Fire Severity Effects on Herpetofaunal Diversity in Florida Scrub Communities of the Lake Wales Ridge

Michelle Lindsay

A thesis submitted in partial fulfillment of the requirements for the degree of Master of Science Department of Biology College of Art and Science University of South Florida

Major Professor: Alison Gainsbury, Ph.D. David Lewis, Ph.D. Neal Halstead, Ph.D.

Date of Approval: June 15, 2021

Keywords: Prescribed burning, Anolis carolinensis, Plestiodon reynoldsi, Microhabitat variables, Scrubby flatwoods, Anaxyrus quercicus

Acknowledgments

I would like to thank my advisor Alison Gainsbury for guiding me through this process and for your invaluable guidance. Your dedication to your students and this field is greatly apparent and appreciated. I would also like to thank my other committee members David Lewis and Neal Halstead for all of their beneficial advice and helping me move forward in this thesis. Thank you to the Biological Sciences department at University of South

Florida and for the USFSP student research grant for helping fund some of my research.

Thank you to the Archbold Biological Station for providing the data and resources for allowing my research to take place in such unprecedented times. Also, the staff at Archbold Biological Station particularly Betsie

Rothermel and Kevin Main. I also want to personally thank my family for supporting me in my continued education and most of all my husband Conner Lindsay for the constant moral support, I would not have been able to do it without him.

Table of Contents

List of Tables ...... ii

List of Figures ...... iii

Abstract ...... iv

Fire Severity Effects on Herpetofaunal Diversity in Florida Scrub Communities of the Lake Wales Ridge ...... 1 Introduction ...... 1

Materials and Methods ...... 5 Study area ...... 5 Data acquisition ...... 5 Fire severity ...... 6 Microhabitat variables ...... 7 Herpetofaunal collection ...... 8 Herpetofaunal species traits ...... 11 Statistical analyses ...... 11

Results ...... 14

Discussion ...... 25 Community composition and fire severity ...... 25 Species diversity and fire severity ...... 27 Species diversity and microhabitat variables ...... 32 Microhabitat variables and fire severity ...... 33 Limitations and future research ...... 34 Conclusions ...... 34

References ...... 35

Appendices ...... 46 Appendix 1 ...... 46 Appendix 2 ...... 47

List of Tables

Table 1 Fire intensity categories for prescribed fires within Archbold Biological Station Reprinted from “Station Fire Management Plan,” by Main and Menges, 1997, Land Management Publication, 97. Used with permission ...... 7

Table 2 Reptile and amphibian traits and their associated functional significance...... 11

Table 3 Total number of species at the different fire severity plots...... 14

List of Figures

Fig. 1 a) Site map of Archbold Biological Station with the study area circled in black. Reprinted from “Station Fire Management Plan,” by Main and Menges, 1997, Land Management Publication, 97. Used with permission, b) Map of plot locations within Archbold Biological Station...... 6

Fig. 2 A schematic diagram of the 8m x 8m (64 m2) plots selected which consisted of a five-meter vegetation transect, four coverboards in each corner of the plot, two camera traps with a drift fence array and visual encounter survey within the plot...... 9

Fig. 3 Drift fence technique with a two-sided camera system in a high severity plot (adapted from Martin et al. 2017)...... 10

Fig. 4 Photographs of some reptiles and amphibians captured utilizing the different sampling methods, from left to right: Sceloporus woodi captured by visual encounter survey, Bufo terrestris sighted underneath a coverboard and Pituophis melanoleucus mugitus captured by a camera trap...... 15

Fig. 5 Species Accumulation Curve: black line represents all reptile species, green line represents amphibian species and red line represents lizard species, recorded in all 12 fire severity plots. .... 16

Fig. 6 Boxplots showing community composition of herpetofauna and fire severity, a) reptile communities, b) lizard communities, c) amphibian communities...... 17

Fig. 7 Boxplots showing species taxonomic diversity and fire severity, a) reptile species richness b) reptile diversity, c) lizard richness, d) lizard diversity, e) amphibian richness, f) amphibian diversity...... 19

Fig. 8 Boxplots showing species functional diversity and fire severity plots. Diversity indices include; functional richness, functional evenness and functional divergence, a) reptile functional diversity, b) lizard functional diversity, c) amphibian functional diversity...... 20

Fig. 9 Reptile diversity related to leaf litter depth standard deviation (SD) a) lizard richness, b) lizard taxonomic diversity, c) reptile functional evenness, d) reptile functional divergence...... 21

Fig. 10 Amphibian diversity related to leaf litter depth (cm), a) amphibian richness, b) amphibian taxonomic diversity, c) amphibian functional evenness, d) amphibian functional divergence...... 22

Fig. 11 Boxplots for microhabitat variables and variability at the different fire severity plots, nonmatching letters identifying significant pairwise differences, a) the standard deviation of litter depth, b) the standard deviation of percent litter, c) the standard deviation of basal area of shrub, d) mean leaf litter depth...... 23

iii

Abstract

Fires, either natural or prescribed, are essential for conserving pyrogenic ecosystems; however, climate change is predicted to increase fire severity that could negatively impact species diversity. Reptile and amphibian species may be particularly at risk given they are ectothermic species. The objective of this study was to better understand the impacts of fire severity on the herpetofaunal communities of the Florida scrub habitats, which is an imperiled ecosystem that sustains over two-thirds of listed reptiles and amphibians in Florida. We conducted a field- based study to test taxonomic and functional herpetofaunal diversity differences across four varying fire severities: unburned, low, medium and high. We also examined the association between herpetofaunal diversity and microhabitat variables. We recorded 549 individuals representing 23 different reptile and amphibian species. The herpetofaunal community was not significantly different between the fire severity plots; however, microhabitat variables and variability were significantly associated with species diversity. Interesting, the endangered sand skink,

Plestiodon reynoldsi, was not recorded at the high severity plots, which could have implications on fire management practices of this federally listed threatened species. In this study, we demonstrate fire severity does not have direct effects but indirect effects on herpetofaunal species diversity in the Florida scrub habitat, increased fire severity impacts on leaf litter has the potential to have detrimental effects on herpetofaunal diversity. This study highlights the importance of fine-scale microhabitat variables as an important indicator for biodiversity conservation.

Introduction

Fires, either natural or prescribed, are essential for conserving pyrogenic ecosystems, such as the

Mediterranean chaparral, Australian mallee, African Savanna and Florida scrub (Archibald et al. 2005, Ashton and

Knipps 2011, Elzer et al. 2013, McCoy et al. 2013, Santos and Cheylan 2013). Over the past centuries fire suppression had been established in many parts of the world resulting in altered disturbance regimes, which leads to biodiversity loss (Hromada et al. 2018). This biodiversity loss alerted managers to the danger of fire suppression, in which many foresters in North America made prescribed burning a management tool to restore ecosystems

(Jacobsen et al. 2020, Perry et al. 2009). Prescribed burning is a common practice throughout the southeastern

United States that aids in ecosystem processes (Menges et al. 2017). It helps restore ground cover, restore vegetation structure, enhance wildlife habitat, and reduce fuel loads (Kuchinke et al. 2020, Menges et al. 2017). Fire management through prescribed burning is used to mimic natural fire regimes through variations in fire behavior, fire-return interval, fire severity and burn season (Duncan et al. 2015, Menges et al. 2017, McCoy et al. 2013). Fire severity is a factor that is of particular interest since climate change is predicted to increase the severity of fires, by increasing fuel loads, changing wind speeds, reducing rainfall, and increasing natural fire ignitions (Penman et al.

2015, Zuniga et al. 2021). All of which could cause detrimental impacts on species diversity (Hossack et al. 2009).

To conserve pyrogenic ecosystems biodiversity, we need to better understand the effects of prescribed burning, especially the effects of fire severity on native biodiversity (Menges et al. 2017).

Fire intensity is the energy output from the fire, whereas fire severity is a measure of how fire intensity alters ecosystems (Keeley 2009). Fire severity quantifies the loss of organic matter aboveground through measures of biomass loss, such as plant mortality, scorch height, and loss of leaf litter (Kuchinke et al. 2020). These parameters will vary depending on the ecosystem (Keeley 2009). Roberts et al. (2008) demonstrated the abundance of small mammal species in California was directly related to fire severity due to altered vegetation structure and habitat heterogeneity. The abundance of deer mice decreased with fire severity, due to the high fire severity altering their important food source of conifers. Fontaine and Kennedy (2012) conducted a meta-analysis on the short-term taxonomic response of vertebrates to fire severity. They showed positive responses of birds to higher severity fires 1

with some mammal species responding positively and other mammal species responding negatively. They recommended higher intensity fires should be implemented by land managers when burning a landscape to create a mosaic of mixed severity. Interestingly, they had insufficient data on reptile and amphibian species to include in their analysis due to high severity data being insufficient. Most studies on the effects of fire severity predominately focus on changes to the vegetation and bird species responses (Fontaine and Kennedy 2012). The effects of fire severity on mammals are relatively limited; however, even less is known about how fire severity influences reptile and amphibian species (Fontaine and Kennedy 2012). A better understanding of how fire severity impacts these ectothermic species will further aid in conservation management practices to maintain the native biodiversity (Price et al. 2010).

Reptile and amphibian species are important to conserve due to their vital ecological functions in an ecosystem, such as their role as mesopredators (Timm et al. 2020). Both reptile and amphibian species act as predators to insects, but in turn serve as prey for birds (Greenberg and Waldrop 2008, Leavitt and Schalk 2018).

They are not only an important element of biological diversity but help support the biological diversity of vertebrates (Greenberg and Waldrop 2008). Reptiles and amphibians might be particularly sensitive to fire altering their environment because they are ectothermic. Ectothermic species regulate body temperature by relying on ambient temperatures, which is essential for locomotion, reproduction, digestion, and escape (Elzer et al. 2013).

Elzer et al. (2013) found small changes in canopy cover may result in profound changes to thermoregulatory opportunities for reptiles by reducing the thermal quality of reptile’s habitats. Alternatively, Erwin and Stasiak

(1979) documented fire, by prescribed burning, caused minimal direct effects on reptile and amphibian communities given most are able to avoid fire by escaping or burrowing. However, indirect effects due to altered reproduction, diet, and thermal performance are not well studied (Mushinsky 1985). Most reptiles and amphibians, when compared to other taxa such as birds and mammals, have low dispersal ability, small home ranges, and thermoregulatory restraints that increases their vulnerability to habitat degradation (Bohm et al. 2013). Thus, one expects fire severity to increase habitat degradation and cause indirect effects on herpetofaunal communities. Given reptiles and amphibians are critical ecosystem components, it is important to better understand the impacts of fire severity on these ectothermic species (Timm et al. 2020).

Understanding the impact of fire severity on the herpetofauna community is vital for Florida scrub habitats.

The Florida scrub habitat is an important pyrogenic ecosystem because it is a globally threatened ecosystem that sustains over two-thirds of listed reptiles and amphibians in Florida (Meshaka and Layne 2002). Florida scrub habitats are rapidly declining, having lost over 60% of its original extent due to agriculture and urban development, and include a diverse and endemic herpetofauna (Ashton and Knipps 2011, Meshaka and Layne 2002, McCoy and

Mushinsky 1994). Most studies to date focus on herpetofaunal diversity changes driven by time since fire and fire frequency (Ashton and Knipps 2011, Halstead 2007, McCoy et al. 2013). Within Florida scrub, researchers demonstrated no difference in reptile and amphibian species richness with time since fire; however, the abundance of one species, Plestiodon reynoldsi (Florida sand skink), did show variation with the largest abundance in burn sites of more than 17 years intervals. This result is most likely due to leaf litter accumulation since the litter may contain important prey items for the Florida sand skink to utilize (Ashton and Knipps 2011, McCoy et al. 2013). Mushinsky

(1985) found greater herpetofaunal diversity was bimodal with peaks occurring at an annual fire frequency and every 7 years. Whereas the lowest diversity occurred at sites with a fire frequency of every 2 years. Halstead (2007) found fire frequency had notable effects on the community composition of reptiles and amphibians in Sandhill habitats. Halstead (2007) noted the community composition in Sandhill habitats was positively correlated with herbaceous ground cover and leaf litter. However, the effects of fire severity on herpetofaunal diversity in Florida scrub communities of the Lake Wales Ridge remains poorly understood (Mushinsky 1985, Hossack et al. 2009).

The objective of this study was to better understand reptile and amphibian species responses to varying fire severity in the Florida scrub communities of the Lake Wales Ridge. The goal of this study was to provide land managers with information on the impacts of fire severity on reptile and amphibian species in the Florida scrub habitat since it is an imperiled ecosystem that harbors many threatened and endangered reptiles and amphibians in

Florida (Meshaka and Layne 2002). Furthermore, this information will help improve mitigating strategies for biodiversity conservation in the face of climate change with predicted intensifying fire regimes, including fire severity (Hossack et al. 2009, Zuniga et al. 2021). The novelty of this research was to test the following hypotheses:

1) Species composition was expected to shift based on fire severity, because fire severity impacts the

succession of vegetation communities. Some species such as Aspidoscelis sexlineata (six-lined

racerunner) and Sceloporus woodi (Florida scrub lizard) are documented to be common in habitats that 3

are opened due to fire, thus is expected to be the most dominant species in the higher fire severity

plots. The dominant species in the lower severity plots are expected to be Anolis carolinensis (green

anole) due to habitat preferences for dense, large shrubs.

2) Reptile and amphibian taxonomic and functional diversity was predicted to be largest at the lower fire

severity plots. This is due to higher fire severity plots causing increased ground openness and shrub

scorch height that may reduce thermal quality for reptile and amphibian species.

3) Reptile and amphibian taxonomic and functional diversity was expected to be positively associated

with some microhabitat variables and variability and negatively associated with others (e.g., leaf litter

vs. bare ground). The more microhabitats within the fire severity plots the more species should be able

to coexist.

4) Microhabitat variables and variability were also expected to shift based on fire severity due to higher

severity fires causing the majority of the vegetation to be consumed by the fire. Microhabitat variables

and variability were predicted to decrease with increasing fire severity with variables being the largest

at the lower severity plots and smallest at the higher severity plots.

Materials and Methods

Study area

This study was conducted at Archbold Biological Station. This research facility is 5,193 acres found in the central Florida city of Venus, in Highlands county (27°10’50” N, 81°21’00” W) (Abrahamson et al. 1948, McCoy et al. 2013). It is located in the Florida scrub on the Lake Wales Ridge and is home to many endemic species

(https://www.archbold-station.org). The Lake Wales Ridge is a geomorphological feature that runs down central peninsular Florida (Weekley et al. 2008). The ridge is composed of a beach dune system and a historical shoreline that dates to the Pleistocene and it is considered a global hotspot for biodiversity (Weekley et al. 2008). The southern part of the ridge is a region that encompasses a vast number of endemic species compared to that of other areas of the ridge (Abrahamson et al. 1984). Fires in this station have been reported starting at the 1920’s and the station has been continuously mapping all fires since 1967. Archbold Biological Station and the adjoining Archbold reserve, which is an additional 3,648 acres, is also home to 19 federally listed species, 50 species of reptile, and 23 species of amphibians, which makes this a valuable area for conducting conservation research

(https://www.archbold-station.org).

Data acquisition

Sampling took place between July 22-August 4, August 26-September 8, and September 23-October 6,

2020 at Archbold Biological Station (Fig. 1a). There was a total of twelve plots sampled, each consisting of an 8m x

8m square (64 m2). The plots were selected based on varying fire severity, at least 300 meters apart from one another and with a 100-meter buffer zone with a similar severity index due to fires being patchy in nature. I selected three plots with high burn severity, three plots with medium burn severity, three plots with low burn severity and three plots that remained unburned, resulting in a total of 12 plots (Fig. 1b). Each of the 12 plots were sampled during the three different sampling periods. All sampling plots consisted of similar time since last fire and fire

frequency. Each plot selected was burned in the past 0-2 years, with a fire return interval of every four years and given a severity index of either high fire severity, medium or low fire severity.

A B

Fig. 1 a) Site map of Archbold Biological Station with the study area circled in black. Reprinted from “Station Fire

Management Plan,” by Main and Menges, 1997, Land Management Publication, 97. Used with permission, b) Map of plot locations within Archbold Biological Station.

Fire severity

The fire severity of each plot was determined by Archbold Biological Station’s staff by their intensive fire mapping program. After fires occurred, the intensity of each fire was measured visually by vegetation indicators

(Table 1) and all fires were all mapped into GIS within two months of the burn (Main and Menges 1997). Aerial flights were also conducted over several burned areas to determine accurate intensity boundaries and ground truthed

(Main and Menges 1997). For this study, plots were selected based on the intensity of the last fire that occurred.

Plots consisting of a fire intensity measure of >3 were selected as the high severity plots, whereas plots comprising a fire intensity measure of 2 were interpreted as medium severity, and plots consisting of a fire intensity score of <1 were selected as low severity plots. The plots that were selected a fire intensity of 0 were considered unburned.

Table 1. Fire intensity categories for prescribed fires within Archbold Biological Station. Reprinted from “Station

Fire Management Plan,” by Main and Menges, 1997, Land Management Publication, 97. Used with permission.

FIRE DESCRIPT SURFACE DEAD LEAVES (0-2 M) INTENSITY ION LITTER PALMETTO LEAF TWIGS BLADES 0 Unburned Unburned All Unburned, (Green) Not Consumed Not Consumed (Green) Patches Some Leaves Not Mainly Not 1 Light Scorched- (Brown) Unburned Consumed (Green) Consumed Mainly Not Leaf Blades Partially 2 Moderate Consumed Leaves Scorched (Brown) Consumed Consumed Leaf Blades Completely Small Twigs 3 Heavy Consumed All Burned Off Consumed-Petioles May Consumed Remain

Microhabitat variables

At each sampling plot, we measured the following variables: percent leaf litter, percent bareground, leaf litter depth (cm), canopy openness (%), shrub height (m), basal area of shrub (cm), number of fallen logs, and distance to pond (m). One-five meter transect was placed randomly with ground stakes, rope and a measuring tape.

Along each transect, a 50cm-by-50cm vegetation quadrat was placed at the start of the transect and every meter throughout the transect to obtain five vegetation measurements for each plot. Percent leaf litter and percent bare ground within the quadrat were estimated to the nearest 5% within the vegetation quadrat. The depth of the leaf litter was procured using a ruler in the deepest part of the litter within the quadrat and estimated to the nearest millimeter.

The canopy openness was measured in the same location and determined by a spherical densiometer, which

consisted of twenty-four ¼” squares (Lemmon 1957). The densiometer was placed 1.3m above the ground and each square of the grid was then equally divided into four smaller squares. The number of squares that are covered by the canopy was then be multiplied by 1.04 to produce the overstory density percentage (Lemmon 1957).

Shrub height was also measured along the transect using a height pole. A six-foot PVC pole was utilized indicating height measurements throughout. The scorch height of each of those measured shrubs were also determined with the height pole. Scorch height was based off visual burn marks in the base of the shrub or tree.

Basal area of shrub was measured at 1.3m above the ground, utilizing a measuring tape. The number of fallen logs along the transect was included, with any fallen log that is 2 meters from the transect. Distance to nearest pond was ascertained from a map using QGIS (QGIS Development Team 2018), by taking the point from the plot center to the center of the nearest pond or wetland.

Herpetofaunal collection

At each sampling plot, we collected reptile and amphibians with the following sampling techniques: timed visual encounter surveys, coverboards, and drift fence camera trap arrays. Timed Visual Encounter Surveys (VES) were utilized to detect the presence of reptile and amphibian species. Each visual encounter survey consisted of two people actively searching the plot for herpetofaunal species. During each of the surveys, the plots were intensely sampled for a 45 minutes period and remained consistent for all 12 plots for all sampling periods. The plots were set-up with marked flags to establish the boundary of the plot to be searched (Fig. 2). Each person started in a different corner of the plot and walked in a clockwise circle pattern slowly while they made their way to the center of the plot. Plants, ground, leaf litter and logs were actively and diligently searched. Animals were recorded only if they are within the quadrat or two meters just outside of the quadrat, to limit any potential escaped animals. Every reptile and amphibian species found was captured manually, if possible. The animal was then placed in a five-gallon bucket for morphological data collection.

For each animal capture, date, time, plot number (#), species identification (ID), snout-vent length (to the nearest millimeter), sex with the exception of the skinks, and age class (juvenile, adult) were all recorded. Snout- vent length was obtained by a ruler to the nearest millimeter, when the animal was in hand. Sex of the animal was determined if species had observed sexually dimorphic characteristics. Age class was determined by length and 8

distinct adult characteristics. Reptiles were given an individual ID with nontoxic nail polish, while pictures were used for amphibians for pattern identification. Animals that were seen but not captured were still recorded for presence/absence data. Plestiodon reynoldsi tracks were also used for the presence of individuals within the plot.

Surveys were conducted either in the morning when reptiles are most active during the summer (08:00-

11:00) or in the late afternoon (17:00-20:00). Surveys were conducted at least twice for each plot per sampling period to make sure all reptile and amphibian species were collected. Coverboards were checked during visual encounter surveys, in which they were flipped over to check for any reptile or amphibian species underneath that may be seeking refuge. Plots were also randomized to remove any possible bias with time of activity. Plots that were sampled once in the morning were sampled again at least four days later and sampled in the afternoon. If multiple plots were sampled on the same day, the plots were switched the next survey to reduce any possible sampling bias.

= Camera trap array

= Vegetation transect

= Coverboard

= Observer path

For each plot, one camera trap array was set-up for a thirty-six-day sampling period. Camera traps were arranged in a single-fence drift array, consisting of a three-meter-long silt fence that aided in funneling species into the cameras’ view (Fig. 3). For each animal found on the retrieved camera data; plot # including specified severity index, transect #, camera trap # and species found were recorded. Individuals’ identification for species were manageable due to distinct markings or characteristics. Any reptile or amphibian species captured within 60 minutes of the paired camera were only counted as one individual. This adapted drift fence technique is sensitive to the detection of ectotherms unlike many other camera trap methods (Martin et al. 2017).

Fig. 3 Drift fence technique with a two-sided camera system in a high severity plot (adapted from Martin et al.

2017).

Herpetofaunal species traits

For each captured species, trait data was collected in the field or collated from the literature to test for functional diversity (Table 2). In this study, trait types such as, resource quantity, habitat use, and foraging mode were selected based on important reptile and amphibian measures that affects their performance and fitness. Body size, primary habitat strata, shelter use, reproductive mode, and activity pattern were all examined. Body size is utilized as an important trait since it is related to thermal tolerance and dispersal ability (Hu et al. 2016). Primary habitat, shelter use and activity patterns are indicators of habitat preference and microclimate tolerance with reproductive mode referencing reproductive potential (Hu et al. 2016).

Table 2. Reptile and amphibian traits and their associated functional significance.

Trait Type Trait Functional significance Resource Quantity (RQ) Snout-vent length, Body size is utilized as an important trait since it is related Reproductive mode to thermal tolerance and dispersal ability and reproductive mode is indicative to reproduction potential (Hu et al. 2016). Categories for reproductive mode: ovoviviparous or oviparous

Habitat Use (HU) Primary habitat strata, Indicative of habitat preference (Hu et al. 2016). Shelter use Categories for primary habitat strata: terrestrial, arboreal or fossorial Categories for shelter use: leaf litter, crevices, coarse woody debris, sand

Foraging Mode (FM) Time of Activity Indicative of microclimate tolerance (Hu et al. 2016). Categories: diurnal, nocturnal, or crepuscular

Statistical analyses

To test for species composition shifts based on fire severity, an analysis of similarities (ANOSIM: Clarke

1993) was performed. ANOISM is a non-parametric statistical test that gives an R-value and pairwise comparisons between fire severity plots, which represents similarity in species abundance and reptile and amphibian composition between fire severities. The R-values range from -1 and +1, with values closer to 1 indicating significantly different communities (Santos and Cheylan 2013). Differences between reptile and amphibian abundance and composition among fire severity plots were quantified with the Bray-Curtis similarity index. The significance of R was

determined by permuting the grouping vector 999 times to acquire the null-model R distribution (Santos and

Cheylan 2013).

To test if reptile and amphibian taxonomic and functional diversity were largest at the lower fire severity plots, species richness and Shannon diversity index was quantified for taxonomic diversity (Magurran 1988). To represent functional diversity; functional richness (FRci), functional evenness (FEve) and functional divergence

(FDiv) were quantified using the distance-based functional diversity indices of the FD package (Laliberté et al.

2014) utilizing a Gower distance matrix (Díaz-García et al. 2017). Functional Richness indicates the level to which species within a community have occupied functional trait space, with greater numbers indicating species occupying greater trait space which could indicate a buffer against environmental fluctuations (Díaz-García et al. 2017, Hu et al. 2016). Functional Evenness is a measure examining the distribution of species in functional trait space (Hu et al.

2016). The values for functional evenness ranges between 0 and 1, with the lower numbers indicates lower evenness. Lower functional evenness demonstrates there is an under-utilization of resources (Mason et al. 2005).

Functional divergence indicates the degree in which the dispersion of community functional traits is maximized with the distribution of abundances (Díaz-García et al. 2017). Functional divergence also ranges between 0 and 1, with larger numbers indicating higher divergence, which could indicate increased ecosystem function (Mason et al.

2005). To test if there was a significant difference in reptile and amphibian taxonomic diversity, as well as, functional diversity between the different fire severity plots an analysis of variance (ANOVA) was performed.

To examine if reptile and amphibian taxonomic and functional diversity was positively associated with microhabitat variables, quadratic regressions were performed given assumptions of normality were violated.

Quadratic regressions were run for all microhabitat variables as the predictor variables and diversity of both reptile and amphibian species as response variables. The mean and standard deviation of all microhabitat variables in each plot were also quantified to test if species diversity was responding to certain microhabitat variables.

To test if there was a significant difference in microhabitat variables and variability between the fire severity plots, an ANOVA was performed on all microhabitat variables. A post hoc Tukey test for pairwise comparisons was also conducted. Species accumulation curves were run to visualize species captured based on sampling effort. A statistical significance of alpha value 0.05 was used to reject the null hypotheses, and all

statistical tests were performed in R 3.6.2 (R Core Team 2019). ANOSIM and species accumulation curve was conducted with the vegan package (Oksanen et al. 2019). The ggplot2 package (Wickham 2016) was used for data visualization.

Results

A total of 549 individuals were captured representing 23 different reptile and amphibian species across the

12 sampling plots (Table 3). A total of 15 individual snakes were recorded consisting of 9 different species and a total of 235 lizards from 5 different species. The most abundant reptile species recorded was Anolis carolinensis

(green anole), comprising 41.6% of total reptiles. There were 278 individual amphibians representing 9 different species. The most abundant species of amphibians was Hyla squirella (squirrel tree frog), comprising 52.8% of all amphibian individuals recorded. Anaxyrus quercicus (oak toad) was the second most abundant amphibian species comprising of 33.1% of all amphibian individuals recorded. Across the sampled plots, species richness ranged from

0 to 8 for reptile species and 2 to 6 for amphibian species. Reptile taxonomic diversity ranged from 0 to 1.73 and amphibian taxonomic diversity ranged from 0.67 to 1.56. Below are photographs of some reptiles and amphibians captured utilizing the different sampling methods (Fig. 4).

Table 3. Total number of species at the different fire severity plots.

Species Unburned Low Medium High

Snakes Cemophora coccinea 0 0 0 1 Coluber constrictor 1 2 1 0 Crotalus adamanteus 0 0 0 1 Diadophis punctatus 1 0 1 0 Drymarchon couperi 0 0 1 0 Masticophis flagellum 1 0 1 0 Micrurus fulvius 0 1 1 0 Pituophis guttatus 1 0 0 0 Pituophis melanoleucus 0 0 1 0 Lizards Anolis carolinensis 25 22 30 24 Aspidoscelis sexlineata 4 8 4 5 Plestiodon inexpectatus 4 8 7 2 Plestiodon reynoldsi 1 2 1 0

Sceloporus woodi 30 17 19 22 Amphibians Acris gryllus dorsalis 0 0 0 1 Anaxyrus quercicus 9 32 19 32 Anaxyrus terrestris 0 7 5 6 Eleutherodactylus planirostris 1 2 9 0 Gastrophryne carolinensis 1 1 1 0 Hyla femoralis 2 4 11 6 Hyla gratiosa 0 1 0 0 Hyla squirella 11 23 36 77 Lithobates capito 1 1 0 0

Fig. 4 Photographs of some reptiles and amphibians captured utilizing the different sampling methods, from left to right: Sceloporus woodi captured by visual encounter survey, Anaxyrus terrestris sighted underneath a coverboard and Pituophis melanoleucus mugitus captured by a camera trap.

Species accumulation curve

There is a plateau to the species accumulation curve for both the recorded amphibian and lizard species sampled within the plots. However, the species accumulation curve for all recorded reptile species does not plateau, indicating that not all snake species present in the plots were captured (Fig. 5).

Fig. 5 Species Accumulation Curve: black line represents all reptile species, green line represents amphibian species and red line represents lizard species, recorded in all 12 fire severity plots.

Community composition and fire severity

There were no significant differences between reptile communities across fire severity plots (R=-0.21, p=0.97; Fig. 6a). Lizard and amphibian communities also were not significantly different between the communities across fire severity plots (R=-0.18; p=0.92; Fig. 6b, R=0.07 p=0.28; Fig. 6c, respectively). The low R-value indicates that the communities are relatively similar in composition. Across the fire severity plots, the low to medium severity plots were the most similar compared with the high severity plots being the most dissimilar to all other severity plots.

A B

Fig. 6 Boxplots showing community composition of herpetofauna and fire severity, a) reptile communities, b) lizard communities, c) amphibian communities.

Species diversity and fire severity

There were no statistically significant differences between species diversity across the fire severity plots.

The reptile and amphibian taxonomic diversity revealed a trend with fire severity. Reptiles had the largest species richness (F-value=1.27, p=0.35; Fig. 7a) and taxonomic diversity at the medium fire severity plots (F-value=1.10, p=0.40; Fig. 7b). Lizard richness (F-value=1.02, p=0.43; Fig. 7c) and taxonomic diversity also revealed a similar pattern, with the greatest diversity and richness occurring at the medium severity plots (F-value=0.92, p=0.47; Fig.

7d). For both reptile and lizard taxonomic richness and diversity, the high severity plot had the lowest measurements

compared to the unburned, low, and medium severity plots. Amphibian species richness (F-value=2.08, p=0.18; Fig.

7e) and taxonomic diversity were largest at the low fire severity plots and lowest at the unburned plots (F-value=

0.51, p=0.68; Fig. 7f).

For functional diversity, reptile, lizard and amphibian responses revealed a similar pattern to taxonomic diversity. For reptile species, functional richness had the largest value at the high fire severity plot and the largest median at the medium severity plots (F-value=0.19, p=0.90; Fig. 8a). Reptile functional evenness was largest in the medium fire severity plot along with the low severity plot (F-value=2.53, p=0.13; Fig. 8a). Both the unburned and high severity plots showed a lower functional evenness for reptiles comparatively. Reptile functional divergence was also largest in the medium severity plot but showed similar divergence between all the different severity plots except in the high severity plots (F-value=0.80, p=0.53; Fig. 8a).

For lizard species, functional richness was largest among the medium severity plots and lowest at the unburned and high severity plots (F-value=1.22, p=0.36; Fig. 8b). Functional evenness for lizards was largest at the medium severity plot and lowest at the high severity plot (F-value=2.67, p=0.12; Fig. 8b). Lizard functional divergence was largest at the medium severity plot followed by the low severity plot, with the lowest divergence at the unburned and high severity plots (F-value=0.26, p=0.85; Fig. 8b). For all three functional diversity indices for lizards, the unburned and high severity plots had the lowest measurements compared to the low and medium severity plots.

Amphibian functional diversity showed functional richness was largest among the lower severity plots (F- value=0.73, p=0.56; Fig. 8c). Functional evenness was largest for amphibians among the medium severity plots (F- value=2.67, p=0.12; Fig. 8c). Amphibian functional divergence did not show a trend across the different fire severity plots (F-value=2.13, p=0.17; Fig. 8c). For all three functional diversity indices for amphibians, the unburned plots had the lowest measurements compared to the low, medium and high severity plots.

A 8 B 1.5 6

1.0 4

0.5 2 Reptile Diversity Reptile Richness

0 0.0 Un Lo Med Hi Un Lo Med Hi Fire Severity Fire Severity

C 5 D

4 1.0

2 0.5 Lizard Diversity Lizard Richness 1

0 0.0 Un Lo Med Hi Un Lo Med Hi Fire Severity Fire Severity 1.6 E 6 F

1.4 5

1.2 4 1.0 3

Amphibian Diversity 0.8 Amphibian Richness 2 Un Lo Med Hi Un Lo Med Hi Fire Severity Fire Severity

Reptile Functional Diversity Lizard Functional Diversity Amphibian Functional Diversity

A B 0.20 C 0.20

0.2 0.15 0.15

0.10 0.10 0.1 0.05 0.05 Functional Richness Functional Richness Functional Richness 0.0 0.00 0.00 Un Lo Med Hi Un Lo Med Hi Un Lo Med Hi Fire Severity Fire Severity Fire Severity

0.75 0.75 0.75

0.50 0.50 0.50

0.25 0.25 0.25 Functional Evenness Functional Evenness Functional Evenness 0.00 0.00 0.00 Un Lo Med Hi Un Lo Med Hi Un Lo Med Hi Fire Severity Fire Severity Fire Severity

1.00 1.00

0.75 0.75 0.75

0.50 0.50 0.50

0.25 0.25 0.25 Functional Divergence Functional Divergence Functional Divergence 0.00 0.00 0.00 Un Lo Med Hi Un Lo Med Hi Un Lo Med Hi Fire Severity Fire Severity Fire Severity

Fig. 8 Boxplots showing species functional diversity and fire severity. Diversity indices include; functional richness, functional evenness and functional divergence, a) reptile functional diversity, b) lizard functional diversity, c) amphibian functional diversity.

Species diversity and microhabitat variables

The sole microhabitat variable associated with species diversity was leaf litter. All other measured variables did not show any relationship with species diversity. Lizard species richness (R2=0.48, p=0.05; Fig. 9a) and taxonomic diversity were significantly associated with leaf litter depth standard deviation (R2=0.63, p=0.01; Fig.

9b). Reptile functional richness (R2=0.59, p=0.01; Fig. 9c) and divergence was also significantly associated with leaf litter depth standard deviation (R2=0.78, p<0.01; Fig. 9d).

Amphibian richness (R2=0.37, p=0.12; Fig. 10a) was not related to mean leaf litter depth, however taxonomic diversity was related to mean leaf litter depth at the different fire severity plots (R2=0.66, p<0.01;

Fig.10b). Amphibian functional evenness (R2=0.82, p<0.01; Fig. 10c) and functional divergence was associated with mean leaf litter depth (R2=0.57, p=0.02; Fig. 10d). Interestingly, all species diversity measures were greatest at the plots with an intermediate depth of leaf litter. Even though leaf litter depth was the only microhabitat variable associated with species diversity; it explains a large variation given the large R2 values.

A 5 B

4 1.0

2 0.5 Lizard Diversity Lizard Richness 1

0 0.0

0.0 0.5 1.0 1.5 0.0 0.5 1.0 1.5 Litter Depth (SD) Litter Depth (SD) C 0.9 D 1.00

0.6 0.75

0.50 0.3

0.25 0.0 Reptile Functional Eveness

Reptile Functional Divergence 0.00

−0.3 0.0 0.5 1.0 1.5 0.0 0.5 1.0 1.5 Litter Depth (SD) Litter Depth (SD)

Fig. 9 Reptile diversity related to leaf litter depth standard deviation (SD) a) lizard richness, b) lizard taxonomic diversity, c) reptile functional evenness, d) reptile functional divergence.

1.6 A 6 B

1.2

3 0.8 Amphibian Diversity Amphibian Richness

0.4 0 1 2 3 4 0 1 2 3 4 Litter Depth (cm) Litter Depth (cm) C D 1.25

0.75 1.00

0.75 0.50

0.50 0.25

0.25 0.00 Amphibian Functional Evenness Amphibian Functional Divergence 0.00 0 1 2 3 4 0 1 2 3 4 Litter Depth (cm) Litter Depth (cm)

Fig. 10 Amphibian diversity related to leaf litter depth (cm), a) amphibian richness, b) amphibian taxonomic diversity, c) amphibian functional evenness, d) amphibian functional divergence.

Microhabitat variables and fire severity

Among the microhabitat measurements that were collected in the field, there was a statistically significant difference between fire severity plots for leaf litter depth and fire severity (F-value=4.43, p=0.04; Fig. 11a). A post hoc Tukey test showed that the unburned plots and the high severity plots differed significantly at p<.05, the other fire severity plots were not significantly different. The standard deviation of percent litter was also observed to be statistically different between fire severity plots with the greatest variability at the medium severity plots (F-

value=8.01, p<0.01; Fig. 11b). A post hoc Tukey test showed the unburned plots and the medium severity plots differed significantly at p<.05. The high severity plots and the medium severity plots also differed significantly at p<.05. The high severity plots and the low severity plots differed significantly at p<.05. There were no statistically significant differences between fire severity plots for basal area of shrub (F-value=1.61, p=0.26; Fig.11c). The mean litter depth also showed statistically significant differences between fire severity showing a similar pattern to the standard deviation of leaf litter depth (F-value=14.13, p<0.01; Fig. 11d). A post hoc Tukey test showed the unburned plots and the medium severity plots differed significantly at p<.05, as well as the high severity plots compared to the low severity plots and the unburned plots. All four microhabitat variables were lowest in the high fire severity plots.

A a B b 40 1.5 ab 30 ab ab 1.0 20 a b 0.5 10 Litter Depth (SD) Percent Litter (SD) Percent c

0.0 0 Un Lo Med Hi Un Lo Med Hi Fire Severity Fire Severity C a D a 4 4 a ab 3 a 3 bc 2 2 c

Basal Area (SD) a Litter Depth (cm) 1 1

0 Un Lo Med Hi Un Lo Med Hi Fire Severity Fire Severity

Fig. 11 Boxplots for microhabitat variables at the different fire severity plots, nonmatching letters identifying significant pairwise differences, a) the standard deviation of litter depth, b) the standard deviation of percent litter, c) the standard deviation of basal area of shrub, d) mean leaf litter depth.

Discussion

The herpetofaunal species in the Florida scrub at Archbold Biological Station may not be responding directly to fire severity but instead to the indirect effects of fire severity, such as microhabitat variables. This coincides with the findings of previous researchers that examined herpetofaunal responses to time since fire and fire frequency. They demonstrated species responded to the following microhabitat variables: bare ground, canopy cover, and percent leaf litter cover (Ashton and Knipps 2011, Halstead 2007, Moseley et al. 2003, Mushinsky 1985).

Although, we show community composition, species taxonomic and functional diversity across the fire severity plots were not significantly different; the trends provide support for the intermediate disturbance hypothesis for reptile and lizard diversity. The “intermediate disturbance hypothesis” (IDH; Connell 1978) states the largest diversity will occur at the intermediate frequency of disturbance because dominant species reduce diversity in habitats with low disturbance and few species are able to endure high disturbances. The IDH encompasses not only environments subject to disturbances at an intermediate frequency but also at an intermediate intensity, such as fire intensity (Davis and Moritz 2013). Research on species responses to disturbances following the IDH have been relatively mixed, as Moi et al. (2020) discovered when conducting a literature review on the IDH. They found 499 published articles in support for the IDH and 138 that found no support, with articles not supporting the IDH to be increasing with time (Moi et al. 2020). However, there may be other drivers, such as habitat heterogeneity, that are influencing species responses to disturbance (Moi et al. 2020). In this study, species diversity demonstrated weak support for the IDH. Further support for the IDH was provided with species diversity peaking at intermediate depths of leaf litter.

Community composition and fire severity

In this study, there were no significant differences in reptile, lizard and amphibian community composition across fire severity plots. However, the high severity plots showed a trend with the greatest dissimilarity compared to the unburned, low and medium fire severity plots. Snake species contributed to the dissimilarity between reptile communities, with two-thirds of the high fire severity plots with no recorded snake species. Furthermore, two

threatened snake species, Drymarchon couperi (eastern indigo) and Pituophis melanoleucus mugitus (Florida pinesnake), were only located at medium fire severity plots. The most common snake species, Coluber constrictor

(black racer), only had four individuals captured with one recorded at an unburned plot, two at a low severity plot, and one at a medium severity plot. None were captured at the high fire severity plot, which was interesting given this species has a broad habitat range (Meshaka and Layne 2002). However, Rochester et al. (2010) also discovered similar results with C. constrictor with no individuals captured following a wildfire in California when this species was detected prior to the fire. This was potentially due to habitat preference and loss of cover. In our study, snake species varied between and within the fire severity plots causing the overall similarity in species composition across the fire severity plots.

The lizard species dissimilarities at the high severity plots are attributed to the greater abundance of

Sceloporus woodi, in addition to one-third of the high fire severity plots having no recorded lizard species. As expected, S. woodi was captured in greater abundances at higher severity plots since this species tends to prefer more open habitats (McCoy et al. 2013). In addition, two lizard species were also aiding in the dissimilarity of the high fire severity plots, Plestiodon inexpectatus was found in low abundance in the high fire severity plots compared to the other fire severity plots and Plestiodon reynoldsi was not present at any of the high severity plots. These findings are consistent with previous studies that observed these species preferring unburned habitats (McCoy et al.

2013). This preference is likely due to the high severity plots burning most of the surface litter and the delayed time interval for the leaf litter layer to return to a proportion that is sufficient for these species (Ashton and Knipps 2011,

McCoy et al. 2013). It is important to remember the dissimilarities observed in the high severity plots were a trend and not significant differences. This is most likely due to pyrogenic ecosystems selecting for species that adapt to fire. Ashton and Knipps (2011) observed similar patterns for reptile community composition when comparing time since fire in rosemary scrub habitats.

The large abundance of Hyla squirella and Anaxyrus quercicus within the higher severity plots were causing the amphibian dissimilarity. It was surprising to see the large abundance of certain species at the high severity plots, particularly H. squirella. This species seemed to prefer the new growth of palmettos for refugia that occurred at the higher severity burn plots in this study. They also may be responding to other environmental factors associated to burned habitats, such as greater prey abundance and modified predator-prey interactions as seen in this 26

study, with low abundance and richness of snake species occurring in the high fire severity plots. A. quercicus was also captured at higher severity burn plots, with both species yielding a larger proportion of juveniles. Both these species may also be utilizing these higher severity plots for an optimal growth temperature since higher severity burn plots may be more susceptible to higher temperatures on ground level and in burrows. Hossack et al. (2009) observed optimal temperatures for Bufo boreas (boreal toads) and found that high severity plots had the preferred temperature range for this species compared to that of unburned plots. A. quercicus is also able to avoid desiccation by burrowing, so is most likely able to withstand high severity plots (Moseley et al. 2003).

Although species composition was not significantly different between fire severity plots, the trends found in this study were also documented in tallgrass prairies by Wilgers and Horne (2006). The researchers showed the annual burning regime to be the most dissimilar herpetofaunal community composition compared to the long unburned, 4-year burned and unburned plots (Wilgers and Horne 2006). Pinto et al. (2018) also observed significant changes in herpetofaunal community composition to time-since-fire intervals in the Mediterranean, with notable differences between medium burnt transects and recently burned transects. While Rochester et al. (2010) found changes in herpetofaunal communities in California following a wildfire in chaparral and coastal sage, but no differences in grassland and woodland plots. Warren-Thomas et al. (2013) also documented no change in frog community composition between burnt and unburnt habitats in the forests of Peru. Overall, the results in this study support previous studies comparing herpetofaunal composition responses to fire (Ashton and Knipps 2011, Warren-

Thomas et al. 2013).

Species diversity and fire severity

In this study, there also were no significant differences in reptile, lizard and amphibian taxonomic nor functional diversity across fire severity plots. Once again, a repeated trend occurred for reptile and lizard taxonomic and functional diversity with the largest diversity at the medium fire severity plots. The lowest richness and diversity for reptile species occurred at the high severity plots, potentially due to reduced thermal quality for reptiles to utilize or another factor, like predator-prey abundance (Ferreira et al. 2016). Many of the snake species documented within this study are active foragers, which could increase their vulnerability to predators such as birds in the high severity plots, where fire tends to consume all vegetation leaving a more open habitat (Richter et al.

2019, Webb and Shine 2007). This was documented by Webb and Shine (2007) when examining the abundance of two snake species, Hoplocephalus bungaroides (broad-headed snake) and Cryptophis nigrescens (small-eyed snake), after a wildfire in Australia. They found the abundance of active foraging snake species decreased significantly following a wildfire (Webb and Shine 2007). This could account for the lower reptile taxonomic richness and diversity at the high fire severity plots since the majority of snakes captured were active foragers and not recorded in the high severity plots where activity seems less favorable; however, many of the snake species that were captured prefer open habitats so the intermediate severity plots may serve as the more favorable sites. Our results support those of Howze and Smith (2021), they observed snake site selection to burn return interval in the longleaf pine ecosystem. They documented snakes that prefer open habitats, such as Crotalus adamanteus (eastern diamondback), Pituophis melanoleucus (Florida pine snake), and Masticophis flagellum (eastern coachwhip), tend to prefer areas with an intermediate burn return interval. All recorded species avoided unburned habitats and favored an intermediate burn return interval of every 2 to 2.5 years potentially due to increased basking areas and prey availability (Howze and Smith 2021).

There also was a vast amount of variation in diversity in the high severity plots, which could be a result of a fire measure that was not measured in this study, such as fire seasonality. Greenberg et al. (2019) conducted a study in Florida longleaf pine sandhill habitats recording herpetofaunal short term responses to burn season and found

Plestiodon inexpectatus had increased capture rates two years after growing season burns compared to dormant season burns. Whereas other herpetofaunal species, like Anaxyrus quercicus showed decreased captures rate two years after growing season burns; thus, season of burn may be affecting certain species within these high severity plots in scrubby flatwood communities (Greenberg et al. 2019). Due to the weak trend observed in this study with a peak of diversity at the medium fire severity plots, reptile species are most likely weakly responding to the IDH.

Ashton and Knipps (2011) support these findings having documented Asphidoscelis sexlineata having greater abundance at sites with an intermediate time since fire in rosemary scrub habitats, since this species is most likely following habitat structure differences, such as percent bare ground. They found a polynomial relationship with time since fire and percent bare ground with the largest percent bare ground at the intermediate levels of time since fire, thus the intermediate disturbance frequency is causing the larger abundance of this species at the intermediate levels since they also prefer open habitats (Ashton and Knipps 2011). However, not all documented 28

species supported the IDH, with many species responding differently to disturbances in different types of habitats, such as sandhill and sand pine scrub habitats (Ashton and Knipps 2011, Mushinsky 1985). Lindenmayer et al.

(2008) documented reptile taxonomic responses to wildfire in Australia and found no relationships between time since fire, number of fires, or the severity of fires and that reptile species were associated with vegetation. Hu et al.

(2013) also observed the same pattern in temperate forests with reptile species evenness responding instead to fine scale vegetation structure. Thus, fire seems to impact species differently depending on the species and the fire’s characteristics for example, fires can enhance regrowth of vegetation which could increase foraging opportunities of herbivores and insectivorous reptiles (Roe et al. 2019). That could also be the reason for seeing smaller taxonomic diversity at the unburned plot since only few lizard species, such as Anolis carolinensis, are able to specialize in these undisturbed habitats due to habitat preference and species-specific behaviors, thus fire at an intermediate level creates more opportunities for more species to coexist (Davis and Moritz 2013, Roe et al. 2019).

Reptile functional diversity appears to have also followed the same pattern as taxonomic diversity, with the largest values occurring at the medium fire severity plots and the lowest at the unburned plots. Reptile functional diversity did not differ significantly between fire severity plots, which was also seen in by Hu et al. (2016). They attributed their findings to the communities being relatively species-poor with four dominant generalist species with similar traits. This could also be the case for this study, since there were four lizard species dominating the communities that had relatively similar traits. However, Chergui et al. (2020) showed functional responses of reptiles are habitat dependent, with functional richness increasing with fire in pine plantations but did not change in cork oak forests. Thus, the Florida scrub may be a habitat that does not show any significant differences in functional diversity across fire severity, potentially due to species being adapted to these pyrogenic habitats (Ashton and Knipps, 2011), which selects for species with similar fire-adapted traits (Gainsbury and Colli 2019). Although, reproductive mode was used in this study, other fecundity measures may lead to different results since some studies observed species with larger clutch sizes responding linearly to disturbance (Hu et al. 2020).

Amphibian taxonomic diversity showed a different pattern with the largest diversity at the lower fire severity plots. Amphibians were expected to respond differently than reptiles due to their dispersal patterns, restricted geographical ranges, and complex life histories all which could make them particularly vulnerable to fire and management practices (Pilliod et al. 2003). Amphibians have moist skin that is permeable, so most rely on 29

moisture or wet habitats to survive since most amphibian species find refugia in moist environments like riparian zones and half of their life cycle consists of developing in the water (Greenberg and Waldrop 2008). This could be an indicator of why amphibian diversity does not seem to change when it comes to fire. Amphibians have also been shown to be rather resilient to fires and this study also demonstrates that with the higher richness and diversity at the burned plots compared to the unburned plots (Munoz et al. 2019). This contradicts what Hromada et al. (2018) discovered when comparing amphibian responses to unburned and burned plots in an oak/hickory forest. They found higher abundance of amphibian species at the unburned plots however these responses were most likely not due to prescribed fire but another environmental factor, such as wetland accessibility.

As expected, there was an increase in amphibian taxonomic diversity at the lower fire severity plots in this study, however due to the peak in diversity at the low fire severity plot amphibian species are most likely not following the IDH. Westgate et al. (2012) supports these findings when they studied anuran responses to fire in southeastern Australia and whether or not they follow the IDH. Their results indicated no evidence that anuran species richness was related to fire return interval at an intermediate level of disturbance, since there was neither an increase nor decrease in species richness at differing fire return intervals (Westgate et al. 2012). On the other hand,

Schurbon and Fauth (2004) discovered that one amphibian species in particular, Hyla femoralis (pinewoods treefrog), had the largest abundance at an intermediate level of years since burn in forests of South Carolina, stating that these species were fire-dependent thus for the increase at the intermediate level. Ultimately more research needs to be done on the IDH and fire severity. It is also hard to determine what is actually considered an intermediate level of natural disturbance, in this case it could be either considered a low or medium severity plots or neither since this study was conducted after prescribed burns instead of natural fires (Lindenmayer et al. 2008). Especially, given that species from the genera Bufo and Scaphiopus were observed to have larger capture rates following a prescribed burn compared to that of a wildfire (Brown 2013).

Amphibian functional richness showed a similar trend to taxonomic diversity, with the largest values for functional richness occurring at the low fire severity plot. For functional evenness, the medium severity plots had the largest functional evenness values with functional divergence not showing much of a trend between the different fire severity plots. However, amphibian functional diversity was not significantly different between the different fire severity plots potentially due to the plots being relatively patchy following prescribed burning (Hu et al. 2020). This 30

also coincides with Hu et al. (2020) finding no differences in amphibian functional diversity from time since fire between plots. Also, since vegetation tends to grow back relatively quickly, with new shoots sprouting out almost immediately following a fire and taking around 1-4 years to return to pre-burn condition, perhaps amphibian species are able to recolonize from adjacent areas and because only few burrowing species are able to persist at the high levels of intensity, species will recolonize in the lower fire severity plots, thus the larger diversity (Abrahamson et al. 1984, Westgate et al. 2012). Although, amphibian taxonomic and functional diversity was smallest at the unburned plots, these unburned plots also had the greatest variation in diversity compared with the other fire severity plots, potentially due to the variation in lichen abundance only in the unburned plots, which could account for the observed variation since, lichen is a significant element of ground cover than can also become interlaced with insects and leaf litter providing refuge for both prey and predators (Hawkes and Menges 2003).

It is interesting to note that functional responses for both reptile and amphibian species were following the same pattern as taxonomic diversity. Species richness and functional richness resembled the same pattern, which has been seen in other studies exploring at both taxonomic and functional responses of species. Tsianou and Kallimanis

(2019) studied amphibian diversity across environmental drivers in Europe. They documented functional richness of amphibians decreased with higher latitudes and mirrored species taxonomic richness due to their restricted thermal ranges (Tsianou and Kallimanis 2019). Hu et al. (2020) had similar conclusions when looking at herpetofaunal diversity to time since fire; however, attributed it to species richness remaining relatively constant between disturbance ages thus not seeing a difference between taxonomic species richness and functional richness.

Functional richness was also relatively low for reptile, lizard and amphibian species in this study, which may indicate a lower buffer against environmental fluctuation (Díaz-García et al. 2017, Hu et al. 2016). However, the lower functional richness may also be due to the low number of species captured within the plots since taxonomic species richness has a strong effect on functional richness (Cadotte et al. 2011, Tsianou and Kallimanis 2019). This may not always be the case given that Santillan et al. (2019) observed bird species richness in Ecuador to increase with forest fragmentation, while functional richness decreased with fragmentation at low elevations showing a decoupled response. Although, elevation appears to be the contributing factor for the observed response in their study (Santillan et al. 2019), responses between taxonomic and functional diversity metrics can vary between habitats, disturbances, and environmental gradients (Cadotte et al. 2011).

Both functional evenness and functional divergence for reptile and amphibian species yielded a trend of greater functional diversity in both the medium and low plots, respectively. These species are most likely utilizing all the resources in these intermediate burnt habitats. Biswas and Mallik (2010) also came to the same conclusions when observing different anthropogenic disturbances to both species’ diversity and functional diversity of upland plant communities in Canada. Both species taxonomic diversity and functional diversity were following the same pattern with abundance peaking at intermediate levels of disturbance intensity, thus also following the IDH (Biswas and Mallik 2010). Although, there was not significant differences in functional diversity and fire severity, the traits used in this study were good representation of traits for these species in the Florida scrub at Archbold Biological

Station since these are traits that have been used in many functional diversity studies for reptile and amphibians

(Chergui et al. 2020, Hu et al. 2016, Tsianou and Kallimanis 2019) providing support that the herpetofauna species diversity is not responding to the direct effects of fire, but instead to the indirect effects of fire. Ultimately, the low to medium burn plots most likely provide the greatest thermal heterogeneity and microhabitats, in particular leaf litter depth, for reptiles and amphibians to utilize resulting in greater richness and diversity (Hossack et al. 2009).

Species diversity and microhabitat variables

Of the microhabitat variables that were measured in this study, leaf litter depth appeared to be influencing species diversity in this study. Leaf litter depth has been documented throughout different biogeographic regions, such as Costa Rica, India, and Uganda, to be correlated with herpetofaunal diversity (Balaji et al. 2014, Fauth et al.

1989, Vonesh 2001, Whitfield et al. 2014). Leaf litter accumulation is an important food source for small lizards feeding on arthropods (Hu et al. 2013). It can also provide important shelter use for many species and protection from predators, such as birds (Whitfield et al. 2014). It also provides a buffer of high humidity and stable temperatures (Whitfield et al. 2014). Leaf litter may also promote optimal habitats for egg laying for many of these reptile species (Whitfield et al. 2014). Reptile taxonomic and functional diversity was expected to be positively associated with microhabitat variables, since researchers have documented the more microhabitats for species to utilize the greater the diversity (Price et al. 2010). However, for lizards, species richness and diversity along with reptile functional evenness and divergence all appeared to be greatest at the plots with medium leaf litter depth standard deviation.

Amphibian taxonomic and functional diversity was also expected to be positively associated with microhabitat variables, however amphibian diversity, functional evenness and functional divergence responded directly to the depth of the leaf litter and showed greatest diversity at the intermediate depths. Loss of leaf litter and cover in microhabitats could also increase the risk of desiccation in amphibians by increasing exposure to higher temperatures and reducing moisture (Pilliod et al. 2003). This desiccation could lead to increased physiological stress and predation (Pilliod et al. 2003). Reduction in moisture has been a large factor in decreased amphibian abundance thus, amphibians may not be as prevalent in high severity plots where microhabitats and moisture tend to decline (Moseley et al. 2003). Distance to pond was not a significant covariate for amphibian species responses, but some species may be able to reproduce in ephemeral sources of water (Greenberg et al. 2018).

Richter et al. (2019) found greater richness and diversity of plants in medium severity plots compared to low and high severity plots in mixed conifer forests. They attributed this response to species being adapted to frequent low severity fires and these species not being able to respond to high severity fires causing a decrease in diversity (Richter et al. 2019). Reptile and amphibian species may also be responding to the complexity of vegetation with the greatest microhabitat variability at an intermediate level. The species in this study may not prefer certain types of general homogenous habitats, but more of a balance between open and closed habitats (McCoy et al.

2013). This may account for why species diversity was greatest at the intermediate microhabitat variable level.

Especially, given that greater leaf litter cover has resulted in lower diversity of the understory, which leads to homogenization as seen by Richter et al. (2019).

Microhabitat variables and fire severity

Microhabitat variables and variability decreased with increasing fire severity with variables being the largest at the lower severity plots and smallest at the higher severity plots but only for the standard deviation of leaf litter and the mean leaf litter depth, which were the only variables significantly influencing species diversity. Fire severity effecting microhabitat variables have also been documented by Costa et al. (2020), comparing habitat structure to fire severity. They found that unburned plots had increased canopy cover, tree density and leaf-litter weight while those habitat structure variables decreased with increasing fire severity. However, they also found that some variables, such as ground exposure, increased with fire severity then dropped at the high fire severity plots

(Costa et al. 2020). Percent litter showed a different pattern with increased variability with increasing fire severity, but then dropped at the high severity plots. Basal area also followed the same pattern as percent litter, although it was not significantly different between the fire severity plots. This is most likely the cause of high severity fires causing everything to be consumed resulting in decreased microhabitat variables, and in turn decreased species diversity (Richter et al. 2019). That is also why litter depth is lowest at the high severity plots. The unburned and high severity plots were responding similarly, but species appeared to prefer the intermediate variables occurring in the low to medium fire severity plots. Ultimately, the different fire severity plots are significantly influencing microhabitat variables and variability that are important factors for herpetofaunal diversity.

Limitations and future research

This research was only conducted over a single breeding season with limited sampling of plots due to both monetary and time constraints. Due to the limited sampling period, this was an underrepresentation of the herpetofaunal community, particularly the snake species. Further research is needed to better understand the effects of fire severity on snakes. Studies should also be done over longer sampling periods, utilizing other sampling methods such as pitfall and funnel traps to capture temporal shifts. Sampling considerably more sites may also aid in reducing the observed variability. Future studies should also collect the distance of potential breeding sources for amphibians, since these parameters could influence some of the observed species’ responses. Especially, since

Robertson et al. (2018). observed increased population sizes of amphibian species with increases fire intensity due to potential increases in nutrient availability in the Florida scrub. Fire frequency, fire history, seasonality and time since fire should all be examined in conjunction with fire severity in future studies to further understand the relationship of fire on herpetofaunal diversity.

Conclusion

Although, it appears that fire severity does not have significant direct effects on herpetofaunal species diversity in the Florida scrub habitats of the Lake Wales Ridge, it does have indirect effects on microhabitat variables that are important for these communities. Increased fire severity could impact leaf litter depth which could have detrimental effects on herpetofaunal diversity. In addition, there are particular species that may be more

vulnerable to fire severity in the face of climate change. The threatened Plestiodon reynoldsi is one of these species and should continue to be closely monitored. This study highlights the importance of fine-scale microhabitat variables as an important indicator for biodiversity conservation. The current prescribed fire management at

Archbold Biological Station should continue to burn these areas frequently with varying low to medium fire intensities to aid multiple herpetofaunal species by providing adequate microhabitats which in turn, will increase richness and diversity of native species and species of concern.

References

Abrahamson, W., Johnson, A., Layne, J., and Peroni, P. (1984). Vegetation of the Archbold Biological Station,

Florida: An Example of the Southern Lake Wales Ridge. Florida Scientist, 47 (4): 209-250.

Archbold Biological Station. (2018). About Us. Retrieved from http://www.archbold-station.org/

Archibald, S., Bond, W., Stock, W., and Fairbanks, D. (2005). Shaping the Landscape: Fire-Grazer Interactions in

an African Savanna. Ecological Applications, 15, 96-109. http://doi.org/10.1890/03-5210

Ashton, K. and Knipps, A. (2011). Effects of Fire History on Amphibian and Reptile Assemblages in Rosemary

Scrub. Journal of Herpetology, 45 (4), 497-503.

Balaji, D., Raachakonda, S., and Rao, S. (2014). Drivers of reptile and amphibian assemblages outside the protected

areas of Western Ghats, India. Journal for Nature Conservation, 22(4), 337-341.

http://doi.org/10.1016/j.jnc.2014.03.004

Biswas, S. and Mallik, A. (2010). Disturbance effects on species diversity and functional diversity in riparian and

upland plant communities. Ecology, 91 (1), 28-35.

Bohm, M., Collen, B., Baillie, J., Bowles, P., Chanson, J., Cox, N., … Zug, G. (2013). The conservation status of

the world’s reptiles. Biological Conservation, 157, 372-385. https://doi.org/10.1016/j.biocon.2012.07.015 35

Brown, D. (2013). Impacts of Low, Moderate and High Severity Fire on Herpetofauna and their Habitat in a

Southern USA Mixed Pine/hardwood Forest. Graduate Thesis and Dissertations

Cadotte, M., Carscadden, K., and Mirotchnick, N. (2011). Beyond species: functional diversity and the maintenance

of ecological processes and services. Journal of Applied Ecology, 48, 1079-1087.

https://doi.org/10.1111/j.1365-2664.2011.02048.x

Chergui, B., Pleguezuelos, J., Fahd, S., and Santos, X. (2020). Modelling functional response of reptiles to fire in

two Mediterranean forest types. Science of the Total Environment, 732, 139205.

http://doi.org/10.1016/j.scitotenv.2020.139205

Clarke, K. (1993). Non-parametric multivariate analyses of changes in community structure. Austral Ecology, 18

(1), 117-143. http://doi.org/10.1111/j.1442-9993.1993.tb00438.x

Costa, B., Pantoja, D., Sousa, H., Queiroz, T., and Colli, G. (2020). Long-term, fire-induced changes in habitat

structure and microclimate affect Cerrado lizard communities. Biodiversity and Conservation, 29, 1659-

1681.

Connell, J. (1978). Diversity in Tropical Rain Forests and Coral Reefs. Science, 199 (4335), 1302-1310.

Davis, F. and Moritz, M. (2013). Disturbance, Mechanism of. Encyclopedia of Biodiversity, (2), 562-567.

http://doi.org/10.1016/B978-0-12-384719-5.00034-4

Díaz-García, J., Pineda, E., Barrera, F., and Moreno, C. (2017). Amphibian species and functional diversity as

indicators of restoration success in tropical montane forest. Biodivers Conserv, 26, 2569-2589.

http://doi.org/10.1007/s10531-017-1372-2

Duncan, B., Schmalzer, P., Breininger, D, and Stolen, E. (2015). Comparing fuels reduction and patch mosaic fire

regimes for reducing fire spread potential: A spatial modeling approach. Ecological Modelling, 314, 90-99.

http://doi.org/10.1016/j.ecolmodel.2015.07.013

Elzer, A., Pike, D., Webb, J., Hammill, K., Bradstock, R., and Shine, R. (2013). Forest-fire regimes affect

thermoregulatory opportunities for terrestrial ectotherms. Austral Ecology, 38, 190-198.

http://doi.org/10.1111/j.1442-9993.2012.02391.x

Erwin, W. and Stasiak, R. (1979). Vertebrate Mortality During the Burning of a Reestablished Prairie in Nebraska.

The American Midland Naturalist, 101(1), 247-249. http://doi.org/10.2307/2424922

Fauth, J, Crother, B., and Slowinski, J. (1989). Elevational Patterns of Species Richness, Evenness, and Abundance

of the Costa Rican Leaf-Litter Herpetofauna. Biotropica, 21(2), 178-185.

Ferreira, C., Santos, X., and Carretero, M. (2016). Does ecophysiology mediate reptile resonses to fire regimes?

Evidence from Iberian lizards. PeerJ 4, 2107. http://doi.org/10.7717/peerj.2107

Fontaine, J. and Kennedy, P. (2012). Meta-analysis of avian and small-mammal response to fire severity and fire

surrogate treatments in U.S. fire-prone forests. Ecological Applications (22), 5, 1547-1561.

http://doi.org/10.1890/12-0009.1

Gainsbury, A. and Colli, G. (2019). Phylogenetic community structure as an ecological indicator of anthropogenic

disturbance for endemic lizards in a biodiversity hotspot. Ecological Indicators, 103, 766-773.

http://doi.org/10.1016/j.ecolind.2019.03.008

Greenberg, C., Zarnoch, S., and Austin, J. (2019). Short-term response to season of burn by amphibians and reptiles

in a Florida longleaf pine-wiregrass sandhill, Can.J. For. Res., 49, 1580-1589.

http://doi.org/10.1139/cjfr-2019-0219

Greenberg, C., Moorman C., Matthews-Snoberger, C., Waldrop, T., Simon, D., Heh, A., and Hagan, D. (2018).

Long-Term Herpetofaunal Response to Repeated Fuel Reduction Treatments. The Jornal of Wildlife

Management, 82(3), 553-565. http://doi.org/10.1002/jwmg.21402

Greenberg, C. and Waldrop, T. (2008). Short-term response of reptiles and amphibians to prescribed fire and

mechanical fuel reduction in a southern Appalachian upland hardwood forest. Forest Ecology and

Management, 255, 2883-2893. http://doi.org/10.1016/j.foreco.2008.01.064

Halstead, N. (2007). Long term effects of prescribed fire on reptile and amphibian communities in Florida sandhill.

Graduate Thesis and Dissertations. http://scholarcommons.usf.edu/etd/2199

Hawkes, C. and Menges, E. (2003). Effects of Lichens on Seedling Emergence in a Xeric Florida Shrubland.

Southeastern Naturalist, 2 (2), 223-234.

Hossack, B., Eby, L., Guscio, G., Corn, P. (2009). Thermal Characteristics of Amphibian Microhabitats

in a Fire-disturbed Landscape. Forest Ecology and Management, 258, 1414-1421.

http://doi.org/10.1016/j.foreco.2009.06.043

Howze, J. and Smith, L. (2021). The influence of prescribed fire on site selection in snakes in the longleaf pine

ecosystem. Forest Ecology and Management, 481. http://doi.org/10.1016/j.foreco.2020.118703

Hromada, S., Howey, C., Dickinson, M., Perry, R., Roosenburg, W., and Gienger, C.M. (2018). Response of reptile

and amphibian communities to the reintroduction of fire in an oak/hickory forest. Forest Ecology and

Management, 428, 1-13. http://doi.org/10.1016/j.foreco.2018.06.018

Hu, Y., Doherty, T., and Jessop, T. (2020). How influential are squamate reptile traits in explaining population

responses to environmental disturbance? Wildlife Research, 47 (3), 249-259.

http://doi.org/10.1071/WR19064

Hu, Y., Kelly, L., Gillespie, G. and Jessop, T. (2016). Lizard response to forest fire and timber harvesting:

Complementary insights from species and community approaches. Forest Ecology and Management, 379,

206-215. http://doi.org/10.1016/j.foreco.2016.07.040

Hu, Y., Magaton, S., Gillespie, G. and Jessop, T. (2013). Small reptile community responses to rotational logging.

Biological Conservation, 166, 76-83. http://doi.org/10.1016/j.biocon.2013.05.019

Hu, Y., Urlus, J., Gillespie, G., Letnic, M., and Jessop, T. (2013). Evaluating the role of fire disturbance in

structuring small reptile communities in temperate forests. Biological Conservation, 22, 1949-1963.

http://doi.org/10.1007/s10531-013-0519-z

Jacobsen, C., Brown, D., Flint, W., Schuler, J., and Schuler, T. (2020). Influence of Prescribed fire and forest

structure on woodland salamander abundance in the central Appalachians, USA. Forest Ecology and

Management, 468. http://doi.org/10.1016/j.foreco.2020.118185

Keeley, J. (2009). Fire intensity, fire severity and burn severity: a brief review and suggested usage. International

Journal of Wildland Fire, 18, 116-126. Doi:10.1071/WF07049

Kuchinke, D., Stefano, J., Sitters, H., Loyn, R., Gell, P., and Palmer, G. (2020). Prescribed burn severity has

minimal effect on common bird species in a fire-prone forest ecosystem. Forest Ecology and Management,

475. http://doi.org/10.106/j.foreco.2020.118437

Laliberté, E., Legendre, P., and Shipley, B. (2014). FD: measuring functional diversity from multiple traits, and

other tools for functional ecology. R package version 1.0-12.

Leavitt, D. and Schalk, C. (2018). Functional perspectives on the dynamics of desert lizard

assemblages. Journal of Arid Environments, 150, 34-41. http://doi.org/10.1016/j.jaridenv.2017.11.014

Lemmon, P. (1957). A New Instrument for Measuring Forest Overstory Density. Journal of Forestry, 55(9): 667-

668.

Lindenmayer, D., Wood, J., MacGregor, C., Michael, D., Cunningham, C., Crane, M., Montague-Drake, R., Brown,

D., Muntz, R., and Driscoll, D. (2008). How Predictable are reptile responses to wildfire? Oikos, 117 (7),

1086-1097. http://doi.org/10.1111/j.2008.0030-1299.16683.x.

Magurran, A. (1988) Ecological Diversity and Its Measurement. Springer Netherlands. Retrieved from

http://search.ebscohost.com.ezproxy.lib.usf.edu

Main, K. and Menges, E. (1997). Station Fire Management Plan. Land Management Publication, 97.

Martin, S., Rautsaw, R., Robb, F., Bolt, M., Parkinson, C., and Seigel, R. (2017). Set AHDriFT: Applying Game

Cameras to Drift Fences for Surveying Herpetofauna and Small Mammals. Wildlife Society Bulletin, 41 (4),

804-809. http://doi.org/10.1002/wsb.805

Mason, N., Mouillot, W., Lee, W., and Wilson, J. (2005). Functional richness, functional evenness and functional

divergence: the primary components of functional diversity. OIKOS, 111: 112-118.

McCoy, E., Britt, E., Catenazzi, A. and Mushinsky, H. (2013). Fire and Herpetofaunal Diversity in the Florida Scrub

Ecosystem. Natural Areas Journal, 33, (3): 316-326. http://doi.org/10.3375/043.033.0310

McCoy, E. and Mushinsky, H. (1994). Effects of Fragmentation on the Richness of Vertebrates in the Florida Scrub

Habitat. Ecology, 75 (2), 446-457.

Menges, E., Main, K., Pickert, R., and Ewing, K. (2017). Evaluating a Fire Management Plan for Fire Regime Goals

in a Florida Landscape. Natural Areas Journal, 37 (2), 212-227.

Meshaka, W. and Layne, J. (2002). Herpetofauna of a Long-Unburned Sandhill Habitat in South-Central Florida.

Florida Scientist, 65 (1), 35-50. http://doi.org/131.247.112.3

Moi, D., Garcia-Rios, R., Hong, Z., Daquila, B., and Roger, M. (2020). Intermediate Disturbance Hypothesis in

Ecology: A Literature Review, Annales Zoologici Fennici, 57 (1-6), 67-78.

http://doi.org/10.5735/086.057.0108

Moseley, K., Castleberry, S., and Schweitzer, S. (2003). Effects of prescribed fire on herpetofauna in bottomland

hardwood forests. Southeastern Naturalist, 2(4), 475-486.

Munoz, A., Felicisimo, A., and Santos, X. (2019). Assessing the resistance of a breeding amphibian community to a

large wildlife. Acta Oecologica, 99, 103439. http://doi.org/10.1016/j.actao.2019.06.002

Mushinsky, H. (1985). Fire and The Florida Sandhill Herpetofaunal Community: With Special Attention to

Responses of Cnemidophorus sexlineatus. Herpetologica, 41 (3), 333-342

Oksanen, J., Blanchet, G., Friendly, M., Kindt, R., Legendre, P., McGlinn, D., Minchin, P., O'Hara, R., Simpson, G.,

Solymos, P., Stevens, H., Szoecs, E., and Wagner, H. (2019). vegan: Community Ecology Package. R

package version 2.5-6. https://CRAN.R-project.org/package=vegan

Penman, T., Keith, D., Elith, J., Mahony, M., Tingley, R, Baumgartner, J., and Regan, T. Interactive effects of

climate change and fire on metapopulation viability of a forest-dependent frog in south-eastern Australia.

Biological Conservation, 190, 142-153. http://doi.org/10.1016/j.biocon.2015.05.020

Perry, R., Rudolph, D., and Thill, R. (2009). Reptile and Amphibian Responses to Restoration of Fire-Maintained

Pine Woodlands. Restoration Ecology (17): 6, 917-927. http://doi.org/10.1111/j.1526-100X.2009.00521.x

Pilliod, D., Bury, R., Hyde, E., Pearl, C., and Corn, P. (2003). Fire and amphibians in North America. Forest

Ecology and Management, 178, 163-181. http://doi.org/10.1016/S0378-1127(03)00060-4

Pinto, T., Moreira, B., Freitas, H., and Santos, X. (2018). The role of fire history, land-use, and vegetation structure

on the response of Mediterranean lizards to fire. Fire Ecology and Management, 419-420, 139- 145.

http://doi.org/10.1016/j.foreco.2018.03.029

Price, B., Kutt, A.S., and McAlpine, C.A. (2010). The importance of fine-scale savanna heterogeneity for reptiles

and small mammals. Biological Conservation, 143, 2504-2513. http://doi.org/10.1016/j.biocon.2010.06.017

QGIS Development Team (2018). QGIS Geographic Information System. Open Source Geospatial Foundation

Project. http://qgis.osgeo.org

R Core Team (2019). R: A language and environment for statistical computing. R Foundation for Statistical

Computing, Vienna, Austria. URL: https://www.R-project.org/

Richter, C., Rejmanek, M., Miller, J., Welch, K., Weeks, J., and Safford, H. (2019). The species diversity x fire

severity relationship is hump-shaped in semiarid yellow pine and mixed conifer forests. Ecosphere, 10 (10),

e02881. http://doi.org/10.1002/ecs2.2882

Roberts, S., Wagtendonk, J., Miles, A., Kelt, D., and Lutz, J. (2008). Modeling the Effects of Fire Severity and

Spatial Complexity on Small Mammals in Yosemite National Park, California. Fire Ecology (4), 2, 83-104.

http://doi.org/10.4996/fireecology.0402083

Robertson, J., Fitzpatrick, S., Rothermel, B., and Chan, L. (2018). Fire does not strongly affect genetic diversity or

structure of a common treefrog in the endangered Florida Scrub. Journal of Heredity, 109 (3),243-252.

http://doi.org/10.1093/jhered/esx088

Rochester, C., Brehme, C., Clark, D., Stokes, D., Hathaway, S., and Fisher, R. (2010). Reptile and Amphibian

Responses to Large-Scale Wildfires in Southern California. Journal of Herpetology, 44(3), 333-351.

Roe, J., Wild, K., and Chavez, M. (2019). Responses of a forest-dwelling terrestrial turtle, Terrapene Carolina, to

prescribed fire in a Longleaf Pine ecosystem. Forest Ecology and Management, 432, 949-956.

http://doi.org/10.1016/j.foreco.2018.10.026

Santillan, V., Quitian, M., Tinoco, B., Zarate, E., Schleuning, M., Bohning-Gaesa, K., and Neuschulz, E. (2019).

Different responses of taxonomic and functional bird diversity to forest fragmentation across an elevational

gradient. Oecologia, 189, 863-873.

Santos, X and Cheylan, M. (2013). Taxonomic and functional response of a Mediterranean reptile assemblage to a

repeated fire regime. Biological Conservation, 168, 90-98. http://doi.org/10.1016/j.biocon.2013.09.008

Schurbon, J. and Fauth, J. (2004). Fire as Friend and Foe of Amphibians: A Reply. Conservation Biology, 18(4),

1156-1159.

Taylor, J., Ellis, M., Williams, N., and Kloecker, U. (2020). Responses of Birds and Reptiles in Warrumbungle

National Park after the Extensive 2013 Wildfire. Proceedings of the Linnean Society of New South Wales,

141.

Timm, S., Wolf, A., Gao, X., and Kellner, K. (2020). Assessing multi-scale habitat relationships and responses to

forest management for cryptic and uncommon herpetofauna in the Missouri Ozarks, USA. Forest Ecology

and Management, 460. http://doi.org/10.1016/j.foreco.2020.117892

Tsianou, M. and Kallimanis, A. (2019). Geographical patterns and environmental drivers of functional diversity and

trait space of amphibians of Europe. Ecological Research, 35, 123-138.

http://doi.org/10.1111/1440-1703.12069

Vonesh, J. (2001). Patterns of Richness and Abundance in a Tropical African Leaf-litter Herpetofauna. Biotropica,

33(3), 502-510.

Warren-Thomas, E., Menton, M., Huaman, J., Vargas, R., Wadley, E., Price, N., and Axmacher, J. (2013). Frog

Communities in Fire-Disturbed Forests of the Peruvian Amazon. Herpetological Bulletin, 126, 14-24.

Webb, J. and Shine, R. Different Effects of an Intense Wildfire on Survival of Sympatric Snakes. The Journal of

Wildlife Managament, 72 (6), 1394-1398. http://doi.org/10.2193/2007-515

Weekley, C., Menges, E., and Pickert, R. (2008). An Ecological Map of Florida’s Lake Wales Ridge: A New

Boundary Delineation and an Assessment of Pos- Columbian Habitat Loss. Florida Scientist, 71(1):45-64.

Westgate, M., Driscoll, D., and Lindenmayer, D. (2012). Can the intermediate hypothesis and information on

species traits predict anuran responses to fire? Oikos, 121, 1516-1524.

http://doi.org/10.1111/j.1600-0706.2011.19863.x

Whitfield, S., Reider, K., Greenspan, S, and Donnelly, M. (2014). Litter Dynamics Regulate Population Densities in

a Declining Terrestrial Herpetofauna, Copeia, 3, 454-461. http://doi.org/10.1643/CE-13-061

Wickham, H. ggplot2: Elegant Graphics for Data Analysis. Springer-Verlag New York, 2016.

Wilgers, D. and Horne, E. (2006). Effects of Different Burn Regimes on Tallgrass Prairie Herpetofaunal Species

Diversity and Community Composition in the Flint Hill, Kansas. Journal of Herpetology, 40 (1), 73-84.

Zuniga, A., Rau, J., Jaksic, F., Vergara, P., Encina-Montoya, F., and Fuentes-Ramirez, A. (2021). Rodent

assemblage composition as indicator of fire severity in a protected area of south-central Chile. Austral

Ecology, 46, 249-260. http://doi.org/10.1111/aec.12975

Appendices

Appendix 1

Appendix 2

MEMORANDUM TO: Michelle Mark,

FROM: Farah Moulvi, MSPH, IACUC Coordinator Institutional Animal Care & Use Committee Research Integrity & Compliance DATE: 6/10/2020 Fire Severity Effects on Herpetofaunal Diversity in Florida Scrub Communities of the PROJECT TITLE: Lake Wales Ridge

USF department, institute, center, etc. FUNDING SOURCE:

IACUC PROTOCOL #: W IS00008013 PROTOCOL STATUS: APPROVED

The Institutional Animal Care and Use Committee (IACUC) reviewed your application requesting the use of animals in research for the above-entitled study. The IACUC APPROVED your request to use the following animals in your protocol for a one-year period beginning 6/10/2020:

Reptiles (Any age or sex. Mostly small bodied reptiles 500 will be sampled. )

Amphibians (Any amphibians that we are able to see.) 200

Please take note of the following:

• IACUC approval is granted for a one-year period at the end of which, an annual renewal form must be submitted for years two (2) and three (3) of the protocol through the eIACUC system. After three years, all continuing studies must be completely re-described in a new electronic application and submitted to IACUC for review.

• All modifications to the IACUC-Approved Protocol must be approved by the IACUC prior to initiating the modification. Modifications can be submitted to the IACUC for review and approval as an Amendment or Procedural Change through the eIACUC system. These changes must be within the scope of the original research hypothesis, involve the original species and justified in writing. Any change in the IACUC-approved protocol that does not meet the latter definition is considered a major protocol change and requires the submission of a new application.

• All costs invoiced to a grant account must be allocable to the purpose of the grant. Costs allocable to one protocol may not be shifted to another in order to meet deficiencies caused by overruns, or for other reasons convenience. Rotation of charges among protocols by month without establishing that the rotation schedule credibly reflects the relative benefit to each protocol is unacceptable.

• The PI must assist the IACUC with tracking wild animal field research activities. The PI must report episodes of wild animal use, the approximate range of taxa, and the approximate numbers of animals encountered or used at intervals appropriate to the study but at least once a year.