dispersal-scenario = "yearly-treat-diff" [
+ let cache-mean -2.4986 + 0.08858 * disp-dist + random-normal 0 1.663
+ set pcache (1 + exp(-1 * cache-mean)) ^ (-1)]
+
+ if dispersal-scenario = "fixedaverage" or dispersal-scenario = "treat-diff"[
+ let cache-mean -2.4986 + 0.08858 * disp-dist
+ set pcache (1 + exp(-1 * cache-mean)) ^ (-1)]
+ if dispersal-scenario = "sensitivity" [
+ let cache-mean (item 4 sens-params) + 0.08858 * disp-dist
+ set pcache (1 + exp(-1 * cache-mean)) ^ (-1)]
+ if dispersal-scenario = "custom" [set pcache cache-prob]
+
+ ;Check if acorn is cached and if it survives
+ ifelse random-float 1 < pcache [
+ set cached TRUE
+ if random-float 1 < (item 4 acorn-params) [die]
+ ] [
+ set cached FALSE
+ if random-float 1 < (item 3 acorn-params) [die]
+ ]
+ ]
+ ;;check if eaten
+ [if random-float 1 < (item 5 acorn-params) [die]]
+end
+
+
+to germinate-mast
+ let temp species
+ ifelse cached = TRUE [
+ ;Based on Haas and Heske 2005 (0.77 for buried)
+ set germ germ-prob]
+ ;;Based on:
+ ;;Lombardo & McCarthy 2009 (0.26 germ for weeviled)
+ ;;and Haas and Heske 2005 (0.09 for surfaced)
+ [ifelse weevil = TRUE [set germ 0.09 * 0.26]
+ [set germ 0.09]]
+ if random-float 1 < germ [
+ hatch-oaks 1 [
+ set species temp
+ set age 0
+ set size 1
+ set seedling TRUE
+ set hidden? TRUE
+ set height 0.092 ;based on HEE seedlings
+ set indeffect (random-normal 0 item 1 growth-params)
+ ]]
+ die
+end
+
+
+to grow-seedling ;Based on HEE data
+ set light (1 - [shade] of patch-here)
+ let sp-ef 0
+ if species = "WO" [set sp-ef 1]
+ ;Calculate probability of being browsed
+ let pbrowse 0
+ if browse-scenario = "custom" or browse-scenario = "fixedaverage"[set pbrowse prob-browsed]
+ if browse-scenario = "hee"[
+ let br-mean -4.521 + 0.158 * (height * 100) - 0.00148 * (height * 100 * height * 100) + 0.372 * sp-ef + br-year-ef
+ set pbrowse (1 + exp(-1 * br-mean)) ^ (-1)
+ ]
+ if browse-scenario = "sensitivity"[
+ let br-mean (item 7 sens-params) + 0.158 * (height * 100) - 0.00148 * (height * 100 * height * 100) + 0.372 * sp-ef
+ set pbrowse (1 + exp(-1 * br-mean)) ^ (-1)
+ ]
+ ifelse random-float 1 < pbrowse [set browsed 1][set browsed 0]
+
+ let growthpred ((item 0 growth-params) + indeffect
+ + (item 2 growth-params) * browsed + (item 3 growth-params) * (1 - light)
+ + (item 4 growth-params) * sp-ef + (random-normal 0 item 5 growth-params))
+
+ set actual-growth (inv-neglog growthpred) / 100
+
+ if actual-growth < 0 [set actual-growth 0]
+
+ set height (height + actual-growth)
+
+ if height > 1.37 [ ;When seedling reaches dbh 0.01, height = 2.039
+
+ set dbh max list 0 ((-1.88 + 0.01372 * (height * 100)) / 100)
+ init-params ;Convert to sapling
+ ]
+end
+
+
+to check-seedling-survival ;Based on HEE data
+
+ if age = 0 [if random-float 1 > 0.567 [die]]
+
+ let sp-ef 0
+ if species = "WO" [set sp-ef 1]
+
+ let survpred ((item 0 surv-params) + (item 1 surv-params) * sp-ef + (item 2 surv-params)
+ * (1 - light) + (item 3 surv-params) * (min list 5 age))
+
+ let psurv (1 + exp(-1 * survpred)) ^ (-1)
+ if random-float 1 > psurv [die]
+
+ set age age + 1
+end
+
+
+;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;
+;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;
+
+
+;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;
+;;;;;;;; Light Calculation Procedures ;;;;;;;;
+;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;
+
+to calc-light
+ ask patches [set pcolor brown set stem-dens 0 set shade 0]
+ foreach sort-by [[height] of ?1 > [height] of ?2] turtles with [seedling = FALSE] [
+ ask ? [
+ let start-ycor ycor
+ let temp canopy-density
+ set light (1 - mean [shade] of patches in-radius max list 3 canopy-radius)
+ set density [stem-dens] of patch-here ;Stem density will penalize growth
+ ask patches in-radius density-dep [set stem-dens stem-dens + 1] ;Radius calibrated to prevent giant overshot in basal area over time
+
+ ;draw initial canopy
+ ;Shape of shade projection based on ShadeMotion simulation for 39 degrees latitude
+ ifelse canopy-radius < 2 [
+ ;Calculate overlapping shade instead of summing up leaf-weight; mathematically equivalent and fewer calculations to make.
+ ask patches in-radius canopy-radius [set shade (shade + (1 - shade) * temp) ]]
+ [
+ set ycor (ycor + canopy-radius / 2)
+ ask patches in-radius canopy-radius [
+ ;Interior circle
+ set in-layer1 TRUE
+ set shade (shade + (1 - shade) * temp)]
+
+ ;smaller ellipse (minus the initial canopy)
+ set ycor (ycor + canopy-radius / 2.5)
+ let test1 in-ellipse (canopy-radius * 2) (canopy-radius * 1.1) 1
+ ask test1 [
+ set in-layer2 TRUE
+ set shade (shade + (1 - shade) * temp * 0.5)]
+
+ ;larger ellipse (minus the smaller ellipse)
+ set ycor (ycor + canopy-radius / 4.5)
+ let test2 in-ellipse (canopy-radius * 3) (canopy-radius * 1.5) 2
+ ask test2 [
+ set shade (shade + (1 - shade) * temp * 0.25)]
+
+ ;reset layering
+ let test3 in-ellipse (canopy-radius * 3) (canopy-radius * 1.5) 3
+ ask test3 [set in-layer1 FALSE set in-layer2 FALSE]
+ ]
+
+ ;move tree back to starting location
+ set ycor start-ycor
+
+
+ ]]
+end
+
+;Modified from Wilensky, U. (2003) Netlogo Fur Model
+to-report in-ellipse [x-radius y-radius layer] ;; patch procedure
+
+ if layer = 1 [
+ report other patches in-radius (max list x-radius y-radius)
+ with [1.0 >= ((xdistance myself ^ 2) / (x-radius ^ 2)) +
+ ((ydistance myself ^ 2) / (y-radius ^ 2)) and in-layer1 = FALSE]]
+ if layer = 2 [
+ report other patches in-radius (max list x-radius y-radius)
+ with [1.0 >= ((xdistance myself ^ 2) / (x-radius ^ 2)) +
+ ((ydistance myself ^ 2) / (y-radius ^ 2))
+ and in-layer2 = FALSE and in-layer1 = FALSE]]
+ if layer = 3 [
+ report other patches in-radius (max list x-radius y-radius)
+ with [1.0 >= ((xdistance myself ^ 2) / (x-radius ^ 2)) +
+ ((ydistance myself ^ 2) / (y-radius ^ 2))]]
+end
+
+
+to-report xdistance [other-patch] ;; patch procedure
+ let xdiff abs (pxcor - [pxcor] of other-patch)
+ report min (list xdiff (world-width - xdiff))
+end
+
+to-report ydistance [other-patch] ;; patch procedure
+ let ydiff abs (pycor - [pycor] of other-patch)
+ report min (list ydiff (world-height - ydiff))
+end
+
+;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;
+;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;
+
+
+
+;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;
+;;;;;;;; Utility Procedures ;;;;;;;;
+;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;
+
+to-report random-weibull [Wshape Wscale]
+ let R random-float 1
+ report Wscale * (- ln(R)) ^ (1 / Wshape)
+end
+
+to check-in-core ;checks if agent is in the core or buffer area
+ ifelse xcor < xcutoff and xcor > (-1 * xcutoff) and ycor < ycutoff and ycor > (-1 * ycutoff)[set in-core TRUE][set in-core FALSE]
+end
+
+to-report inv-neglog [x] ;neg-log transformation function
+ ifelse x <= 1 [report (1 - exp(-1 * x))][report (exp(x) - 1)]
+end
+
+to calc-site-quality
+ ifelse manual-site-qual = TRUE [
+
+ set sitequal-woak sqwoak
+ set sitequal-boak sqboak
+ set sitequal-maple sqmap
+ set sitequal-poplar sqpop
+
+ ][
+ ;based bon Botkin 1993 and Holm 2012 with some guesses
+ ;white oak
+ let fT degree-days-index DegDays 1977 5894
+ let fWT saturation-index wt-dist 0.933
+ let fN nitrogen-index available-N "intermediate"
+ let fWL wilt-index wilt 0.45
+ set sitequal-woak fT * fWT * fN * fWL
+ ;black oak
+ set fT degree-days-index DegDays 2068 5421
+ set fWT saturation-index wt-dist 0.933
+ set fN nitrogen-index available-N "tolerant"
+ set fWL wilt-index wilt 0.45
+ set sitequal-boak fT * fWT * fN * fWL
+ ;maple
+ set fT degree-days-index DegDays 2000 6300
+ set fWT saturation-index wt-dist 0.567
+ set fN nitrogen-index available-N "intermediate"
+ set fWL wilt-index wilt 0.35
+ set sitequal-maple fT * fWT * fN * fWL
+ ;poplar
+ set fT degree-days-index DegDays 2171 6363
+ set fWT saturation-index wt-dist 0.544 ;;based on white spruce/red maple (similar moisture tolerance) ??
+ set fN nitrogen-index available-N "intolerant"
+ set fWL wilt-index wilt 0.245 ;;based on white spruce/red maple (similar moisture tolerance) ??
+ set sitequal-poplar fT * fWT * fN * fWL
+ ]
+end
+
+to init-stand [dens-adjust space-adjust n-oak n-maple n-poplar n-sap-oak n-sap-maple n-sap-poplar]
+ ;Simulate overstory of new forest stand
+ ;Create adult trees
+ create-oaks dens-adjust * n-oak [ ;dens-adjust adds more trees when there is a buffer
+ ifelse random-float 1 > 0.5 [set species "WO"][set species "BO"]
+ set dbh (random-normal 45.75 5) / 100
+ init-params
+ set age round (Amax * (dbh * 100) / Dmax)
+ ifelse space-adjust = TRUE [
+ loop [
+ setxy random-xcor random-ycor
+ if count oaks in-radius 7 < 2 [stop]]]
+ [setxy random-xcor random-ycor]
+ ]
+ create-maples dens-adjust * n-maple [
+ set dbh (random-normal 40.8 5) / 100
+ init-params
+ set age round (Amax * (dbh * 100) / Dmax)
+ ifelse space-adjust = TRUE [
+ loop [
+ setxy random-xcor random-ycor
+ if count turtles in-radius 8 < 2 [stop]]]
+ [setxy random-xcor random-ycor]
+ ]
+ create-poplars dens-adjust * n-poplar [
+ set dbh (random-normal 45.07 5) / 100
+ init-params
+ set age round (Amax * (dbh * 100) / Dmax)
+ ifelse space-adjust = TRUE [
+ loop [
+ setxy random-xcor random-ycor
+ if count turtles in-radius 8 < 2 [stop]]]
+ [setxy random-xcor random-ycor]
+ ]
+
+ calc-light
+
+ ;Create saplings
+ create-oaks dens-adjust * n-sap-oak [
+ set dbh max list 0.015 ((random-normal 14.9 5) / 100)
+ ifelse random-float 1 > 0.5 [set species "WO"][set species "BO"]
+ init-params
+ set age round (Amax * (dbh * 100) / Dmax)
+ loop [
+ setxy random-xcor random-ycor
+ if mean [shade] of patches in-radius 2 < 0.6 [stop]]
+ ]
+ create-maples dens-adjust * n-sap-maple [
+ set dbh max list 0.015 ((random-normal 10.3 5) / 100)
+ init-params
+ set age round (Amax * (dbh * 100) / Dmax)
+ setxy random-xcor random-ycor
+ ]
+ create-poplars dens-adjust * n-sap-poplar [
+ set dbh max list 0.015 ((random-normal 14.9 5) / 100)
+ init-params
+ set age round (Amax * (dbh * 100) / Dmax)
+ loop [
+ setxy random-xcor random-ycor
+ if mean [shade] of patches in-radius 2 < 0.6 [stop]]
+ ]
+
+ calc-light
+ ask turtles [check-in-core]
+
+end
+
+to init-params ;initialize new sapling oaks with parameter values
+ ifelse dbh < 0.1 [set shape "square"] [set size 2 set shape "tree"]
+ set hidden? FALSE
+ set actual-growth 0.02
+ set min-increment 0.01
+ set growth-mortality 0.369
+ set age 1
+ set seedling FALSE
+ if breed = oaks [
+ ifelse species = "WO" [;White oak based on Botkin 1993 and Holm 2012
+ set color white
+ set Dmax 100 set Hmax 3800 ;based on HEE data
+ set Amax 400 set intrinsic-mortality 4.0 / Amax ;Based on 2% reaching max age; common to all species
+ set b2 2 * (Hmax - 137) / (Dmax) set b3 (Hmax - 137) / (Dmax ^ 2)
+ set C 1.75
+ set G 104
+ set light-tol "intermediate"
+ set acorn-mean random-exponential 10 ;;mean oak production / m2 based on HEE histogram
+ ][set color black ;Black oak based on Botkin 1993 and Holm 2012
+ set Dmax 100 set Hmax 3800 ;based on HEE data
+ set Amax 300 set intrinsic-mortality 4.0 / Amax ;Based on 2% reaching max age; common to all species
+ set b2 2 * (Hmax - 137) / (Dmax) set b3 (Hmax - 137) / (Dmax ^ 2)
+ set C 1.75
+ set G 122
+ set light-tol "intermediate"
+ set acorn-mean random-exponential 10 ;;mean oak production / m2 based on HEE histogram
+ ]]
+ if breed = maples [ ;Sugar maple based bon Botkin 1993 and Holm 2012
+ set color sky
+ set Dmax 100 set Hmax 3350 ;set Dmax 170
+ set Amax 400 set intrinsic-mortality 4.0 / Amax
+ set b2 2 * (Hmax - 137) / (Dmax) set b3 (Hmax - 137) / (Dmax ^ 2)
+ set C 1.57
+ set G 118.7
+ set light-tol "high"
+ ]
+ if breed = poplars [ ;Tulip poplar based on Holm 2012 with some guesses
+ set color red
+ set Dmax 100 set Hmax 4000
+ set Amax 300 set intrinsic-mortality 4.0 / Amax
+ set b2 2 * (Hmax - 137) / (Dmax) set b3 (Hmax - 137) / (Dmax ^ 2)
+ set C 1.75 ;assumed to be similar to oak
+ set G 140
+ set light-tol "low"
+ ]
+ allometerize dbh
+
+end
+
+to calc-global-vars ;Calculate global reporter values
+
+ let adjust (x-core * y-core) / 10000
+
+ set basal-area (sum [ba] of turtles with [dbh >= 0.015 and in-core = TRUE]) / adjust
+ set basal-area-ovs (sum [ba] of turtles with [dbh >= 0.3 and in-core = TRUE]) / adjust
+ set basal-area-ft basal-area * 4.356
+ set ba-oak (sum [ba] of oaks with [dbh >= 0.015 and in-core = TRUE]) / adjust
+ set ba-oak-ovs (sum [ba] of oaks with [dbh >= 0.3 and in-core = TRUE]) / adjust
+ set ba-map (sum [ba] of maples with [dbh >= 0.015 and in-core = TRUE]) / adjust
+ set ba-map-ovs (sum [ba] of maples with [dbh >= 0.3 and in-core = TRUE]) / adjust
+ set ba-pop (sum [ba] of poplars with [dbh >= 0.015 and in-core = TRUE]) / adjust
+ set ba-pop-ovs (sum [ba] of poplars with [dbh >= 0.3 and in-core = TRUE]) / adjust
+ set dens (count turtles with [dbh >= 0.015 and in-core = TRUE]) / adjust
+ set dens-ovs (count turtles with [dbh >= 0.3 and in-core = TRUE]) / adjust
+ set dens-ac dens * 0.40477
+ set dens-oak (count oaks with [dbh >= 0.015 and in-core = TRUE]) / adjust
+ set dens-oak-ovs (count oaks with [dbh >= 0.3 and in-core = TRUE]) / adjust
+ set dens-map (count maples with [dbh >= 0.015 and in-core = TRUE]) / adjust
+ set dens-map-ovs (count maples with [dbh >= 0.3 and in-core = TRUE]) / adjust
+ set dens-pop (count poplars with [dbh >= 0.015 and in-core = TRUE]) / adjust
+ set dens-pop-ovs (count poplars with [dbh >= 0.3 and in-core = TRUE]) / adjust
+ ifelse basal-area > 0 [
+ set qdbh sqrt(basal-area * adjust / (0.0000785 * count turtles with [dbh >= 0.015 and in-core = TRUE]))
+ set qdbh-in 2 * (sqrt(mean [ba] of turtles with [dbh >= 0.015 and in-core = TRUE] / pi)) * 39.37
+ set prop-oak (ba-oak / basal-area)
+ set prop-tol (ba-map / basal-area)
+ set prop-intol (ba-pop / basal-area)
+ ][set qdbh 0 set qdbh-in 0 set prop-oak 0 set prop-tol 0 set prop-intol 0]
+ set qdbh-ovs sqrt(basal-area-ovs * adjust / (0.0000785 * (max list 1 count turtles with [dbh >= 0.3 and in-core = TRUE])))
+ set qdbh-oak sqrt(ba-oak * adjust / (0.0000785 * (max list 1 count oaks with [dbh >= 0.015 and in-core = TRUE])))
+ set qdbh-oak-ovs sqrt(ba-oak-ovs * adjust / (0.0000785 * (max list 1 count oaks with [dbh >= 0.3 and in-core = TRUE])))
+ set qdbh-map sqrt(ba-map * adjust / (0.0000785 * (max list 1 count maples with [dbh >= 0.015 and in-core = TRUE])))
+ set qdbh-map-ovs sqrt(ba-map-ovs * adjust / (0.0000785 * (max list 1 count maples with [dbh >= 0.3 and in-core = TRUE])))
+ set qdbh-pop sqrt(ba-pop * adjust / (0.0000785 * (max list 1 count poplars with [dbh >= 0.015 and in-core = TRUE])))
+ set qdbh-pop-ovs sqrt(ba-pop-ovs * adjust / (0.0000785 * (max list 1 count poplars with [dbh >= 0.3 and in-core = TRUE])))
+ set total-seedlings (count turtles with [seedling = TRUE and in-core = TRUE]) / adjust
+ set new-seedlings (count turtles with [seedling = TRUE and age = 1 and in-core = TRUE]) / adjust
+ set new-bo-seedlings (count turtles with [seedling = TRUE and species = "BO" and age = 1 and in-core = TRUE]) / adjust
+ set new-wo-seedlings (count turtles with [seedling = TRUE and species = "WO" and age = 1 and in-core = TRUE]) / adjust
+ set seedlings-class1 (count turtles with [seedling = TRUE and height <= 0.3 and in-core = TRUE]) / adjust
+ set seedlings-class2 (count turtles with [seedling = TRUE and height > 0.3 and height <= 0.6 and in-core = TRUE]) / adjust
+ set seedlings-class3 (count turtles with [seedling = TRUE and height > 0.6 and height <= 1.4 and in-core = TRUE]) / adjust
+ set seedlings-class123 (count turtles with [seedling = TRUE and height <= 1.4 and in-core = TRUE]) / adjust
+ set seedbo-class123 (count turtles with [seedling = TRUE and species = "BO" and height <= 1.4 and in-core = TRUE]) / adjust
+ set seedwo-class123 (count turtles with [seedling = TRUE and species = "WO" and height <= 1.4 and in-core = TRUE]) / adjust
+ set seedlings-class4 (count oaks with [height > 1.4 and dbh < 0.015 and in-core = TRUE]) / adjust
+ set seedbo-class4 (count oaks with [height > 1.4 and species = "BO" and dbh < 0.015 and in-core = TRUE]) / adjust
+ set seedwo-class4 (count oaks with [height > 1.4 and species = "WO" and dbh < 0.015 and in-core = TRUE]) / adjust
+ set total-acorns round (acorn-count / adjust)
+ set total-bo-acorns round (bo-acorn-count / adjust)
+ set total-wo-acorns round (wo-acorn-count / adjust)
+ ifelse count oaks with [dbh > 0.2 and in-core = TRUE] > 0 [set acorns-pertree (total-acorns / count oaks with [dbh > 0.2 and in-core = TRUE])]
+ [set acorns-pertree 0]
+ ifelse total-acorns > 0 [set pct-germ (new-seedlings / total-acorns)][set pct-germ 0]
+ ifelse total-bo-acorns > 0 [set pct-bo-germ (new-bo-seedlings / total-bo-acorns)][set pct-bo-germ 0]
+ ifelse total-wo-acorns > 0 [set pct-wo-germ (new-wo-seedlings / total-wo-acorns)][set pct-wo-germ 0]
+ set regen-dens (count oaks with [height >= 1.4 and dbh < 0.015 and in-core = TRUE]) / adjust
+ set regen-stump-dens (count oaks with [sprout? = TRUE and height >= 1.4 and dbh < 0.015 and in-core = TRUE]) / adjust
+
+end
+
+to color-patches
+ if shade > 0 [set pcolor lime + 1]
+ if shade > 0.1 [set pcolor lime]
+ if shade > 0.2 [set pcolor lime - 1]
+ if shade > 0.3 [set pcolor green + 2]
+ if shade > 0.5 [set pcolor green + 1]
+ if shade > 0.7 [set pcolor green]
+ if shade > 0.8 [set pcolor green - 1]
+ if shade > 0.9 [set pcolor green - 2]
+ if shade > 0.95 [set pcolor green - 3]
+end
+
+
+;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;
+;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;
+
+
+
+;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;
+;;;;;;;; Harvesting Procedures ;;;;;;;;
+;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;
+
+to conduct-harvest
+
+ if harvest-type = "none" [stop]
+
+ if (ticks + 1) != harvest-year [stop]
+
+ let adjust (x-core * y-core) / 10000
+
+ if harvest-type = "clearcut" [
+ ask turtles with [dbh > 0.01 and in-core = TRUE] [create-sprout]
+ ]
+
+ if harvest-type = "singletree" [
+ set basal-area (sum [ba] of turtles with [dbh >= 0.01 and in-core = TRUE]) / adjust
+ set harvest-year harvest-year + 20
+ if basal-area >= 25 [
+ loop [
+ let potential one-of turtles with [dbh >= 0.10 and in-core = TRUE]
+ ask potential [create-sprout]
+ set basal-area (sum [ba] of turtles with [dbh >= 0.01 and in-core = TRUE]) / adjust
+ if basal-area <= 25 [stop]
+ ]
+ ]
+ ]
+
+ if harvest-type = "shelterwood" [
+ ifelse shelter-phase = 1 [
+ set shelter-phase 2
+ ask turtles with [breed != oaks and dbh <= 0.254 and in-core = TRUE] [die] ;treated with herbicide
+ set harvest-year (harvest-year + 7)
+ ]
+ [
+ ifelse shelter-phase = 2 [
+ set shelter-phase 3
+ set harvest-year (harvest-year + 8)
+ if basal-area >= 16.1 [
+ loop [
+ let potential min-one-of turtles with [age > 20 and in-core = TRUE] [light]
+ ask potential [ifelse breed = oaks [create-sprout][die]]
+ set basal-area (sum [ba] of turtles with [dbh >= 0.01 and in-core = TRUE]) / adjust
+ if basal-area <= 16.1 [stop]
+ ]
+ ]
+ ]
+ [
+ ask turtles with [age > 20 and in-core = TRUE] [create-sprout]
+ set shelter-phase 1
+ ]
+
+ ]]
+
+end
diff --git a/netlogostyle.tex b/netlogostyle.tex
new file mode 100644
index 0000000..faad242
--- /dev/null
+++ b/netlogostyle.tex
@@ -0,0 +1,18 @@
+\lstdefinelanguage{NetLogo}{
+ morekeywords=[1]{to,to-report,end, extensions,to-report,globals,breed,directed-link-breed, link-breed},
+ extendedchars=true,
+ breaklines=true,
+ breakatwhitespace=true,
+ basicstyle=\linespread{0.5}\scriptsize,
+ columns=fullflexible,
+ tabsize=4,
+ keywordstyle=[1]\bfseries,
+ morekeywords=[2]{let,set,loop, if, report, foreach, print, tick, while, every, ifelse, ask},
+ keywordstyle=[2]\bfseries,
+ alsoletter={-,?,.},
+ morekeywords=[3]{position, not, reverse, fput, substring, length, word, timer,is-number?,filter,first, bf, butfirst, last, empty?, n-of, color, who, shape},
+ keywordstyle=[3]\bfseries,
+ comment=[l]{\;},
+ string=[d]{"},
+}
+
diff --git a/notoccite.sty b/notoccite.sty
new file mode 100644
index 0000000..af4ec3b
--- /dev/null
+++ b/notoccite.sty
@@ -0,0 +1,43 @@
+% notoccite.sty no t.o.c. cite Jul 20, 2000
+% Donald Arseneau asnd@triumf.ca TRIUMF, Vancouver, Canada,
+% This is unrestricted software contributed to the public domain.
+%
+% Ordinarily, cites used in titles or figure captions also appear in
+% the table of contents and list of figures. If you then run bibtex
+% using the unsrt (unsorted) style, they get numbered starting from 1,
+% not the number they should have in the main text.
+%
+% A good option is to avoid cites in titles, and to specify optional
+% caption text without cites:
+% \caption [Picture of a bird.]{Picture of a bird \cite{audobon}.}
+%
+% If you must use moving cites, you could manage them by deleting toc
+% and lof files, then running latex once, then bibtex. However, the
+% following definition fixes the problem so you don't need to worry
+% about that.
+%
+% *NOTE* This definition works for the ordinary LaTeX definitions
+% for \cite and others (\addtocontents, \label) but it may well
+% fail when used with various packages for citations or cross
+% references.
+%
+% It works by locally setting \@fileswfalse, which is something like
+% \nofiles, but \@fileswfalse does not affect \label or \addtocontents.
+% \nofiles does most of its work by redefining \protected@write, and
+% neither \addtocontents nor \label check for \if@filesw. \cite *does*
+% check \if@filesw.
+
+\ProvidesPackage{notoccite}[2000/07/20]
+\def\@starttoc#1{%
+ \begingroup
+ \@fileswfalse
+ \makeatletter
+ \@input{\jobname.#1}%
+ \endgroup
+ \if@filesw
+ \expandafter\newwrite\csname tf@#1\endcsname
+ \immediate\openout \csname tf@#1\endcsname \jobname.#1\relax
+ \fi
+ \@nobreakfalse
+}
+
diff --git a/pulongtable.sty b/pulongtable.sty
new file mode 100644
index 0000000..f816cdd
--- /dev/null
+++ b/pulongtable.sty
@@ -0,0 +1,446 @@
+%%
+%% pulongtable.sty 2006-09-14 11:37 Mark Senn http://www.ecn.purdue.edu/~mark
+%%
+%% THIS is pulongtable.sty. It is based on longtable.sty.
+%%
+%% This is file `longtable.sty',
+%% generated with the docstrip utility.
+%%
+%% The original source files were:
+%%
+%% longtable.dtx (with options: `package')
+%%
+%% This is a generated file.
+%%
+%% Copyright 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003
+%% The LaTeX3 Project and any individual authors listed elsewhere
+%% in this file.
+%%
+%% This file was generated from file(s) of the Standard LaTeX `Tools Bundle'.
+%% --------------------------------------------------------------------------
+%%
+%% It may be distributed and/or modified under the
+%% conditions of the LaTeX Project Public License, either version 1.3
+%% of this license or (at your option) any later version.
+%% The latest version of this license is in
+%% http://www.latex-project.org/lppl.txt
+%% and version 1.3 or later is part of all distributions of LaTeX
+%% version 2003/12/01 or later.
+%%
+%% This file may only be distributed together with a copy of the LaTeX
+%% `Tools Bundle'. You may however distribute the LaTeX `Tools Bundle'
+%% without such generated files.
+%%
+%% The list of all files belonging to the LaTeX `Tools Bundle' is
+%% given in the file `manifest.txt'.
+%%
+%% File: longtable.dtx Copyright (C) 1990-2001 David Carlisle
+\NeedsTeXFormat{LaTeX2e}[1995/06/01]
+\ProvidesPackage{pulongtable}
+ [2004/02/01 v4.11 Multi-page Table package (DPC)]
+\def\LT@err{\PackageError{longtable}}
+\def\LT@warn{\PackageWarning{longtable}}
+\def\LT@final@warn{%
+ \AtEndDocument{%
+ \LT@warn{Table \@width s have changed. Rerun LaTeX.\@gobbletwo}}%
+ \global\let\LT@final@warn\relax}
+\DeclareOption{errorshow}{%
+ \def\LT@warn{\PackageInfo{longtable}}}
+\DeclareOption{pausing}{%
+ \def\LT@warn#1{%
+ \LT@err{#1}{This is not really an error}}}
+\DeclareOption{set}{}
+\DeclareOption{final}{}
+\ProcessOptions
+\newskip\LTleft \LTleft=\fill
+\newskip\LTright \LTright=\fill
+\newskip\LTpre \LTpre=\bigskipamount
+\newskip\LTpost \LTpost=\bigskipamount
+\newcount\LTchunksize \LTchunksize=20
+\let\c@LTchunksize\LTchunksize
+\newdimen\LTcapwidth \LTcapwidth=4in
+\newbox\LT@head
+\newbox\LT@firsthead
+\newbox\LT@foot
+\newbox\LT@lastfoot
+\newcount\LT@cols
+\newcount\LT@rows
+\newcounter{LT@tables}
+\newcounter{LT@chunks}[LT@tables]
+%\ifx\c@table\undefined
+% \newcounter{table}
+% \def\fnum@table{\tablename~\thetable}
+%\fi
+%\ifx\tablename\undefined
+% \def\tablename{Table}
+%\fi
+\newtoks\LT@p@ftn
+\mathchardef\LT@end@pen=30000
+\def\longtable{%
+ \par
+ \ifx\multicols\@undefined
+ \else
+ \ifnum\col@number>\@ne
+ \@twocolumntrue
+ \fi
+ \fi
+ \if@twocolumn
+ \LT@err{longtable not in 1-column mode}\@ehc
+ \fi
+ \begingroup
+ \@ifnextchar[\LT@array{\LT@array[x]}}
+\def\LT@array[#1]#2{%
+ \refstepcounter{table}\stepcounter{LT@tables}%
+ \if l#1%
+ \LTleft\z@ \LTright\fill
+ \else\if r#1%
+ \LTleft\fill \LTright\z@
+ \else\if c#1%
+ \LTleft\fill \LTright\fill
+ \fi\fi\fi
+ \let\LT@mcol\multicolumn
+ \let\LT@@tabarray\@tabarray
+ \let\LT@@hl\hline
+ \def\@tabarray{%
+ \let\hline\LT@@hl
+ \LT@@tabarray}%
+ \let\\\LT@tabularcr\let\tabularnewline\\%
+ \def\newpage{\noalign{\break}}%
+ \def\pagebreak{\noalign{\ifnum`}=0\fi\@testopt{\LT@no@pgbk-}4}%
+ \def\nopagebreak{\noalign{\ifnum`}=0\fi\@testopt\LT@no@pgbk4}%
+ \let\hline\LT@hline \let\kill\LT@kill\let\caption\LT@caption
+ \@tempdima\ht\strutbox
+ \let\@endpbox\LT@endpbox
+ \ifx\extrarowheight\@undefined
+ \let\@acol\@tabacol
+ \let\@classz\@tabclassz \let\@classiv\@tabclassiv
+ \def\@startpbox{\vtop\LT@startpbox}%
+ \let\@@startpbox\@startpbox
+ \let\@@endpbox\@endpbox
+ \let\LT@LL@FM@cr\@tabularcr
+ \else
+ \advance\@tempdima\extrarowheight
+ \col@sep\tabcolsep
+ \let\@startpbox\LT@startpbox\let\LT@LL@FM@cr\@arraycr
+ \fi
+ \setbox\@arstrutbox\hbox{\vrule
+ \@height \arraystretch \@tempdima
+ \@depth \arraystretch \dp \strutbox
+ \@width \z@}%
+ \let\@sharp##\let\protect\relax
+ \begingroup
+ \@mkpream{#2}%
+ \xdef\LT@bchunk{%
+ \global\advance\c@LT@chunks\@ne
+ \global\LT@rows\z@\setbox\z@\vbox\bgroup
+ \LT@setprevdepth
+ \tabskip\LTleft \noexpand\halign to\hsize\bgroup
+ \tabskip\z@ \@arstrut \@preamble \tabskip\LTright \cr}%
+ \endgroup
+ \expandafter\LT@nofcols\LT@bchunk&\LT@nofcols
+ \LT@make@row
+ \m@th\let\par\@empty
+ \everycr{}\lineskip\z@\baselineskip\z@
+ \LT@bchunk}
+\def\LT@no@pgbk#1[#2]{\penalty #1\@getpen{#2}\ifnum`{=0\fi}}
+\def\LT@start{%
+ \let\LT@start\endgraf
+ \endgraf\penalty\z@\vskip\LTpre
+ \dimen@\pagetotal
+ \advance\dimen@ \ht\ifvoid\LT@firsthead\LT@head\else\LT@firsthead\fi
+ \advance\dimen@ \dp\ifvoid\LT@firsthead\LT@head\else\LT@firsthead\fi
+ \advance\dimen@ \ht\LT@foot
+ \dimen@ii\vfuzz
+ \vfuzz\maxdimen
+ \setbox\tw@\copy\z@
+ \setbox\tw@\vsplit\tw@ to \ht\@arstrutbox
+ \setbox\tw@\vbox{\unvbox\tw@}%
+ \vfuzz\dimen@ii
+ \advance\dimen@ \ht
+ \ifdim\ht\@arstrutbox>\ht\tw@\@arstrutbox\else\tw@\fi
+ \advance\dimen@\dp
+ \ifdim\dp\@arstrutbox>\dp\tw@\@arstrutbox\else\tw@\fi
+ \advance\dimen@ -\pagegoal
+ \ifdim \dimen@>\z@\vfil\break\fi
+ \global\@colroom\@colht
+ \ifvoid\LT@foot\else
+ \advance\vsize-\ht\LT@foot
+ \global\advance\@colroom-\ht\LT@foot
+ \dimen@\pagegoal\advance\dimen@-\ht\LT@foot\pagegoal\dimen@
+ \maxdepth\z@
+ \fi
+ \ifvoid\LT@firsthead\copy\LT@head\else\box\LT@firsthead\fi\nobreak
+ \output{\LT@output}}
+\def\endlongtable{%
+ \crcr
+ \noalign{%
+ \let\LT@entry\LT@entry@chop
+ \xdef\LT@save@row{\LT@save@row}}%
+ \LT@echunk
+ \LT@start
+ \unvbox\z@
+ \LT@get@widths
+ \if@filesw
+ {\let\LT@entry\LT@entry@write\immediate\write\@auxout{%
+ \gdef\expandafter\noexpand
+ \csname LT@\romannumeral\c@LT@tables\endcsname
+ {\LT@save@row}}}%
+ \fi
+ \ifx\LT@save@row\LT@@save@row
+ \else
+ \LT@warn{Column \@width s have changed\MessageBreak
+ in table \thetable}%
+ \LT@final@warn
+ \fi
+ \endgraf\penalty -\LT@end@pen
+ \endgroup
+ \global\@mparbottom\z@
+ \pagegoal\vsize
+ \endgraf\penalty\z@\addvspace\LTpost
+ \ifvoid\footins\else\insert\footins{}\fi}
+\def\LT@nofcols#1&{%
+ \futurelet\@let@token\LT@n@fcols}
+\def\LT@n@fcols{%
+ \advance\LT@cols\@ne
+ \ifx\@let@token\LT@nofcols
+ \expandafter\@gobble
+ \else
+ \expandafter\LT@nofcols
+ \fi}
+\def\LT@tabularcr{%
+ \relax\iffalse{\fi\ifnum0=`}\fi
+ \@ifstar
+ {\def\crcr{\LT@crcr\noalign{\nobreak}}\let\cr\crcr
+ \LT@t@bularcr}%
+ {\LT@t@bularcr}}
+\let\LT@crcr\crcr
+\let\LT@setprevdepth\relax
+\def\LT@t@bularcr{%
+ \global\advance\LT@rows\@ne
+ \ifnum\LT@rows=\LTchunksize
+ \gdef\LT@setprevdepth{%
+ \prevdepth\z@\global
+ \global\let\LT@setprevdepth\relax}%
+ \expandafter\LT@xtabularcr
+ \else
+ \ifnum0=`{}\fi
+ \expandafter\LT@LL@FM@cr
+ \fi}
+\def\LT@xtabularcr{%
+ \@ifnextchar[\LT@argtabularcr\LT@ntabularcr}
+\def\LT@ntabularcr{%
+ \ifnum0=`{}\fi
+ \LT@echunk
+ \LT@start
+ \unvbox\z@
+ \LT@get@widths
+ \LT@bchunk}
+\def\LT@argtabularcr[#1]{%
+ \ifnum0=`{}\fi
+ \ifdim #1>\z@
+ \unskip\@xargarraycr{#1}%
+ \else
+ \@yargarraycr{#1}%
+ \fi
+ \LT@echunk
+ \LT@start
+ \unvbox\z@
+ \LT@get@widths
+ \LT@bchunk}
+\def\LT@echunk{%
+ \crcr\LT@save@row\cr\egroup
+ \global\setbox\@ne\lastbox
+ \unskip
+ \egroup}
+\def\LT@entry#1#2{%
+ \ifhmode\@firstofone{&}\fi\omit
+ \ifnum#1=\c@LT@chunks
+ \else
+ \kern#2\relax
+ \fi}
+\def\LT@entry@chop#1#2{%
+ \noexpand\LT@entry
+ {\ifnum#1>\c@LT@chunks
+ 1}{0pt%
+ \else
+ #1}{#2%
+ \fi}}
+\def\LT@entry@write{%
+ \noexpand\LT@entry^^J%
+ \@spaces}
+\def\LT@kill{%
+ \LT@echunk
+ \LT@get@widths
+ \expandafter\LT@rebox\LT@bchunk}
+\def\LT@rebox#1\bgroup{%
+ #1\bgroup
+ \unvbox\z@
+ \unskip
+ \setbox\z@\lastbox}
+\def\LT@blank@row{%
+ \xdef\LT@save@row{\expandafter\LT@build@blank
+ \romannumeral\number\LT@cols 001 }}
+\def\LT@build@blank#1{%
+ \if#1m%
+ \noexpand\LT@entry{1}{0pt}%
+ \expandafter\LT@build@blank
+ \fi}
+\def\LT@make@row{%
+ \global\expandafter\let\expandafter\LT@save@row
+ \csname LT@\romannumeral\c@LT@tables\endcsname
+ \ifx\LT@save@row\relax
+ \LT@blank@row
+ \else
+ {\let\LT@entry\or
+ \if!%
+ \ifcase\expandafter\expandafter\expandafter\LT@cols
+ \expandafter\@gobble\LT@save@row
+ \or
+ \else
+ \relax
+ \fi
+ !%
+ \else
+ \aftergroup\LT@blank@row
+ \fi}%
+ \fi}
+\let\setlongtables\relax
+\def\LT@get@widths{%
+ \setbox\tw@\hbox{%
+ \unhbox\@ne
+ \let\LT@old@row\LT@save@row
+ \global\let\LT@save@row\@empty
+ \count@\LT@cols
+ \loop
+ \unskip
+ \setbox\tw@\lastbox
+ \ifhbox\tw@
+ \LT@def@row
+ \advance\count@\m@ne
+ \repeat}%
+ \ifx\LT@@save@row\@undefined
+ \let\LT@@save@row\LT@save@row
+ \fi}
+\def\LT@def@row{%
+ \let\LT@entry\or
+ \edef\@tempa{%
+ \ifcase\expandafter\count@\LT@old@row
+ \else
+ {1}{0pt}%
+ \fi}%
+ \let\LT@entry\relax
+ \xdef\LT@save@row{%
+ \LT@entry
+ \expandafter\LT@max@sel\@tempa
+ \LT@save@row}}
+\def\LT@max@sel#1#2{%
+ {\ifdim#2=\wd\tw@
+ #1%
+ \else
+ \number\c@LT@chunks
+ \fi}%
+ {\the\wd\tw@}}
+\def\LT@hline{%
+ \noalign{\ifnum0=`}\fi
+ \penalty\@M
+ \futurelet\@let@token\LT@@hline}
+\def\LT@@hline{%
+ \ifx\@let@token\hline
+ \global\let\@gtempa\@gobble
+ \gdef\LT@sep{\penalty-\@medpenalty\vskip\doublerulesep}%
+ \else
+ \global\let\@gtempa\@empty
+ \gdef\LT@sep{\penalty-\@lowpenalty\vskip-\arrayrulewidth}%
+ \fi
+ \ifnum0=`{\fi}%
+ \multispan\LT@cols
+ \unskip\leaders\hrule\@height\arrayrulewidth\hfill\cr
+ \noalign{\LT@sep}%
+ \multispan\LT@cols
+ \unskip\leaders\hrule\@height\arrayrulewidth\hfill\cr
+ \noalign{\penalty\@M}%
+ \@gtempa}
+\def\LT@caption{%
+ \noalign\bgroup
+ \@ifnextchar[{\egroup\LT@c@ption\@firstofone}\LT@capti@n}
+\def\LT@c@ption#1[#2]#3{%
+ \LT@makecaption#1\fnum@table{#3}%
+ \def\@tempa{#2}%
+ \ifx\@tempa\@empty\else
+ {\let\\\space
+ \addcontentsline{lot}{table}{\protect\numberline{\thetable}{#2}}}%
+ \fi}
+\def\LT@capti@n{%
+ \@ifstar
+ {\egroup\LT@c@ption\@gobble[]}%
+ {\egroup\@xdblarg{\LT@c@ption\@firstofone}}}
+\def\LT@makecaption#1#2#3{%
+ \LT@mcol\LT@cols c{\hbox to\z@{\hss\parbox[t]\LTcapwidth{%
+ \sbox\@tempboxa{#1{#2: }#3}%
+ \ifdim\wd\@tempboxa>\hsize
+ #1{#2: }#3%
+ \else
+ \hbox to\hsize{\hfil\box\@tempboxa\hfil}%
+ \fi
+ \endgraf\vskip\baselineskip}%
+ \hss}}}
+\def\LT@output{%
+ \ifnum\outputpenalty <-\@Mi
+ \ifnum\outputpenalty > -\LT@end@pen
+ \LT@err{floats and marginpars not allowed in a longtable}\@ehc
+ \else
+ \setbox\z@\vbox{\unvbox\@cclv}%
+ \ifdim \ht\LT@lastfoot>\ht\LT@foot
+ \dimen@\pagegoal
+ \advance\dimen@-\ht\LT@lastfoot
+ \ifdim\dimen@<\ht\z@
+ \setbox\@cclv\vbox{\unvbox\z@\copy\LT@foot\vss}%
+ \@makecol
+ \@outputpage
+ \setbox\z@\vbox{\box\LT@head}%
+ \fi
+ \fi
+ \global\@colroom\@colht
+ \global\vsize\@colht
+ \vbox
+ {\unvbox\z@\box\ifvoid\LT@lastfoot\LT@foot\else\LT@lastfoot\fi}%
+ \fi
+ \else
+ \setbox\@cclv\vbox{\unvbox\@cclv\copy\LT@foot\vss}%
+ \@makecol
+ \@outputpage
+ \global\vsize\@colroom
+ \copy\LT@head\nobreak
+ \fi}
+\def\LT@end@hd@ft#1{%
+ \LT@echunk
+ \ifx\LT@start\endgraf
+ \LT@err
+ {Longtable head or foot not at start of table}%
+ {Increase LTchunksize}%
+ \fi
+ \setbox#1\box\z@
+ \LT@get@widths
+ \LT@bchunk}
+\def\endfirsthead{\LT@end@hd@ft\LT@firsthead}
+\def\endhead{\LT@end@hd@ft\LT@head}
+\def\endfoot{\LT@end@hd@ft\LT@foot}
+\def\endlastfoot{\LT@end@hd@ft\LT@lastfoot}
+\def\LT@startpbox#1{%
+ \bgroup
+ \let\@footnotetext\LT@p@ftntext
+ \setlength\hsize{#1}%
+ \@arrayparboxrestore
+ \vrule \@height \ht\@arstrutbox \@width \z@}
+\def\LT@endpbox{%
+ \@finalstrut\@arstrutbox
+ \egroup
+ \the\LT@p@ftn
+ \global\LT@p@ftn{}%
+ \hfil}
+\def\LT@p@ftntext#1{%
+ \edef\@tempa{\the\LT@p@ftn\noexpand\footnotetext[\the\c@footnote]}%
+ \global\LT@p@ftn\expandafter{\@tempa{#1}}}%
+\endinput
+%%
+%% End of file `longtable.sty'.
diff --git a/puthesis.cls b/puthesis.cls
new file mode 100644
index 0000000..e7f14e0
--- /dev/null
+++ b/puthesis.cls
@@ -0,0 +1,1605 @@
+%
+% puthesis.cls 2014-07-07 Mark Senn http://engineering.purdue.edu/~mark
+%
+% INDEX: Purdue University thesis document class
+%
+% DESCRIPTION:
+%
+% This is a LaTeX document class for Purdue University theses.
+%
+% USAGE:
+%
+% See http://engineering.purdue.edu/~mark/puthesis for more information.
+%
+
+\NeedsTeXFormat{LaTeX2e}
+\ProvidesClass{puthesis}[2017/07/07 Purdue thesis class]
+\usepackage{ifthen}
+\newcommand{\ifthen}[2]{\ifthenelse{#1}{#2}{}}
+\usepackage{endnotes}
+\usepackage{pulongtable}
+\usepackage{rotating}
+\usepackage{lscape}
+
+\makeatletter
+
+\def\@@number{\string##}
+
+\newcount{\@@i}
+\newcounter{@@tcount}
+\newcounter{@@volume}
+\newcounter{last@@volume}
+\newcounter{save@@page}
+\newlength{\@@captionwidth}
+\newlength{\@@parindent} \setlength{\@@parindent}{\parindent}
+\newlength{\@@padding}
+\newlength{\@@tlength}
+\newlength{\@@ulength}
+\newcommand{\@@dept}{unknown}
+\newcommand{\set@@dept}[1]{\renewcommand{\@@dept}{#1}}
+
+\newcommand{\@@Repeat}[2]{%
+ \@@i=0
+ \loop
+ \ifnum\@@i<#2
+ #1
+ \advance\@@i by 1
+ \repeat
+}
+
+\newcommand{\@@blankpage}{%
+ \clearpage
+ \mbox{}\clearpage
+}
+
+\newcommand{\articlepages}[1]{%
+ \@@Repeat{\@@blankpage}{#1}
+}
+
+\newif\if@@more
+\@@moretrue
+\newcommand{\articleinclude}[1]{%
+ \def\@@t{#1}
+ \@@i=0
+ \loop
+ \advance\@@i by 1
+ \def\@@u{\@@t\the\@@i.eps}
+ \immediate\write16{\@@u}
+ \IfFileExists{\@@u}{\noindent\includegraphics[width=\textwidth]{\@@u}\newpage}{\@@morefalse}
+ \if@@more\repeat
+}
+
+\DeclareOption{iupuiece}{\set@@dept{iupuiece}}
+\DeclareOption{aae}{\set@@dept{aae}}
+\DeclareOption{agecon}{\set@@dept{agecon}}
+\DeclareOption{agry}{\set@@dept{agry}}
+\DeclareOption{ansc}{\set@@dept{ansc}}
+\DeclareOption{che}{\set@@dept{che}}
+\DeclareOption{ce}{\set@@dept{ce}}
+\DeclareOption{edci}{\set@@dept{edci}}
+\DeclareOption{ece}{\set@@dept{ece}}
+\DeclareOption{hsci}{\set@@dept{hsci}}
+\DeclareOption{ie}{\set@@dept{ie}}
+\DeclareOption{ling}{\set@@dept{ling}}
+\DeclareOption{mgmt}{\set@@dept{mgmt}}
+\DeclareOption{mse}{\set@@dept{mse}}
+\DeclareOption{me}{\set@@dept{me}}
+\DeclareOption{ne}{\set@@dept{ne}}
+\DeclareOption{chem}{\set@@dept{chem}}
+\DeclareOption{cs}{\set@@dept{cs}}
+\DeclareOption{eas}{\set@@dept{eas}}
+\DeclareOption{math}{\set@@dept{math}}
+\DeclareOption{phys}{\set@@dept{phys}}
+\DeclareOption{stat}{\set@@dept{stat}}
+\newcommand{\pendnotes}{}
+\DeclareOption{endnotes}{%
+ \let\footnote=\endnote
+ \renewcommand{\pendnotes}{%
+ \newpage
+ \begingroup
+%% ! \renewcommand{\baselinestretch}{0.9}
+ \setlength{\parindent}{0pt}
+ \setlength{\parskip}{1.5ex}
+ \renewcommand{\enotesize}{\normalsize}
+ \theendnotes
+ \endgroup
+ }
+}
+\newcommand{\@@type}{unknown}
+\newcommand{\set@@type}[1]{\renewcommand{\@@type}{#1}}
+\DeclareOption{bypass}{\set@@type{bypass}}
+\DeclareOption{dissertation}{\set@@type{dissertation}}
+\DeclareOption{prelim}{\set@@type{prelim}}
+\DeclareOption{thesis}{\set@@type{thesis}}
+\newboolean{@@uglyheadings}
+\setboolean{@@uglyheadings}{false}
+\DeclareOption{uglyheadings}{\setboolean{@@uglyheadings}{true}}
+\newboolean{@@unset}
+\newcommand{\@@optionbibstyle}{}
+\newcommand{\set@@optionbibstyle}[1]{\renewcommand{\@@optionbibstyle}{#1}\addtocounter{@@tcount}{1}}
+\DeclareOption{abbrv}{\set@@optionbibstyle{abbrv}}
+\DeclareOption{abbrvnat}{\set@@optionbibstyle{abbrvnat}}
+\DeclareOption{aer}{\set@@optionbibstyle{aer}}
+\DeclareOption{agsm}{\set@@optionbibstyle{agsm}}
+\DeclareOption{aip}{\set@@optionbibstyle{aip}}
+\DeclareOption{alpha}{\set@@optionbibstyle{alpha}}
+\DeclareOption{ama}{\set@@optionbibstyle{ama}}
+\DeclareOption{apacite}{\set@@optionbibstyle{apacite}}
+\DeclareOption{apalike}{\set@@optionbibstyle{apalike}}
+\DeclareOption{astron}{\set@@optionbibstyle{astron}}
+\DeclareOption{chicago}{\set@@optionbibstyle{chicago}}
+\DeclareOption{ieee}{\set@@optionbibstyle{ieee}}
+\DeclareOption{IEEEtran}{\set@@optionbibstyle{IEEEtran}}
+\DeclareOption{jfm}{\set@@optionbibstyle{jfm}}
+\DeclareOption{jfm2}{\set@@optionbibstyle{jfm2}}
+\DeclareOption{kluwer}{\set@@optionbibstyle{kluwer}}
+\DeclareOption{phaip}{\set@@optionbibstyle{phaip}}
+\DeclareOption{plain}{\set@@optionbibstyle{plain}}
+\DeclareOption{plainnat}{\set@@optionbibstyle{plainnat}}
+\DeclareOption{unsrt}{\set@@optionbibstyle{unsrt}}
+\DeclareOption{unsrtnat}{\set@@optionbibstyle{unsrtnat}}
+\newif\if@openbib
+\@openbibfalse
+%
+\newboolean{@@nonchapterblankpages}
+\setboolean{@@nonchapterblankpages}{false}
+\newcommand{\nonchapterblankpages}{\setboolean{@@nonchapterblankpages}{true}}
+\newcommand{\nononchapterblankpages}{\setboolean{@@nonchapterblankpages}{false}}
+\DeclareOption{nonchapterblankpages}{\nonchapterblankpages}
+\DeclareOption{nononchapterblankpages}{\nononchapterblankpages}
+%
+\newboolean{@@chapterblankpages}
+\setboolean{@@chapterblankpages}{true}
+\newcommand{\chapterblankpages}{\setboolean{@@chapterblankpages}{true}}
+\newcommand{\nochapterblankpages}{\setboolean{@@chapterblankpages}{false}}
+\DeclareOption{chapterblankpages}{\chapterblankpages}
+\DeclareOption{nochapterblankpages}{\nochapterblankpages}
+%
+\newboolean{@@coversheets}
+\setboolean{@@coversheets}{true}
+\newcommand{\coversheets}{\setboolean{@@coversheets}{true}}
+\newcommand{\nocoversheets}{\setboolean{@@coversheets}{false}}
+\DeclareOption{coversheets}{\coversheets}
+\DeclareOption{nocoversheets}{\nocoversheets}
+%
+\DeclareOption{miser}{
+ \nononchapterblankpages
+ \nochapterblankpages
+ \nocoversheets
+}
+\DeclareOption{nomiser}{
+ \nonchapterblankpages
+ \chapterblankpages
+ \coversheets
+}
+
+\newboolean{number@@all@@volumes}
+\setboolean{number@@all@@volumes}{false}
+\DeclareOption{numberallvolumes}{\setboolean{number@@all@@volumes}{true}}
+\DeclareOption{nonumberallvolumes}{\setboolean{number@@all@@volumes}{false}}
+
+%\DeclareOptions*(\PassOptionsToClass{\CurrentOption}{report}}
+ % Used to count number of citation/bibliography styles used.
+\setcounter{@@tcount}{0}
+\ProcessOptions
+\newcommand{\IW}[1]{\immediate\write16{#1}}
+{
+ \catcode`\+=13
+ \def+{\space}
+ \ifthen{\equal{\@@dept}{unknown}}
+ {
+ \IW{}
+ \IW{You must specify an option for your school or department, for example,}
+ \IW{++++\string\documentclass[aae]{puthesis}}
+ \IW{The available school and department options are defined at}
+ \IW{++++http://engineering.purdue.edu/\string~mark/puthesis/\@@number Options}
+ \IW{ABORTING...}
+ \IW{}
+ \stop
+ }
+ \ifthen{\equal{\@@type}{unknown}}
+ {
+ \IW{}
+ \IW{You must specify an option for your type of document, for example,}
+ \IW{++++\string\documentclass[aae,dissertation]{puthesis}}
+ \IW{The available document types are}
+ \IW{++++bypass+++++++++Master's Bypass Report}
+ \IW{++++dissertation+++Ph.D. dissertation}
+ \IW{++++prelim+++++++++Ph.D. Preliminary Report}
+ \IW{++++thesis+++++++++Master's Thesis}
+ \IW{ABORTING...}
+ \IW{}
+ \stop
+ }
+ \ifnum\value{@@tcount}>1
+ \IW{}
+ \IW{You may specify only one citation/bibliography style from the below list:}
+ \IW{++++abbrv}
+ \IW{++++abbrvnat}
+ \IW{++++agsm}
+ \IW{++++aip}
+ \IW{++++alpha}
+ \IW{++++ama}
+ \IW{++++apacite}
+ \IW{++++apalike}
+ \IW{++++astron}
+ \IW{++++chicago}
+ \IW{++++ieee}
+ \IW{++++IEEEtran}
+ \IW{++++jfm}
+ \IW{++++jfm2}
+ \IW{++++kluwer}
+ \IW{++++phaip}
+ \IW{++++plain}
+ \IW{++++plainnat}
+ \IW{++++unsrt}
+ \IW{++++unsrtnat}
+ \IW{See}
+ \IW{++++http://engineering.purdue.edu/\string~mark/puthesis/\@@number Options}
+ \IW{for more details.}
+ \IW{ABORTING...}
+ \IW{}
+ \stop
+ \fi
+}
+\LoadClass[12pt,twoside]{report}[2004/02/16]
+
+% Got these lines using "grep dotted report.cls" and changed
+% as needed for new \@dottedtocline \vskip parameter.
+\renewcommand*\l@section{\@dottedtocline{1}{\smalltocskip}{1.5em}{2.3em}}
+\renewcommand*\l@subsection{\@dottedtocline{2}{\smalltocskip}{3.8em}{3.2em}}
+\renewcommand*\l@subsubsection{\@dottedtocline{3}{\smalltocskip}{7.0em}{4.1em}}
+\renewcommand*\l@paragraph{\@dottedtocline{4}{\smalltocskip}{10em}{5em}}
+\renewcommand*\l@subparagraph{\@dottedtocline{5}{\smalltocskip}{12em}{6em}}
+\renewcommand*\l@figure{\@dottedtocline{1}{\bigtocskip}{0em}{2.3em}}
+\renewcommand*\l@table{\@dottedtocline{1}{\bigtocskip}{0em}{2.3em}}
+
+%{\obeyspaces
+% \gdef\@@Identification{{
+% \IW{}
+% \IW{ ||}
+% \IW{ see || below}
+% \IW{ ||}
+% \IW{ \string\\ || //}
+% \IW{ \string\\||//}
+% \IW{ \string\\//}
+% \IW{+----------------------------------------------------------------------------+}
+% \IW{| This document is using the 2012/11/30 version of puthesis.cls. It was |}
+% \IW{| probably read in because you used \string\documentclass[...]{puthesis}. |}
+% \IW{| |}
+% \IW{| Only the latest version of this file available from |}
+% \IW{| http://engineering.purdue.edu/\string~mark/puthesis |}
+% \IW{| is supported. Put it in the same subdirectory as your thesis. |}
+% \IW{| |}
+% \IW{| This file may be being read from your thesis directory or somewhere else. |}
+% \IW{| To find out which is being used search for ``puthesis.cls'' near the |}
+% \IW{| beginning of this log. If it has ``(puthesis.cls'' or ``(./puthesis.cls'' |}
+% \IW{| it's from your thesis directory, otherwise it is being read form somewhere |}
+% \IW{| else. |}
+% \IW{+----------------------------------------------------------------------------+}
+% \IW{ //\string\\}
+% \IW{ //||\string\\}
+% \IW{ // || \string\\}
+% \IW{ ||}
+% \IW{ see || above}
+% \IW{ ||}
+% \IW{}
+%}}
+%}
+
+\def\AppendixFigure{\relax}
+\def\AppendixTable{\relax}
+\def\@@bibname{REFERENCES}
+\def\@@deptbibstyle{unsrt}
+\def\@@coversheetspace{\vfill}
+\def\@@evenfoot{}
+\def\@@evenhead{\hfil\textrm{\thepage}}
+\newboolean{@@figurecaptions}
+ \setboolean{@@figurecaptions}{false}
+\def\@@oddfoot{}
+\def\@@oddhead{\hfil\textrm{\thepage}}
+\def\@@sectionbaselinestretch{\relax}
+\def\@@sectionseries{\bfseries}
+\def\@@startthebibliography{\coversheet{\@@bibname}}
+\def\@@startvita{\coversheet{\@@vitaname}}
+\def\@@subsectionbaselinestretch{\relax}
+\def\@@subsectionseries{\bfseries}
+\def\@@subsubsectionbaselinestretch{\relax}
+\def\@@subsubsectionseries{\bfseries}
+\def\@@t{\relax}
+\def\@@thebibliographyparsep{\relax}
+
+\ifthen{\equal{iupuiece}{\@@dept}}
+ {
+ \def\@@makechapterhead#1{
+ \ifthenelse{\boolean{@@inappendix}}
+ {\large\bf APPENDIX \thechapter\\\uppercase{#1}}
+ {\large\bf\thechapter. \uppercase{#1}}
+ }
+ \def\@@makechapterheadspacea{\vspace*{0.6truein}} % by trial and error
+ \def\@@makechapterheadspaceb{\vspace{0.5\baselineskip}}
+ \def\@@makeschapterheadspacea{\vspace*{0.6truein}} % by trial and error
+ \def\@@makeschapterheadspaceb{\vspace*{0.5\baselineskip}}
+ \def\@@startvita{\relax}
+ }
+\ifthen{\equal{aae}{\@@dept}}
+ {
+ \def\fnum@table{\tablename~\thetable.~}
+ \setboolean{@@figurecaptions}{true}
+ \def\@@makechapterhead#1{\large\bf\thechapter. #1}
+ \def\@@makechapterheadspacea{\vspace*{0.5truein}}
+ \def\@@makechapterheadspaceb{\vspace{0.5\baselineskip}}
+ \def\@@makeschapterheadspacea{\vspace*{0.5truein}}
+ \def\@@makeschapterheadspaceb{\vspace*{0.5\baselineskip}}
+ }
+\ifthen{\equal{agecon}{\@@dept}}
+ {
+ \def\@@makechapterhead#1{\rm\uppercase{Chapter \thechapter. #1}}
+ \def\@@makechapterheadspacea{\vspace*{0.6truein}}
+ \def\@@makechapterheadspaceb{}
+ \def\@@makeschapterheadspacea{\vspace*{0.6truein}}
+ \def\@@makeschapterheadspaceb{\vspace*{0.5\baselineskip}}
+ }
+%
+% The agry option is being built to be most like the plain Purdue standards.
+% It was originally based on the CS format.
+\ifthen{\equal{agry}{\@@dept}}
+ {
+ \def\@@makechapterhead#1{\thechapter\quad\uppercase{#1}}
+ \def\@@makechapterheadspacea{\vspace*{0.6truein}} % by trial and error
+ \def\@@makechapterheadspaceb{\vspace{0.5\baselineskip}}
+ \def\@@makeschapterheadspacea{\vspace*{0.6truein}} % by trial and error
+ \def\@@makeschapterheadspaceb{\vspace*{0.5\baselineskip}}
+ \def\@@sectionseries{}
+ \def\@@subsectionseries{}
+ \def\@@subsubsectionseries{}
+ }
+\ifthen{\equal{ansc}{\@@dept}}
+ {
+ \def\@@makechapterhead#1{
+ \ifthenelse{\boolean{@@inappendix}}
+ {\large\bf APPENDIX \thechapter\\\uppercase{#1}}
+ {\large\bf\thechapter. \uppercase{#1}}
+ }
+ \def\@@makechapterheadspacea{\vspace*{0.5truein}}
+ \def\@@makechapterheadspaceb{\vspace{0.5\baselineskip}}
+ \def\@@makeschapterheadspacea{\vspace*{0.5truein}}
+ \def\@@makeschapterheadspaceb{\vspace*{0.5\baselineskip}}
+ \def\@@startvita{\relax}
+ }
+\ifthen{\equal{che}{\@@dept}}
+ {
+ \def\@@makechapterhead#1{\thechapter. \uppercase{#1}}
+ \def\@@makechapterheadspacea{\vspace*{0.65625truein}}
+ \def\@@makechapterheadspaceb{\vspace{0.5\baselineskip}}
+ \def\@@makeschapterheadspacea{\vspace*{0.65625truein}}
+ \def\@@makeschapterheadspaceb{\vspace*{0.5\baselineskip}}
+ \def\@@sectionbaselinestretch{\renewcommand{\baselinestretch}{1}}
+ \def\@@sectionseries{}
+ \def\@@subsectionbaselinestretch{\renewcommand{\baselinestretch}{1}}
+ \def\@@subsectionseries{\relax}
+ \def\@@subsubsectionbaselinestretch{\renewcommand{\baselinestretch}{1}}
+ \def\@@subsubsectionseries{\relax}
+ }
+\ifthen{\equal{ce}{\@@dept}}
+ {
+ \def\fnum@table{\tablename~\thetable.~}
+ \setboolean{@@figurecaptions}{true}
+ \def\@@makechapterhead#1{\large\rm\thechapter. \uppercase{#1}}
+ \def\@@makechapterheadspacea{\vspace*{0.5truein}}
+ \def\@@makechapterheadspaceb{\vspace{0.5\baselineskip}}
+ \def\@@makeschapterheadspacea{\vspace*{0.5truein}}
+ \def\@@makeschapterheadspaceb{\vspace*{0.5\baselineskip}}
+ \def\@@subsectionseries{\normalfont}
+ \def\@@subsubsectionseries{\normalfont}
+ }
+\ifthen{\equal{edci}{\@@dept}}
+ {
+ \def\@@deptbibstyle{apalike}
+ \def\@@makechapterhead#1{\large\bf\thechapter. #1}
+ \def\@@makechapterheadspacea{\vspace*{0.5truein}}
+ \def\@@makechapterheadspaceb{\vspace{0.5\baselineskip}}
+ \def\@@makeschapterheadspacea{\vspace*{0.5truein}}
+ \def\@@makeschapterheadspaceb{\vspace*{0.5\baselineskip}}
+ \def\@@startthebibliography{}
+ \def\@@startvita{\relax}
+ }
+\ifthen{\equal{ece}{\@@dept}}
+ {
+ \def\@@deptbibstyle{IEEEtran}
+ \renewcommand*\l@section{\@dottedtocline{1}{\bigtocskip}{1.5em}{2.3em}}
+ \renewcommand*\l@subsection{\@dottedtocline{2}{\bigtocskip}{3.8em}{3.2em}}
+ \renewcommand*\l@subsubsection{\@dottedtocline{3}{\bigtocskip}{7.0em}{4.1em}}
+ \renewcommand*\l@paragraph{\@dottedtocline{4}{\bigtocskip}{10em}{5em}}
+ \renewcommand*\l@subparagraph{\@dottedtocline{5}{\bigtocskip}{12em}{6em}}
+ \renewcommand*\l@figure{\@dottedtocline{1}{\bigtocskip}{1.5em}{2.3em}}
+ \def\@@makechapterhead#1{{\large\bf\thechapter. \uppercase{#1}}}
+ \def\@@makechapterheadspacea{\vspace*{0.6truein}} % by trial and error
+ \def\@@makechapterheadspaceb{\vspace{0.5\baselineskip}}
+ \def\@@makeschapterhead#1{{\large\bf\uppercase{#1}}}
+ \def\@@makeschapterheadspacea{\vspace*{0.6truein}} % by trial and error
+ \def\@@makeschapterheadspaceb{\vspace*{0.5\baselineskip}}
+ \nochapterblankpages
+ }
+\ifthen{\equal{hsci}{\@@dept}}
+ {
+ \def\@@makechapterhead#1{
+ \ifthenelse{\boolean{@@inappendix}}
+ {\large\bf APPENDIX \thechapter\\\uppercase{#1}}
+ {\large\bf\thechapter. \uppercase{#1}}
+ }
+ \def\@@makechapterheadspacea{\vspace*{0.6truein}} % by trial and error
+ \def\@@makechapterheadspaceb{\vspace{0.5\baselineskip}}
+ \def\@@makeschapterheadspacea{\vspace*{0.6truein}} % by trial and error
+ \def\@@makeschapterheadspaceb{\vspace*{0.5\baselineskip}}
+ \nochapterblankpages
+ }
+\ifthen{\equal{ie}{\@@dept}}
+ {
+ \def\@@makechapterhead#1{
+ \ifthenelse{\boolean{@@inappendix}}
+ {\large\bf APPENDIX \thechapter\\\uppercase{#1}}
+ {\large\bf\thechapter. \uppercase{#1}}
+ }
+ \def\@@makechapterheadspacea{\vspace*{0.6truein}} % by trial and error
+ \def\@@makechapterheadspaceb{\vspace{0.5\baselineskip}}
+ \def\@@makeschapterheadspacea{\vspace*{0.6truein}} % by trial and error
+ \def\@@makeschapterheadspaceb{\vspace*{0.5\baselineskip}}
+ \nochapterblankpages
+ }
+\ifthen{\equal{ling}{\@@dept}}
+{
+ \def\@@deptbibstyle{unified}
+ \renewcommand*\l@section{\@dottedtocline{1}{\bigtocskip}{1.5em}{2.3em}}
+ \renewcommand*\l@subsection{\@dottedtocline{2}{\bigtocskip}{3.8em}{3.2em}}
+ \renewcommand*\l@subsubsection{\@dottedtocline{3}{\bigtocskip}{7.0em}{4.1em}}
+ \renewcommand*\l@paragraph{\@dottedtocline{4}{\bigtocskip}{10em}{5em}}
+ \renewcommand*\l@subparagraph{\@dottedtocline{5}{\bigtocskip}{12em}{6em}}
+ \renewcommand*\l@figure{\@dottedtocline{1}{\bigtocskip}{1.5em}{2.3em}}
+ \def\@@makechapterhead#1{{\thechapter. \uppercase{#1}}}
+ \def\@@makechapterheadspacea{\vspace*{0.6truein}} % by trial and error
+ \def\@@makechapterheadspaceb{\vspace{0.5\baselineskip}}
+ \def\@@makeschapterhead#1{{\uppercase{#1}}}
+ \def\@@makeschapterheadspacea{\vspace*{0.6truein}} % by trial and error
+ \def\@@makeschapterheadspaceb{\vspace*{0.5\baselineskip}}
+ \nochapterblankpages
+}
+\ifthen{\equal{mgmt}{\@@dept}}
+ {
+ \def\@@makechapterhead#1{\large\bf\thechapter. #1}
+ \def\@@makechapterheadspacea{\vspace*{0.5truein}}
+ \def\@@makechapterheadspaceb{\vspace{0.5\baselineskip}}
+ \def\@@makeschapterheadspacea{\vspace*{0.5truein}}
+ \def\@@makeschapterheadspaceb{\vspace*{0.5\baselineskip}}
+ }
+\ifthen{\equal{mse}{\@@dept}}
+ {
+ \def\AppendixFigure{\addtocontents{lof}{Appendix Figure\hfil}}
+ \def\AppendixTable{\addtocontents{lot}{Appendix Table\hfil}}
+ \def\@@coversheetspace{\vspace*{0.53125truein}}
+ \def\@@makechapterhead#1{
+ \rm
+ \ifthen{\boolean{@@inappendix}}
+ {APPENDIX }
+ \thechapter. \uppercase{#1}
+ }
+ \def\@@makechapterheadspacea{\vspace*{0.65625truein}}
+ \def\@@makechapterheadspaceb{\vspace{0.5\baselineskip}}
+ \def\@@makeschapterheadspacea{\vspace*{0.65625truein}}
+ \def\@@makeschapterheadspaceb{\vspace*{0.5\baselineskip}}
+ \def\@@sectionbaselinestretch{\renewcommand*{\baselinestretch}{1.0}}
+ \def\@@subsectionbaselinestretch{\renewcommand*{\baselinestretch}{1.0}}
+ \def\@@subsectionseries{\normalfont}
+ \def\@@subsubsectionseries{\normalfont}
+ \def\@@thebibliographyparsep{\parsep 6pt}
+ }
+\ifthen{\equal{me}{\@@dept}}
+ {
+ \ifthenelse{\equal{\@@optionbibstyle}{jfm}}
+ {\def\@@deptbibstyle{jfm}}
+ {\ifthenelse{\equal{\@@optionbibstyle}{jfm2}}
+ {\def\@@deptbibstyle{jfm2}}
+ {\def\@@deptbibstyle{pumeunsrt}}
+ }
+ \def\fnum@table{\tablename~\thetable.~}
+ \setboolean{@@figurecaptions}{true}
+ \def\@@makechapterhead#1{\rm\thechapter. \uppercase{#1}}
+ \def\@@makechapterheadspacea{\vspace*{0.5truein}}
+ \def\@@makechapterheadspaceb{\vspace{0.5\baselineskip}}
+ \def\@@makeschapterheadspacea{\vspace*{0.5truein}}
+ \def\@@makeschapterheadspaceb{\vspace*{0.5\baselineskip}}
+ \nochapterblankpages
+ }
+\ifthen{\equal{ne}{\@@dept}}
+ {
+ \def\@@makechapterhead#1{
+ \ifthenelse{\boolean{@@inappendix}}
+ {\large\bf APPENDIX \thechapter\\\uppercase{#1}}
+ {\large\bf\thechapter. \uppercase{#1}}
+ }
+ \def\@@makechapterheadspacea{\vspace*{0.6truein}} % by trial and error
+ \def\@@makechapterheadspaceb{\vspace{0.5\baselineskip}}
+ \def\@@makeschapterheadspacea{\vspace*{0.6truein}} % by trial and error
+ \def\@@makeschapterheadspaceb{\vspace*{0.5\baselineskip}}
+ \nochapterblankpages
+ }
+\ifthen{\equal{chem}{\@@dept}}
+ {
+ % The vertical spacing between floats appearing either at the
+ % top or at the bottom of a page for double column floats in
+ % two-column page format.
+ \setlength{\dblfloatsep}{36pt}
+ % The vertical spacing between floats and text, for both top
+ % and bottom floats for double column floats in two-column page
+ % format.
+ \setlength{\dbltextfloatsep}{36pt}
+ % The vertical spacing between floats appearing either at the
+ % top or at the bottom of a page.
+ \setlength{\floatsep}{36pt}
+ % The vertical spacing above and below a float that appears in
+ % the middle of a text page with the h placement argument.
+ \setlength{\intextsep}{36pt}
+ \def\@@makechapterhead#1{\thechapter. #1}
+ \def\@@makechapterheadspacea{\vspace*{0.7truein}}
+ \def\@@makechapterheadspaceb{\vspace*{36pt}}
+ \def\@@makeschapterheadspacea{\vspace*{0.7truein}}
+ \def\@@makeschapterheadspaceb{\vspace*{36pt}}
+ % The vertical spacing between floats and text, for both top
+ % and bottom floats.
+ \setlength{\textfloatsep}{36pt}
+ }
+\ifthen{\equal{cs}{\@@dept}}
+ {
+ \def\@@makechapterhead#1{\thechapter\quad\uppercase{#1}}
+ \def\@@makechapterheadspacea{\vspace*{0.6truein}} % by trial and error
+ \def\@@makechapterheadspaceb{\vspace{0.5\baselineskip}}
+ \def\@@makeschapterheadspacea{\vspace*{0.6truein}} % by trial and error
+ \def\@@makeschapterheadspaceb{\vspace*{0.5\baselineskip}}
+ \def\@@sectionseries{}
+ \def\@@subsectionseries{}
+ \def\@@subsubsectionseries{}
+ }
+\ifthen{\equal{eas}{\@@dept}}
+ {
+ \def\@@makechapterhead#1{\thechapter\quad\uppercase{#1}}
+ \def\@@makechapterheadspacea{\vspace*{0.5truein}}
+ \def\@@makechapterheadspaceb{\vspace{0.5\baselineskip}}
+ \def\@@makeschapterheadspacea{\vspace*{0.5truein}}
+ \def\@@makeschapterheadspaceb{\vspace*{0.5\baselineskip}}
+ }
+\ifthen{\equal{math}{\@@dept}}
+ {
+ \def\@@makechapterhead#1{\large\bf\thechapter. #1}
+ \def\@@makechapterheadspacea{\vspace*{0.5truein}}
+ \def\@@makechapterheadspaceb{\vspace{0.5\baselineskip}}
+ \def\@@makeschapterheadspacea{\vspace*{0.5truein}}
+ \def\@@makeschapterheadspaceb{\vspace*{0.5\baselineskip}}
+ }
+\ifthen{\equal{phys}{\@@dept}}
+ {
+ \def\@@makechapterhead#1{\large\bf\thechapter. #1}
+ \def\@@makechapterheadspacea{\vspace*{0.5truein}}
+ \def\@@makechapterheadspaceb{\vspace{0.5\baselineskip}}
+ \def\@@makeschapterheadspacea{\vspace*{0.5truein}}
+ \def\@@makeschapterheadspaceb{\vspace*{0.5\baselineskip}}
+ }
+\ifthen{\equal{stat}{\@@dept}}
+ {
+ \def\@@makechapterhead#1{\large\rm\thechapter. \uppercase{#1}}
+ \def\@@makechapterheadspacea{\vspace*{0.5truein}}
+ \def\@@makechapterheadspaceb{\vspace{0.5\baselineskip}}
+ \def\@@makeschapterheadspacea{\vspace*{0.5truein}}
+ \def\@@makeschapterheadspaceb{\vspace*{0.5\baselineskip}}
+ }
+\AtBeginDocument{
+ \usepackage{notoccite}
+}
+\AtEndDocument{
+%% \@@Identification
+%% \write\@auxout{\string\newcounter{\string\csname@@@appendix\endcsname}}
+%% \write\@auxout{\string\setcounter{\string\csname@@@appendix\endcsname}{\arabic{@@appendix}}}
+}
+\newenvironment{cland}
+ {\begin{landscape}\hbox\bgroup\hss\vbox\bgroup}
+ {\egroup\hss\egroup\end{landscape}}
+
+\newenvironment{lquotation}
+ {\begin{quotation}\renewcommand{\baselinestretch}{1}\reset@font}
+ {\end{quotation}}
+
+\newcommand{\coversheet}[1]{
+ {
+ \ifthen{\boolean{@@coversheets}}
+ {
+ \ifthenelse{\boolean{@@chapterblankpages}}
+ {\clearpage} %% ! {\cleardoublepage}
+ {\clearpage}
+ \pagestyle{empty}
+ \mbox{}
+ \@@coversheetspace
+ \begin{center}
+ #1
+ \end{center}
+ \vfill
+ \newpage
+ \addtocounter{page}{-1}
+ }
+ \@@NotTableOfContents
+ }
+}
+\setlength{\paperheight}{11truein}
+\setlength{\paperwidth}{8.5truein}
+\renewcommand*{\baselinestretch}{1.5}
+\setlength{\evensidemargin}{0.5truein}
+\setlength{\oddsidemargin}{0.5truein}
+\setlength{\textheight}{8.75truein}
+\setlength{\textwidth}{6truein}
+\clubpenalty=10000
+\widowpenalty=10000
+\displaywidowpenalty=10000
+\ifthen{\equal{mse}{\@@dept}}
+ {\setlength{\parindent}{0.40625truein}}
+%\setlength\parskip{0\p@ \@plus \p@}
+\pagenumbering{roman}
+\newcommand{\@@NotTableOfContents}{%
+ \renewcommand*{\@evenhead}{\@@evenhead}
+ \renewcommand*{\@oddhead}{\@@oddhead}
+ \renewcommand*{\@evenfoot}{\@@evenfoot}
+ \renewcommand*{\@oddfoot}{\@@oddfoot}
+}
+\@@NotTableOfContents
+\setlength{\topmargin}{-0.5truein}
+\addtolength{\topmargin}{0.04875truein}
+\settoheight{\headheight}{- ivx1 -}
+\setlength{\headsep}{0.5truein}
+\addtolength{\headsep}{-\headheight}
+\addtolength{\headsep}{-0.0625truein}
+\newcommand*{\@@TitleAuthor}{\relax}
+\newcommand*{\@@AbstractAuthor}{\relax}
+\renewcommand*{\author}[2]{%
+ \renewcommand*{\@@TitleAuthor}{#1}%
+ \renewcommand*{\@@AbstractAuthor}{#2}%
+}
+\newcommand*{\@@Campus}{\relax}
+\newcommand*{\@@Campus@Input}{}
+\newcommand*{\campus}[1]{%
+ \renewcommand\@@Campus@Input{#1}
+ \renewcommand{\@@t}{Fort Wayne} \ifthen{\equal{#1}{\@@t}}{\edef\@@Campus{\@@t}}
+ \renewcommand{\@@t}{Hammond} \ifthen{\equal{#1}{\@@t}}{\edef\@@Campus{\@@t}}
+ \renewcommand{\@@t}{Indianapolis} \ifthen{\equal{#1}{\@@t}}{\edef\@@Campus{\@@t}}
+ \renewcommand{\@@t}{West Lafayette} \ifthen{\equal{#1}{\@@t}}{\edef\@@Campus{\@@t}}
+ \renewcommand{\@@t}{Westville} \ifthen{\equal{#1}{\@@t}}{\edef\@@Campus{\@@t}}
+}
+
+
+\newcommand*{\@@TitleDegree}{\relax}
+\newcommand*{\@@AbstractDegree}{\relax}
+\newcommand*{\@@DegreeMonth}{\relax}
+\newcommand*{\@@DegreeYear}{\relax}
+\newcommand*{\pudegree}[4]{%
+ \renewcommand*{\@@TitleDegree}{#1}%
+ \renewcommand*{\@@AbstractDegree}{#2}%
+ \renewcommand*{\@@DegreeMonth}{#3}%
+ \renewcommand*{\@@DegreeYear}{#4}%
+}
+\newcommand*{\@@MajorProf}{\relax}
+\newcommand*{\majorprof}[1]{\renewcommand*{\@@MajorProf}{Major Professor: #1}}
+\newcommand*{\majorprofs}[1]{\renewcommand*{\@@MajorProf}{Major Professors: #1}}
+\newcommand*{\@@Title}{\relax}
+\renewcommand*{\title}[1]{\renewcommand*{\@@Title}{#1}}
+\renewcommand*{\maketitle}
+ {
+ \ifthen{\equal{\@@Campus@Input}{}}
+ {
+ \IW{You must specify which campus your degree is from, for example,}
+ \IW{\space\space\space\space\string\campus{West Lafayette}}
+ \IW{See}
+ \IW{\space\space\space\space http://engineering.purdue.edu/\string~mark/puthesis/\@@number campus}
+ \IW{for the valid choices.}
+ \IW{ABORTING...}
+ \IW{}
+ \stop
+ }
+ \ifthen{\equal{\@@Campus}{\relax}}
+ {
+ \IW{The campus specified in your}
+ \IW{\space\space\space\space\string\campus{\@@Campus@Input}}
+ \IW{command is not valid.}
+ \IW{See}
+ \IW{\space\space\space\space http://engineering.purdue.edu/\string~mark/puthesis/\@@number campus}
+ \IW{for the valid choices.}
+ \IW{ABORTING...}
+ \IW{}
+ \stop
+ }
+ {
+ \renewcommand*{\baselinestretch}{2} \reset@font
+ \setcounter{save@@page}{\value{page}}
+ \begin{titlepage}
+ \mbox{}
+ \vfil
+ \vfil
+ \begin{center}
+ \uppercase\expandafter{\@@Title}
+ \end{center}
+ \ifthen{\(\boolean{number@@all@@volumes} \and \value{last@@volume}>1\) \or \value{@@volume}>1}
+ {
+ \begin{center}
+ VOLUME \the@@volume
+ \end{center}
+ }
+ \vfil
+ \begin{center}
+ \ifthen{\equal{bypass}{\@@type}}{A Master's Bypass Report\\}
+ \ifthen{\equal{dissertation}{\@@type}}{A Dissertation\\}
+ \ifthen{\equal{prelim}{\@@type}}{A Preliminary Report\\}
+ \ifthen{\equal{thesis}{\@@type}}{A Thesis\\}
+ Submitted to the Faculty\\
+ of\\
+ Purdue University\\
+ by\\
+ \@@TitleAuthor
+ \end{center}
+ \vfil
+ \begin{center}
+ In Partial Fulfillment of the\\
+ Requirements for the Degree\\
+ of\\
+ \@@TitleDegree
+ \end{center}
+ \ifthenelse{\equal{mse}{\@@dept}}
+ {}
+ {\vfil}
+ \begin{center}
+ \@@DegreeMonth\ \@@DegreeYear\\
+ Purdue University\\
+ \@@Campus, Indiana
+ \end{center}
+ \ifthenelse{\equal{mse}{\@@dept}}
+ {}
+ {\vfil\vfil}
+ \end{titlepage}
+ \setcounter{page}{\value{save@@page}}
+ \ifthen{\value{@@volume}=0 \or \value{@@volume}=1}
+ {\setcounter{page}{2}}
+ }
+ }
+\newenvironment{dedication}%
+ {%
+ \newpage
+ \mbox{}
+ \vfil
+ \begin{center}%
+ }%
+ {%
+ \end{center}%
+ \vfil
+ \eject
+ \@@NotTableOfContents
+ }
+\newboolean{@@inother}
+\setboolean{@@inother}{false}
+ % #1 "next" or "odd": start on next or next odd page?
+ % #2 what to print at top of page
+ % #3 "y" or "n": put in table of contents?
+ % #4 amount of extra space to put after heading at top of page
+\newcommand{\@@nonchapter}[4]{{%
+ \@@NotTableOfContents
+ \bgroup
+ \setboolean{@@inother}{true}
+ \renewcommand{\large}{}%
+ \renewcommand{\bf}{}%
+ \ifthenelse{\equal{ce}{\@@dept}}
+ {\chapter*{\uppercase{#2}}}
+ {\chapter*{#2}}
+ \ifthen{\equal{y}{#3}}
+ {
+ \ifthenelse{\equal{ce}{\@@dept}}
+ {\addcontentsline{toc}{chapter}{\uppercase{#2}}}
+ {\addcontentsline{toc}{chapter}{#2}}
+ }
+ \egroup
+ \vspace{#4}
+}}
+\newenvironment{acknowledgments}%
+ {\@@nonchapter{next}{ACKNOWLEDGMENTS}{n}{0pt}}%
+ {}
+\newenvironment{preface}%
+ {\@@nonchapter{next}{PREFACE}{n}{0pt}}%
+ {}
+\renewcommand*{\tableofcontents}{
+ \@@nonchapter{odd}{TABLE OF CONTENTS}{n}{0pt}
+ {\leftskip=0pt \noindent\hbox to\textwidth{\hfil Page}\par}
+ {%
+% \renewcommand*{\@oddhead}{%
+% \hfil\textrm{\hfil oddhead a \thepage}%
+% \renewcommand*{\@oddhead}{\hfil oddhead b \thepage}%
+% }
+% \renewcommand*{\@evenhead}{%
+% \hfil\textrm{\hfil evenhead a \thepage}%
+% \renewcommand*{\@evenhead}{\hfil evenhead b \thepage}%
+% }
+ \output={
+ \let \par \@@par
+ \ifnum \outputpenalty<-\@M
+ \@specialoutput
+ \else
+ \@makecol
+ \@opcol
+ \@startcolumn
+ \@whilesw \if@fcolmade \fi
+ {%
+ \@opcol\@startcolumn}%
+ \fi
+ \ifnum \outputpenalty>-\@Miv
+ \ifdim \@colroom<1.5\baselineskip
+ \ifdim \@colroom<\textheight
+ \@latex@warning@no@line{Text page \thepage\space
+ contains only floats}%
+ \@emptycol
+ \else
+ \global \vsize \@colroom
+ \fi
+ \else
+ \global \vsize \@colroom
+ \fi
+ \else
+ \global \vsize \maxdimen
+ \fi
+ {\leftskip=0pt \noindent\hbox to\textwidth{\hfil Page}\par}
+ }
+ \renewcommand{\baselinestretch}{1}\reset@font
+ \@starttoc{toc}
+ }
+}
+\ifthen{\equal{che}{\@@dept}\or \equal{ece}{\@@dept}\or \equal{hsci}{\@@dept}\or \equal{me}{\@@dept}}
+ {%
+ \def\numberline#1{\hb@xt@\@tempdima{#1\hfil}}
+ \renewcommand*\l@figure{\@dottedtocline{1}{\bigtocskip}{0em}{2.3em}}
+ \renewcommand*\l@table{\@dottedtocline{1}{\bigtocskip}{0em}{2.3em}}
+ }
+\def\bigtocskip{0.5\baselineskip plus.2\p@}
+\def\smalltocskip{0pt}
+\def\@dottedtocline#1#2#3#4#5#6{%
+ \ifnum #1>\c@tocdepth
+ \else
+ \vskip #2
+ {%
+ \leftskip #3
+ \rightskip \@tocrmarg
+ \parfillskip -\rightskip
+ \parindent #3
+ \@afterindenttrue
+ \interlinepenalty\@M
+ \leavevmode
+ \@tempdima #4
+ \advance\@tempdima \@@padding
+ \advance\leftskip \@tempdima
+ \hbox{}\hskip -\leftskip
+ #5\nobreak
+ \leaders\hbox{$\m@th \mkern \@dotsep mu.\mkern \@dotsep mu$}\hfill
+ \nobreak
+ \renewcommand{\@pnumwidth}{1.55em}
+ \setlength{\@@tlength}{\@pnumwidth}
+ \settowidth{\@@ulength}{\reset@font \rm #6}
+ \ifdim \@@ulength>\@@tlength
+ \hbox to\@@ulength{\hfil\reset@font \rm #6}\par
+ \else
+ \hbox to\@@tlength{\hfil\reset@font \rm #6}\par
+ \fi
+ }%
+ \fi
+}
+\renewcommand*{\listoftables}{
+ \@@nonchapter{next}{LIST OF TABLES}{y}{0pt}
+ {\leftskip=0pt \noindent\hbox to\textwidth{Table\hfil Page}\par}
+ {%
+ \output={
+ \let \par \@@par
+ \ifnum \outputpenalty<-\@M
+ \@specialoutput
+ \else
+ \@makecol
+ \@opcol
+ \@startcolumn
+ \@whilesw \if@fcolmade \fi
+ {%
+ \@opcol\@startcolumn}%
+ \fi
+ \ifnum \outputpenalty>-\@Miv
+ \ifdim \@colroom<1.5\baselineskip
+ \ifdim \@colroom<\textheight
+ \@latex@warning@no@line{Text page \thepage\space
+ contains only floats}%
+ \@emptycol
+ \else
+ \global \vsize \@colroom
+ \fi
+ \else
+ \global \vsize \@colroom
+ \fi
+ \else
+ \global \vsize \maxdimen
+ \fi
+ {\leftskip=0pt \noindent\hbox to\textwidth{Table\hfil Page}\par}
+ }
+ \renewcommand{\baselinestretch}{1}\reset@font
+ \@starttoc{lot}
+ }
+}
+\renewcommand{\listoffigures}{
+ \@@nonchapter{next}{LIST OF FIGURES}{y}{0pt}
+ {\leftskip=0pt \noindent\hbox to\textwidth{Figure\hfil Page}\par}
+ {%
+ \output={
+ \let \par \@@par
+ \ifnum \outputpenalty<-\@M
+ \@specialoutput
+ \else
+ \@makecol
+ \@opcol
+ \@startcolumn
+ \@whilesw \if@fcolmade \fi
+ {%
+ \@opcol\@startcolumn}%
+ \fi
+ \ifnum \outputpenalty>-\@Miv
+ \ifdim \@colroom<1.5\baselineskip
+ \ifdim \@colroom<\textheight
+ \@latex@warning@no@line{Text page \thepage\space
+ contains only floats}%
+ \@emptycol
+ \else
+ \global \vsize \@colroom
+ \fi
+ \else
+ \global \vsize \@colroom
+ \fi
+ \else
+ \global \vsize \maxdimen
+ \fi
+ {\leftskip=0pt \noindent\hbox to\textwidth{Figure\hfil Page}\par}
+ }
+ \renewcommand{\baselinestretch}{1}\reset@font
+ \@starttoc{lof}
+ }
+}
+\newcommand{\@@startlist}[1]{
+ \@@nonchapter{odd}{#1}{y}{0pt}%
+ \setlength{\LTleft}{\parindent}%
+ \setlength{\LTright}{0truein}%
+ %% ! change to \setlength[\@@tlength}{\textwidth - \LTleft - \LTright - 2*\tabcolsep - 1truein} later
+ \setlength{\@@tlength}{\textwidth}
+ \addtolength{\@@tlength}{-\LTleft}
+ \addtolength{\@@tlength}{-\LTright}
+ \addtolength{\@@tlength}{-\tabcolsep}
+ \addtolength{\@@tlength}{-\tabcolsep}
+ \addtolength{\@@tlength}{-1truein}
+ %% ! Get real width of first column.
+ %% ! \setcounter{\@@tcount}{\c@LT@tables}
+ %% ! \addtocounter(\@@tcount){1}
+ %% ! \def\stuff{\csname LT@\romannumeral\@@tcount\endcsname}
+ %% ! \showthe\stuff
+ %% ! \addtolength{\@@tlength}{-?}
+}
+\newenvironment{symbols}%
+ {%
+ \@@startlist{SYMBOLS}
+ \begin{longtable}{lp{\@@tlength}}%
+ }
+ {\end{longtable}}
+\newenvironment{abbreviations}
+ {
+ \@@startlist{ABBREVIATIONS}
+ \begin{longtable}{lp{\@@tlength}}%
+ }
+ {\end{longtable}}
+\newenvironment{nomenclature}
+ {
+ \@@startlist{NOMENCLATURE}
+ \begin{longtable}{lp{\@@tlength}}%
+ }
+ {\end{longtable}}
+\renewenvironment{glossary}
+ {
+ \@@startlist{GLOSSARY}
+ \begin{longtable}{lp{\@@tlength}}%
+ }
+ {\end{longtable}}
+\renewenvironment{abstract}%
+ {%
+ \@@nonchapter{next}{ABSTRACT}{y}{0pt}
+ {%
+ \renewcommand*{\baselinestretch}{1.0}\reset@font
+ \vspace*{\baselineskip}
+ \vbox{
+ \noindent
+ \@@AbstractAuthor~\@@AbstractDegree,
+ Purdue University,
+ \@@DegreeMonth\ %
+ \@@DegreeYear.
+ {%
+ \renewcommand*{\\}{}%
+ \@@Title.
+ }
+ \ifthenelse{\isundefined{\@@OverrideMajorProf}}
+ {\@@MajorProf\ifthen{\equal{me}{\@@dept}}{, School of Mechanical Engineering}.}
+ {\@@OverrideMajorProf}
+ }
+ \vspace*{\baselineskip}
+ \par
+ }
+ }%
+ {\par}
+\renewcommand{\chapter}{%
+ \ifthenelse{
+ \( \boolean{@@inother} \and \boolean{@@nonchapterblankpages} \)
+ \or \( \not \boolean{@@inother} \and \boolean{@@chapterblankpages} \)
+ }
+ {\cleardoublepage}
+ {\clearpage}
+ \@@NotTableOfContents
+ \global\@topnum\z@
+ \@afterindentfalse
+ \secdef \@chapter \@schapter
+}
+\newboolean{@@inchapters}
+\setboolean{@@inchapters}{false}
+\def\@chapter[#1]#2{%
+ \ifnum \c@secnumdepth >\m@ne
+ \refstepcounter{chapter}%
+ \typeout{\@chapapp\space\thechapter.}%
+ \setboolean{@@unset}{true}
+ \ifthen{\equal{agecon}{\@@dept}}
+ {\setboolean{@@unset}{false}\addcontentsline{toc}{chapter}{\protect\uppercase{\@chapapp\ \thechapter. #1}}}
+ \ifthen{\equal{ce}{\@@dept}}
+ {\setboolean{@@unset}{false}\addcontentsline{toc}{chapter}{\protect\numberline{\thechapter}\uppercase{#1}}}
+ \ifthen{\equal{me}{\@@dept}}
+ {\setboolean{@@unset}{false}\addcontentsline{toc}{chapter}{\protect\numberline{\thechapter.}\uppercase{#1}}}
+ \ifthen{\equal{mse}{\@@dept}}
+ {\setboolean{@@unset}{false}\addcontentsline{toc}{chapter}{\protect\uppercase{\@chapapp\ \thechapter. #1}}}
+ \ifthen{\boolean{@@unset}}
+ {\addcontentsline{toc}{chapter}{\protect\numberline{\thechapter}#1}}
+ \else
+ \ifthenelse{\equal{ce}{\@@dept}}
+ {\addcontentsline{toc}{chapter}{\uppercase{#1}}}
+ {\addcontentsline{toc}{chapter}{#1}}
+ \fi
+ \chaptermark{#1}%
+% \addtocontents{lof}{\protect\addvspace{10\p@}}%
+% \addtocontents{lot}{\protect\addvspace{10\p@}}%
+ \ifthenelse{\equal{ce}{\@@dept}}
+ {\@makechapterhead{\uppercase{#2}}}
+ {\@makechapterhead{#2}}
+ \@afterheading
+ \ifthen{\not \boolean{@@inchapters}}
+ {
+ \pagenumbering{arabic}%
+ \@@inchapterstrue
+ }
+}
+\def\@makechapterhead#1{%
+ \@@makechapterheadspacea
+ {\centering
+ \@@makechapterhead{#1}
+ \endgraf
+ }
+ \@@makechapterheadspaceb
+}
+\def\@makeschapterhead#1{%
+ \@@makeschapterheadspacea
+ {\centering
+ \ifthenelse{\equal{ce}{\@@dept}}
+ {\large\bf\uppercase{#1}}
+ {\large\bf #1}
+ \endgraf
+ }
+ \@@makeschapterheadspaceb
+}
+\def\@startsection#1#2#3#4#5#6{%
+ \if@noskipsec
+ \leavevmode
+ \fi
+ \par
+ \@tempskipa #4\relax
+ \@afterindenttrue
+ \ifdim \@tempskipa <\z@
+ \@tempskipa -\@tempskipa
+ \fi
+ \if@nobreak
+ \everypar{}%
+ \else
+ \addpenalty{\@secpenalty}%
+ \addvspace{\@tempskipa}%
+ \fi
+ \@ifstar
+ {\@ssect{#3}{#4}{#5}{#6}}%
+ {\@dblarg{\@sect{#1}{#2}{#3}{#4}{#5}{#6}}}%
+}
+\newenvironment{cabstract}{\begin{quote}\textbf{Abstract}\quad}{\end{quote}}
+\newtheorem{definition}{Definition}[section]
+\newtheorem{observation}{Observation}[section]
+\newsavebox{\proofbox}
+\sbox{\proofbox}{\rule{7pt}{7pt}}
+\newtheorem{Proof}{Proof}
+\renewcommand{\theProof}{}
+\newenvironment{proof}{\begin{Proof}\rm}{\hfill \usebox{\proofbox} \end{Proof}}
+\newtheorem{proposition}{Proposition}[section]
+\newtheorem{theorem}{Theorem}[section]
+\renewcommand{\section}{%
+ \@startsection
+ {section}%
+ {1}%
+ {\z@}%
+ {24pt}%
+ {12pt}%
+ {\@@sectionbaselinestretch\reset@font\@@sectionseries}%
+}
+\ifthen{\boolean{@@uglyheadings}}
+ {
+ \def\undertext#1{$\underline{\smash{\hbox{#1}}}$}
+ \def\section{
+ \vspace{36truept}
+ \secdef{\@@section}{\@@ssection}
+ }
+ \def\@@section[#1]#2{
+ \refstepcounter{section}
+ \sectionmark{#1}
+ \hbox to \textwidth{\hss\undertext{\thesection\ #2}\hss}\nopagebreak
+ \addcontentsline{toc}{section}{\protect\numberline{\csname thesection\endcsname}#1}
+ \nopagebreak
+ }
+ \def\@@ssection#1{
+ \sectionmark{#1}
+ \hbox to \textwidth{\hss\undertext{#1}\hss}\nopagebreak
+ \nopagebreak
+ }
+ }
+\renewcommand{\subsection}{%
+ \@startsection
+ {subsection}%
+ {2}%
+ {\z@}%
+ {24pt}% \@plus -1ex \@minus -.2ex}%
+ {12pt}% 1.5ex \@plus .2ex}%
+ {\@@subsectionbaselinestretch\reset@font\@@subsectionseries}%
+}
+\ifthen{\boolean{@@uglyheadings}}
+ {
+ \renewcommand{\subsection}[1]{
+ \refstepcounter{subsection}
+ \vspace{36truept}
+ \subsectionmark{#1}
+ \hbox to \textwidth{\hss\thesubsection\ #1\hss}\nopagebreak
+ \addcontentsline{toc}{subsection}{\protect\numberline{\csname thesubsection\endcsname}#1}
+ \nopagebreak
+ }
+ }
+\renewcommand{\subsubsection}{
+ \@startsection
+ {subsubsection}%
+ {3}%
+ {\z@}%
+ {24pt}% \@plus -1ex \@minus -.2ex}%
+ {12pt}% 1.5ex \@plus .2ex}%
+ {\@@subsubsectionbaselinestretch\reset@font\@@subsubsectionseries}%
+}
+\ifthen{\equal{mse}{\@@dept}}
+ {
+ \def\undertext#1{$\underline{\hbox{#1}}$}
+ \newbox{\@@Strut}
+ \renewcommand{\subsubsection}[1]{
+ \refstepcounter{subsubsection}
+ \vspace{6truept}
+ \subsubsectionmark{#1}
+ \noindent \undertext{\vphantom{gjpqy}\thesubsubsection. #1}\newline\nopagebreak
+ \addcontentsline{toc}{subsubsection}{\protect\numberline{\csname thesubsubsection\endcsname}#1}%
+ \nopagebreak\indent
+ }
+ }
+\ifthen{\boolean{@@uglyheadings}}
+ {
+ \renewcommand{\subsubsection}[1]{
+ \refstepcounter{subsubsection}
+ \vspace{36truept}
+ \subsubsectionmark{#1}
+ \noindent \undertext{\thesubsubsection\ #1}\newline\nopagebreak
+ \addcontentsline{toc}{subsubsection}{\protect\numberline{\csname thesubsubsection\endcsname}#1}
+ \nopagebreak
+ }
+ }
+\ifthen{\boolean{@@uglyheadings}}
+ {
+ \setcounter{tocdepth}{4}
+ \newcommand*{\l@subsubsubsection}{\@dottedtocline{4}{\smalltocskip}{11.2em}{5.0em}}
+ \def\subsubsubsectionmark#1{}
+ \newcounter{subsubsubsection}[subsubsection]
+ \renewcommand{\thesubsubsubsection}{\thesubsubsection.\arabic{subsubsubsection}}
+ \newcommand{\subsubsubsection}[1]{
+ \refstepcounter{subsubsubsection}
+ \vspace{36truept}
+ \subsubsubsectionmark{#1}
+ \noindent \thesubsubsubsection. #1\newline\nopagebreak
+ \addcontentsline{toc}{subsubsubsection}{\protect\numberline{\csname thesubsubsubsection\endcsname}#1}
+ \nopagebreak
+ }
+ }
+\def\@sect#1#2#3#4#5#6[#7]#8{%
+ \ifnum #2>\c@secnumdepth
+ \let\@svsec\@empty
+ \else
+ \refstepcounter{#1}%
+ \edef\@svsec{\csname the#1\endcsname\hskip 1em}%
+ \fi
+ \ifthen{\equal{che}{\@@dept}}
+ {%
+ \ifnum #2=3
+ \refstepcounter{#1}%
+ \edef\@svsec{\csname the#1\endcsname\hskip 1em}%
+ \fi
+ }%
+ \@tempskipa #5\relax
+ \ifdim \@tempskipa>\z@
+ \begingroup
+ #6\relax
+ \@hangfrom{\hskip #3\relax\@svsec}%
+ {\interlinepenalty \@M #8\par}%
+ \endgroup
+ \csname #1mark\endcsname{#7}%
+ \addcontentsline{toc}{#1}{%
+ \ifnum #2>\c@secnumdepth
+ \else
+ \protect\numberline{\csname the#1\endcsname}%
+ \fi
+ #7%
+ }%
+ \else
+ \def\@svsechd{%
+ #6%
+ \hskip #3\relax
+ \@svsec #8\csname #1mark\endcsname
+ {#7}%
+ \addcontentsline{toc}{#1}{%
+ \ifnum #2>\c@secnumdepth
+ \else
+ \protect\numberline{\csname the#1\endcsname}
+ \fi
+ #7%
+ }%
+ }%
+ \fi
+ \@xsect{#5}%
+}
+\renewcommand*{\l@chapter}{\@dottedtocline{0}{\bigtocskip}{0em}{1.4em}}
+\ifthen{\equal{ece}{\@@dept}\or \equal{hsci}{\@@dept}}
+ {\renewcommand{\figurename}{Fig.}}
+\renewcommand{\fnum@figure}{\figurename~\thefigure.}
+\renewcommand{\fnum@table}{\tablename~\thetable.}
+\newboolean{@@centercaption}
+\renewcommand{\caption}{%
+ \setboolean{@@centercaption}{true}
+ \refstepcounter\@captype \@dblarg{\@caption\@captype}%
+}
+\newcommand\bcaption{%
+ \setboolean{@@centercaption}{true}
+ \refstepcounter\@captype \@dblarg{\@caption\@captype}%
+}
+\long\def\@makecaption#1#2{%
+ \vspace*{\abovecaptionskip}
+ \ifthenelse{\boolean{@@centercaption}} % center caption
+ {
+ \setlength{\@@captionwidth}{\textwidth}
+ \addtolength{\@@captionwidth}{-4\@@parindent}
+ \ifthenelse{\equal{figure}{\@captype} \or \boolean{@@figurecaptions}}
+ {
+ \sbox\@tempboxa{#1 #2}%
+ \renewcommand{\baselinestretch}{1.0}\reset@font
+ \ifdim \wd\@tempboxa >\hsize
+ \centerline{\parbox[t]{\@@captionwidth}{#1 #2}}
+ \else
+ \centerline{#1 #2}%
+ \fi
+ }
+ {
+ \renewcommand{\baselinestretch}{1.0}\reset@font
+ \centerline{#1}
+ \sbox\@tempboxa{#2}%
+ \ifdim \wd\@tempboxa >\hsize
+ \centerline{\parbox[t]{\@@captionwidth}{#2}}
+ \else
+ \centerline{#2}%
+ \fi
+ }
+ \vspace*{\belowcaptionskip}
+ }
+ {
+ \setlength{\topsep}{0pt}
+ \setlength{\parskip}{0pt}
+ \setlength{\partopsep}{0pt}
+ \begin{quote}
+ \renewcommand{\\}{}
+ \renewcommand{\baselinestretch}{1.0}\reset@font
+ \ifthenelse{\equal{aae}{\@@dept}}
+ {#1. #2\par}
+ {
+ \ifthenelse{\equal{table}{\@captype}}
+ {\hfil\strut #1\hfil\break #2\par}
+ {#1 #2\par}
+ }
+ \end{quote}
+ }
+ \vskip\belowcaptionskip
+}
+\setlength\belowcaptionskip{5\p@}
+\long\def\@makefntext#1{%
+ \baselineskip=12pt
+ \noindent
+ \@makefnmark #1%
+}
+\ifthenelse{\equal{}{\@@optionbibstyle}}
+ {\newcommand{\@@bibstyle}{\@@deptbibstyle}}
+ {\newcommand{\@@bibstyle}{\@@optionbibstyle}}
+
+\ifthen{\equal{abbrv}{\@@bibstyle}}
+ {
+ \renewcommand{\bibname}{{\normalsize\rm REFERENCES}}
+ \bibliographystyle{abbrv}
+ }
+
+\ifthen{\equal{abbrvnat}{\@@bibstyle}}
+ {
+ \usepackage[sort]{natbib}
+ \renewcommand{\bibname}{{\normalsize\rm REFERENCES}}
+ \bibliographystyle{abbrvnat}
+ }
+
+\ifthen{\equal{aer}{\@@bibstyle}}
+ {
+ \usepackage{harvard}
+ \bibliographystyle{aer}
+ }
+
+
+\ifthen{\equal{agsm}{\@@bibstyle}}
+ {
+ \usepackage{cite}
+ \usepackage{harvard}
+ \bibliographystyle{agsm}
+ }
+
+\ifthen{\equal{aip}{\@@bibstyle}}
+ {
+ \usepackage{cite}
+ \usepackage{revtex}
+ \renewcommand{\bibname}{{\normalsize\rm REFERENCES}}
+ \bibliographystyle{aip}
+ }
+
+\ifthen{\equal{alpha}{\@@bibstyle}}
+ {
+ \usepackage{cite}
+ \bibliographystyle{alpha}
+ }
+
+\ifthen{\equal{ama}{\@@bibstyle}}
+{
+ \usepackage{cite}
+ \bibliographystyle{ama}
+}
+
+\ifthen{\equal{apacite}{\@@bibstyle}}
+ {
+ \usepackage{apacite}
+ \bibliographystyle{apacite}
+ }
+
+\ifthen{\equal{apalike}{\@@bibstyle}}
+ {
+ \usepackage{apalike}
+ \bibliographystyle{apalike}
+ }
+
+\ifthen{\equal{astron}{\@@bibstyle}}
+ {
+ \usepackage{astron}
+ \bibliographystyle{astron}
+ }
+
+\ifthen{\equal{chicago}{\@@bibstyle}}
+ {
+ \usepackage{natbib}
+ \bibliographystyle{chicago}
+ }
+
+\ifthen{\equal{ieee}{\@@bibstyle}}
+ {
+ \usepackage{cite}
+ \bibliographystyle{ieee}
+ }
+
+\ifthen{\equal{IEEEtran}{\@@bibstyle}}
+ {
+ \usepackage{cite}
+ \bibliographystyle{IEEEtran}
+ }
+
+\ifthen{\equal{jfm}{\@@bibstyle}}
+ {
+ \usepackage{natbib}
+ \bibliographystyle{jfm}
+ }
+
+\ifthen{\equal{jfm2}{\@@bibstyle}}
+ {
+ \usepackage{natbib}
+ \bibliographystyle{jfm2}
+ }
+
+\ifthen{\equal{kluwer}{\@@bibstyle}}
+ {
+ \usepackage{harvard}
+ \bibliographystyle{kluwer}
+ }
+
+\ifthen{\equal{phaip}{\@@bibstyle}}
+ {
+ \bibliographystyle{phaip}
+ }
+
+\ifthen{\equal{pumeunsrt}{\@@bibstyle}}
+ {
+ \usepackage{cite}
+ \bibliographystyle{pumeunsrt}
+ }
+
+\ifthen{\equal{plain}{\@@bibstyle}}
+ {
+ \usepackage{cite}
+ \bibliographystyle{plain}
+ }
+
+\ifthen{\equal{plainnat}{\@@bibstyle}}
+ {
+ \usepackage{natbib}
+ \renewcommand{\bibname}{{\normalsize\rm REFERENCES}}
+ \bibliographystyle{plainnat}
+ }
+
+\ifthen{\equal{unified}{\@@bibstyle}}
+{
+ \usepackage{natbib}
+ \bibliographystyle{unified}
+ \def\@@bibname{{\normalsize\rm REFERENCES}}
+ \nochapterblankpages
+}
+
+\ifthen{\equal{unsrt}{\@@bibstyle}}
+ {
+ \usepackage{cite}
+ \bibliographystyle{unsrt}
+ }
+
+\ifthen{\equal{unsrtnat}{\@@bibstyle}}
+ {
+ \usepackage{natbib}
+ \renewcommand{\bibname}{{\normalsize\rm REFERENCES}}
+ \bibliographystyle{unsrtnat}
+ }
+
+\renewenvironment{thebibliography}[1]
+ {
+ \@@startthebibliography
+ \@@nonchapter{odd}{\@@bibname}{y}{24pt}
+ \list
+ {\@biblabel{\arabic{enumiv}}}%
+ {%
+ \settowidth\labelwidth{\@biblabel{#1}}%
+ \leftmargin\labelwidth
+ \advance\leftmargin\labelsep
+ \if@openbib
+ \advance\leftmargin\bibindent
+ \itemindent -\bibindent
+ \listparindent \itemindent
+ \parsep \z@
+ \fi
+ \@@thebibliographyparsep
+ \usecounter{enumiv}%
+ \let\p@enumiv\@empty
+ \renewcommand{\theenumiv}{\arabic{enumiv}}%
+ \baselineskip=12pt
+ }%
+ \if@openbib
+ \renewcommand{\newblock}{\par}
+ \else
+ \renewcommand{\newblock}{\hskip .11em \@plus.33em \@minus.07em}%
+ \fi
+ \sloppy\clubpenalty4000\widowpenalty4000%
+ \sfcode`\.=\@m
+ }
+ {}
+\newboolean{@@inappendix}
+\newcommand{\@@appendixname}{Appendix}
+\renewcommand{\appendix}{\par
+ \setboolean{@@inappendix}{true}
+ \setcounter{chapter}{0}%
+ \setcounter{section}{0}%
+ \coversheet{\uppercase\expandafter{\@@appendixname}}
+ \renewcommand{\@chapapp}{\appendixname}%
+ \renewcommand{\thechapter}{\Alph{chapter}}
+}
+\newcommand{\@@appendicesname}{Appendices}
+\newcommand{\appendices}{\par
+ \setboolean{@@inappendix}{true}
+ \setcounter{chapter}{0}%
+ \setcounter{section}{0}%
+ \coversheet{\uppercase\expandafter{\@@appendicesname}}
+ \renewcommand{\@chapapp}{\appendixname}%
+ \renewcommand{\thechapter}{\Alph{chapter}}
+}
+\newcommand{\@@vitaname}{VITA}
+\newenvironment{vita}
+ {
+ \@@startvita
+ \@@nonchapter{odd}{\@@vitaname}{y}{0pt}%
+ }
+ {}
+\def\@starttoc#1{%
+ \begingroup
+ \@input{\jobname.#1}%
+ \if@filesw \expandafter\newwrite\csname tf@#1\endcsname
+ \immediate\openout \csname tf@#1\endcsname \jobname.#1\relax
+ \fi
+ \global\@nobreakfalse
+ \endgroup
+}
+\newcommand{\num}[1]{\ensuremath{#1}}
+\newcommand{\ten}[1]{\ensuremath{{}\cdot 10^{#1}}}
+
+\setcounter{topnumber}{10}
+\renewcommand{\topfraction}{1}
+\setcounter{bottomnumber}{10}
+\renewcommand{\bottomfraction}{1}
+\setcounter{totalnumber}{10}
+\renewcommand{\textfraction}{0}
+\renewcommand{\floatpagefraction}{0}
+
+\setlength{\floatsep}{18pt plus 3pt minus 3pt}
+\setlength{\textfloatsep}{30pt plus 3pt minus 6pt}
+\setlength{\intextsep}{18pt plus 3pt minus 3pt}
+\setlength{\dblfloatsep}{18pt plus 3pt minus 3pt}
+\setlength{\dbltextfloatsep}{30pt plus 3pt minus 6pt}
+
+\newcommand{\Baselinestretch}[1]{\renewcommand{\baselinestretch}{#1}\reset@font}
+
+\def\volume{
+ \addtocounter{@@volume}{1}
+ \write\@auxout{\string\setcounter{last@@volume}{\the@@volume}}
+ \ifthen{\(\boolean{number@@all@@volumes} \and \value{last@@volume}>1\) \or \value{@@volume}>1}
+ {
+ \addtocontents{toc}{\string\vskip\string\baselineskip \noindent VOLUME \the@@volume}
+ }
+ \maketitle
+}
+
+\ifthen{\equal{me}{\@@dept}}
+ {
+ \ifthenelse{\equal{\@@optionbibstyle}{jfm}}
+ {\def\@@deptbibstyle{jfm}}
+ {\ifthenelse{\equal{\@@optionbibstyle}{jfm2}}
+ {\def\@@deptbibstyle{jfm2}}
+ {\def\@@deptbibstyle{pumeunsrt}}
+ }
+ \def\fnum@figure{\figurename~\thefigure.~}
+ \def\fnum@table{\tablename~\thetable.~}
+ \setboolean{@@figurecaptions}{true}
+ \def\@@makechapterhead#1{\rm\thechapter. \uppercase{#1}}
+ \def\@@makechapterheadspacea{\vspace*{0.5truein}}
+ \def\@@makechapterheadspaceb{\vspace{0.5\baselineskip}}
+ \def\@@makeschapterheadspacea{\vspace*{0.5truein}}
+ \def\@@makeschapterheadspaceb{\vspace*{0.5\baselineskip}}
+ \nochapterblankpages
+ }
+
+\makeatother
+
+\raggedbottom
diff --git a/references.bib b/references.bib
new file mode 100644
index 0000000..d1fb6d7
--- /dev/null
+++ b/references.bib
@@ -0,0 +1,2619 @@
+@article{Abra90,
+author = {Abrams, M. D.},
+title = {Adaptations and responses to drought in \textit{{Q}uercus} species of {N}orth {A}merica},
+journal = {Tree Physiology},
+volume = {7},
+year = {1990},
+pages = {227-238},
+}
+
+@article{Abra92,
+author = {Abrams, M. D.},
+title = {Fire and the development of oak forests},
+journal = {BioScience},
+volume = {42},
+year = {1992},
+pages = {346-353},
+}
+
+@article{Abra94,
+author = {Abrams, M. D. and Kubiske, M. E. and Mostoller, S. A.},
+title = {Relating wet and dry year ecophysiology t oleaf structure in contrasting temperate tree species},
+journal = {Ecology},
+volume = {75},
+year = {1994},
+pages = {123-133},
+}
+
+@article{Abra96,
+author = {Abrams, M. D.},
+title = {Distribution, historical development and ecophysiological attributes of oak species in the eastern {U}nited {S}tates},
+journal = {Annales des Sciences Forestieres},
+volume = {53},
+year = {1996},
+pages = {927-939},
+}
+
+@article{Abra03,
+author = {Abrams, M. D.},
+title = {Where has all the white oak gone?},
+journal = {BioScience},
+volume = {53},
+year = {2003},
+pages = {927-939},
+}
+
+@article{Abra05,
+author = {Abrams, M. D.},
+title = {Prescribing fire in eastern oak forests: is time running out?},
+journal = {Northern Journal of Applied Forestry},
+volume = {22},
+year = {2005},
+pages = {190-196},
+}
+
+@article{Abra12,
+author = {Abrams, M. D. and Johnson, S. E.},
+title = {Long-term impacts of deer exclosures on mixed-oak forest composition at the {V}alley {F}orge {N}ational {H}istorical {P}ark, {P}ennsylvania, {USA}},
+journal = {Journal of the Torrey Botanical Society},
+volume = {139},
+year = {2012},
+pages = {167-180},
+}
+
+@article{Adam01,
+author = {Adams, A. S. and Rieske, L. K.},
+title = {Herbivory and fire influence white oak (\textit{{Q}uercus alba} {L}.) seedling vigor},
+journal = {Forest Science},
+volume = {47},
+year = {2001},
+pages = {331-337},
+}
+
+@article{Adam03,
+author = {Adams, A. S. and Rieske, L. K.},
+title = {Prescribed fire affects white oak seedling phytochemistry: implications for insect herbivory},
+journal = {Forest Ecology and Management},
+volume = {176},
+year = {2003},
+pages = {37-47},
+}
+
+@article{Adle10,
+author = {Adler, P. B. and Ellner, S. P. and Levine, J. M.},
+title = {Coexistence of perennial plants: an embarassment of niches},
+journal = {Ecology Letters},
+volume = {13},
+year = {2010},
+pages = {1019-1029},
+}
+
+@article{Aldr05,
+author = {Aldrich, P. R. and Parker, G. R. and Romero-Severson, J. and Michler, C. H.},
+title = {Confirmation of oak recruitment failure in {I}ndiana old-growth forest: 75 years of data},
+journal = {Forest Science},
+volume = {51},
+year = {2005},
+pages = {406-416},
+}
+
+@article{Alle10,
+author = {Allen, C. D. and Macalady, A. K. and Chenchouni, H. and Bachelet, D. and McDowell, N. and Vennetier, M. and Kitzberg, T. and Rigling, A. and Breshears, D. D. and Hogg, E. H. and Gonzalez, P. and Fensham, R. and Zhang, Z. and Castro, J. and Demidova, N. and Lim, J. and Allard, G. and Running, S. W. and Semerci, A. and Cobb, N.},
+title = {A global overview of drought and heat-induced tree mortality reveals emerging climate change risks for forests},
+journal = {Forest Ecology and Management},
+volume = {259},
+year = {2010},
+pages = {660-684},
+}
+
+@article{Alth97,
+author = {Althoff, D. P. and Storm, G. L. and Dewalle, D. R.},
+title = {Daytime habitat selection by cottontails in central {P}ennsylvania},
+journal = {Journal of Wildlife Management},
+volume = {61},
+year = {1997},
+pages = {450-459},
+}
+
+@article{Alve88,
+author = {Alverson, W. S. and Waller, D. M. and Solheim, S. L.},
+title = {Forests too deer: edge effects in northern {W}isconsin},
+journal = {Conservation Biology},
+volume = {2},
+year = {1988},
+pages = {348-358},
+}
+
+@article{Ande92,
+author = {Andersson, C.},
+title = {The effect of weevil and fungal attacks on the germination of \textit{Quercus robur} acorns},
+journal = {Forest Ecology and Management},
+volume = {50},
+year = {1992},
+pages = {247-251},
+}
+
+@article{Andr14,
+author = {Andruk, C. M. and Schwope, C. and Fowler, N. L.},
+title = {The joint effects of fire and herbivory on harwood regeneration in the central {T}exas woodlands},
+journal = {Forest Ecology and Management},
+volume = {334},
+year = {2014},
+pages = {193-200},
+}
+
+@techreport{Arno77,
+author = {Arno, S. F. and Sneck, K. M.},
+title = {A method for determining fire history in coniferous forests in the {M}ountain {W}est},
+year = {1977},
+type = {Gen. Tech. Rep.},
+number = {INT-GTR-42},
+institution = {U.S. Department of Agriculture, Forest Service, Intermountain Forest and Range Experiment Station},
+address = {Ogden, UT},
+}
+
+@article{Baha85,
+author = {Bahari, Z. A. and Pallardy, S. G. and Parker, W. C.},
+title = {Photosynthesis, water relations, and drought adaptation in six woody species of oak-hickory forests in central {M}issouri},
+journal = {Forest Science},
+volume = {31},
+year = {1985},
+pages = {557-569},
+}
+
+@article{Bart96,
+author = {Barton, A. M. and Gleeson, S. K.},
+title = {Ecophysiology of seedlings of oaks and red maple across a topographic gradient in eastern {K}entucky},
+journal = {Forest Science},
+volume = {42},
+year = {1996},
+pages = {335-342},
+}
+
+@article{Bass01,
+author = {Basset, Y. and Charles, E. and Hammond, D. and Brown, V. K.},
+title = {Short-term effects of canopy openness on insect herbivores in a rain forest in {G}uyana},
+journal = {Journal of Applied Ecology},
+volume = {38},
+year = {2001},
+pages = {1045-1059},
+}
+
+@techreport{Beck70,
+author = {Beck, D. E.},
+title = {Effects of competition on survival and height growth of red oak seedlings},
+year = {1970},
+type = {Res. Pap.},
+number = {SE-56},
+institution = {U.S. Department of Agriculture, Forest Service, Southeastern Forest Experiment Station},
+address = {Asheville, NC},
+}
+
+@book{Beck81,
+author = {Beck, D. E. and Della-Bianca, L.},
+title = {Yellow-poplar: characteristics and management},
+publisher = {USDA Forest Service Agriculture HAndbook No. 583},
+year = {1981},
+}
+
+@article{Beck86,
+author = {Beck, D. E. and Hooper, R. M.},
+title = {Development of a southern {A}ppalachian hardwood stand after clearcutting},
+journal = {Southern Journal of Applied Forestry},
+volume = {10},
+year = {1986},
+pages = {168-172},
+}
+
+@article{Beck03,
+author = {Beckage, B. and Clark, J. S.},
+title = {Seedling survival and growth of three forest tree species: the role of spatial heterogeneity},
+journal = {Ecology},
+volume = {84},
+year = {2003},
+pages = {1849-1861},
+}
+
+@article{Beie90,
+author = {Beier, P. and McCullough, D. R.},
+title = {Factors influencing white-tailed deer activity patterns and habitat use},
+journal = {Wildlife Monographs},
+volume = {109},
+year = {1990},
+pages = {3-51},
+}
+
+@article{Bell05,
+author = {Bellocq, M. I. and Jones, C. and Dey, D. C. and Turgeon, J. J.},
+title = {Does the shelterwood method to regenerate oak forests affect acorn production and predation?},
+journal = {Forest Ecology and Management},
+volume = {205},
+year = {2005},
+pages = {311-323},
+}
+
+@article{Bels86,
+author = {Belsky, A. J.},
+title = {Does herbivory benefit plants? {A} review of the evidence},
+journal = {American Naturalist},
+volume = {127},
+year = {1986},
+pages = {870-892},
+}
+
+@article{Berg04,
+author = {Berger, A. L. and Puettmann, K. J. and Host, G. E.},
+title = {Harvesting impacts on soil and understory vegetation: the influence of season of harvest and within-site disturbance patterns on clear-cut aspen stands in {M}innesota},
+journal = {Canadian Journal of Forestry Research},
+volume = {34},
+year = {2004},
+pages = {2159-2168},
+}
+
+@article{Berg11,
+author = {Bergeron, P. and Reale, D. and Humphries, M. M. and Garant, D.},
+title = {Anticipation and tracking of pulsed resources drive population dynamics in eastern chipmunks},
+journal = {Ecology},
+volume = {92},
+year = {2011},
+pages = {2027-2034},
+}
+
+@book{Beta90,
+author = {Betancourt, J. L. and Van Devender, T. R. and Martin, P. S.},
+title = {Packrat middens: the last 40,000 years of biotic change},
+publisher = {University of Arizona Press},
+year = {1990},
+address = {Tucson, AZ},
+}
+
+@article{Bide15,
+author = {Bideau, E. and Maublanc, M. and Picot, D. and Hamard, J. and Ballon, P. and Gerard, J.},
+title = {Short-term browsing by roe deer has little effect on survival and growth of sessile oak seedlings},
+journal = {Scandinavian Journal of Forest Research},
+doi = {10.1080/02827581.2015.1054873},
+year = {2015},
+}
+
+@article{Bier92,
+author = {Bierregaard, R. O. and Lovejoy, T. E. and Kapos, V. and dos Santos, A. and Hutchings, R. W.},
+title = {The biological dynamics of tropical rainforest fragments},
+journal = {BioScience},
+volume = {42},
+year = {1992},
+pages = {859-866},
+}
+
+@article{Bigg99,
+author = {Bigger, S. D. and Marvier, M. A.},
+title = {How different would a world without herbivory be? {A} search for generality in ecology},
+journal = {Integrative Biology},
+volume = {1},
+year = {1999},
+pages = {60-67},
+}
+
+@article{Blym77,
+author = {Blymyer, M. J. and Mosby, H. S.},
+title = {Deer utilization of clearcuts in southwestern {V}irginia},
+journal = {Southern Journal of Applied Forestry},
+volume = {3},
+year = {1977},
+pages = {10-13},
+}
+
+@inbook{Bona92,
+author = {Bonan, G. B.},
+booktitle = {A systems analysis of the global boreal forest},
+editor = {Shugart, H. H. and Leemans, R. and Bonan, G. B.},
+title = {A simulation analysis of environmental factors and ecological processes in North American boreal forests},
+publisher = {Cambridge University Press},
+year = {1992},
+address = {Cambridge, UK},
+pages = {404-427},
+}
+
+@article{Bond02,
+author = {Bond, B. T. and Burger, L. W. and Leopold, B. D. and Jones, J. C. and Godwin, K. D.},
+title = {Habitat use by cottontail rabbits across multiple spatial scales in {M}ississippi},
+journal = {Journal of Wildlife Management},
+volume = {66},
+year = {2002},
+pages = {1171-1178},
+}
+
+@book{Botk93,
+author = {Botkin, D. B.},
+title = {Forest dynamics: an ecological model},
+publisher = {Oxford University Press},
+year = {1993},
+address = {New York, NY},
+}
+
+@article{Boul09,
+author = {Boulanger, V. and Baltzinger, C. and Sa\"{\i}d, S. and Ballon, P. and Picard, J. F. and Dupouey, J. L.},
+title = {Ranking temperate woody species along a gradient of browsing by deer},
+journal = {Forest Ecology and Management},
+volume = {258},
+year = {2009},
+pages = {1397-1406},
+}
+
+@article{Bowe93,
+author = {Bowers, M. A. and Ellis, A.},
+title = {Load size variation in the eastern chipmunk, \textit{{T}amias striatus}: the importance of distance from burrow and canopy cover},
+journal = {Ethology},
+volume = {94},
+year = {1993},
+pages = {72-82},
+}
+
+@article{Bowe93b,
+author = {Bowers, M. A. and Jefferson, J. L. and Kuebler, M. G.},
+title = {Variation in giving-up densities of foraging chipmunks (\textit{{T}amias striatus}) and squirrels (\textit{{S}ciurus carolinensis})},
+journal = {Oikos},
+volume = {66},
+year = {1993},
+pages = {229-236},
+}
+
+@article{Broo98,
+author = {Brooks, S. P. and Gelman, A.},
+title = {General methods for monitoring convergence of iterative simulations},
+journal = {Journal of Computational Graphics and Statistics},
+volume = {7},
+year = {1998},
+pages = {434-455},
+}
+
+@article{Bros99,
+author = {Brose, P. and Van Lear, D. and Cooper, R.},
+title = {Using shelterwood harvests and prescribed fire to regenerate oak stands on productive upland sites},
+journal = {Forest Ecology and Management},
+volume = {113},
+year = {1999},
+pages = {125-141},
+}
+
+@article{Bros01,
+author = {Brose, P. and Schuler, T. and Van Lear, D. and Berst, J.},
+title = {Bringing fire back: the changing regimes of the {A}ppalachian mixed-oak forests},
+journal = {Journal of Forestry},
+volume = {99},
+year = {2001},
+pages = {30-35},
+}
+
+@techreport{Bros11,
+author = {Brose, P.},
+title = {Fate of the acorn crop at {C}lear {C}reek {S}tate {F}orest, {P}ennsylvania},
+year = {2011},
+type = {Gen. Tech. Rep.},
+number = {NRS-P-78},
+institution = {U.S. Department of Agriculture, Forest Service, Northern Research Station},
+address = {Newtown Square, PA},
+}
+
+@article{Brow88,
+author = {Brown, J. S. and Kotler, B. P. and Smith, R. J. and Wirtz, W. O.},
+title = {The effects of owl predation on the foraging behavior of heteromyid rodents},
+journal = {Oecologia},
+volume = {76},
+year = {1988},
+pages = {408-415},
+}
+
+@article{Buck98,
+author = {Buckley, D. S. and Sharik, T. L. and Isebrands, J. G.},
+title = {Regeneration of northern red oak: positive and negative effects of competitor removal},
+journal = {Ecology},
+volume = {79},
+year = {1998},
+pages = {65-78},
+}
+
+@article{Bugm01a,
+author = {Bugmann, H.},
+title = {A review of forest gap models},
+journal = {Climatic Change},
+volume = {51},
+year = {2001},
+pages = {239-305},
+}
+
+@article{Bugm01b,
+author = {Bugmann, H. and Wullschleger, S. D. and Price, D. T. and Ogle, K. and Clark, D. F. and Solomon, A. M.},
+title = {Comparing the performance of forest gap models in {N}orth {A}merica},
+journal = {Climatic Change},
+volume = {51},
+year = {2001},
+pages = {349-388},
+}
+
+@article{Bund91,
+author = {Bundy, P. P. and Alm, A. A. and Baughman, M. J.},
+title = {Red oak regeneration following "scarification" and harvesting: a case study},
+journal = {Northern Journal of Applied Forestry},
+volume = {8},
+year = {1991},
+pages = {173-174},
+}
+
+@article{Cade00,
+author = {Cadenasso, M. L. and Pickett, S. T. A.},
+title = {Linking forest edge structure to edge function: mediation of herbivore damage},
+journal = {Journal of Ecology},
+volume = {88},
+year = {2000},
+pages = {31-44},
+}
+
+@techreport{Carm13,
+author = {Carman, S. F.},
+title = {Indiana forest management history and practices},
+year = {2013},
+type = {Gen. Tech. Rep.},
+number = {NRS-P-108},
+institution = {U.S. Department of Agriculture, Forest Service, Northern Forest Experiment Station},
+address = {Newtown Square, PA},
+pages = {12-23},
+}
+
+@article{Cast00,
+author = {Castleberry, S. B. and Ford, W. M. and Miller, K. V. and Smith, W. P.},
+title = {Influences of herbivory and canopy opening size on forest regeneration in a southern bottomland hardwood forest},
+journal = {Forest Ecology and Management},
+volume = {131},
+year = {2000},
+pages = {57-64},
+}
+
+@article{Cast02,
+author = {Castleberry, N. L. and Castleberry, S. B. and Ford, W. M. and Wood, P. B. and Mengak, M. T.},
+title = {Allegheny woodrat (\textit{Neotoma magister}) food habits in the {C}entral {A}ppalachians},
+journal = {American Midland Naturalist},
+volume = {147},
+year = {2002},
+pages = {80-92},
+}
+
+@book{Casw01,
+author = {Caswell, H.},
+title = {Matrix population models},
+publisher = {Sinauer Associates},
+year = {2001},
+address = {Sunderland, MA},
+}
+
+@article{Caug15,
+author = {Caughlin, T. T. and Ferguson, J. M. and Lichstein, J. W. and Zuidema, P. A. and Bunyavejchewin, S. and Levey, D. J.},
+title = {Loss of animal seed dispersal increases extinction risk in a tropical tree species due to pervasive negative density dependence across life stages},
+journal = {Proceedings of the Royal Society of London B},
+volume = {282},
+year = {2015},
+pages = {1-9},
+}
+
+@article{Chri55,
+author = {Christisen, D. M.},
+title = {Yield of seed by oaks in the {M}issouri {O}zarks},
+journal = {Journal of Forestry},
+volume = {6},
+year = {1955},
+pages = {439-441},
+}
+
+@article{Clar99,
+author = {Clark, J. S. and Silman, M. and Kern, R. and Macklin, E. and HilleRisLambers, J.},
+title = {Seed dispersal near and far: patterns across temperate and tropical forests},
+journal = {Ecology},
+volume = {80},
+year = {1999},
+pages = {1475-1494},
+}
+
+@article{Clar94,
+author = {Clarke, M. F. and Kramer, D. L.},
+title = {Scatter-hoarding by a larder-hoarding rodent: intraspecific variation in the hoarding behavior of the eastern chipmunk, \textit{{T}amias striatus}},
+journal = {Animal Behavior},
+volume = {48},
+year = {1994},
+pages = {299-308},
+}
+
+@article{Clin93,
+author = {Clinton, B. D. and Boring, L. R. and Swank, W. T.},
+title = {Canopy gap characteristics and drought influences in oak frests of the {C}oweeta {B}asin},
+journal = {Ecology},
+volume = {74},
+year = {1993},
+pages = {1551-1558},
+}
+
+@article{Clot07,
+author = {Clotfelter, E. D. and Pederson, A. B. and Cranford, J. A. and Ram, N. and Snajdr, E. A. and Nolan, V. and Ketterson, E. D.},
+title = {Acorn mast drives long-term dynamics of rodent and songbird populations},
+journal = {Oecologia},
+volume = {154},
+year = {2007},
+pages = {493-503},
+}
+
+@inbook{Conn71,
+author = {Connell, J. H.},
+booktitle = {Dynamics of Populations},
+editor = {McShea, W. J. and Healy, W.},
+title = {On the role of natural enemies in preventing competitive exclusion in some marine animals and rain forest trees},
+publisher = {PUDOC},
+year = {1971},
+address = {Wageningen, Netherlands},
+pages = {298-312},
+}
+
+@article{Cote04,
+author = {C{\^o}t{\'e}, S. D. and Rooney, T. P. and Tremblay, J. P. and Dussault, C. and Waller, D. M.},
+title = {Ecological impacts of deer overabundance},
+journal = {Annual Review of Ecology, Evolution, and Systematics},
+volume = {35},
+year = {2004},
+pages = {113-147},
+}
+
+@article{Coul12,
+author = {Coulson, T.},
+title = {Integral projection models, their construction and use in posing hypotheses in ecology},
+journal = {Oikos},
+volume = {121},
+year = {2012},
+pages = {1337-1350},
+}
+
+@article{Craw95,
+author = {Crawley, M. J. and Long, C. R.},
+title = {Alternate bearing, predator satiation and seedling recruitment in \textit{{Q}uercus robur} {L}.},
+journal = {Journal of Ecology},
+volume = {83},
+year = {1995},
+pages = {683-696},
+}
+
+@article{Crim10,
+author = {Crimmins, S. M. and Edwards, J. W. and Ford, W. M. and Keyser, P. D. and Crum, J. M.},
+title = {Browsing patterns of white-tailed deer following increased timber harvest and a decline in population density},
+journal = {International Journal of Forestry Research},
+volume = {Article ID 592034},
+year = {2010},
+pages = {1-7},
+}
+
+@article{Crow88,
+author = {Crow, T. R.},
+title = {Reproductive mode and mechanisms for self-replacement of northern red oak (\textit{{Q}uercus rubra}) - a review},
+journal = {Forest Science},
+volume = {34},
+year = {1988},
+pages = {19-40},
+}
+
+@article{Crow92,
+author = {Crow, T. R.},
+title = {Population dynamics and growth patterns for a cohort of northern red oak (\textit{{Q}uercus rubra}) seedlings},
+journal = {Oecologia},
+volume = {91},
+year = {1992},
+pages = {192-200},
+}
+
+@article{Dale01,
+author = {Dale, V. H. and Joyce, L. A. and McNulty, S. and Neilson, R. P. and Ayres, M. P. and Flannigan, M. D. and Hanson, P. J. and Irland, L. C. and Lugo, A. E. and Peterson, C. J. and Simberloff, D. and Swanson, F. J. and Stocks, B. J. and Wotton, B. M.},
+title = {Climate change and forest disturbances},
+journal = {BioScience},
+volume = {51},
+year = {2001},
+pages = {723-734},
+}
+
+@article{Dalg12,
+author = {Dalgleish, H. J. and Swihart, R. K.},
+title = {American chestnut past and future: implications for resource pulses and consumer populations of eastern {U}.{S}. forests},
+journal = {Restoration Ecology},
+volume = {20},
+year = {2012},
+pages = {490-497},
+}
+
+@article{Dalg15,
+author = {Dalgleish, H. J. and Lichti, N. I. and Schmedding, N. and Swihart, R. K.},
+title = {Exposure to herbivores increases seedling growth and survival of {A}merican chestnut (\textit{{C}astanea dentata}) through decreased interspecific competition in canopy gaps},
+journal = {Restoration Ecology},
+volume = {DOI: 10.111/rec.12223},
+year = {2015},
+pages = {},
+}
+
+@article{Davi93,
+author = {Davidson, D. W.},
+title = {The effects of herbivory and granivory on terrestrial plant succession},
+journal = {Oikos},
+volume = {68},
+year = {1993},
+pages = {23-35},
+}
+
+@article{Demp77,
+author = {Dempster, A. P. and Laird, N. M. and Rubin, D. B.},
+title = {Maximum likelihood from incomplete data via the \textit{{EM}} algorithm},
+journal = {Journal of the Royal Statistical Society},
+volume = {39},
+year = {1977},
+pages = {1-38},
+}
+
+@phdthesis{Dey91,
+title = {A comprehensive {O}zark regenerator},
+author = {Dey, D.},
+school = {University of Missouri},
+address = {Columbia, MO},
+year = {1991},
+}
+
+@inbook{Dey02,
+author = {Dey, D. C.},
+booktitle = {Oak forest ecosystems: ecology and management for wildlife},
+editor = {McShea, W. J. and Healy, W.},
+title = {The ecological basis for oak silviculture in eastern {N}orth {A}merica},
+publisher = {Johns Hopkins University Press},
+year = {2002},
+address = {Baltimore, MD},
+pages = {60-79},
+}
+
+@article{Dey08,
+author = {Dey, D. C. and Jacobs, D. and McNabb, K. and Miller, G. and Baldwin, V. and Foster, G.},
+title = {Artificial regeneration of major oak (\textit{Quercus}) species in the eastern {U}nited {S}tates - a review of the literature},
+journal = {Forest Science},
+volume = {54},
+year = {2008},
+pages = {77-106},
+}
+
+@article{Dey09,
+author = {Dey, D. C. and Spetich, M. A. and Weigel, D. R. and Johnson, P. S. and Graney, D. L. and Kabrick, J. M.},
+title = {A suggested approach for design of oak (\textit{{Q}uercus {L}}.) regeneration research considering regional differences},
+journal = {New Forests},
+volume = {37},
+year = {2009},
+pages = {123-135},
+}
+
+@article{Dobs15,
+author = {Dobson, A. and Blossey, B.},
+title = {Earthworm invasion, white‐tailed deer and seedling establishment in deciduous forests of north‐eastern {N}orth {A}merica},
+journal = {Journal of Ecology},
+volume = {103},
+year = {2015},
+pages = {153-164},
+}
+
+@article{Down44,
+author = {Downs, A. A. and McQuilkin, W. E.},
+title = {Seed production of southern {A}ppalachian oaks},
+journal = {Journal of Forestry},
+volume = {42},
+year = {1944},
+pages = {913-920},
+}
+
+@article{Elli05,
+author = {Ellison, A. M. and Bank, M. S. and Clinton, B. D. and Coblurn, E. A. and Elliott, K. and Ford, C. R. and Foster, D. R. and Kloeppel, B. D. and Knoepp, J. D. and Lovett, G. M. and Mohan, J. and Orwig, D. A. and Rodenhouse, N. L. and Sobczak, W. V. and Stinson, K. A. and Stone, J. K. and Swan, C. M. and Thompson, J. and VonHolle, B. and Webster, J. R.},
+title = {Loss of foundation species: consequences for the structure and dynamics of forested ecosystems},
+journal = {Frontiers in Ecology and the Environment},
+volume = {3},
+year = {2005},
+pages = {479-486},
+}
+
+@article{Elln06,
+author = {Ellner, S. P. and Rees, M.},
+title = {Integral projection models for species with complex demography},
+journal = {American Naturalist},
+volume = {167},
+year = {2006},
+pages = {410-428},
+}
+
+@article{Enri95,
+author = {Enright, N. J. and Franco, M. and Silvertown, J.},
+title = {Comparing plant life histories using elasticity analysis: the importance of life span and the number of life-cycle stages},
+journal = {Oecologia},
+volume = {104},
+year = {1995},
+pages = {79-84},
+}
+
+@inbook{Feen76,
+author = {Feeny, P.},
+booktitle = {Recent advances in phytochemistry volume 10},
+editor = {Wallace, J. W. and Mansell, R. L.},
+title = {Plant apparency and chemical defense},
+publisher = {Johns Hopkins University Press},
+year = {1976},
+address = {New York, NY},
+pages = {1-40},
+}
+
+@article{Fork00,
+author = {Forkner, R. E. and Hunter, M. D.},
+title = {What goes up must come down? {N}utrient addition and predation pressure on oak herbivores},
+journal = {Ecology},
+volume = {81},
+year = {2000},
+pages = {1588-1600},
+}
+
+@article{Fote01,
+author = {Fotelli, M. N. and Gebler, A. and Peuke, A. D. and Rennenberg, H.},
+title = {Drought affects the competitive interactions between \textit{{F}agus sylvatica} seedlings an early successional species, \textit{{R}ubus fruticosus}: responses of growth, water status, and \textsuperscript{13}{C} composition},
+journal = {New Phytologist},
+volume = {151},
+year = {2001},
+pages = {427-435},
+}
+
+@techreport{Fral03,
+author = {Fralish, J. S.},
+title = {The {C}entral {H}ardwood {F}orest: its boundaries and physiographic provinces},
+year = {2003},
+number = {NC-GTR-234},
+type = {Gen. Tech. Rep.},
+institution = {U.S. Department of Agriculture, Forest Service, North Central Forest Experiment Station},
+address = {St. Paul, MN},
+}
+
+@techreport{Fral04,
+author = {Fralish, J. S.},
+title = {The keystone role of oak and hickory in the {C}entral {H}ardwood {F}orest},
+number = {SRS-GTR-73},
+type = {Gen. Tech. Rep.},
+institution = {U.S. Department of Agriculture, Forest Service, Southern Research Station},
+address = {Asheville, NC},
+year = {2004},
+pages = {78-86},
+}
+
+@article{Fran90,
+author = {Franklin, J. F. and Bledsoe, C. S. and Callahan, J. T.},
+title = {Contributions of the long-term ecological research program},
+journal = {BioScience},
+volume = {40},
+year = {1990},
+pages = {509-523},
+}
+
+@techreport{Galf91,
+author = {Galford, J. R. and Auchmoody, L. R. and Smith, H. C. and Walters, R. S.},
+title = {Insects affecting establishment of northern red oak seedlings in central {P}ennsylvania},
+number = {NE-148},
+type = {Gen. Tech. Rep.},
+institution = {U.S. Department of Agriculture, Forest Service, Northeastern Forest Experiment Station},
+address = {Radnor, PA},
+year = {1991},
+pages = {271-280},
+}
+
+@article{Garc02,
+author = {Garc\'{\i}a, D. and Ba\~{n}uelos, M. J. and Houle, G.},
+title = {Differential effects of acorn burial and litter cover on \textit{{Q}uercus rubra} recruitment at the limit of its range in eastern {N}orth {A}merica},
+journal = {Canadian Journal of Botany},
+volume = {80},
+year = {2002},
+pages = {1115-1120},
+}
+
+@book{Gelm04,
+author = {Gelman, A. and Carlin, J. B. and Stern, H. S. and Dunson, D. B. and Vehtari, A. and Rubin, D. B.},
+title = {Bayesian data analysis},
+publisher = {Chapman and Hall},
+year = {2004},
+address = {Boca Raton, FL},
+pagetotal = {675}
+}
+
+@techreport{Gibs72,
+author = {Gibson, L. P.},
+title = {Insects that damage white oak acorns},
+number = {NE-RP-220},
+type = {Res. Paper},
+institution = {U.S. Department of Agriculture, Forest Service, Northeastern Research Station},
+address = {Upper Darby, PA},
+year = {1972},
+pages = {1-7},
+}
+
+@techreport{Gibs82,
+author = {Gibson, L. P.},
+title = {Insects that damage northern red oak acorns},
+number = {NE-RP-492},
+type = {Res. Paper},
+institution = {U.S. Department of Agriculture, Forest Service, Northeastern Research Station},
+address = {Upper Darby, PA},
+year = {1982},
+pages = {1-6},
+}
+
+@article{Gohe03,
+author = {Goheen, J. R. and Swihart, R. K.},
+title = {Food-hoarding behavior of gray squirrels and {N}orth {A}merican red squirrels in the central hardwoods region: implications for forest regeneration},
+journal = {Canadian Journal of Zoology},
+volume = {81},
+year = {2003},
+pages = {1636-1639},
+}
+
+@inbook{Gree02,
+author = {Greenberg, C. H. and Parresol, B. R},
+booktitle = {Oak forest ecosystems: ecology and management for wildlife},
+editor = {McShea, W. J. and Healy, W.},
+title = {Dynamics of acorn production by five species of {S}outhern {A}ppalachian oaks},
+publisher = {Johns Hopkins University Press},
+year = {2002},
+address = {Baltimore, MD},
+pages = {149-172}
+}
+
+@book{Grim05,
+author = {Grimm, V. and Railsback, S. F.},
+title = {Individual-based modeling and ecology},
+publisher = {Princeton University Press},
+year = {2005},
+address = {Princeton, NJ},
+}
+
+@article{Grim06,
+author = {Grimm, V. and Berger, U. and Bastiansen, F. and Eliassen, S. and Ginot, V. and Giske, J. and Goss-Custard, J. and Grand, T. and Heinz, S. K. and Huse, G. and Huth, A. and Jepsen, J. U. and Jorgensen, C. and Mooij, W. M. and Muller, B. and Pe'er G. and Piou, C. and Railsback, S. F. and Robbins, A. M. and Robbins, M. M. and Rossmanith, E. and Ruger, N. and Strand, E. and Souissi, S. and Stillman, R. A. and Vabo, R. and Visser, U. and DeAngelis, D. L.},
+title = {A standard protol for describing individual-based and agent-based models},
+journal = {Ecological Modelling},
+volume = {198},
+year = {2006},
+pages = {115-126},
+}
+
+@article{Grim10,
+author = {Grimm, V. and Berger, U. and DeAngelis, D. L. and Polhill, J. G. and Giske, J. and Railsback, S. F.},
+title = {The {ODD} protocol: a review and first update},
+journal = {Ecological Modelling},
+volume = {221},
+year = {2010},
+pages = {2760-2768},
+}
+
+@article{Guar98,
+author = {Guariguata, M. R. and Pinard, M. A.},
+title = {Ecological knowledge of regeneration from seed in neotropical forest trees: implications for natural forest management},
+journal = {Forest Ecology and Management},
+volume = {112},
+year = {1998},
+pages = {87-99},
+}
+
+@article{Gust13,
+author = {Gustafson, E. J. and Sturtevant, B. R.},
+title = {Modeling forest mortality caused by drought stress: implications for climate change.},
+journal = {Ecosystems},
+volume = {16},
+year = {2013},
+pages = {60-74},
+}
+
+@techreport{Guye06,
+author = {Guyette, R. P. and Dey, D. C. and Stambaugh, M. C. and Muzika, R.},
+title = {Fire scars reveal variability and dynamics of eastern fire regimes},
+year = {2006},
+type = {Gen. Tech. Rep.},
+number = {NRS-P-1},
+institution = {U.S. Department of Agriculture, Forest Service, Northern Research Station},
+address = {Newtown Square, PA},
+pages = {20-39},
+}
+
+@article{Haas05,
+author = {Haas, J. P. and Heske, E. J.},
+title = {Experimental study of the effects of mammalian acorn predators on red oak acorn survival and germination},
+journal = {Journal of Mammalogy},
+volume = {86},
+year = {2005},
+pages = {1015-1021},
+}
+
+@article{Hadj96,
+author = {Hadj-Chikh, L. and Steele, M. and Smallwood, P.},
+title = {Caching decisions by grey squirrels: a test of the handling time and perishability hypothesis},
+journal = {Animal Behavior},
+volume = {52},
+year = {1996},
+pages = {941-948},
+}
+
+@article{Hanl98,
+author = {Hanley, M. E.},
+title = {Seedling herbivory, community composition and plant life history traits},
+journal = {Perspectives in Plant Ecology, Evolution and Systematics},
+volume = {1},
+year = {1998},
+pages = {191-205},
+}
+
+@article{Harm99,
+author = {Harmer, R.},
+title = {Survival and new shoot production by artificially browsed seedlings of ash, beech, oak, and sycamore grown under different levels of shade},
+journal = {Forest Ecology and Management},
+volume = {116},
+year = {1999},
+pages = {39-50},
+}
+
+@article{Harp99,
+author = {Harpole, D. N. and Haas, C. A.},
+title = {Effects of seven silvicultural treatments on terrestrial salamanders},
+journal = {Forest Ecology and Management},
+volume = {114},
+year = {1999},
+pages = {349-356},
+}
+
+@article{Harp05,
+author = {Harper, K. A. and MacDonald, S. E. and Burton, P. J. and Chen, J. and Brosofske, K. D. and Saunders, S. C. and Euskirchen, E. S. and Roberts, D. and Jaiteh, M. S. and Esseen, P.},
+title = {Edge influence on forest structure and composition in fragmented landscapes},
+journal = {Conservation Biology},
+volume = {19},
+year = {2005},
+pages = {768-782},
+}
+
+@book{Haul09,
+author = {Haulton, S.},
+title = {Indiana {D}ivision of {F}orestry {P}roperties section wildlife habitat program 2009 annual report},
+publisher = {Indiana Division of Forestry},
+year = {2009},
+address = {Indianapolis, IN},
+note = {Available online at http://www.indiana.gov/dnr/forestry/files/fo-WildlifeHabAnnRep.pdf},
+}
+
+@inbook{Heal97,
+author = {Healy, W. M.},
+booktitle = {The science of overabundance: deer ecology and population management},
+editor = {McShea, W. J. and Underwood, H. B. and Rappole, J. H.},
+title = {Influence of deer on the structure and composition of oak forests in central {M}assachusetts},
+publisher = {Smithsonian Institution Press},
+year = {1997},
+address = {Washington, DC},
+pages = {249-266},
+}
+
+@article{Heal99,
+author = {Healy, W. M. and Lewis, A. M. and Boose, E. F.},
+title = {Variation of red oak acorn production},
+journal = {Forest Ecology and Management},
+volume = {116},
+year = {1999},
+pages = {1-11},
+}
+
+@article{Hirs12,
+author = {Hirsch, B. T. and Kays, R. and Pereira, V. E. and Jansen, P. A.},
+title = {Directed seed dispersal towards areas with low conspecific tree density by a scatter-hoarding rodent},
+journal = {Ecology Letters},
+volume = {15},
+year = {2012},
+pages = {1423-1429},
+}
+
+@article{Holm13,
+author = {Holm, J. A. and Thompson, J. R. and McShea, W. J. and Bourg, N. A.},
+title = {Interactive effects of chronic deer browsing and canopy gap disturbance on forest successional dynamics},
+journal = {Ecosphere},
+volume = {4},
+year = {2013},
+pages = {art144},
+}
+
+@techreport{Hoov13,
+author = {Hoover, W. L.},
+title = {Value of the {M}organ-{M}onroe-{Y}ellowwood state forest complex},
+year = {2013},
+type = {Gen. Tech. Rep.},
+number = {NRS-P-108},
+institution = {U.S. Department of Agriculture, Forest Service, Northern Forest Experiment Station},
+address = {Newtown Square, PA},
+pages = {287-313},
+}
+
+@article{Hors03,
+author = {Horsley, S. B. and Stout, S. L. and DeCalesta, D. S.},
+title = {White-tailed deer impact on the vegetation dynamics of a northern hardwood forest},
+journal = {Ecological Applications},
+volume = {13},
+year = {2003},
+pages = {98-118},
+}
+
+@article{Howe82,
+author = {Howe, H. F. and Smallwood, J. S.},
+title = {Ecology of seed dispersal},
+journal = {Annual Review of Ecology and Systematics},
+volume = {13},
+year = {1982},
+pages = {201-228},
+}
+
+@article{Huen87,
+author = {Huenneke, L. F. and Marks, P. L.},
+title = {Stem dynamics of the shrub \textit{{A}lnus incana} spp. \textit{rugosa}: transition matrix models},
+journal = {Ecology},
+volume = {68},
+year = {1987},
+pages = {1234-1242},
+}
+
+@article{Hulm99,
+author = {Hulme, P. E. and Borelli, T.},
+title = {Variability in post-dispersal seed predation in deciduous woodland: relative importance of location, seed species, burial, and density},
+journal = {Plant Ecology},
+volume = {145},
+year = {1999},
+pages = {149-156},
+}
+
+@article{Hunt91,
+author = {Huntly, N.},
+title = {Herbivores and the dynamics of communities and ecosystems},
+journal = {Annual Review of Ecology and Systematics},
+volume = {22},
+year = {1991},
+pages = {447-503},
+}
+
+@book{Indi84,
+author = "{Indiana Department of Natural Resources}",
+title = {Indiana forest soils handbook},
+publisher = {Indiana Department of Natural Resources, Division of Forestry},
+year = {1984},
+address = {Indianapolis, IN},
+}
+
+@book{Jack09,
+author = {Jackman, S.},
+title = {Bayesian analysis for the social sciences},
+publisher = {John Wiley and Sons},
+year = {2009},
+address = {West Sussex, UK},
+pagetotal = {598},
+}
+
+@article{Jans04,
+author = {Jansen, P. A. and Bongers, F. and Hemerik, L.},
+title = {Seed mass and mast seeding enhance dispersal by a neotropical scatter-hoarding rodent},
+journal = {Ecological Monographs},
+volume = {74},
+year = {2004},
+pages = {569-589},
+}
+
+@article{Janz70,
+author = {Janzen, D. H.},
+title = {Herbivores and the number of tree species in tropical forests},
+journal = {American Naturalist},
+volume = {104},
+year = {1970},
+pages = {501-528},
+}
+
+@article{Janz71,
+author = {Janzen, D. H.},
+title = {Seed predation by animals},
+journal = {Annual Review of Ecology and Systematics},
+volume = {2},
+year = {1971},
+pages = {465-492},
+}
+
+@article{Jeff06,
+author = {Jeffries, J. M. and Marquis, R. J. and Forkner, R. E.},
+title = {Forest age influences oak insect herbivore community structure, richness, and density},
+journal = {Ecological Applications},
+volume = {16},
+year = {2006},
+pages = {901-912},
+}
+
+@article{Jenk98,
+author = {Jenkins, M. A. and G. R. Parker},
+title = {Composition and diversity of woody vegetation in silvicultural openings of southern {I}ndiana forests},
+journal = {Forest Ecology and Management},
+volume = {109},
+year = {1998},
+pages = {57-74},
+}
+
+@article{Jink12,
+author = {Jinks, R. L. and Parratt, M. and Morgan G.},
+title = {Preference of granivorous rodents for seeds of 12 temperate tree and shrub species used in direct sowing},
+journal = {Forest Ecology and Management},
+volume = {278},
+year = {2012},
+pages = {71-79},
+}
+
+@article{John84,
+author = {Johnson, P.S.},
+title = {Responses of northern red oak to three overstory treatments},
+journal = {Canadian Journal of Forest Research},
+volume = {14},
+year = {1984},
+pages = {536-542},
+}
+
+@article{John95,
+author = {Johnson, A. S. and Hale, P. E. and Ford, W. M. and Wentworth, J. M. and French, J. R. and Anderson, O. F. and Pullen, G. B.},
+title = {White-tailed deer foraging in relation to successional stage, overstory type, and management of southern {A}ppalachian forests},
+journal = {American Midland Naturalist},
+volume = {133},
+year = {1995},
+pages = {18-35},
+}
+
+@techreport{John94,
+author = {Johnson, P. S.},
+title = {How to manage oak forests for acorn production},
+year = {1994},
+type = {Tech. Brief},
+number = {TB-NC-1},
+institution = {U.S. Department of Agriculture, Forest Service, North Central Forest Experiment Station},
+address = {St. Paul, MN},
+pages = {},
+}
+
+@book{John09,
+author = {Johnson, P. S. and Shifley, S. R. and Rogers, R.},
+title = {The ecology and silviculture of oaks},
+publisher = {CABI Publishing},
+year = {2009},
+address = {New York, NY},
+pages = {},
+pagetotal = {600}
+}
+
+@article{Jong08,
+author = {Jongejans, E. and Skarpaas, O. and Shea, K.},
+title = {Dispersal, demography, and spatial population models for conservation and control management},
+journal = {Perspectives in Plant Ecology, Evolution, and Systematics},
+volume = {9},
+year = {2008},
+pages = {153-170},
+}
+
+@article{Jong11,
+author = {Jongejans, E. and Shea, K. and Skarpaas, O. and Kelly, D. and Ellner, S. P.},
+title = {Importance of individual and environmental variation for invasive species spread: a spatial integral projection model},
+journal = {Ecology},
+volume = {92},
+year = {2011},
+pages = {86-97},
+}
+
+@techreport{Kalb13,
+author = {Kalb, R. A. and Mycroft, C. J.},
+title = {The {H}ardwood {E}cosystem {E}xperiment: goals, design, and implementation},
+year = {2013},
+type = {Gen. Tech. Rep.},
+number = {NRS-P-108},
+institution = {U.S. Department of Agriculture, Forest Service, Northern Forest Experiment Station},
+address = {Newtown Square, PA},
+pages = {36-59},
+}
+
+@article{Kell13,
+author = {Kellner, K. F. and Urban, N. A. and Swihart, R. K.},
+title = {Short-term responses of small mammals to timber harvest in the {US} {C}entral {H}ardwood {F}orest region},
+journal = {Journal of Wildlife Management},
+volume = {77},
+year = {2013},
+pages = {1650-1663},
+}
+
+@article{Kell14,
+author = {Kellner, K. F. and Riegel, J. K. and Swihart, R. K.},
+title = {Effects of silvicultural disturbance on acorn infestation and removal},
+journal = {New Forests},
+volume = {45},
+year = {2014},
+pages = {265-281},
+}
+
+@article{Kene99,
+author = {Kenefic, L. S. and Nyland, R. D.},
+title = {Sugar maple height-diameter and age-diameter relationships in an uneven-aged northern hardwood stand},
+journal = {Northern Journal of Applied Forestry},
+volume = {16},
+year = {1999},
+pages = {43-47},
+}
+
+@manual{Kell15,
+author = {Kellner, K. F.},
+title = {jags{UI}: a wrapper around rjags to streamline {JAGS} analyses},
+year = {2015},
+note = {R package version 1.3.7},
+}
+
+@book{Kery10,
+author = {K\'{e}ry, M.},
+title = {Introduction to {WINBUGS} for ecologists},
+publisher = {Elsevier Academic Press},
+year = {2010},
+address = {Burlington, MA},
+}
+
+@article{Kirk90,
+author = {Kirkland, G. L.},
+title = {Patterns of initial small mammal community change after clearcutting of temperate {N}orth {A}merican forests},
+journal = {Oikos},
+volume = {59},
+year = {1990},
+pages = {313-320},
+}
+
+@article{Kitt95,
+author = {Kittredge, D. B. and Ashton, P. M.},
+title = {Impact of deer browsing on regeneration in mixed stands in southern {N}ew {E}ngland},
+journal = {Northern Journal of Applied Forestry},
+volume = {12},
+year = {1995},
+pages = {115-120},
+}
+
+@article{Kotl91,
+author = {Kotler, B. P. and Brown, J. S. and Hasson, O.},
+title = {Factors affecting gerbil foraging behavior and rate of owl predation},
+journal = {Ecology},
+volume = {72},
+year = {1991},
+pages = {2249-2260},
+}
+
+@article{Koen00,
+author = {Koenig, W. D. and Knops, J. H.},
+title = {Patterns of annual seed production by {N}orthern {H}emisphere trees: a global perspective},
+journal = {The American Naturalist},
+volume = {155},
+year = {2000},
+pages = {59-69},
+}
+
+@inbook{Koen02,
+author = {Koenig, W. and Knops, J.},
+booktitle = {Oak forest ecosystems: ecology and management for wildlife},
+editor = {McShea, W. J. and Healy, W.},
+title = {The behavioral ecology of masting in oaks},
+publisher = {Johns Hopkins University Press},
+year = {2002},
+address = {Baltimore, MD},
+pages = {129-148},
+}
+
+@article{Kubi94,
+author = {Kubiske, M. E. and Abrams, M. D.},
+title = {Ecophysiological analysis of temperate woody species on contrasting sites during wet and dry years},
+journal = {Oecologia},
+volume = {98},
+year = {1994},
+pages = {303-312},
+}
+
+@article{Lafo02,
+author = {Lafon, C. W. and Speer, J. H.},
+title = {Using dendrochronology to identify major ice storm events in oak forests of southwestern Virginia},
+journal = {Climate Research},
+volume = {20},
+year = {2002},
+pages = {41-54},
+}
+
+@article{Lars98,
+author = {Larsen, D. R. and Johnson, P. S.},
+title = {Linking the ecology of natural oak regeneration to silviculture},
+journal = {Forest Ecology and Management},
+volume = {106},
+year = {1998},
+pages = {1-7},
+}
+
+@article{Lesl45,
+author = {Leslie, P. H.},
+title = {On the use of matrices in certain population mathematics},
+journal = {Biometrika},
+volume = {33},
+year = {1945},
+pages = {183-212},
+}
+
+@article{Lhot03,
+author = {Lhotka, J. M. and Zaczek, J. J.},
+title = {Effects of scarification disturbance on the seedling and midstory layer in a successional mixed-oak forest},
+journal = {Northern Journal of Applied Forestry},
+volume = {20},
+year = {2003},
+pages = {85-91},
+}
+
+@article{Lian05,
+author = {Liang, J. and Buongiorno, J. and Monserud, R. A.},
+title = {Growth and yield of all-aged {D}ouglas-fir western hemlock forest stands: a matrix model with stand diversity effects},
+journal = {Canadian Journal of Forest Research},
+volume = {35},
+year = {2005},
+pages = {2368-2381},
+}
+
+@article{Lian13,
+author = {Liang, J. and Picard, N.},
+title = {Matrix models of forest dynamics: an overview and outlook},
+journal = {Forest Science},
+volume = {59},
+year = {2013},
+pages = {359-378},
+}
+
+@phdthesis{Lich12,
+title = {Implications of context-dependent scatterhoarding for seed survival and dispersal in {N}orth {A}merican oaks (\textit{Quercus})},
+author = {Lichti, N.},
+school = {Purdue University},
+address = {West Lafayette, IN},
+year = {2012},
+}
+
+@article{Lich14,
+author = {Lichti, N. I. and Steele, M. A. and Zhang, H. and Swihart, R. K.},
+title = {Mast species composition alters seed fate in {N}orth {A}merican rodent-dispersed hardwoods},
+journal = {Ecology},
+volume = {95},
+year = {2014},
+pages = {1746-1758},
+}
+
+@article{Lini86,
+author = {Linit, M. J. and Johnson, P. S. and McKinney, R. A. and Kearby, W. H.},
+title = {Insects and leaf area losses of planted northern red oak seedlings in an {O}zark forest},
+journal = {Forest Science},
+volume = {32},
+year = {1986},
+pages = {11-20},
+}
+
+@article{Liu95,
+author = {Liu, J. and Ashton, P. S.},
+title = {Individual-based simulation models for forest succession and management},
+journal = {Forest Ecology and Management},
+volume = {73},
+year = {1995},
+pages = {157-175},
+}
+
+@article{Loft90,
+author = {Loftis, D. L.},
+title = {A shelterwood method for regenerating red oak in the southern {A}ppalachians},
+journal = {Forest Science},
+volume = {36},
+year = {1990},
+pages = {917-929},
+}
+
+@article{Lomb08,
+author = {Lombardo, J. A. and McCarthy, B. C.},
+title = {Silvicultural treatment effects on oak seed production and predation by acorn weevils in southeastern {O}hio},
+journal = {Forest Ecology and Management},
+volume = {255},
+year = {2008},
+pages = {2566-2576},
+}
+
+@article{Lomb09,
+author = {Lombardo, J. A. and McCarthy, B. C.},
+title = {Seed germination and seedling vigor of weevil-damaged acorns of red oak},
+journal = {Canadian Journal of Forest Research},
+volume = {39},
+year = {2009},
+pages = {1600-1605},
+}
+
+@article{Lori94,
+author = {Lorimer, C. G. and Chapman, J. W. and Lambert, W. D.},
+title = {Tall understorey vegetation as a factor in the poor development of oak seedlings beneath mature stands},
+journal = {Journal of Ecology},
+volume = {82},
+year = {1994},
+pages = {227-237},
+}
+
+@article{Lusk07,
+author = {Lusk, J. L. and Swihart, R. K. and Goheen, J. R.},
+title = {Correlates of interspecific synchrony and interannual variation in seed production by deciduous trees},
+journal = {Forest Ecology and Management},
+volume = {242},
+year = {2007},
+pages = {656-670},
+}
+
+@article{Macd83,
+author = {MacDonald, J. E. and Powell, G. R.},
+title = {Relationships between stump sprouting and parent-tree diameter in sugar maple in the 1\textsuperscript{st} year following clearcutting},
+journal = {Canadian Journal of Forest Research},
+volume = {13},
+year = {1983},
+pages = {390-394},
+}
+
+@article{Mank99,
+author = {Mankin, P. C. and Warner, R. E.},
+title = {A regional model of the eastern cottontail and land-use changes in {I}llinois},
+journal = {Journal of Wildlife Management},
+volume = {63},
+year = {1999},
+pages = {956-963},
+}
+
+@article{Mans98,
+author = {Manson, R. H. and Stiles, E. W.},
+title = {Links between microhabitat preferences and seed predation by small mammals in old fields},
+journal = {Oikos},
+volume = {82},
+year = {1998},
+pages = {37-50},
+}
+
+@techreport{Marq76,
+author = {Marquis, D. A. and Eckert, P. L. and Roach, B. A.},
+title = {Acorn weevils, rodents, and deer all contribute to oak-regeneration difficulties in {P}ennsylvania},
+number = {NE-RP-356},
+type = {Res. Paper},
+institution = {U.S. Department of Agriculture, Forest Service, Northeastern Forest Experiment Station},
+address = {Upper Darby, PA},
+year = {1976},
+pages = {1-8},
+}
+
+@techreport{Marq81,
+author = {Marquis, D. A.},
+title = {Effect of deer browsing on timber production in {A}llegheny hardwood forests of northwestern {P}ennsylvania},
+number = {NE-475},
+type = {Res. Paper},
+institution = {U.S. Department of Agriculture, Forest Service, Northeastern Forest Experiment Station},
+address = {Radnor, Pennsylvania},
+year = {1981},
+}
+
+@article{Marq94,
+author = {Marquis, R. J. and Whelan, C. J.},
+title = {Insectivorous birds increase growth of white oak through consumption of leaf-chewing insects},
+journal = {Ecology},
+volume = {75},
+year = {1994},
+pages = {2007-2014},
+}
+
+@book{Mart61,
+author = {Martin, A. C. and Zim, H. S. and Nelson, A. L.},
+title = {American wildlife and plants: a guide to wildlife food habits},
+publisher = {Dover Publishers},
+year = {1961},
+address = {New York, NY},
+}
+
+@article{Mart87,
+author = {Martin, U. and Pallardy, S. G. and Bahari, Z. A.},
+title = {Dehydration tolerance of leaf tissues of six woody angiosperm species},
+journal = {Physiologia Plantarum},
+volume = {69},
+year = {1987},
+pages = {182-186},
+}
+
+@article{Mart08,
+author = {Martin, B. and Shao, G. and Swihart, R. K. and Parker, G. R. and Tang, L.},
+title = {Implications of shared edge length between land cover types for landscape quality: a case study in {M}idwestern {U.S.A.}},
+journal = {Landscape Ecology},
+volume = {23},
+year = {2008},
+pages = {391-402},
+}
+
+@article{Mcca94,
+author = {McCarthy, B. C.},
+title = {Experimental studies of hickory recruitment in a wooded hedgerow and forest},
+journal = {Bulletin of the Torrey Botanical Club},
+volume = {121},
+year = {1994},
+pages = {240-250},
+}
+
+@article{Mcew11,
+author = {McEwan, R. W. and Dyer, J. M. and Pederson, N.},
+title = {Multiple interacting ecosystem drivers: toward an encompassing
+hypothesis of oak forest dynamics across eastern {N}orth {A}merica},
+journal = {Ecography},
+volume = {34},
+year = {2011},
+pages = {244-256},
+}
+
+@book{Mcna94,
+author = {McNab, W. H. and Avers, P. E.},
+title = {Ecological subregions of the {U}nited {S}tates: section descriptions ({WO-WSA-5})},
+publisher = {United States Department of Agriculture, Forest Service},
+year = {1994},
+address = {Washington, DC},
+}
+
+@article{Mcph93,
+author = {McPherson, G. R.},
+title = {Effects of herbivory and herb interference on oak establishment in a semi-arid temperate savanna},
+journal = {Journal of Vegetation Science},
+volume = {4},
+year = {1993},
+pages = {687-692},
+}
+
+@article{Mcsh00,
+author = {McShea, W. J.},
+title = {The influence of acorn crops on annual variation in rodent and bird populations},
+journal = {Ecology},
+volume = {81},
+year = {2000},
+pages = {228-238},
+}
+
+@article{Mcsh93,
+author = {McShea, W. J. and Schwede, G.},
+title = {Variable acorn crops: responses of white-tailed deer and other mast consumers},
+journal = {Journal of Mammalogy},
+volume = {74},
+year = {1993},
+pages = {999-1006},
+}
+
+@book{Mcsh02,
+author = {McShea, W. J. and Healy, W. M.},
+title = {Oak forest ecosystems: ecology and management for wildlife},
+publisher = {Johns Hopkins University Press},
+year = {2002},
+address = {Baltimore, MD},
+pagetotal = {448}
+}
+
+@article{Mcsh07,
+author = {McShea, W. J. and Healy, W. M. and Devers, P. and Fearer, T. and Koch, F. H. and Stauffer, D. and Waldon, J.},
+title = {Forestry matters: decline of oaks will impact wildlife in hardwood forests},
+journal = {Journal of Wildlife Management},
+volume = {71},
+year = {2007},
+pages = {1717-1728},
+}
+
+@article{Megr01,
+author = {Megrey, B. A. and Hinckley, S.},
+title = {Effect of turbulence on feeding of larval fishes: a sensitivity analysis using an individual-based model},
+journal = {Journal of Marine Science},
+volume = {58},
+year = {2001},
+pages = {1015-1029},
+}
+
+@article{Mein02,
+author = {Meiners, S. J. and Martinkovic, M. J.},
+title = {Survival of and herbivore damage to a cohort of \textit{{Q}uercus rubra} planted across a forest-old-field edge},
+journal = {American Midland Naturalist},
+volume = {147},
+year = {2002},
+pages = {247-255},
+}
+
+@article{Mill09,
+author = {Miller, B. F. and Campbell, T. A. and Laseter, B. R. and Ford, W. M. and Miller, K. V.},
+title = {White-tailed deer herbivory and timber harvesting rates: implications for regeneration success},
+journal = {Forest Ecology and Management},
+volume = {258},
+year = {2009},
+pages = {1067-1072},
+}
+
+@article{Mish10,
+author = {Mishra, V. and Cherkauer, K. A. and Shukla, S.},
+title = {Assessment of drought due to historic climate variability and projected future climate change in the midwestern {U}nited {S}tates},
+journal = {Journal of Hydrometeorology},
+volume = {11},
+year = {2010},
+pages = {46-68},
+}
+
+@book{Mlad99,
+author = {Mladenoff, D. J. and Baker, W. L.},
+title = {Spatial modeling of forest landscape change: approaches and applications},
+publisher = {Cambridge University Press},
+year = {1999},
+address = {Cambridge, UK},
+}
+
+@article{Mlad04,
+author = {Mladenoff, D. J.},
+title = {{LANDIS} and forest landscape models},
+journal = {Ecological Modelling},
+volume = {180},
+year = {2004},
+pages = {7-19},
+}
+
+@article{Moor06,
+author = {Moore, J. E. and Swihart, R. K.},
+title = {Nut selection by captive blue jays: importance of availability and implications for seed dispersal},
+journal = {The Condor},
+volume = {108},
+year = {2006},
+pages = {377-388},
+}
+
+@article{Moor07,
+author = {Moore, J. E. and McEuen, A. B. and Swihart, R. K. and Contreras, T. A. and Steele, M. A.},
+title = {Determinants of seed removal distance by scatter-hoarding rodents in deciduous forests},
+journal = {Ecology},
+volume = {88},
+year = {2007},
+pages = {2529-2540},
+}
+
+@article{Moor08,
+author = {Moore, J. E. and Swihart, R. K.},
+title = {Factors affecting the relationship between seed removal and seed mortality},
+journal = {Canadian Journal of Zoology},
+volume = {86},
+year = {2008},
+pages = {378-385},
+}
+
+@article{Morr08,
+author = {Morrissey, R. C. and Jacobs, D. F. and Seifert, J. R. and Fischer, B. C. and Kershaw, J. A.},
+title = {Competitive success of natural oak regeneration in clearcuts during the stem exclusion stage},
+journal = {Canadian Journal of Forest Research},
+volume = {38},
+year = {2008},
+pages = {1419-1430},
+}
+
+@article{Muno11,
+author = {Mu\~{n}oz, A. and Bonal, R.},
+title = {Linking seed dispersal to cache protection strategies},
+journal = {Journal of Ecology},
+volume = {99},
+year = {2011},
+pages = {1016-1025},
+}
+
+@book{Nati14,
+author = "{National Climatic Data Center}",
+title = {Monthly climatological summary, 2011-2012},
+publisher = {NOAA National Climatic Data Center},
+year = {2014},
+address = {Asheville, NC},
+url = {http://www.ncdc.noaa.gov}
+}
+
+@book{Nati15,
+author = "{National Drought Mitigation Center}",
+title = {Tabular data archive for {M}organ and {M}onroe counties in {I}ndiana},
+publisher = {National Drought Mitigation Center},
+year = {2015},
+address = {Lincoln, NE},
+url = {http://www.droughtmonitor.unl.edu}
+}
+
+@article{Neff68,
+author = {Neff, D. J.},
+title = {The pellet-group count technique for big game trend, census, and distribution: a review},
+journal = {Journal of Wildlife Management},
+volume = {32},
+year = {1968},
+pages = {597-614},
+}
+
+@article{Newm88,
+author = {Newman, J. A. and Recer, G. M. and Zwicker, S. M. and Caraco, T.},
+title = {Effects of predation hazard on foraging "constraints": patch use strategies in grey squirrels},
+journal = {Oikos},
+volume = {53},
+year = {1988},
+pages = {93-97},
+}
+
+@article{Niin06,
+author = {Niinemets, U. and Valladares, F.},
+title = {Tolerance to shade, drought, and waterlogging of temperate northern hemisphere trees and shrubs},
+journal = {Ecological Monographs},
+volume = {76},
+year = {2006},
+pages = {521-547},
+}
+
+@article{Nixo80,
+author = {Nixon, C. M. and McClain, M. W. and Donohoe, R. W.},
+title = {Effects of clear-cutting on gray squirrels},
+journal = {Journal of Wildlife Management},
+volume = {44},
+year = {1980},
+pages = {403-412},
+}
+
+@article{Nowa96,
+author = {Nowack, D. J.},
+title = {Notes: estimating leaf area and leaf biomass of open-grown deciduous urban trees},
+journal = {Forest Science},
+volume = {42},
+year = {1996},
+pages = {504-507},
+}
+
+@article{Nowa08,
+author = {Nowacki, G. J. and Abrams, M. D.},
+title = {The demise of fire and “mesophication” of forests in the eastern {U}nited {S}tates},
+journal = {BioScience},
+volume = {58},
+year = {2008},
+pages = {123-138},
+}
+
+@article{Oliv81,
+author = {Oliver, C. D.},
+title = {Forest development in {N}orth {A}merica following major disturbances},
+journal = {Forest Ecology and Management},
+volume = {3},
+year = {1981},
+pages = {153-168},
+}
+
+@book{Oliv96,
+author = {Oliver, C. D. and Larson, B. C.},
+title = {Forest stand dynamics},
+publisher = {Wiley},
+year = {1996},
+address = {New York, NY},
+pagetotal = {520}
+}
+
+@inproceedings{Olso15a,
+author = {Olson, Z. H. and MacGowan, B. J.},
+title = {Survival of timber rattlesnakes: investigating individual, environmental, and ecological effects},
+publisher = {Paper presented at the 75th Midwest Fish and Wildlife Conference, 9 February 2015, Indianapolis, IN},
+year = {2015},
+}
+
+@article{Olso15b,
+author = {Olson, M. G. and Wolf, A. J. and Jensen, R. G.},
+title = {Influence of forest management on acorn production in the southeastern {M}issouri {O}zarks: early results of a long-term ecosystem experiment},
+journal = {Open Journal of Forestry},
+volume = {5},
+year = {2015},
+pages = {568},
+}
+
+@article{Ostf96,
+author = {Ostfeld, R. S. and Jones, C. G. and Wolff, J. O.},
+title = {Of mice and mast},
+journal = {BioScience},
+volume = {46},
+year = {1996},
+pages = {323-330},
+}
+
+@article{Ostf97,
+author = {Ostfeld, R. S. and Manson, R. H. and Canham, C. D.},
+title = {Effects of rodents on survival of tree seeds and seedlings invading old fields},
+journal = {Ecology},
+volume = {78},
+year = {1997},
+pages = {1531-1542},
+}
+
+@article{Paca93,
+author = {Pacala, S. W. and Canham, C. D. and Silander, J. A.},
+title = {Forest models defined by field measurements. {I}. {T}he design of a northeastern forest simulator},
+journal = {Canadian Journal of Forest Research},
+volume = {23},
+year = {1993},
+pages = {1980-1988},
+}
+
+@article{Paca96,
+author = {Pacala, S. W. and Canham, C. D. and Saponara, J. and Silander, J. A. and Kobe, R. K. and Ribbens, E.},
+title = {Forest models defined by field measurements. {II}. {E}stimation, error analysis, and dynamics},
+journal = {Ecological Monographs},
+volume = {66},
+year = {1996},
+pages = {1-43},
+}
+
+@article{Paqu06,
+author = {Paquette, A. and Bouchard, A. and Cogliastro, A.},
+title = {Successful under-planting of red oak and black cherry in early-successional deciduous shelterwoods of {N}orth {A}merica},
+journal = {Annals of Forest Science},
+volume = {63},
+year = {2006},
+pages = {823-831},
+}
+
+@article{Park08,
+author = {Parker, W. C. and Dey, D. C.},
+title = {Influence of overstory density on ecophysiology of red oak (\textit{{Q}uercus rubra}) and sugar maple (\textit{{A}cer saccharum}) seedlings in central {O}ntario shelterwoods},
+journal = {Tree Physiology},
+volume = {28},
+year = {2008},
+pages = {797-804},
+}
+
+@article{Pere08,
+author = {P\'{e}rez-Ramos, I. M. and Urbieta, I. R. and Mara\~{n}\'{o}n, T. and Zavala, M. A. and Kobe, R. K.},
+title = {Seed removal in two coexisting oak species: ecological consequences of seed size, plant cover and seed drop timing},
+journal = {Oikos},
+volume = {117},
+year = {2008},
+pages = {1386-1396},
+}
+
+@techreport{Perr04,
+author = {Perry, R. W. and Thill, R. E. and Tappe, P. A. and Peitz, D. G.},
+title = {Initial response of individual soft-mast producing plants to different forest regeneration strategies in the {O}uachita {M}ountains},
+number = {SRS-GTR-74},
+type = {Gen. Tech. Rep.},
+institution = {U.S. Department of Agriculture, Forest Service, Southern Research Station},
+address = {Asheville, NC},
+year = {2004},
+pages = {60-70},
+}
+
+@article{Pitk99,
+author = {Pitk{\"a}nen, A. and Huttunen, P.},
+title = {A 1300-year forest-fire history at a site in eastern {F}inland based on charcoal and pollen records in laminated lake sediment},
+journal = {The Holocene},
+volume = {9},
+year = {1999},
+pages = {311-320},
+}
+
+@inproceedings{Plum03,
+author = {Plummer, M.},
+title = {{JAGS}: a program for analyasis of {B}ayesian graphical models using {G}ibbs sampling},
+booktitle = {Proceedings of the 3rd International Workship on Distributed Statistical Computing},
+year = {2003},
+publisher = {Technische Universit{\"a}t Wien},
+address = {Vienna, Austria},
+}
+
+@article{Pric94,
+author = {Price, O. and Bowman, D.},
+title = {Fire-stick forestry: a matrix model in support of skillful fire management of \textit{{C}allitris intratropica} {RT} {B}aker by north {A}ustralian aborigines},
+journal = {Journal of Biogeography},
+volume = {21},
+year = {1994},
+pages = {573-580},
+}
+
+@article{Quer06,
+author = {Quero, J. L. and Villar, R. and Mara{\~n}{\'o}n, T. and Zamora, R.},
+title = {Interactions of drought and shade effects on seedlings of four \textit{{Q}uercus} species: physiological and structural leaf responses},
+journal = {New Phytologist},
+volume = {170},
+year = {2006},
+pages = {819-834},
+}
+
+@article{Quin00,
+author = {Quinby, P. A.},
+title = {First-year impacts of shelterwood logging on understory vegetation in an old-growth pine stand in central {O}ntario, {C}anada},
+journal = {Environmental Conservation},
+volume = {27},
+year = {2000},
+pages = {229-241},
+}
+
+@manual{Rcor15,
+author = {{R Development Core Team}},
+title = {R: A Language and Environment for Statistical Computing},
+organization = {R Foundation for Statistical Computing},
+address = {Vienna, Austria},
+year = {2015},
+url = {http://www.R-project.org/},
+}
+
+@techreport{Rath08,
+author = {Rathfon, R. A. and Lichti, N. I. and Swihart, R. K.},
+title = {Disking and mid- and understory removal following an above average acorn crp in three mature oak forests in southern {I}ndiana},
+year = {2008},
+type = {Gen. Tech. Rep.},
+number = {NRS-P-24},
+institution = {U.S. Department of Agriculture, Forest Service, Northern Research Station},
+address = {Newtown Square, PA},
+pages = {59-69},
+}
+
+@article{Reyn06,
+author = {Reynolds-Hogland, M. J. and Mitchell, M. S. and Powell, R. A.},
+title = {Spatio-temporal availability of soft mast in clearcuts in the {S}outhern {A}ppalachians},
+journal = {Forest Ecology and Management},
+volume = {237},
+year = {2006},
+pages = {103-114},
+}
+
+@article{Rich13,
+author = {Richardson, K. B. and Lichti, N. I. and Swihart, R. K.},
+title = {Acorn-foraging preferences of four species of free-ranging avian seed predators in eastern decidious forests},
+journal = {The Condor},
+volume = {115},
+year = {2013},
+pages = {863-873},
+}
+
+@article{Roge98,
+author = {Rogers, R. and Johnson, P. S.},
+title = {Approaches to modeling natural regeneration in oak-dominated forests},
+journal = {Forest Ecology and Management},
+volume = {106},
+year = {1998},
+pages = {45-54},
+}
+
+@article{Roon03,
+author = {Rooney, T. P. and Waller, D. M.},
+title = {Direct and indirect effects of white-tailed deer in forest ecosystems},
+journal = {Forest Ecology and Management},
+volume = {181},
+year = {2003},
+pages = {165-176},
+}
+
+@article{Rose12,
+author = {Rose, A. K. and Greenberg, C. H. and Fearer, T. M.},
+title = {Acorn production prediction models for five common oak species of the eastern {U}nited {S}tates},
+journal = {Journal of Wildlife Management},
+volume = {76},
+year = {2012},
+pages = {750-758},
+}
+
+@article{Ross05,
+author = {Rossell, C. R. and Gorsira, B. and Patch, S.},
+title = {Effects of white-tailed deer on vegetation structure and woody seedling composition in three forest types on the {P}iedmont {P}lateau},
+journal = {Forest Ecology and Management},
+volume = {210},
+year = {2005},
+pages = {415-424},
+}
+
+@article{Royo08,
+author = {Royo, A. A. and Carson, W. P.},
+title = {Direct and indirect effects of a dense understory on tree seedling recruitment in temperate forests: habitat mediated predation versus competition},
+journal = {Canadian Journal of Forest Research},
+volume = {38},
+year = {2008},
+pages = {1634-1645},
+}
+
+@article{Russ01,
+author = {Russell, F. and Leland, D. B. and Fowler, N. L.},
+title = {Effects of white-tailed deer (\textit{{O}docoileus virginianus}) on plants, plant populations and communities: a review},
+journal = {American Midland Naturalist},
+volume = {146},
+year = {2001},
+pages = {1-26},
+}
+
+@techreport{Sand72,
+author = {Sander, I. L.},
+title = {Size of advance reproduction: key to growth following harvest cutting},
+year = {1972},
+type = {Res. Pap.},
+number = {NC-79},
+institution = {U.S. Department of Agriculture, Forest Service},
+}
+
+@article{Saun91,
+author = {Saunders, D. A. and Hobbs, R. J. and Margules, C. R.},
+title = {Biological consequences of ecosystem fragmentation: a review},
+journal = {Conservation Biology},
+volume = {5},
+year = {1991},
+pages = {18-32},
+}
+
+@techreport{Saun13,
+author = {Saunders, M. and Arseneault, J.},
+title = {Pre-treatment analysis of woody vegetation composition and structure on the {H}ardwood {E}cosystem {E}xperiment research units},
+year = {2013},
+type = {Gen. Tech. Rep.},
+number = {NRS-P-108},
+institution = {U.S. Department of Agriculture, Forest Service, Northern Forest Experiment Station},
+address = {Newtown Square, PA},
+}
+
+@article{Schl93,
+author = {Schlesinger, R. C. and Sander, I. L. and Davidson, K. R.},
+title = {Oak regeneration potential increased by shelterwood treatments},
+journal = {Northern Journal of Applied Forestry},
+volume = {10},
+year = {1993},
+pages = {149-153},
+}
+
+@article{Schm08,
+author = {Schmidt, K. A. and Ostfeld, R. S.},
+title = {Numerical and behavioral effects within a pulse-driven system: consequences for shared prey},
+journal = {Ecology},
+volume = {89},
+year = {2008},
+pages = {635-646},
+}
+
+@article{Schn02,
+author = {Schnurr, J. L. and Ostfeld, R. S. and Canham, C. D.},
+title = {Direct and indirect effects of masting on rodent populations and tree seed survival},
+journal = {Oikos},
+volume = {96},
+year = {2002},
+pages = {402-410},
+}
+
+@article{Schu10,
+author = {Schupp, E. W. and Jordano, P. and Gomez J. M.},
+title = {Seed dispersal effectiveness: a conceptual review},
+journal = {New Phytologist},
+volume = {188},
+year = {2010},
+pages = {333-353},
+}
+
+@article{Seag01,
+author = {Seagle, S. W. and Liang, S. Y.},
+title = {Application of a forest gap model for prediction of browsing effects on riparian forest succession},
+journal = {Ecological Modelling},
+volume = {144},
+year = {2001},
+pages = {213-229},
+}
+
+@techreport{Sher02,
+author = {Sherrif, S. L.},
+title = {Missouri {O}zark {F}orest {E}cosystem {P}roject: the experiment},
+year = {2002},
+type = {Gen. Tech. Rep.},
+number = {NC-227},
+institution = {U.S. Department of Agriculture, Forest Service, North Central Forest Experiment Station},
+address = {St. Paul, MN},
+}
+
+@article{Shug77,
+author = {Shugart, H. H. and West, D. C.},
+title = {Development of an {A}ppalachian deciduous forest succession model and its application to assessment of the impact of the chestnut blight},
+journal = {Journal of Environmental Management},
+volume = {5},
+year = {1977},
+pages = {161-179},
+}
+
+@article{Smal01,
+author = {Smallwood, P. D. and Steele, M. A. and Faeth, S. H.},
+title = {The ultimate basis of the caching preferences of rodents, and the oak-dispersal syndrome: tannins, insects, and seed germination},
+journal = {American Zoologist},
+volume = {41},
+year = {2001},
+pages = {840-851},
+}
+
+@inproceedings{Soma13,
+author = {Somarriba, E. and Quesada, F. and Mallek, M.},
+title = {Shade {M}otion simulation of shade patterns of trees on a plot},
+publisher = {Paper presented at Cocoa Agroforestry: Sustainability and Environment, Yaounde, Cameroon},
+year = {2013},
+url = {http://hdl.handle.net/123456789/298},
+}
+
+@article{Sork84,
+author = {Sork, V. L.},
+title = {Examination of seed dispersal and survival in red oak, \textit{{Q}uercus rubra} ({F}agaceae), using metal-tagged acorns},
+journal = {Ecology},
+volume = {65},
+year = {1984},
+pages = {1020-1022},
+}
+
+@article{Sork93a,
+author = {Sork, V. L.},
+title = {Evolutionary ecology of mast-seeding in temperate and tropical oaks (\textit{{Q}uercus} spp.)},
+journal = {Vegetatio},
+volume = {107},
+year = {1993},
+pages = {133-147},
+}
+
+@article{Sork93b,
+author = {Sork, V. L. and Bramble, J. and Sexton, O.},
+title = {Ecology of mast-fruiting in three species of {N}orth {A}merican oaks},
+journal = {Ecology},
+volume = {74},
+year = {1993},
+pages = {528-541},
+}
+
+@book{Spee10,
+author = {Speer, J. H.},
+title = {Fundamentals of tree-ring research},
+publisher = {University of Arizona Press},
+year = {2010},
+address = {Tucson, AZ},
+}
+
+@article{Stan98,
+author = {Stange, E. E. and Shea, K. L.},
+title = {Effects of deer browsing, fabric mats, and tree shelters on \textit{{Q}uercus rubra} seedlings},
+journal = {Restoration Ecology},
+volume = {6},
+year = {1998},
+pages = {29-34},
+}
+
+@article{Steel96,
+author = {Steele, M. A. and Hadj-Chikh, L. Z. and Hazeltine, J.},
+title = {Caching and feeding decisions by \textit{{S}ciurus carolinensis}: responses to weevil-infested acorns},
+journal = {Journal of Mammalogy},
+volume = {77},
+year = {1996},
+pages = {305-314},
+}
+
+@article{Steel01,
+author = {Steele, M. A. and Turner, G. and Smallwood, P. D. and Wolff, J. O. and Radillo, J.},
+title = {Cache management by small mammals: experimental evidence for the significant of acorn-embryo excision},
+journal = {Journal of Mammalogy},
+volume = {82},
+year = {2001},
+pages = {35-42},
+}
+
+@inbook{Steel02,
+author = {Steele, M. A. and Smallwood, P. D.},
+booktitle = {Oak forest ecosystems: ecology and management for wildlife},
+editor = {McShea, W. J. and Healy, W.},
+title = {Acorn dispersal by birds and mammals},
+publisher = {Johns Hopkins University Press},
+year = {2002},
+address = {Baltimore, MD},
+}
+
+@article{Steel06,
+author = {Steele, M. A. and Manierre, S. and Genna, T. and Contreras, T. A. and Smallwood, P. D. and Pereira, M. E.},
+title = {The innate basis of food-hoarding decisions in gray squirrels: evidence for behavioral adaptations to the oaks},
+journal = {Animal Behavior},
+volume = {71},
+year = {2006},
+pages = {155-160},
+}
+
+@article{Steel14,
+author = {Steele, M. A. and Contreras, T. A. and Hadj-Chikh L. Z. and Agosta, S. J. and Smallwood, P. D. and Tomlinson, C. N.},
+title = {Do scatter hoarders trade off increased predation risks for lower rates of cache pilferage?},
+journal = {Behavioral Ecology},
+volume = {},
+year = {2013},
+pages = {art107},
+}
+
+@article{Steel15,
+author = {Steele, M. A. and Rompre, G. and Stratford, J. A. and Zhang, H. and Suchocki, M. and Marino, S.},
+title = {Scatterhoarding rodents favor higher predation risks for cache sites: the potential for predators to influence the dispersal process},
+journal = {Integrative Zoology},
+volume = {10},
+year = {2015},
+pages = {257-266},
+}
+
+@techreport{Stei95,
+author = {Steiner, K. C.},
+title = {Autumn predation of northern red oak seed crops},
+year = {1995},
+type = {Gen. Tech. Rep.},
+number = {NE-197},
+institution = {U.S. Department of Agriculture, Forest Service, Northeastern Forest Experiment Station},
+address = {Radnor, PA},
+pages = {489-494},
+}
+
+@article{Suar08,
+author = {Suarez, M. L. and Kitzberger, T.},
+title = {Recruitment patterns following a severe drought: long-term compositional shifts in {P}atagonian forests},
+journal = {Canadian Journal of Forest Research},
+volume = {38},
+year = {2008},
+pages = {3002-3010},
+}
+
+@article{Summ11,
+author = {Summerville, K. S.},
+title = {Managing the forest for more than the trees: effects of experimental timber harvest on forest {L}epidoptera},
+journal = {Ecological Applications},
+volume = {21},
+year = {2011},
+pages = {806-816},
+}
+
+@article{Sund15,
+author = {Sundaram, M. and Willoughby, J. R. and Lichti, N. I. and Steele, M. A. and Swihart, R. K.},
+title = {Segregating the effects of seed traits and common ancestry of hardwood trees on {E}astern gray squirrel foraging decisions},
+journal = {PLoS One},
+volume = {10},
+year = {2015},
+pages = {e0130942},
+}
+
+@article{Summ02,
+author = {Summerville, K. S. and Crist, T. O.},
+title = {Effects of timber harvest on forest {L}epidoptera},
+journal = {Ecological Applications},
+volume = {12},
+year = {2002},
+pages = {820-835},
+}
+
+@mastersthesis{Swai13,
+author = {Swaim, J. T.},
+title = {Stand development and competitive ability of oak (\textit{Quercus} spp.) following silvicultural clearcutting on the {H}oosier {N}ational {F}orest},
+school = {Purdue University},
+year = {2013},
+address = {West Lafayette, IN},
+pages = {},
+}
+
+@inbook{Swan88,
+author = {Swank, W. T. and Crossley, D. A.},
+booktitle = {Forest hydrology and ecology at {C}oweeta},
+editor = {Swank, W. T. and Crossley, D. A.},
+title = {Introduction and site description},
+publisher = {Springer Verlag},
+year = {1988},
+address = {New York, NY},
+pages = {3-16},
+}
+
+@article{Swih01,
+author = {Swihart, R. K. and Bryant, J. P.},
+title = {Importance of biogeography and ontogeny of woody plants in winter herbivory by mammals},
+journal = {Journal of Mammalogy},
+volume = {82},
+year = {2001},
+pages = {1-21},
+}
+
+@article{Tcha01,
+author = {Tchabovsky, A. V. and Krasnov, B. R. and Khokhlova, I. S. and Shenbrot, G. I.},
+title = {The effect of vegetation cover on vigiliance and foraging tactics in the fat sand rat \textit{{P}sammomys obesus}},
+journal = {Journal of Ethology},
+volume = {19},
+year = {2001},
+pages = {105-113},
+}
+
+@manual{Thie14,
+author = {Thiele, J. C.},
+title = {{RN}et{L}ogo: provides an interface to the agent-based modelling platform},
+year = {2014},
+note = {R package version 1.0-1},
+}
+
+@article{Thom00,
+author = {Thompson, R. S. and Anderson, K. H.},
+title = {Biomes of western {N}orth {A}merica at 18,000, 6000, and 0 \textsuperscript{14}{C} yr \textsc{BP} reconstructed from pollen and packrat midden data},
+journal = {Journal of Biogeography},
+volume = {27},
+year = {2000},
+pages = {555-584},
+}
+
+@article{Tier85,
+author = {Tierson, W. C. and Mattfeld, G. F. and Sage, R. W.},
+title = {Seasonal movements and home ranges of white-tailed deer in the {A}dirondacks},
+journal = {Journal of Wildlife Management},
+volume = {49},
+year = {1985},
+pages = {760-769},
+}
+
+@techreport{True53,
+author = {True, R. P.},
+title = {Studies on sprout reproduction of yellow poplar as related to decay},
+year = {1953},
+type = {Curr. Rep.},
+number = {3},
+institution = {West Virginia University Agricultural Experiment Station},
+address = {Morgantown, WV},
+}
+
+@book{Urba90,
+author = {Urban, D. L.},
+title = {A versitile model to simulate forest pattern: a user's guide to {ZELIG} version 1.0},
+publisher = {University of Virginia},
+year = {1990},
+address = {Charlottesville, VA},
+}
+
+@inbook{Urba92,
+author = {Urban, D. L. and Shugart, H. H.},
+booktitle = {Plant succession, theory, and prediction},
+editor = {Glenn-Lewin, D. C. and Peet, R. K. and Veblin, T. T.},
+title = {Individual based models of forest succession},
+publisher = {Chapman and Hall},
+year = {1992},
+address = {London, UK},
+pages = {249-292},
+}
+
+@article{Vall08,
+author = {Valladares, F. and Niinemets, U.},
+title = {Shade tolerance, a key plant feature of complex nature and consequences},
+journal = {Annual Review of Ecology, Evolution, and Systematics},
+volume = {39},
+year = {2008},
+pages = {237-257},
+}
+
+@article{vanh97,
+author = {van Hees, A. F. M.},
+title = {Growth and morphology of pedunculate oak (\textit{{Q}uercus robur} {L}) and beech (\textit{{F}agus sylvatica} {L}) seedlings in relation to shading and drought},
+journal = {Annals of Forest Science},
+volume = {54},
+year = {1997},
+pages = {9-18},
+}
+
+@book{Vand90,
+author = {Vander Wall, S. B.},
+title = {Food hoarding in animals},
+publisher = {University of Chicago Press},
+year = {1990},
+address = {Chicago, IL},
+pagetotal = {453}
+}
+
+@article{Vand01,
+author = {Vander Wall, S. B.},
+title = {The evolutionary ecology of nut dispersal},
+journal = {Botanical Review},
+volume = {67},
+year = {2001},
+pages = {74-117},
+}
+
+@article{Vand05,
+author = {Vander Wall, S. B. and Kuhn, K. M. and Beck, M. J.},
+title = {Seed removal, seed predation, and secondary dispersal},
+journal = {Ecology},
+volume = {86},
+year = {2005},
+pages = {801-806},
+}
+
+@article{Vand10,
+author = {Vander Wall, S. B.},
+title = {How plants manipulate the scatter-hoarding of seed-dispersing animals},
+journal = {Philosophical Transactions of the Royal Society B},
+volume = {365},
+year = {2010},
+pages = {989-997},
+}
+
+@article{Verc98,
+author = {VerCauteren, K. C. and Hyngstrom, S. E.},
+title = {Effects of agricultural activities and hunting on home ranges of female white-tailed deer},
+journal = {Journal of Wildlife Management},
+volume = {62},
+year = {1998},
+pages = {280-285},
+}
+
+@inbook{Verc05,
+author = {VerCauteren, K. C. and Dolbeer, R. A. and Gese, E. M.},
+booktitle = {Techniques for wildlife investigations and management},
+editor = {Braun, C. E.},
+title = {Identification and management of wildlife damage},
+publisher = {The Wildlife Society},
+year = {2005},
+address = {Bethesda, MD},
+pages = {740-778},
+}
+
+@mastersthesis{Verm53,
+author = {Verme, L. J.},
+title = {Production and utilization of acorns in {C}linton {C}ounty, {M}ichigan},
+school = {Michigan State University},
+year = {1953},
+address = {East Lansing, MI},
+pagetotal = {77},
+}
+
+@article{Voel08,
+author = {voelker, S. L. and Muzika, R. and Guyette, R. P.},
+title = {Individual tree and stand level influences on the growth, vigor, and decline of red oaks in the {O}zarks},
+journal = {Forest Science},
+volume = {54},
+year = {2008},
+pages = {8-20},
+}
+
+@article{Wake09,
+author = {Wakeland, B. and Swihart, R. K.},
+title = {Ratings of white-tailed deer preferences for woody browse in {I}ndiana},
+journal = {Proceedings of the Indiana Academy of Science},
+volume = {118},
+year = {2009},
+pages = {96-101},
+}
+
+@article{Wall97,
+author = {Waller, D. M. and Alverson, W. S.},
+title = {The white-tailed deer: a keystone herbivore},
+journal = {Wildlife Society Bulletin},
+volume = {25},
+year = {1997},
+pages = {217-226},
+}
+
+@article{Weig02,
+author = {Weigel, D. R. and Peng, C. Y.},
+title = {Predicting stump sprouting and competitive success of five oak species in southern {I}ndiana},
+journal = {Canadian Journal of Forest Research},
+volume = {32},
+year = {2002},
+pages = {703-712},
+}
+
+@techreport{Wend75,
+author = {Wendel, G. W.},
+title = {Stump sprout growth and quality of several {A}ppalachian hardwood species after clearcutting},
+year = {1975},
+type = {Res. Pap.},
+number = {NE-329},
+institution = {U.S. Department of Agriculture, Forest Service, Northeastern Forest Experiment Station},
+address = {Upper Darby, PA},
+}
+
+@article{Whit89,
+author = {Whitmore, T. C.},
+title = {Canopy gaps and the two major groups of forest trees},
+journal = {Ecology},
+volume = {70},
+year = {1989},
+pages = {536-538},
+}
+
+@article{Whit05,
+author = {Whittaker, J. and Whitehead, C. and Somers, M.},
+title = {The neglog transformation and quantile regression for the analysis of a large credit scoring database},
+journal = {Journal of the Royal Statistical Society Series C},
+volume = {54},
+year = {2005},
+pages = {863-878},
+}
+
+@manual{Wile99,
+author = {Wilensky, U.},
+title = {{N}et{L}ogo},
+organization = {Northwestern University Center for Connected Learning and Computer-Based Modeling},
+year = {1999},
+address = {Evanston, IL},
+url = {http://ccl. northwestern.edu/netlogo},
+}
+
+@article{Will85,
+author = {Williamson, S. J. and Hirth, D. H.},
+title = {An evaluation of edge use by white-tailed deer},
+journal = {Wildlife Society Bulletin},
+volume = {13},
+year = {1985},
+pages = {252-257},
+}
+
+@inbook{Wirt08,
+author = {Wirth, R. and Meyer, S. T. and Leal, I. R. and Tabarelli, M.},
+booktitle = {Progress in botany},
+editor = {L{\"u}ttge, U. and Matyssek, R. and Ramos, C. and Miguel, F.},
+title = {Plant herbivore interactions at the forest edge},
+publisher = {Springer},
+year = {2008},
+address = {Berlin, Germany},
+pages = {423-448},
+}
+
+@article{Wolf96,
+author = {Wolff, J. O.},
+title = {Population fluctuations of mast-eating rodents are correlated with production of acorns},
+journal = {Journal of Mammalogy},
+volume = {77},
+year = {1996},
+pages = {850-856},
+}
+
+@article{Wood05,
+author = {Wood, M. D.},
+title = {Tannin and lipid content of acorns in scatterhoards and larderhoards},
+journal = {Northeastern Naturalist},
+volume = {12},
+year = {2006},
+pages = {463-472},
+}
+
+@article{Xiao15,
+author = {Xiao, Z. and Krebs, C. J.},
+title = {Modeling the costs and benefits of seed scatterhoarding to plants},
+journal = {Ecosphere},
+volume = {6},
+year = {2015},
+pages = {art53},
+}
+
+@article{Xu08,
+author = {Xu, C. and Gertner, G. Z.},
+title = {Uncertainty and sensitivity analysis for models with correlated parameters},
+journal = {Reliability Engineering and System Safety},
+volume = {93},
+year = {2008},
+pages = {1563-1573},
+}
+
+@article{Yahn82,
+author = {Yahner, R. H.},
+title = {Microhabitat use by small mammals in farmstead shelterbelts},
+journal = {Journal of Mammalogy},
+volume = {63},
+year = {1982},
+pages = {440-445},
+}
+
+@article{Zamo14,
+author = {Zamora, R. and Matias, L.},
+title = {Seed dispersers, seed predators, and browsers act synergistically as biotic filters in a mosiac landscape},
+journal = {PLoS One},
+volume = {9},
+year = {2014},
+pages = {e107385},
+}
+
+@article{Zhen00,
+author = {Zheng, D. and Chen, J. and Song, B. and Xu, M. and Sneed, P. and Jenson, R.},
+title = {Effects of silvicultural treatments on summer forest microclimates in southeastern {M}issouri {O}zarks},
+journal = {Climate Research},
+volume = {15},
+year = {2000},
+pages = {45-59},
+}
+
+@article{Zuid09,
+author = {Zuidema, P. A. and Brienen, R. J. W. and During, H. J. and Guneralp, B.},
+title = {Do persistantly fast-growing juviniles contribute disproportionately to population growth? {A} new analysis tool for matrix models and its application to rainforest trees},
+journal = {American Naturalist},
+volume = {174},
+year = {2009},
+pages = {709-719},
+}
+
+@article{Zuid10,
+author = {Zuidema, P. A. and Jongejans, E. and Chien, P. D. and During, H. J. and Schieving, F.},
+title = {Integral projection models for trees: a new parameterization method and validation of model output},
+journal = {Journal of Ecology},
+volume = {98},
+year = {2010},
+pages = {345-355},
+}
diff --git a/tables/tab2-1.tex b/tables/tab2-1.tex
new file mode 100644
index 0000000..6214a41
--- /dev/null
+++ b/tables/tab2-1.tex
@@ -0,0 +1,23 @@
+
+\begin{table}[htbp]
+ \centering
+ \caption{Parameter estimates for covariate effects on probability of acorn removal from exclosure boxes (squirrels were excluded from exclosure M) and probability of the unremoved acorns remaining uneaten. Results are based on separate logistic regression models.}
+ \begin{tabular}{rcccc}
+ \toprule
+ \multicolumn{1}{l}{\textbf{Parameter}} & \textbf{Estimate} & \textbf{SE} & \textit{\textbf{Z}} & \textit{\textbf{p}} \\
+ \midrule
+ \multicolumn{1}{l}{\textit{Removal Probability}} & & & & \\
+ \multicolumn{1}{l}{Intercept} & 2.39 & 0.07 & 34.91 & \textless 0.01 \\
+ \multicolumn{1}{l}{White oak} & -1.04 & 0.06 & -16.33 & \textless 0.01 \\
+ \multicolumn{1}{l}{Shelterwood} & 0.51 & 0.06 & 8.00 & \textless 0.01 \\
+ \multicolumn{1}{l}{Exclosure M} & -0.80 & 0.06 & -12.36 & \textless 0.01 \\
+ & & & & \\
+ \multicolumn{1}{l}{\textit{Survival Probability}} & & & & \\
+ \multicolumn{1}{l}{Intercept} & -1.30 & 0.15 & -8.73 & \textless 0.01 \\
+ \multicolumn{1}{l}{White oak} & -1.05 & 0.13 & -7.93 & \textless 0.01 \\
+ \multicolumn{1}{l}{Shelterwood} & 0.46 & 0.13 & 3.63 & \textless 0.01 \\
+ \multicolumn{1}{l}{Exclosure M} & 1.62 & 0.15 & 10.48 & \textless 0.01 \\
+ \bottomrule
+ \end{tabular}%
+ %\label{tab:addlabel}%
+\end{table}% \ No newline at end of file
diff --git a/tables/tab2-2.tex b/tables/tab2-2.tex
new file mode 100644
index 0000000..9288e4a
--- /dev/null
+++ b/tables/tab2-2.tex
@@ -0,0 +1,18 @@
+
+\begin{table}[htbp]
+ \centering
+ \caption{Parameter estimates for effects of microsite type (relative to open forest floor) on acorn survival probability, based on logistic regression.}
+ \begin{tabular}{rccc}
+ \toprule
+ \multicolumn{1}{l}{\textbf{Parameter}} & \textbf{Estimate} & \textbf{SE} & \textit{\textbf{p}} \\
+ \midrule
+ \multicolumn{1}{l}{Intercept} & -1.04 & 0.12 & \textless 0.01 \\
+ \multicolumn{1}{l}{Base of tree} & -0.69 & 0.25 & \textless 0.01 \\
+ \multicolumn{1}{l}{Bush/dense vegetation} & -2.18 & 1.51 & 0.03 \\
+ \multicolumn{1}{l}{Log/coarse woody debris} & -1.71 & 0.28 & \textless 0.01 \\
+ \multicolumn{1}{l}{Tree stump} & -1.23 & 0.89 & 0.10 \\
+ \multicolumn{1}{l}{Fallen treetop} & -1.53 & 0.88 & 0.03 \\
+ \bottomrule
+ \end{tabular}%
+ %\label{tab:addlabel}%
+\end{table}%
diff --git a/tables/tab3-1.tex b/tables/tab3-1.tex
new file mode 100644
index 0000000..efa7033
--- /dev/null
+++ b/tables/tab3-1.tex
@@ -0,0 +1,24 @@
+
+\begin{sidewaystable}[htbp]
+ \centering
+ \caption{Parameter estimates from hierarchical proportional odds logistic regression models of browse damage (index) by white-tailed deer, eastern cottontail rabbits, and insects on oak seedlings. The \textit{f} statistic is the fraction of the sampled posterior distribution with the same sign as the mean; this value reflects the level of certainty that the parameter estimate is positive or negative. Parameters with 95\% credible intervals that did not overlap zero are bolded.}
+ \begin{tabular}{rrrrrrrrrr}
+ \toprule
+ \multicolumn{1}{c}{\textbf{}} & \multicolumn{3}{c}{\textbf{Rabbit}} & \multicolumn{3}{c}{\textbf{Deer}} & \multicolumn{3}{c}{\textbf{Insects}} \\
+ \multicolumn{1}{l}{\textbf{Parameter}} & \multicolumn{1}{c}{\textbf{Mean}} & \multicolumn{1}{c}{\textbf{95\% CI}} & \multicolumn{1}{c}{\textit{\textbf{f}}} & \multicolumn{1}{c}{\textbf{Mean}} & \multicolumn{1}{c}{\textbf{95\% CI}} & \multicolumn{1}{c}{\textit{\textbf{f}}} & \multicolumn{1}{c}{\textbf{Mean}} & \multicolumn{1}{c}{\textbf{95\% CI}} & \multicolumn{1}{c}{\textit{\textbf{f}}} \\
+ \midrule
+ \multicolumn{1}{l}{\textit{Random Effects}} & & & & & & & & & \\
+ \multicolumn{1}{l}{Plot SD} & \multicolumn{1}{c}{1.10} & \multicolumn{1}{c}{(0.02, 2.87)} & \multicolumn{1}{c}{} & \multicolumn{1}{c}{0.32} & \multicolumn{1}{c}{(0.01, 0.82)} & \multicolumn{1}{c}{} & \multicolumn{1}{c}{1.11} & \multicolumn{1}{c}{(0.65, 1.87)} & \multicolumn{1}{c}{} \\
+ \multicolumn{1}{l}{Subplot SD} & \multicolumn{1}{c}{2.09} & \multicolumn{1}{c}{(1.13, 3.28)} & \multicolumn{1}{c}{} & \multicolumn{1}{c}{0.39} & \multicolumn{1}{c}{(0.05, 0.76)} & \multicolumn{1}{c}{} & \multicolumn{1}{c}{0.44} & \multicolumn{1}{c}{(0.27, 0.64)} & \multicolumn{1}{c}{} \\
+ \multicolumn{1}{l}{Seedling SD} & \multicolumn{1}{c}{1.62} & \multicolumn{1}{c}{(1.08, 2.27)} & \multicolumn{1}{c}{} & \multicolumn{1}{c}{0.81} & \multicolumn{1}{c}{(0.41, 1.17)} & \multicolumn{1}{c}{} & \multicolumn{1}{c}{0.51} & \multicolumn{1}{c}{(0.33, 0.65)} & \multicolumn{1}{c}{} \\
+ & \multicolumn{1}{c}{} & \multicolumn{1}{c}{} & \multicolumn{1}{c}{} & \multicolumn{1}{c}{} & \multicolumn{1}{c}{} & \multicolumn{1}{c}{} & \multicolumn{1}{c}{} & \multicolumn{1}{c}{} & \multicolumn{1}{c}{} \\
+ \multicolumn{1}{l}{\textit{Fixed Effects}} & \multicolumn{1}{c}{} & \multicolumn{1}{c}{} & \multicolumn{1}{c}{} & \multicolumn{1}{c}{} & \multicolumn{1}{c}{} & \multicolumn{1}{c}{} & \multicolumn{1}{c}{} & \multicolumn{1}{c}{} & \multicolumn{1}{c}{} \\
+ \multicolumn{1}{l}{Distance to edge} & \multicolumn{1}{c}{0.91} & \multicolumn{1}{c}{(-0.29, 2.27)} & \multicolumn{1}{c}{0.93} & \multicolumn{1}{c}{-0.17} & \multicolumn{1}{c}{(-0.53, 0.24)} & \multicolumn{1}{c}{0.81} & \multicolumn{1}{c}{-0.51} & \multicolumn{1}{c}{(-1.09, 0.12)} & \multicolumn{1}{c}{0.96} \\
+ \multicolumn{1}{l}{Competition index} & \multicolumn{1}{c}{\textbf{-0.81}} & \multicolumn{1}{c}{(-1.38, -0.28)} & \multicolumn{1}{c}{1.00} & \multicolumn{1}{c}{\textbf{-0.38}} & \multicolumn{1}{c}{(-0.61, -0.17)} & \multicolumn{1}{c}{1.00} & \multicolumn{1}{c}{0.01} & \multicolumn{1}{c}{(-0.06, 0.09)} & \multicolumn{1}{c}{0.60} \\
+ \multicolumn{1}{l}{Species (white oak)} & \multicolumn{1}{c}{0.58} & \multicolumn{1}{c}{(-0.18, 1.35)} & \multicolumn{1}{c}{0.93} & \multicolumn{1}{c}{\textbf{0.42}} & \multicolumn{1}{c}{(0.02, 0.83)} & \multicolumn{1}{c}{0.98} & \multicolumn{1}{c}{0.12} & \multicolumn{1}{c}{(-0.06, 0.30)} & \multicolumn{1}{c}{0.92} \\
+ \multicolumn{1}{l}{Height} & \multicolumn{1}{c}{\textbf{0.87}} & \multicolumn{1}{c}{(0.30, 1.52)} & \multicolumn{1}{c}{1.00} & \multicolumn{1}{c}{\textbf{1.79}} & \multicolumn{1}{c}{(1.42, 2.18)} & \multicolumn{1}{c}{1.00} & \multicolumn{1}{c}{\textbf{0.25}} & \multicolumn{1}{c}{(0.13, 0.38)} & \multicolumn{1}{c}{1.00} \\
+ \multicolumn{1}{l}{Height\textsuperscript{2}} & \multicolumn{1}{c}{\textbf{-1.20}} & \multicolumn{1}{c}{(-1.96, -0.50)} & \multicolumn{1}{c}{1.00} & \multicolumn{1}{c}{\textbf{-0.81}} & \multicolumn{1}{c}{(-1.10, -0.53)} & \multicolumn{1}{c}{1.00} & \multicolumn{1}{c}{-0.06} & \multicolumn{1}{c}{(-0.19, 0.06)} & \multicolumn{1}{c}{0.85} \\
+ \bottomrule
+ \end{tabular}%
+ %\label{tab:addlabel}%
+\end{sidewaystable}%
diff --git a/tables/tab4-1.tex b/tables/tab4-1.tex
new file mode 100644
index 0000000..4bfa829
--- /dev/null
+++ b/tables/tab4-1.tex
@@ -0,0 +1,27 @@
+% Table generated by Excel2LaTeX from sheet 'Sheet4'
+\setlength{\tabcolsep}{3pt}
+\begin{table}[htbp]
+ \centering
+ \caption{Parameter estimates from hierarchical logistic regression models of planted oak seedling survival over a four year period, for two oak species (black oak \textit{Q. velutina} and white oak \textit{Q. alba}). Parameters with 95\% credible intervals that did not overlap zero are bolded.}
+ \begin{tabular}{rcccc}
+ \toprule
+ & \multicolumn{2}{c}{\textbf{Black Oak}} & \multicolumn{2}{c}{\textbf{White oak}} \\
+ \multicolumn{1}{l}{\textbf{Parameter}} & \textbf{Estimate} & \textbf{95\% CI} & \textbf{Estimate} & \textbf{95\% CI} \\
+ \midrule
+ \multicolumn{1}{l}{Plot-level variation (SD)} & 0.40 & (0.06, 0.96) & 0.31 & (0.04, 0.77) \\
+ \multicolumn{1}{l}{Subplot-level variation (SD)} & 0.72 & (0.49, 1.00) & 0.65 & (0.42, 0.90) \\
+ \multicolumn{1}{l}{Inside clearcut harvest} & \textbf{-1.13} & (-2.07, -0.25) & -0.15 & (-0.94, 0.61) \\
+ \multicolumn{1}{l}{On clearcut edge} & -0.48 & (-1.60, 0.51) & -0.11 & (-1.01, 0.80) \\
+ \multicolumn{1}{l}{Inside shelterwood harvest} & 0.64 & (-0.51, 1.77) & 0.20 & (-0.80, 1.21) \\
+ \multicolumn{1}{l}{Aspect (southwestern)} & \textbf{-0.78} & ( -1.47, -0.02) & 0.20 & (-0.43, 0.82) \\
+ \multicolumn{1}{l}{Time (since planting)} & \textbf{0.57} & (0.39, 0.74) & \textbf{0.52} & (0.36, 0.69) \\
+ \multicolumn{1}{l}{Season (growing season)} & \textbf{-0.94} & (-1.23, -0.60) & \textbf{-1.33} & (-1.66, -0.99) \\
+ \multicolumn{1}{l}{Competition index} & \textbf{-0.34} & (-0.53, -0.16) & \textbf{-0.17} & (-0.34, -0.01) \\
+ \multicolumn{1}{l}{Browsed} & -0.43 & (-0.97, 0.15) & -0.40 & (-0.82, 0.03) \\
+ \multicolumn{1}{l}{Root collar diameter} & \textbf{0.27} & (0.11, 0.41) & \textbf{0.26} & (0.10, 0.40) \\
+ \multicolumn{1}{l}{Sprouting} & \textbf{-0.52} & (-1.05, 0.07) & \textbf{-0.91} & (-1.39, -0.41) \\
+ \bottomrule
+ \end{tabular}%
+ %\label{tab:addlabel}%
+\end{table}%
+\setlength{\tabcolsep}{6pt}
diff --git a/tables/tab4-2.tex b/tables/tab4-2.tex
new file mode 100644
index 0000000..38f8a9a
--- /dev/null
+++ b/tables/tab4-2.tex
@@ -0,0 +1,27 @@
+% Table generated by Excel2LaTeX from sheet 'Sheet5'
+\setlength{\tabcolsep}{3pt}
+\begin{table}[htbp]
+ \centering
+ \caption{Parameter estimates from hierarchical logistic regression models of planted oak seedling growth over a four year period, for two oak species (black oak \textit{Q. velutina} and white oak \textit{Q. alba}). Parameters with 95\% credible intervals that did not overlap zero are bolded.}
+ \begin{tabular}{rcccc}
+ \toprule
+ \textbf{} & \multicolumn{2}{c}{\textbf{Black Oak}} & \multicolumn{2}{c}{\textbf{White Oak}} \\
+ \multicolumn{1}{l}{\textbf{Parameter}} & \textbf{Estimate} & \textbf{95\% CI} & \textbf{Estimate} & \textbf{95\% CI} \\
+ \midrule
+ \multicolumn{1}{l}{Plot-level variation (SD)} & 1.82 & (0.35, 3.54) & 1.42 & (0.06, 3.62) \\
+ \multicolumn{1}{l}{Subplot-level variation (SD)} & 0.9 & (0.06, 2.22) & 2.22 & (0.93, 3.45) \\
+ \multicolumn{1}{l}{Seedling-level variation (SD)} & 0.24 & (0.01, 0.75) & 1.12 & (0.16, 2.77) \\
+ \multicolumn{1}{l}{Residual variation} & 7.45 & (7.09, 7.83) & 10.77 & (10.22, 11.36) \\
+ \multicolumn{1}{l}{Inside clearcut harvest} & 2.49 & (-1.02, 5.79) & \textbf{6.37} & (2.37, 9.70) \\
+ \multicolumn{1}{l}{On clearcut edge} & 0.27 & (-3.25, 3.81) & 1.25 & (-2.93, 4.99) \\
+ \multicolumn{1}{l}{Inside shelterwood harvest} & 0.06 & (-3.58, 3.84) & 1.25 & (-3.31, 5.61) \\
+ \multicolumn{1}{l}{Aspect (southwestern)} & 1.88 & (-0.77, 4.36) & 2.01 & (-0.73, 4.80) \\
+ \multicolumn{1}{l}{Time (since planting)} & \textbf{1.75} & (1.16, 2.35) & \textbf{4.87} & (4.00, 5.77) \\
+ \multicolumn{1}{l}{Competition index} & -0.16 & (-0.77, 0.44) & -0.33 & (-1.26, 0.70) \\
+ \multicolumn{1}{l}{Browsed} & \textbf{-3.29} & (-5.61, -1.07) & \textbf{-3.86} & (-6.12, -1.64) \\
+ \multicolumn{1}{l}{Resprout status (sprout)} & \textbf{-7.75} & (-10.41, -5.16) & \textbf{-8.59} & (-13.45, -3.56) \\
+ \bottomrule
+ \end{tabular}%
+ %\label{tab:addlabel}%
+\end{table}%
+\setlength{\tabcolsep}{6pt}
diff --git a/tables/tab5-1.tex b/tables/tab5-1.tex
new file mode 100644
index 0000000..365f010
--- /dev/null
+++ b/tables/tab5-1.tex
@@ -0,0 +1,19 @@
+
+\begin{table}[htbp]
+ \centering
+ \caption{Values used to initialize the forest stand in each model run. The initial forest included trees in two size categories (midstory and overstory) for each of the three tree categories in the model: oak (white oak \textit{Quercus alba} and black oak \textit{Q. velutina}), shade tolerant (sugar maple \textit{Acer saccharum}) and shade intolerant (tulip poplar \textit{Liriodendron tulipifera}). Each size $\times$ species category was defined by an initial density (per hectare) and diameter at breast height (dbh) distribution. Values are based on forest structure data collected pre-harvest from the Hardwood Ecosystem Experiment \citep{Saun13}.}
+ \begin{tabular}{rrccc}
+ \toprule
+ \textbf{} & \textbf{} & \textbf{} & \multicolumn{2}{c}{\textbf{dbh}} \\
+ \multicolumn{1}{l}{\textbf{Species}} & \textbf{Size Category} & \textbf{Trees ha\textsuperscript{-1}} & \textbf{Mean} & \textbf{SD} \\
+ \midrule
+ \multicolumn{1}{l}{Oaks (both)} & Midstory & 95 & 14.9 & 5 \\
+ & Overstory & 89 & 45.75 & 5 \\
+ \multicolumn{1}{l}{Sugar maple} & Midstory & 499 & 10.3 & 5 \\
+ & Overstory & 11 & 40.8 & 5 \\
+ \multicolumn{1}{l}{Tulip poplar} & Midstory & 163 & 14.9 & 5 \\
+ & Overstory & 9 & 45.07 & 5 \\
+ \bottomrule
+ \end{tabular}%
+ %\label{tab:addlabel}%
+\end{table}%
diff --git a/tables/tab5-2.tex b/tables/tab5-2.tex
new file mode 100644
index 0000000..ede2461
--- /dev/null
+++ b/tables/tab5-2.tex
@@ -0,0 +1,16 @@
+
+\begin{sidewaystable}[htbp]
+ \centering
+ \caption{Yearly mean acorn production parameter \textit{meanAcorn} (per m\textsuperscript{2} canopy area) by black (\textit{Quercus velutina}) and white (\textit{Q. alba}) oaks under three masting scenarios: low mast year, average mast year, and high mast year. Mast production by individual trees each year is randomly selected from an exponential distribution with mean equal to \textit{meanAcorn}. Total acorn production means by average-sized trees (canopy radius 5 m) are presented for comparison.}
+ \begin{tabular}{rcccccc}
+ \toprule
+ & \multicolumn{2}{c}{\textbf{Low Mast Year}} & \multicolumn{2}{c}{\textbf{Average Mast Year}} & \multicolumn{2}{c}{\textbf{High Mast Year}} \\
+ \midrule
+ \multicolumn{1}{l}{\textbf{Species}} & \textit{\textbf{meanAcorn}} & \textbf{Total} & \textit{\textbf{meanAcorn}} & \textbf{Total} & \textit{\textbf{meanAcorn}} & \textbf{Total} \\
+ \multicolumn{1}{l}{Black Oak} & 3.75 & 294 & 11.07 & 869 & 16.58 & 1303 \\
+ \multicolumn{1}{l}{White Oak} & 8.58 & 673 & 11.74 & 922 & 22.34 & 1754 \\
+ \bottomrule
+ \end{tabular}%
+ %\label{tab:addlabel}%
+\end{sidewaystable}%
+
diff --git a/tables/tab5-3.tex b/tables/tab5-3.tex
new file mode 100644
index 0000000..121e537
--- /dev/null
+++ b/tables/tab5-3.tex
@@ -0,0 +1,24 @@
+
+\begin{sidewaystable}[htbp]
+ \centering
+ \caption{Parameter estimates for the contextual forest submodel of SOEL for the four tree species included: black and white oak (\textit{Quercus velutina} and \textit{Q. alba}), sugar maple (\textit{Acer saccharum}), and tulip poplar (\textit{Liriodendron tulipifera}). Parameter estimates were obtained from \citet{Botk93}, \citet{Holm13}, and \citet{Kell14}.}
+ \begin{tabular}{llcccc}
+ \toprule
+ \textbf{Parameter} & \textbf{Description} & \textbf{Black Oak} & \textbf{White Oak} & \textbf{Sugar Maple} & \textbf{Tulip Poplar} \\
+ \midrule
+ DBH$_{\text{max}}$ & Max possible tree DBH (cm) & 100 & 100 & 100 & 100 \\
+ H$_{\text{max}}$ & Max possible tree height (cm) & 3800 & 3800 & 3350 & 4000 \\
+ Age$_{\text{max}}$ & Max possible tree age (years) & 300 & 400 & 400 & 300 \\
+ C & DBH-leaf area relationship & 1.75 & 1.75 & 1.57 & 1.75 \\
+ G & Controls growth rate & 122 & 104 & 119 & 140 \\
+ L & Tolerance for low-light conditions & Intermediate & Intermediate & High & Low \\
+ DEGD$_{\text{min}}$ & Minimum tolerable degree-days & 1977 & 2068 & 2000 & 2171 \\
+ DEGD$_{\text{max}}$ & Maximum tolerable degree-days & 5984 & 5421 & 6300 & 6363 \\
+ N & Tolerance for low soil nitrogen (fertility) & Intermediate & Intermediate & Intermediate & Intermediate \\
+ DT$_{\text{min}}$ & Min tolerable depth to water table (m) & 0.933 & 0.933 & 0.567 & 0.544 \\
+ WILT & Wilting index & 0.45 & 0.45 & 0.35 & 0.245 \\
+ S & Max saplings/year entering per 100 m\textsuperscript{2} & N/A & N/A & 3 & 10 \\
+ \bottomrule
+ \end{tabular}%
+ %\label{tab:addlabel}%
+\end{sidewaystable}%
diff --git a/tables/tab5-4.tex b/tables/tab5-4.tex
new file mode 100644
index 0000000..1dcc28b
--- /dev/null
+++ b/tables/tab5-4.tex
@@ -0,0 +1,27 @@
+% Table generated by Excel2LaTeX from sheet 'Sheet9'
+\begin{sidewaystable}[htbp]
+\footnotesize
+ \centering
+ \caption{SOEL-derived metrics of acorn production by mature oak trees (dbh \textgreater 15.2 cm) based on a 10 year simulation, compared to production estimates in the literature from species in the red oak section (RO) and the white oak section (WO).}
+ \begin{tabular}{lcccccccc}
+ \toprule
+ & & & & \multicolumn{2}{c}{\textbf{Acorns tree\textsuperscript{-1} year\textsuperscript{-1}}} & \multicolumn{3}{c}{\textbf{Acorns ha\textsuperscript{-1} year\textsuperscript{-1}}} \\
+ \textbf{Source} & \textbf{State} & \textbf{Species} & \textbf{Years} & \textbf{Mean} & \textbf{SE} & \textbf{Mean} & \textbf{Min} & \textbf{Max} \\
+ \midrule
+ SOEL & & RO, WO & 10 & 922 & 209 & 90,757 & 6029 & 193,481 \\
+ \citet{Chri55} & MI & WO & 6 & 1100 & 400 & & & \\
+ \citet{Sork93b} & MI & RO & 8 & 1100 & 368 & & & \\
+ \citet{Sork93b} & MI & RO & 8 & 714 & 295 & & & \\
+ \citet{Rose12} & NC & RO & 10 & 680 & 266 & & & \\
+ \citet{Rose12} & NC & WO & 10 & 2595 & 1017 & & & \\
+ \citet{Kell14} & IN & RO & 6 & 424 & 137 & & & \\
+ \citet{Kell14} & IN & WO & 6 & 382 & 147 & & & \\
+ \citet{Bund91} & MN & RO & 1 & & & 151,000 & & \\
+ \citet{Stei95} & PA & RO & 4 & & & 103,236 & 1300 & 490,518 \\
+ \citet{Lhot03} & IL & RO, WO & 1 & & & 212,619 & & \\
+ \citet{Rath08} & IN & RO, WO & 1 & & & 180,214 & & \\
+ \citet{Olso15b} & MO & RO, WO & 15 & & & & \textless 10,000 & 180,000 \\
+ \bottomrule
+ \end{tabular}%
+ %\label{tab:addlabel}%
+\end{sidewaystable}%
diff --git a/tables/tab6-1.tex b/tables/tab6-1.tex
new file mode 100644
index 0000000..b34cd04
--- /dev/null
+++ b/tables/tab6-1.tex
@@ -0,0 +1,25 @@
+
+\begin{table}[htbp]
+\small
+ \centering
+ \caption{Description and data sources for key parameters included in the early oak life history submodel of SOEL.}
+ \begin{tabular}{clcc}
+ \toprule
+ \textbf{Parameter} & \multicolumn{1}{c}{\textbf{Description}} & \multicolumn{1}{c}{\textbf{Source}} & \textbf{Years} \\
+ \midrule
+ \textit{meanAcorn} & Acorns produced per m\textsuperscript{2} canopy & Kellner et al. 2014 & 9 \\
+ \textit{pWeevil} & Probability of weevil infestation & Kellner et al. 2014 & 9 \\
+ \textit{pDispersal} & Probability of acorn dispersal & Chapter 2 & 9 \\
+ \textit{dispDist} & Dispersal kernel scale & Chapter 2 & 4 \\
+ \textit{pCache} & Probability acorn is cached & Chapter 2 & 4 \\
+ \textit{pDispEaten} & Probability acorn eaten $|$ dispersed & Chapter 2 & 4 \\
+ \textit{pUndispEaten} & Probability acorn eaten $|$ not dispersed & Chapter 2 & 4 \\
+ \textit{pGerm} & Probability of acorn germination & Literature sources & N/A \\
+ \textit{pBrowse} & Probability of browse damage & Chapter 3 & 4 \\
+ \textit{meanGr} & Yearly seedling growth & Chapter 4 & 4 \\
+ \textit{pSurv} & Yearly seedling survival & Chapter 4 & 4 \\
+ \bottomrule
+ \end{tabular}%
+ %\label{tab:addlabel}%
+\end{table}%
+
diff --git a/tables/tab6-2.tex b/tables/tab6-2.tex
new file mode 100644
index 0000000..13ad42f
--- /dev/null
+++ b/tables/tab6-2.tex
@@ -0,0 +1,24 @@
+
+\begin{table}[htbp]
+ \centering
+ \small
+ \caption{Summary of regression models fit to each early life history parameter. Predictors included the effect of the midstory removal treatment (TE) and yearly effect of mast availability on acorn parameters (YE-mast) (+ if effect was positive, - if negative, 0 if nonsignificant). Parameters with YE-Random = ``yes'' included random yearly effects. Other included predictors are summarized in the final column. Predictors were selected for inclusion based on significance (\textit{p} \textless 0.05).}
+ \begin{tabular}{ccccc}
+ \toprule
+ \textbf{Parameter} & \textbf{TE} & \textbf{YE-Mast} & \textbf{YE-Random} & \textbf{Other Covariates} \\
+ \midrule
+ \textit{meanAcorn} & 1 & 0 & Yes & N/A \\
+ \textit{pWeevil} & - & - & Yes & N/A \\
+ \textit{pDispersal} & + & 0 & Yes & Species \\
+ \textit{dispDist} & - & - & No & N/A \\
+ \textit{pCache} & 0 & 0 & No & Distance Dispersed \\
+ \textit{pDispEaten} & + & - & No & Species, Cached Status \\
+ \textit{pUndispEaten} & + & - & No & Species \\
+ \textit{pGerm} & N/A & N/A & N/A & Weevil Status, Cached Status \\
+ \textit{pBrowse} & 0 & 0 & Yes & Species, Height, Height\textsuperscript{2} \\
+ \textit{meanGr} & 0 & 0 & No & Light, Species, Browsed \\
+ \textit{pSurv} & 0 & 0 & No & Light, Species, Age \\
+ \bottomrule
+ \end{tabular}%
+ \label{tab:addlabel}%
+\end{table}%
diff --git a/tables/tab6-3.tex b/tables/tab6-3.tex
new file mode 100644
index 0000000..054d670
--- /dev/null
+++ b/tables/tab6-3.tex
@@ -0,0 +1,24 @@
+
+\begin{table}[htbp]
+\setlength{\tabcolsep}{4pt}
+ \centering
+ \caption{Sensitivity (S) and proportion of the sensitivity uncorrelated with other parameters (U) of three different metrics of oak regeneration to key early oak life history parameters in the individual-based model obtained using global sensitivity analysis \citep{Xu08}. Percent emergence is the yearly percent of acorns that ultimately become seedlings; total new seedlings is the sum total of all new seedlings (height \textless 1.4 m) per ha over a seven-year period; and sapling density is total seed-origin saplings (height $\geq$1.4 m and dbh \textless 1.5 cm) at the conclusion of the seven-year period. The range of input values for each parameter were sampled using a correlated-input Latin hypercube approach from the confidence interval around its fitted model intercept.}
+ \begin{tabular}{lcccccc}
+ \toprule
+ & \multicolumn{2}{c}{\textbf{Percent Emergence}} & \multicolumn{2}{c}{\textbf{Total New Seedlings}} & \multicolumn{2}{c}{\textbf{Sapling Density}} \\
+ \textbf{Parameter} & \textbf{S} & \textbf{U} & \textbf{S} & \textbf{U} & \textbf{S} & \textbf{U} \\
+ \midrule
+ \textit{meanAcorn} & 0.074 & 0.00 & 0.448 & 0.18 & 0.247 & 0.08 \\
+ \textit{pWeevil} & 0.028 & 0.00 & 0.049 & 0.00 & 0.014 & 0.09 \\
+ \textit{pDispersal} & 0.059 & 0.12 & 0.012 & 0.25 & 0.047 & 0.06 \\
+ \textit{dispDist} & 0.039 & 0.02 & 0.044 & 0.04 & 0.012 & 0.04 \\
+ \textit{pCache} & 0.863 & 0.64 & 0.484 & 0.41 & 0.290 & 0.22 \\
+ \textit{pDispEaten} & 0.138 & 0.00 & 0.153 & 0.00 & 0.190 & 0 \\
+ \textit{pUndispEaten} & 0.118 & 0.01 & 0.077 & 0.01 & 0.140 & 0.01 \\
+ \textit{pBrowse} & 0.00 & 0.18 & 0.069 & 0.01 & 0.038 & 0.11 \\
+ \textit{meanGr} & 0.114 & 0.00 & 0.129 & 0.00 & 0.295 & 0.15 \\
+ \textit{pSurv} & 0.102 & 0.07 & 0.061 & 0.05 & 0.155 & 0.10 \\
+ \bottomrule
+ \end{tabular}%
+ %\label{tab:addlabel}%
+\end{table}%
diff --git a/tables/tab6-4.tex b/tables/tab6-4.tex
new file mode 100644
index 0000000..12c2127
--- /dev/null
+++ b/tables/tab6-4.tex
@@ -0,0 +1,16 @@
+
+\begin{table}[htbp]
+ \centering
+ \caption{Two-way analysis of variance results for the effects of harvest treatment (none or midstory removal harvest) and mast production scenario prior to harvest on oak seedling (\textless 1.4 m height) density just before harvest and oak sapling ($\geq$1.4 m height, \textless 1.5 cm d.b.h.) density seven years post-harvest. }
+ \begin{tabular}{lccccc}
+ \toprule
+ \textbf{} & \textbf{} & \multicolumn{2}{c}{\textbf{Seedling Density}} & \multicolumn{2}{c}{\textbf{Sapling Density}} \\
+ \textbf{Predictor} & \textbf{d.f.} & \textit{\textbf{F}} & \textit{\textbf{p}} & \textit{\textbf{F}} & \textit{\textbf{p}} \\
+ \midrule
+ Harvest & 2, 525 & 0.078 & 0.925 & 724 & \textless 0.01 \\
+ Scenario & 4, 525 & 972 & \textless 0.01 & 23.3 & \textless 0.01 \\
+ Harvest $\times$ Scenario & 8, 525 & 1.09 & 0.37 & 3.55 & \textless 0.01 \\
+ \bottomrule
+ \end{tabular}%
+ %\label{tab:addlabel}%
+\end{table}%
diff --git a/tables/tab6-5.tex b/tables/tab6-5.tex
new file mode 100644
index 0000000..95400c3
--- /dev/null
+++ b/tables/tab6-5.tex
@@ -0,0 +1,16 @@
+
+\begin{sidewaystable}[htbp]
+ \centering
+ \caption{Two-way analysis of variance results for the effects of harvest treatment (none or midstory removal) and seed predation scenario (control, harvest treatment effects, yearly effects, or both treatment and yearly effects) on four oak regeneration metrics. Total acorns is the sum of all acorns produced over a seven-year period; percent emergence is the yearly percent of acorns that ultimately recruit; total new seedlings is the sum total of all new seedlings (height \textless 1.4 m) per ha over a seven-year period; and sapling density is total seed-origin saplings (height $\geq$1.4 m and dbh \textless 1.5 cm) at the conclusion of the seven year period.}
+ \begin{tabular}{lccccccccc}
+ \toprule
+ \textbf{} & \textbf{} & \multicolumn{2}{c}{\textbf{Total Acorns}} & \multicolumn{2}{c}{\textbf{Percent Emergence}} & \multicolumn{2}{c}{\textbf{Total New Seedlings}} & \multicolumn{2}{c}{\textbf{Sapling Density}} \\
+ \textbf{Predictor} & \textbf{d.f.} & \textit{\textbf{F}} & \textit{\textbf{p}} & \textit{\textbf{F}} & \textit{\textbf{p}} & \textit{\textbf{F}} & \textit{\textbf{p}} & \textit{\textbf{F}} & \textit{\textbf{p}} \\
+ \midrule
+ Harvest & 1, 280 & 0.67 & 0.41 & 12.7 & \textless 0.01 & 3.2 & 0.07 & 532 & \textless 0.01 \\
+ Scenario & 3, 280 & 2.39 & 0.07 & 316 & \textless 0.01 & 140 & \textless 0.01 & 269 & \textless 0.01 \\
+ Harvest $\times$ Scenario & 3, 280 & 2.28 & 0.08 & 5.79 & \textless 0.01 & 1.92 & 0.13 & 3.64 & 0.01 \\
+ \bottomrule
+ \end{tabular}%
+ %\label{tab:addlabel}%
+\end{sidewaystable}%
diff --git a/tables/tab6-6.tex b/tables/tab6-6.tex
new file mode 100644
index 0000000..43f251a
--- /dev/null
+++ b/tables/tab6-6.tex
@@ -0,0 +1,20 @@
+
+\begin{table}[htbp]
+\footnotesize
+ \centering
+ \caption{Linear regression results for the effects of harvest treatment (none, clearcut, or midstory removal) and drought year probability on two oak regeneration metrics. Total new seedlings is the sum total of all newly recruited seedlings (height \textless 1.4 m) per ha over a seven-year period; and sapling density is total seed-origin saplings (height $\geq$1.4 m and dbh \textless 1.5 cm) at the conclusion of the seven-year period.}
+ \begin{tabular}{lcccccc}
+ \toprule
+ \textbf{} & \multicolumn{3}{c}{\textbf{Total New Seedlings}} & \multicolumn{3}{c}{\textbf{Sapling Density}} \\
+ \textbf{Parameter} & \textbf{Estimate} & \textbf{t-value} & \textit{\textbf{p}} & \textbf{Estimate} & \textbf{t-value} & \textit{\textbf{p}} \\
+ \midrule
+ Intercept & 20443 & 90.9 & \textless 0.01 & 617.5 & 23.8 & \textless 0.01 \\
+ Midstory Removal & 145.2 & 0.457 & 0.65 & 139 & 3.79 & \textless 0.01 \\
+ Clearcut & -14922 & -46.9 & \textless 0.01 & 2295 & 62.5 & \textless 0.01 \\
+ Drought Probability & -8949.7 & -24.1 & \textless 0.01 & -652.9 & -15.2 & \textless 0.01 \\
+ Midstory Removal x Drought & 293 & 0.56 & 0.58 & -132.4 & -2.18 & 0.03 \\
+ Clearcut x Drought & 5772.6 & 11 & \textless 0.01 & -2482 & -40.9 & \textless 0.01 \\
+ \bottomrule
+ \end{tabular}%
+ %\label{tab:addlabel}%
+\end{table}%
diff --git a/vita.tex b/vita.tex
new file mode 100644
index 0000000..c785d92
--- /dev/null
+++ b/vita.tex
@@ -0,0 +1,153 @@
+\begin{vita}
+\setlength{\parindent}{0cm}
+\Large
+\textbf{Kenneth F. Kellner}
+
+
+\normalsize
+Pfendler Hall, 715 W State St., West Lafayette, IN 47907
+
+kkellner@purdue.edu
+
+\hrulefill
+
+\large
+\textbf{Education}
+\normalsize
+
+2012-2015 \, \, \, Ph.D., Purdue University, West Lafayette, IN
+
+2011-2013 \, \, \, Graduate Certificate in Applied Statistics, Purdue University
+
+2009-2012 \, \, \, M.S., Purdue University
+
+2005-2009 \, \, \, B.S., Biology, Wheaton College, Wheaton, IL
+
+\vspace{\baselineskip}
+\large
+\textbf{Awards and Grants}
+\normalsize
+
+2015 \, \, \, \, \, \, \, \,Kirkpatrick Memorial Graduate Student Award
+
+2015 \, \, \, \, \, \, \, \,Teaching Academy Graduate Teaching Award
+
+2014 \, \, \, \, \, \, \, \,Fischer Forestry Fund Scholarship
+
+2012 \, \, \, \, \, \, \, \,Purdue Graduate Student Government DEAL Grant
+
+2010, 2011 \, \, \,NSF Graduate Fellowship Honorable Mentions
+
+2009 \, \, \, \, \, \, \, \,Frederick N. Andrews Fellowship
+
+2009 \, \, \, \, \, \, \, \,David Bruce Memorial Research Award
+
+2009 \, \, \, \, \, \, \, \,Honors in Biology, Wheaton College
+
+2009 \, \, \, \, \, \, \, \,Wheaton College Scholastic Honors Society
+
+2008 \, \, \, \, \, \, \, \,Beaver-Schmale Research Award
+
+2007 \, \, \, \, \, \, \, \,Willis A. Reid Student Research Grant
+
+2005 \, \, \, \, \, \, \, \,National Merit Scholarship
+
+2005 \, \, \, \, \, \, \, \,President's Award, Wheaton College
+
+\vspace{\baselineskip}
+\large
+\textbf{Teaching}
+\normalsize
+
+2014 \, \, \, \, \, \, \, \,Instructor, R and Bayesian Analysis for Ecologists
+
+2012, 2014 \, \, \,Teaching Assistant, Quantitative Methods for Ecologists
+
+2012, 2014 \, \, \,Teaching Assistant, Vertebrate Ecology
+
+2010, 2011 \, \, \,Teaching Assistant, Wildlife Habitat Management
+
+2009 \, \, \, \, \, \, \, \,Teaching Assistant, Invertebrate Zoology
+
+2008 \, \, \, \, \, \, \, \,Teaching Assistant, General Ecology
+
+\vspace{\baselineskip}
+\large
+\textbf{Publications}
+\normalsize
+
+\textbf{Kellner KF}. 2015. jagsUI: a wrapper around rjags to streamline JAGS analyses. R package version 1.3.7.
+
+Leonard OD, Moore JW, Riegel JK, Meier AR, Dunning JB, \textbf{Kellner KF}, and RK Swihart. 2015. Effect of variation in forest harvest intensity on winter occupancy of barred owls and eastern screech-owls in decidious forests of the east-central United States. Journal of Field Ornithology 86: 115-129.
+
+\textbf{Kellner KF} and RK Swihart. 2014. Accounting for imperfect detection in ecology: a quantitative review. PloS one 9: e111436.
+
+\textbf{Kellner KF} and RK Swihart. 2014. Changes in small mammal microhabitat use following silvicultural disturbance. American Midland Naturalist 172: 348-358.
+
+\textbf{Kellner KF}, Riegel JK, and RK Swihart. 2014. Effects of silvicultural disturbance on acorn infestation and removal. New Forests 45: 265-281.
+
+\textbf{Kellner KF}, Urban NA, and RK Swihart. 2013. Short-term responses of small mammals to timber harvest in the U.S. central hardwood forest region. Journal of Wildlife Management 77: 1650-1663.
+
+Smyser TJ, Page LK, Johnson SA, Hudson CM, \textbf{Kellner KF}, Rhodes OE, and RK Swihart. 2013. Raccoon roundworm (Baylisascaris procyonis) mitigation via anthelmintic baiting: patterns of bait removal and response in prevalence. Journal of Wildlife Management 77: 1372-1379.
+
+\textbf{Kellner KF}, Riegel JK, Lichti NI, and RK Swihart. 2013. Oak mast production and animal impacts on acorn survival in the Central Hardwoods. Pages 176-190 in Swihart RK, Saunders MR, Kalb RA, Haulton S, and CH Michler. The Hardwood Ecosystem Experiment: a framework for studying responses to forest management. U.S. Department of Agriculture, Forest Service, Northern Forest Experiment Station Gen. Tech. Rep. NRS-P-108.
+
+\textbf{Kellner KF}, Page LK, Downey M, and S McCord. 2012. Prevalence of Baylisascaris procyonis in rural and suburban intermediate host populations. Journal of Wildlife Diseases 48: 1083-1087.
+
+\textbf{Kellner KF}. 2012. Temporal dynamics of mast and small mammals: short-term responses to silviculture. Master’s Thesis. Purdue University, West Lafayette, IN.
+
+Page LK, Beasley JC, Olson ZH, Smyser TJ, Downey M, \textbf{Kellner KF}, McCord SE, Eagan TS, and OE Rhodes Jr. 2011. Reducing Baylisascaris procyonis roundworm larvae in raccoon latrines. Emerging Infectious Diseases 17: 90-93.
+
+Page, LK, Gehrt SD, Cascione A, and \textbf{KF Kellner}. 2009. The relationship between Baylisascaris procyonis prevalence and raccoon population structure. Journal of Parasitology 95: 1314-1320.
+
+Page LK, Anchor C, Luy E, Kron S, Larsen G, Madsen L, \textbf{Kellner KF}, and TJ Smyser. 2009. Backyard raccoon latrines and risk for Baylisascaris procyonis transmission to humans. Emerging Infectious Diseases 15: 1530-1531.
+
+\vspace{\baselineskip}
+\large
+\textbf{Presentations}
+\normalsize
+
+\textbf{Kellner KF} and RK Swihart. 2015 February. Accounting for imperfect detection in ecology: a quantitative review. Oral presentation at the 75th Midwest Fish and Wildlife Conference, Indianapolis, IN.
+
+\textbf{Kellner KF} and RK Swihart. 2015 February. Oak seedling survival and growth in recently harvested hardwood stands. Oral presentation at the 75th Midwest Fish and Wildlife Conference, Indianapolis, IN.
+
+\textbf{Kellner KF} and RK Swihart. 2015 February. Individual-based models at the intersection of wildlife and forestry: modeling impacts of seed predators, herbivores, and competitors on oak regeneration success. Oral presentation at the 75th Midwest Fish and Wildlife Conference, Indianapolis, IN.
+
+\textbf{Kellner KF}. 2015 February. If you give a mouse an acorn: trophic interactions driving the oak life cycle in managed forests. Invited seminar at Wheaton College, Wheaton, IL.
+
+\textbf{Kellner KF} and RK Swihart. 2014 October. Small mammals as predators and dispersers of acorns: effects of silviculture. Oral presentation at the Wildlife Society Annual conference, Pittsburgh, PA.
+
+\textbf{Kellner KF} and RK Swihart. 2013 June. Effects of silviculture on predation and dispersal of oak (Quercus) acorns by small mammals. Oral presentation at the 9th North American Forest Ecology Workshop, Bloomington, IN.
+
+\textbf{Kellner KF} and RK Swihart. 2013 April. Effects of silviculture on predation and dispersal of oak (Quercus) acorns by small mammals. Poster presented at the 2013 FNR Research Symposium, Purdue University, West Lafayette, IN.
+
+\textbf{Kellner KF}. 2013 February. Impacts of climate change on wildlife in Indiana. Invited presentation at 2013 Annual Meeting of the Indiana DNR Division of Nature Preserves, Indianapolis, IN.
+
+\textbf{Kellner KF}, Urban NA, and RK Swihart. 2012 June. Short-term small mammal responses to silviculture in the Central Hardwoods. Poster presented at the 92nd Annual Meeting of the American Society of Mammalogists, Reno, NV.
+
+\textbf{Kellner KF} and RK Swihart. 2012 April. Short-term small mammal responses to silviculture in the Central Hardwoods. Poster presented at the 2012 FNR Research Symposium, Purdue University, West Lafayette, IN.
+
+\textbf{Kellner KF} Urban NA, and RK Swihart. 2012 February. Mast and small mammals: short-term responses to silviculture in southern Indiana. Invited presentation at the Meeting of the Ball State University Wildlife Society, Ball State University, Muncie, IN.
+
+\textbf{Kellner KF} and RK Swihart. 2012 February. Short-term responses of small mammals to silviculture in southern Indiana. Oral presentation at the 2012 Annual Meeting of the Indiana Wildlife Society, Indianapolis, IN.
+
+\textbf{Kellner KF} and RK Swihart. 2011 April. Oak mast production and animal impacts on acorn survival in the Central Hardwood Forest. Poster presented at the 2011 FNR Research Symposium, Purdue University, West Lafayette, IN.
+
+\textbf{Kellner KF}, Urban, NA, and RK Swihart. 2010 November. Responses of shrew populations to even-aged forest management. Poster presented at the 2010 Hardwood Ecosystem Annual Meeting, West Lafayette, IN.
+
+\textbf{Kellner KF}, Maingi J, and M Henry. 2008 July. Effects of urbanization on the hydrology of Southwestern Ohio watersheds: a comparative historical approach. Oral presentation at the 2008 ECO-REU Symposium, Miami University, Oxford, OH.
+
+\textbf{Kellner KF}, Downey MD, and PP Girgis. 2007 August. Prevalence of Baylisascaris procyonis in intermediate host populations as a function of landscape attributes. Poster presented at the 7th Annual Summer Research Program Symposium, Wheaton College, Wheaton, IL.
+
+\textbf{Kellner KF} and A Cascione. 2006 September. The effects of Baylisascaris procyonis on host population dynamics, and prevalence of B. procyonis in intermediate host populations. Oral presentation at the Wheaton College Science Division Faculty Meeting, Wheaton College, Wheaton, IL.
+
+\textbf{Kellner KF} and A Cascione. 2006 August. The effects of Baylisascaris procyonis on host population dynamics, and prevalence of B. procyonis in intermediate host populations. Poster presented at the 6th Annual Summer Research Program Symposium, Wheaton College, Wheaton, IL.
+
+\vspace{\baselineskip}
+\large
+\textbf{Professional Service}
+\normalsize
+
+Reviewer for Ecological Modelling, Biodiversity and Conservation, Landscape Ecology, Western North American Naturalist, Ecological Applications, Forests, American Midland Naturalist, Journal of Mammalogy, Scandinavian Journal of Forest Science
+
+\end{vita}