HERPETOFAUNAL RESPONSES TO PRESCRIBED FIRE IN UPLAND PINE COMMUNITIES OF NORTHEAST FLORIDA By KEITH CHARLES MORIN A THESIS PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLOR IDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF MASTER OF SCIENCE UNIVERSITY OF FLORIDA 2005
Copyright 2005 by Keith Charles Morin
This document is dedicated to my wife, Alexandra. Without her patience, love, encouragement and support I could not have re turned to school to pursue graduate studies in natural science.
iv ACKNOWLEDGMENTS I would like to thank my major advisor, Max Nickerson, for allowing me to study with him and supporting this project and my gr aduate studies. I would also like to thank the other members of my committee, Harv ey Lillywhite and Mel Sunquist, for their support and advice during this process. I thank my parents, Edward H. and Claire B. Morin, for their continued support of my schooling and career. I also appreciate crucial support of this project from my brother, Craig. The Reptile and Amphibian Conservation Corps (RACC) and the Florida Museum of Natural History provided funding for this project. Sam Jones in the Department of Wildlife Ecology and Conservation facilitated transportation. I appreciate the assistance and infrastructure support of the environm ental staff at Camp Blanding, including Jaclyn Wickert, Matthew Corby, Paul Catlett and Phil Hall. I thank the Camp Blanding Range Control personnel for ensuring sa fe access and travel on base. Special thanks to the major field assi stant on this study, R achael Hallman, who volunteered a lot of time and effort for the success of this project.
v TABLE OF CONTENTS page ACKNOWLEDGMENTS.................................................................................................iv LIST OF TABLES............................................................................................................vii LIST OF FIGURES.........................................................................................................viii ABSTRACT....................................................................................................................... ix CHAPTER 1 INTRODUCTION........................................................................................................1 Area of Study................................................................................................................1 Research Problem and Motivation................................................................................3 Objectives..................................................................................................................... 4 Critical Assumptions....................................................................................................4 Structure of Study.........................................................................................................5 2 THEORETICAL FRAMEWORK AND SYSTEM.....................................................7 Review of Fire Ecology................................................................................................7 Herpetofauna and Fire..................................................................................................9 General Effects......................................................................................................9 Herpetofauna.......................................................................................................10 Reptiles................................................................................................................11 Amphibians..........................................................................................................13 Longleaf Pine Community..........................................................................................15 3 STUDY SITE AND METHODS................................................................................17 Study Site....................................................................................................................1 7 Materials and Methods...............................................................................................19 4 RESULTS...................................................................................................................27 General Captures........................................................................................................27 Effects of Fire Treatment on Herpetofauna................................................................29 Effects of Habitat Type on Herpetofauna...................................................................32
vi Effect of Treatment on Habitat Variables...................................................................33 Habitat Variables and Individual Species...................................................................34 5 DISCUSSION.............................................................................................................36 General Captures........................................................................................................36 Effects of Fire Treatment on Herpetofauna................................................................36 Effects of Habitat Type on Herpetofauna...................................................................40 Effect of Treatment on Habitat Variables...................................................................42 Habitat Variables and Individual Species...................................................................43 6 CONCLUSIONS........................................................................................................44 Summary and Conclusions.........................................................................................44 Limitations of Study...................................................................................................45 Suggestions for Research and Management...............................................................46 APPENDIX TRAP SITE CHARACTERISTICS..........................................................49 LIST OF REFERENCES...................................................................................................58 BIOGRAPHICAL SKETCH.............................................................................................67
vii LIST OF TABLES Table page 4-1 Number of captures by species for burned and unburned areas at Camp Blanding Training Site.............................................................................................................28 4-2 Mean, standard errors and P-values of captures per trap, richness and ShannonWeiner index (HÂ’) for herpetofauna, rept iles, anurans, snakes, and lizards in upland pine communities at Ca mp Blanding Training Site.....................................30 4-3 Means, standard errors and P-values fo r habitat structure variables measured at trap sites....................................................................................................................3 1 4-4 Slope direction, R and P-value of simple linear regression analyses with total captures per trap of six common species of herpetofauna as dependent variables and percent bare sand, canopy cover, volum e of leaf litter and vegetation as independent variables...............................................................................................35 A-1 Characteristics of individual trap sites at Camp Blanding Training Site.................49
viii LIST OF FIGURES Figure page 3-1 Location of Camp Blandi ng Training Site in Florida..............................................18 3-2 Overhead diagram of funnel box trap layout...........................................................21 3-3 Photo of trap array in sandhill..................................................................................21 3-4 Side view trap diagram showing subsurface funnel and trap...................................22 3-5 Trap detail showing funnel opening and underground funnel r outing into box.......22 4-1 Rarefaction curves of total species against total captures in burned and unburned treatments at Camp Bl anding Training Site.............................................................29 A-1 Typical unburned and burned logged sa ndhill at Camp Blanding Training Site.....51 A-2 Typical unburned and burned sandh ill community at Camp Blanding Training Site........................................................................................................................... .52 A-3 Typical unburned and burned flatwo ods community at Ca mp Blanding Training Site........................................................................................................................... .53 A-4 Map of north study units B2 and U1 at Camp Blanding Training Site....................54 A-5 Map of Study unit B1 at Ca mp Blanding Training Site...........................................55 A-6 South study units B3, U2 and U3 at Camp Blanding Training Site.........................56 A-7 Relative position of study units at Camp Blanding Training Site............................57
ix Abstract of Thesis Presen ted to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Master of Science HERPETOFAUNAL RESPONSES TO PRESCRIBED FIRE IN UPLAND PINE COMMUNITIES OF NORTHEAST FLORIDA By Keith Charles Morin December 2005 Chair: Max A. Nickerson Major Department: Wildlife Ecology and Conservation Differences in abundance, richness and divers ity of herpetofauna were measured in burned and unburned areas of upland pine ha bitat at Camp Blanding Training Site, a 73,000 acre military and law enforcement training center in northeast Florida. Sampling of herpetofaunal species was done via drif t fence box traps between November 2003 and November 2004. A total of 295 individual captures representing 24 species were recorded in 5,796 trap-days. Measurements of habitat revealed increased bare sand area and a reduction in litter volume and canopy cove r as a result of recent burn treatments. Measurements on vegetative and woody debris cover were similar between treatments. Herpetofauna responded positively to burn treatment in abundance, richness and diversity. Reptiles also ha d greater abundance, richness a nd diversity in burned areas when examined separately. For most herpetofauna burning had the greatest positive effect in flatwoods. No differences we re found in lizard and amphibian abundance, diversity or richness between burned and unbur ned treatments. Amphibian captures were
x limited to anurans, which were more abundant in flatwoods than sandhill blocks. Habitat type had significant effect on herpetofa unal communities, with flatwoods blocks generally exceeding sandhill and logged sandhi ll in abundance, richne ss and diversity of species. Six-lined racerunners ( Cnemidophorus sexlineatus) responded positively to open conditions created by burn treatment, while eastern fence lizards ( Sceloporus undulatus ) did not, showing little difference in a bundance between treatments and a negative relationship with very open conditions. Temporal and resource limitations preclude d the ability to do population estimation or to incorporate a more complex experiment al design necessary to improve inference of these results. Traps were also unable to e ffectively sample some fossorial herpetofauna and tree frogs present in these communities.
1 CHAPTER 1 INTRODUCTION Area of Study This study examined differences in herp etological habitat and populations produced by use of prescribed fire as the disturbance factor (treatment) in upland pine communities at Camp Blanding Training Site (CBTS), FL . Disturbance occurs naturally in many ecosystems at scales from microhabitat to landscape level. Depending on the return frequency and intensity of the disturbance event, effects on flora and fauna can range from destructive to beneficial (C onnell 1978, Wootton 1998, Reice 2001). While extreme events are typically rare and potenti ally detrimental, mode rate disturbance is important in habitat maintenance for the perp etuation of a wide variety of vertebrate species, including reptiles and am phibians (Karr and Freemark 1985). Wildland fire is a specific type of natura l disturbance. Natural systems worldwide have long been shaped by frequent fires in itiated by lightning stor ms and thousands of years of human influence (Komarek 1964, Kemp 1981, Turney et al. 2001). Over millions of years, flora and fauna in chaparra l, savanna, prairie, scrub, sandhill and other ecosystems have adapted to withstand and benef it from these natural fire events. Humans have also learned to contro l fire; extinguishing it or star ting prescribed fires when necessary. Burning can alter and improve ha bitat structure, comm unity composition and nutrient cycles (Wade 1989, Robbins and My ers 1992, Mistry 2000). Land managers commonly use prescribed fire for wildfire fu el reduction, management of pastoral lands and conservation areas, and replication of natu ral disturbance cycles. This use is well
2 documented in the Australian savanna (Russell -Smith 2002), North American pine forest (Wade 1989), South American Trachypogon savanna (San Jose and Farinas 1983), Western US chaparral systems (Biswell 1974 ), and African grassland and savanna regions (Komarek 1972, Swaine et al. 1992) among others. In areas where fire is common, herpetof auna and other animals have generally learned to avoid its potentially detrimenta l effects through behavioral and morphological adaptations. Indeed, mortality following a typi cal burn is very low and most individuals are able to move ahead of the flame front, find moist soil/water/veg etation, or occupy a refugium such as a burrow or stump hole until the danger is past (Komarek 1969, Means and Campbell 1982, Lips 1991). Smith et al. (2001) found that even the Ridge-nosed rattlesnake, Crotalus willardi, which has infrequent exposur e to fire, suffered mortality only from very intense fires. Most of the snakes observed survived low intensity fires without adverse effects. In addition, more radio-tagged snakes were found in underground refugia after the fires than before. Perhaps more important than direct (if minimal) mortality is the responses of herpetological species and communities to the structural and compositional changes of their habitats post fire. While the increas ed habitat openness and patchiness resulting from fire generally benefits reptile species Â’ abundance and diversity (Enge and Marion 1986, Moseley et al. 2003), these same effects can be detrimental to amphibians, which are dependent upon site moisture retained by canopy closure, accumulated leaf litter and microhabitat structure from fallen logs (B ury 1983, McLeod and Gates 1998, Moseley et al. 2003). In addition, the initial benefits to reptile species following fires tend to
3 decrease with a lengthening fire exclusi on period (Means and Campbell 1982, Mushinsky 1985, McLeod and Gates 1998). Research Problem and Motivation The general problem addressed in this st udy is discovering how short-term effects of prescribed fire influence herpetofaunal species in a fire dependent system. I was motivated to address this problem because th ere are still too few studies in this area, despite some initial work already having been done and briefly s ynthesized (Means and Campbell 1982, Russell et al. 1999). Also, the st udy is located in the southeastern United States where prescribed fire is the most impor tant tool used by land managers to replicate natural disturbance regimes in uplands. The other major natural disturbances in this region are hurricanes, which are not reproducible or as frequent as natural fire events. More information about the effects of fire on community components will assist resource professionals in making wise land management decisions. Specific questions that I a ttempted to answer are: How does the recent use of fire affect relative abundance of herpetofauna? How does the recent use of fire in the st udy areas affect richness and diversity of the herpetofauna? What effect do measurable variables of structure have on diversity, abundance and distribution of herpetofauna when this habitat is altered by fire? In response to these questions, I predic t that recently burned areas will have increased abundance and diversity of herpetofauna , and that reptiles wi ll benefit from fire more than amphibians. Amphibians will show little benefit or be adversely affected by open conditions promoted by burning. I expect that species with spec ific adaptations to patchy, open conditions with new growth will show increased abundance, richness and diversity in burned treatments.
4 Objectives The main objective of this study is to advance understanding of the secondary effects of prescribed fire on high pine herp etofaunal communities in the southeastern United States by measuring their responses to the secondary effects of prescribed fire. Means and Campbell (1982), Greenberg et al. (1994) and Litt et al. (2001) studied the relationship of fire use to upl and herpetofauna in Florida; however, the investigators captured few, if any, large snakes. Trap bi as, species behavior, or other methodologies are cited as causes for this trend in such studies (Greenberg et al. 1994, Litt et al. 2001, Moseley et al. 2003). A secondary objective of this study is to improve an existing general herpetofauna monitoring method in order to make it more effective for the capture of large snakes. Critical Assumptions Basic assumptions in this type of comp arative observational study are that other factors affecting measured variables are appr oximately equal in all units or in paired units. Units that were not converted to pl antation use or other major alteration were selected to limit non-fire influences on study units. I assume here that the major difference between these habitats is the fre quency of burning to which they have been exposed in recent years. An attempt to va lidate assumptions about equal fire regimes prior to the known period was made by bl ocking units based on community type and resource use. Florida Natural Areas I nventory (1990) community designations and history of the major resource us e (logging) were used to assi st in dividing the six study units into three blocks of paired units. In each block there is one burned and one unburned study unit that share community type and resource use history. All burned units were subjected to similar burn timing and frequency in the six years prior to
5 sampling. The fire return and timing here is typical of current management strategy for restoring unburned pinelands in the south east. The Camp Blanding Training Site environmental staff burns units on a schedul e determined by management goals and the major natural community type in the block, acco rding to an Integrated Natural Resource Management Plan (INRMP) (King 1998). Theref ore, it is also an assumption here that units with the same resource and community type have had roughly similar exposure to fire prior to the known period since 1999. Structure of Study Despite some limitations, the study shows th at short-term habitat modification by fire use has effects that can be significant for species of herpetofauna in upland pine communities. Fire effects are assessed by presentation and analysis of herpetological abundance and diversity data in burned and unburned units. Chapter 1 introduces the ar ea of study, the research problem, objectives of the study and critical assumptions made during the research process. The structure of the study is also presented here for the or ientation and benefit of the reader. Chapter 2 briefly reviews fire ecolo gy and general effects of burning on herpetological species. Results of some ot her studies investigating this subject are discussed where relevant to the study questi ons and objectives. Th is places the study questions in context and provi des a theoretical focus. Chapter 3 orients the reader to the phys iographic region in which the study takes place and present characteristics of the sp ecific study location, Camp Blanding Training Site. This chapter also reviews study met hods available and discusses methods selected for this study. Statistical tests us ed for analysis are discussed.
6 Capture data for this study and the resu lts of basic analyses are presented in Chapter 4. These include increases in cap tures per trap, richness and diversity for herpetofauna, reptiles, (including snake sp ecies) in burned areas and differences in measurements among different habitat blocks as well. The lack of difference for most comparisons in the lizard and amphibian groups is reported here. Results of measurements on structural habitat variable s are presented, as well as their potential correlation to abundance of some sp ecies captured in the study. In Chapter 5, I discuss the results and their relationship to principles and effects at work in this system. Comparisons to othe r studies on burning and herpetofauna as well as differences between this study and others are presented. Exceptions to the trends and lack of correlations are also discussed. Th is chapter concludes with implications of the results and how this evidence may be app lied to other research and management situations. The final chapter summarizes the study and presents conclusions and status of objectives and hypotheses presented in Chapte r 1. Evidence previously presented for each conclusion is briefly reiterated and the scope and significance of the study is discussed.
7 CHAPTER 2 THEORETICAL FRAMEWORK AND SYSTEM Review of Fire Ecology Traditionally, disturbance was viewed as a factor preventing progression to a stable climax community with predic table components by resetting or regressing seral stages (Odum 1993). However, more recent views pe rceive that periodic and non-catastrophic perturbations will actually enhance the stability and diversity of ecosystems by maintaining habitat heterogeneity (Connell 1978, Sousa 1984, Reice 2001). At opposite ends of this spectrum, in either a static sy stem or a highly disturbed system, competitive exclusion will favor resident or colonist-type species resp ectively, but not both. Connell (1978) first expounded this Intermediate Dist urbance Hypothesis (IDH) using examples from estuarine and tropical forest communities. Longer interval, highly disruptive events also have their place in nature since a grad ient of typical disturbance frequency and intensity exists across different sy stems (Connell 1978, Wootton 1998, Mackey and Currie 2001, Reice 2001). Eventually, a paradigm emerged that mi rrors this disturbance hypothesis in the more specialized area of fire ecology. The idea that pine savanna and other landscapes are maintained and enhanced by periodic patchy fire disturbance events has been proven and embraced by land managers worldwide (Wade 1989, Glitzenstein et al. 1995, Silva 1996, Williams et al. 2002). Frequent larg e-scale fire events are no longer common, however, even in areas of high lightning incide nce. This is primarily due to the way humans have divided and altere d natural landscape composition and processes. Fires that
8 would once have continued to burn until reaching a water body or major change in fuel type are now stopped by roads, ditches, ur ban areas, agricultural lands or fire suppression. For many areas, such suppre ssion was common policy before the 1980s and is still necessary if fire thr eatens critical resources or hum an safety. Implementation of the fire disturbance paradigm today most ofte n requires planned introduction of fire to the system in the form of prescribed burns. With prescribed fire programs, replication of this natural disturbance can be accomplished in a safe manner that does not threaten surrounding communities, managed resources or personnel involved in the burn (Wade 1989). Woody vegetation and fire intolerant species increase at the expense of open space and herbaceous groundcover in savanna systems if fire is reduced or excluded from the system (San Jose and Farinas 1983, Hoffma n 1999, Sparks et al. 1999). Depending on the recurrence and intensity of fire, plant growth and succession will be inhibited, maintained, or promoted. Effects of fire are dependent on successional stage and fire dependence of the system involved (Swaine et al. 1992, Silva 1996, Hoffman 1999). Annual growth and senescence of gra sses and other vegetation causes an accumulation of litter that becomes fine fuel fo r fires. Where there is year-round growth this accumulation can carry a fire after only a few seasons. This litter must be burned off or otherwise removed for seed ing or regrowth to occur. Though a great proportion of some nutrients like nitrogen and sulfur are vo latilized in this pro cess, nitrogen fixation by plants partially compensates for this phenomenon ; together with what is returned to the soil, this is sufficient for continued gr owth (Norman and Wetselaar 1960, Gillon 1983). Other nutrients, such as phosphorous, potassium , magnesium and calcium, are returned to
9 the soil in the form of ash. This aids in the regrowth of plants that then offer higher nutritional value for herbivores in the sy stem (Gillon 1983, Cook 1994). In southeastern high pine systems, recurrent burning promot es flowering and germination of important component plants and trees, including longleaf pine ( Pinus palustris ), wiregrass ( Aristida berychiana ) and a variety of herbs (Anderson and Menges 1997). Fire also controls height and density of midstory trees and sh rubs through heat stress or outright burning (Snyder 1986, Glitzenstein et al. 2003). Long-te rm fire exclusion can result in unchecked growth of overstory and midstory compone nts and unnatural accumula tion of leaf litter and woody debris. This accumulation not only influences ground level habitat structure and retards understory plant growth, but can also result in a catastrophic and uncontrollable wildfire event capable of k illing overstory pines upon eventual ignition (Robbins and Myers 1992, Varner et al. 2000). Herpetofauna and Fire General Effects Fire also modifies the habitat structure fo r herpetological species. Reptiles have a fairly impermeable skin, which prevents exce ssive evaporative water loss, and thus are typically better adapted than amphibians to take advantage of open stand conditions resulting from burning. The use of basking as a method of core temperature maintenance is also more pronounced in reptiles th an amphibians (Brattstrom 1979, Means and Campbell 1982, Enge and Marion 1986, Pough et al. 2000). The effect of fire on food availabilit y, behavior, population st ructure and species composition has been documented in herpetof aunal studies conducted in the fire-adapted systems of tropical savanna (F aria et al. 2004), spinifex gr assland (Letnic et al. 2004), chaparral (Lillywhite and North 1974, Lillywhite 1977), pine forest (Means and
10 Campbell 1982, McLeod and Gates 1998), sandhill (Mushinsky 1985, Litt et al. 2001) and sand pine scrub (Greenberg et al. 1994). There have also been studies that focus on herpetological species and burning in less fire-p rone systems such as eucalypt open forest (Singh et al. 2002, Woin arski et al. 2004), hardwood forest (Kirkland et al. 1995, Moseley et al. 2003), tropical humid forest (Fredericksen and Fredericksen 2002), and montane regions (Smith et al. 2001, Cunni ngham et al. 2002, Brisson et al. 2003). Though research has occurred on most continents, there is a relative paucity of studies on herpetofaunal responses to fire management in Southern African grasslands and savannas (Barbault 1976, Parr and Chown 2003). Herpetofauna Means and Campbell's (1982) work in Flor ida showed significant differences in occurrence of some reptile and amphibian sp ecies between burned pine areas and habitat where fire was long excluded and hardw ood succession had occurred, making the structure denser and more homogeneous, resul ting in loss of important understory plant species and herpetofauna. In another Florid a sandhill community, bucket-trap captures of reptiles and amphibians were greater in a pl ot with a median level of burning when compared to other plots that were either overg rown or left with few trees due to extremes of fire frequency (Mushinsky 1985). Other studies show no difference in overal l herpetological abunda nce in response to burn treatments. This occurs in non-fire prone systems of bottomland hardwood with recurrent winter burning in the treatment ar eas (Moseley et al. 2003), and tropical humid forest that has been logged and then burned four years later (Fredericksen and Fredericksen 2002).
11 An examination of fire effects on richne ss and diversity for herpetofaunal species yields mixed results as well. McLeod and Gate s (1998), Litt et al. (2001), and Moseley et al. (2003), all detect no significant differences in these measures when comparing burned and unburned areas. However, only the study by Litt et al. (2001) was in a fire dependent system. Another study in Florida sandhill (Mushinsky 1985) recorded the highest herpetofauna diversity and richness in a burne d area with a moderate return frequency. Extremes from high disturbance to no disturbance led to lack of heteroge neity in habitat. This would in turn reduce potential niche vari ation for herpetofauna as a whole. Despite this, Fredericksen and Fredericksen (2002) f ound herpetological richne ss to be higher in wildfire areas of Bolivian tr opical humid forest that were also intensively logged. Amphibian species captured there tended to be more adapted to open conditions than most. Recognizing the different needs of reptiles and amphibians, many studies have focused only on amphibians, reptiles or indivi dual species. Indeed, more differences are seen when responses of lower level taxa are l ooked at separately in habitats affected by fire. Reptiles Many reptile species respond positively to alterations in habitat brought about by burning, if they are adapted to more open conditions. Abundance of certain reptile species increases in areas with structur e opened by burning in sandhill (Means and Campbell 1982, Mushinsky 1985, Litt et al. 2001), sand pine ( Pinus clausa ) scrub (Greenberg et al. 1994), temp erate hardwood forest (Moseley et al. 2003), and chaparral (Lillywhite 1977, Cunningham et al. 2002). Mosele y et al.(2003) also found an increase in reptile richness and diversity in the bur ned plots, while the Litt et al. (2001) and Greenberg et al. (1994) studies did not.
12 In addition to changing basic structure, fi re events are unique in having the ability to produce or consume woody debris in forest s, depending on recurrence and intensity. These structures benefit certain lizard speci es for use as escape cover, egg depositories and perch sites, for the latter being preferred to live trees in some cases (Lillywhite and North 1974, Pounds and Jackson 1983, James and MÂ’Closkey 2003). Stump holes created by decay of fire-killed trees also pr ovide refuge for a variety of creatures, including snakes (Means 1985, Tennant 1997). The gopher tortoise ( Gopherus polyphemus ) is an important spec ies in southeastern upland systems. These tortoises typically inhabit well-burned opencanopied pine stands or savannas, digging long burrows inhabite d by many other species, herpetological and otherwise (Diemer 1986, Cox 1987, Aresco and Guyer 1999). They feed on the wide mixture of herbaceous and woody monocots and dicots in such habitats. Important dietary components include seedling pines ( Pinus spp .) and oaks ( Quercus spp.) as well as wiregrass ( Aristida spp.) and other grasses. Fires of semi-regular recurrence promote faster resprout of these species, increasing palatability and nutri tional value (Macdonald and Mushinsky 1988). There are exceptions to this positive trend for reptiles, however, such as avoidance of recently burned areas by some prairie snak es due to lack of cover from predators (Cavitt 2000) and greater abunda nce of fossorial, litter using, or moist microclimateadapted reptiles in unburned ar eas of pine stands (McLeod and Gates 1998), Amazonian savanna (Faria et al. 2004) and tropical eu calypt forest (Singh et al. 2002, Woinarski et al. 2004). Woinarski et al. (2004) also found decreas ed richness of reptiles overall in burned treatments; this was attributed to th e high capture proporti on of litter dwelling
13 skinks and other reptiles maladapted to very open conditions. Immediate post-fire conditions in spinifex ( Triodia spp . ) grasslands of Australia strongly favored two lizard species ( Rhynchoedura ornata and Ctenophorus nuchalis ) tolerant of high body temperatures and open conditions. Lizard density dropped, but species richness was improved on a site that remained unburned for 15 years. At this site, skinks and other lizard species with different thermoregulat ory needs were also present in a more structurally complex plant assemblage (Let nic et al. 2004). Similarly, both Mushinsky (1992) and Greenberg et al. (1994) found a correlation of decreased abundance of the litter-dwelling five-lined skink ( Eumeces inexpectatus ) to increased openness of burned areas in Florida uplands. Amphibians Many amphibians are typically not morphologi cally adapted to deal with the open dry conditions created by burning. Adult frog populations in an Australian forest were greatly reduced by a spring fuel -reduction burning program (D riscoll 1997). In Florida sandhill, Litt et al. (2001) captu red fewer southern toads ( Bufo terrestris ) in burned plots than control plots. McLeod and Gates (1998) found significantly fewer amphibians in a burned pine stand than in an unburned one. It appears burning is not always detrimen tal to amphibians, however. Moseley (et al. 2003) found no difference in amphibian rich ness, diversity or abundance in a winterburned hardwood forest on the southeastern co astal plain of the United States, and more true toads ( Bufo spp.) were captured in burned areas than in unburned ones. Similarly, amphibians in eucalypt open forest (Woinars ki et al. 2004) and tropical humid forest (Fredericksen and Fredericksen 2002) were unaffected by fire disturbance in their habitats. In a different type of distur bance study, amphibian abundance measured in
14 canopy gaps produced by wind damage and salvage logging was not different than abundance in unaltered sites in a southern A ppalachian closed canopy forest (Greenberg 2001). Kirkland et al. (1995) observed an actu al increase in amphibian abundance in a Pennsylvania hardwood tract that was winter burned, though captures were dominated by one species (70.8%), the American toad ( Bufo americanus ). In studies where no difference in amphibian populations was obs erved between treatments, fires may not have been frequent enough or intense enough (a s in winter burning) to produce changes in parameters important to amphibian surv ival. Many of the amphibians captured in these studies were those that have behavioral and morphol ogical adaptations, such as burrowing or moving to moist areas and thic ker or less permeable skin. These enable survival in more open settings commonly enc ountered post fire, with less cover or surface moisture. The reaction of amphibian populations to secondary fire effects has not been well studied in more xeric, fire prone system s, presumably because they are not affected by fire in these systems (Greenberg et al. 1994) or are difficult to sample properly (Means et al. 2004). In all cases it is clear that reactions of specific species to fire depend greatly on individual preferences and adap tations. Only a few herpetol ogical species are adapted to the very low or very high extremes of temp erature, moisture, and cover that are possible within fire dependent communities (Means and Campbell 1982, Mushinsky 1985). When fire improves the heterogeneity of a landscap e, a variety of patch conditions becomes available. These different microhabitat pa tches support and allow for a wider range of species (MacArthur1965). This can include those species that are exceptions to the trend
15 for each group, whether they are litter and wetland dwelling reptiles (McLeod and Gates 1998, Letnic et al. 2004) or amphibians able to tolerate dry conditions for much of their life cycle (Kirkland et al. 1995, Freder icksen and Fredericksen 2002). The synthesis of Russell (1999) and the report by Campbell and Christman (1982) demonstrate that fire can produce a landscape of varied microhabitats and open areas. The arrangement of these elem ents facilitates precise beha vioral thermoregulation, prey capture, and herbivory for the suite of reptil es and amphibians adapted to the upland pine communities of Florida and the southeastern United States. Longleaf Pine Community Longleaf pine ecosystems once covered ove r 25 million hectares (ca. 62 million acres) in the Southeastern United States (Sto ut and Marion 1993). These habitats were found in much of the upland ar eas on the Coastal Plain fr om Eastern Texas east to Southeastern Virginia and peninsular Flor ida. Across this region, a highly diverse assemblage of plant and animal communitie s evolved, including a rich herpetofaunal element (Stout and Marion 1993). As many as 25 amphibian and 51 reptile species occur in mesic and xeric uplands of the Southeastern Coastal Plain (Enge 1997, Williams and Mullin 1987, Labisky and Hovis 1997). Timber harvesting, agricultural use, urban development, and other human influences have reduced the longleaf pine co mmunities in these areas to less than five percent of their original exte nt with concomitant negative im pacts for the flora and fauna reliant on this habitat type. The loss and m odification of upland areas in the southeast was most intense from the pe riod 1880-1930, but is still occurr ing at a high rate in many areas (Outcalt 2000, Outcalt and Sheffield 1996).
16 Over these lands a gradient from wet co astal flatwoods to xeric upland sandhills and mountain ridges support longleaf pine and a variety of understory, midstory and codominant canopy plants and trees. Much of th is area can be classi fied as a seasonal savanna, characterized by low fertility so il, annual droughts and wet storm periods (Gillison 1983, Mistry 2000). Al so, as in many of the worldÂ’s savannas, these storms are characterized by high frequency of lightning that ignites natu ral fires. Most of these thunderstorms occur in the growing season, re aching a peak in June (Komareck 1964). Earlier fires in May have th e potential to burn more area, however (Robbins and Myers 1992).
17 CHAPTER 3 STUDY SITE AND METHODS Study Site Study areas were located on the Camp Blanding Training Site (CBTS) (lat. 29 59Â’00Â” N, long. 82 00'00" W) in Clay County, FL (Fig 3.1). CBTS is a 73,000-acre training and readiness site for the Florid a Army National Guard and is used by many other law enforcement and military agencies. The closest major city is Jacksonville, approximately 35 miles (57 km) to the northeast. In addition to military activity, sections of CTBS are used for public hunt areas, mi neral extraction activities (mining) and forestry operations (King 1998). Rainfall in the area averages 135 cm annually, with most occurring JuneSeptember. Frequent thunderstorms and high incidence of lightning strikes characterize this period. Mean temp eratures vary from 56 F (13.3 C) in the wint er to 81Âº F (27.2 C) in the summer. This region includes upl and soils of the well-drained Penny-KershawOrtega association and poor ly drained Hurricane-LeonMandarin and Leon-MandarinPottsburg associations (King 1998). CTBS supports a variety of high quality ha bitats native to nort h Florida, including sandhill, bottomland hardwood, scrub, wetlands and mesic flatwoods (FNAI 1990). The property supports substantia l biodiversity, including severa l threatened and endangered species. Twenty-three amphibian and 32 reptile species have been recorded at CBTS. Of these, 16 amphibian and 22 reptile species ar e known to occur in upland pine habitats (King 1998).
18 Figure 3-1. Location of Ca mp Blanding Training S ite (blue) in Florida The site has an active prescribed burni ng program administered and implemented by on-base environmental staff. Areas are burned for fuel reduction, endangered species habitat improvement, range clearing for or dnance disposal and for general ecosystem maintenance (personal communication Paul Catl ett 2004). Knowing the effects of this
19 common management event on ecosystem health is important for land managers at this site as well as on other public a nd private conservation lands. Materials and Methods Standard upland herpetological sampling t echniques include transect searches, pitfall traps and funnel traps in association wi th drift fencing (Corn 1994), radio telemetry (Means 1985), and anuran call counts (Zim merman 1994). Use of passive trapping techniques is the most reliable method availa ble to ensure equal sampling effort between treatments. Enge (2001) reviewed pitfall and funnel trapping techniqu es used in several studies in Florida. For all cases examin ed, large and medium snakes were more effectively captured using funne l traps. For this study, I used a modified version of a four-funnel and four-fence box trap design used by Rudolph et al . (1999) in Texas. This trap type was designed to capture large sn akes, but is effective at capturing other terrestrial herpetofauna as well. Traps consisted of a central 3Â’ by 3Â’ by 1Â’ (91 cm x 91 cm x 30 cm) wooden box frame with Â¼Â” (0.63 cm) mesh sides, plywood top and bottom, four funnels and 50Â’ (15.6 m) of drift fences on each side (Fig 3-2). Fences were standard staked erosion control fence, 62 cm in height and buried a minimum of 5 cm belo w ground level. Fences were secured around funnels using cable ties and then funnels were buried with existing substrate (Fig 3-4). Funnels were construc ted of Â¼Â” (6.3 cm) opening hardware cloth and were 3 feet (91 cm) in length. The entrance of the funnel is a semicircle approximately 8 inches (20 cm) in diameter and 5 inches (13 cm ) in height. Each funnel terminates in an opening 2.5 inches (6.5 cm) in diameter 6 inches (15 cm) above the floor of the trap box (Fig 3-5). A two-foot by one-foot section of hardware cloth was fitted into funnel
20 entrances to split the opening. This splitting device is designed to encourage entry into the funnel and prevent animals from occupying the funnel opening instead of entering the trap. Traps were similar to those of Rudolph et al. (1999) with the following alterations: funnels were modified to run underground befo re entering the box and fences were set up in a star pattern leading into the funnels (Figs. 3-2, 3-3). Th is design effectively hides the trap and contents from predators and othe r animals along fence lines or at the funnel entrance (Figs. 3-4, 3-5). Traps were modified in an effort to reduce non-target captures (and mortality) associated with funnel trappi ng. In snake studies using large funnel traps at Avon Park Air Force Range, FL (personal observation) and in east Texas (Rudolph et al. 1999) significant capture of small carnivores, rodents, and birds th at forage or nest on the ground does occur. I believe some of th is non-target catch results from the large straight funnel entrance of these traps (neces sary for large snake capture) coupled with the ability of predators to see herpetofauna and insect prey items captured within. Six stands were selected (Fig A-7) fro m upland burn units using the GIS coverage provided by CBTS environmental staff. Due to the substantial movement ability of snakes, large areas w ith a relatively contiguous habitat type were chosen. Study units ranged from 48 to 78 ha and were selected from contiguous natural upland pine (not plantation) areas with no divisions or differen ce of burn regime within sites. Three units were burned in April-June of 2003; the rema ining three units had not been burned for a minimum of five years (pre-1999). CBTS Enviro nmental Staff assisted in pairing stands with similar burn frequency prior to the tr eatment period. A desi gnation of B1-B3 for burned areas and U1-U3 for unburned areas was assigned (Table A-1).
21 Figure 3-2. Overhead diagra m of funnel box trap layout Figure 3-3. Photo of trap array in sandhill. Funnel Box Fencing Splitter
22 Figure 3-4. Side view trap diagram showing subsurface funnel and trap. Figure 3-5. Trap detail showing funnel ope ning and underground funne l routing into box. Units B1 and U1 are mostly flatwoods (89.1 % and 92%, respectively) with drier sandhill sections embedded in these largely me sic areas (Figs. A-4 and A-5). Units B2, B3, U2 and U3 are composed of xeric longleaf pine ( Pinus palustris )-turkey oak Box Funnel Fencing
23 ( Quercus laevis ) sandhill (Figs. A-4 and A-6). In areas B3 and U3, however, extensive harvest of longleaf pine pr ior to government acquisiti on in 1939 (King 1998, personal communication CBTS forestry personnel 2004) has resulted in a derived habitat dominated by turkey oak with far fewer pi ne present (Fig. A-1). Understory plant diversity in these areas is also reduced, composed primarily of young turkey oak with substantial wiregrass on ly in burned areas and few other sp ecies. Specific characteristics of each trap site are listed in Table A1 in Appendix A. Figures A-1, A-2 and A-3 show typical burned and unburned block types. Flatwoods are characterized by flat, sandy so ils with a well-defined argillic horizon (clay layer), resulting in seasonal topsoil saturation and ephemeral ponds. The understory is dominated by wiregrass ( Aristida berychiana ) and saw palmetto ( Serenoa repens ) with gallberry ( Ilex glabra ), wax myrtle ( Myrica cerifera ) and fetterbush ( Lyonia lucida ) present in significant amounts. Longleaf pine is the primary overstory tree species (Fig. A-3). Sandhill areas have a thicker entisol sand layer, better drainage (argillic horizon generally absent) and more bare soil than flat woods. Slope is typica lly increased as much as 10% compared to flatwoods areas. Unde rstory components include wiregrass, which is more extensive in burned area s, young turkey oaks, runner oak ( Q. pumila ), gopher apple ( Lichania michauxii ) and occasional palmetto. The overstory here contains a significant turkey oak component in addition to taller longleaf pine with occasional blackjack ( Quercus incana) , bluejack ( Quercus marilandica) or sand post oaks ( Quercus margaretta ) present in the midstory (Fig A-2).
24 Three box traps were placed in each sampli ng unit for a total of nine traps in each treatment type. Random trap placement within each unit was accomplished using the random point function from the USGS Animal Movement extension (Hooge et al. 1999) running on Arcview 3.2 (ESRI 1999). Traps were located a minimum of 75 m from the next closest trap to promote independent sampling. Traps were also placed a minimum of 50 m from the edge of the unit or roads to reduce edge effects due to artificial structural changes in habitat. Figures A-1 through A-3 in A ppendix A show trap locations and habitat within each study unit Traps were open continuously during the following periods: November 11 December 19, 2003, January 12-May 30, 2004 and June 20-November 9, 2004. Trapping effort totaled 5,796 trap-days; where one trap -day equaled one trap open for a 24-hour period. Traps were rendered inaccessible during closed periods by blocking funnel access. Water was continuously supplied in each trap with a standard 1-gallon poultry water device. Traps were checked thr ee times per week during open periods. Upon capture, subjects were identified to species, measured for length, weighed, and released. Before release, subjects were injected with a uniquely numbered PIT tag to permanently mark the animal in case of recap ture. Incidents of recapture were low for most species and not included in calculations. Any lizards seen perched on or next to drift fences were captured by ha nd and added to speciesÂ’ totals. Habitat variables were measured at eight plots within a 7,850-mÂ² circle centered on each trap. Measurements were taken approximate ly one year after the fire treatments in burned areas. Two 1 m Â² plots were randomly placed along each of four 50-meter transects originating at each funnel opening. In each plot, percent vegetative cover,
25 percent bare soil, per cent leaf litter cover, and percent coarse woody debris (CWD) cover were estimated. For CWD, only pieces over 1 inch (2.5 cm) diameter were considered. In each plot, mean depth of leaf litter was obt ained by insertion of a thin metal ruler into the litter layer to the soil leve l at four random locations and measuring to the nearest mm. A stem representing the average height of understory vegetation in each plot was measured to the nearest centimeter. Canopy c over above each plot was measured using a standard concave spherical densiometer. For analysis of abundance, richness and diversity, captured species were grouped into taxonomic categories as follows: total he rpetofauna, reptiles, amphibians, anurans, snakes, turtles, and lizards. Individual spec ies with more than 20 captures were examined separately. For each category, abundance was expressed as the total number of captures per trap and richness as the total number of sp ecies per trap. This does not represent true abundance, which would require 100% detec tion of herpetofauna present or markrecapture methods, but rather, reflects re lative abundance of species based on equal detection ability at each trap site. A measure of diversity was calculated for all categories using the Shannon diversity index, HÂ’ (Piel ou 1977). For precision and to eliminate variation according to habitat subtype, data were blocked into logged sandhill, sandhill, and mixed flatwoods as outlined in Table A1. A two-factor analysis of variance (ANOVA) was used to test a ll taxonomic categories and habi tat variables for differences in treatment effect between burned and unburned areas and among habitat blocks. Variables with non-normal distribution were square root or ar csine-square root transformed to meet assumptions of parametric analysis where necessary. Multiple
26 comparisons for determination of differences of means were done via TukeyÂ’s procedure (SAS Institute 1997). Correlation analysis via linea r regression was used to further examine relationships between individual species with over 20 total captures and habitat variables measured. Only habitat variables showing significant diffe rences between treatments or blocks were examined in correlation analysis. All sta tistical analyses were conducted using SAS statistical software (SAS In stitute 1997) or Sigmastat 2.0 (Jandel Corporation 1995).
27 CHAPTER 4 RESULTS General Captures A total of 295 individuals representing 24 species was captured in 5,796 trap-days (Table 4-1). Eighteen species of reptile a nd six species of amphibi ans (all anuran) were captured. Two species of anurans and six sp ecies of reptiles were captured only in burned areas, though all were represented by few captures. The relatively rare southern hognose snake ( Heterodon simus ) was captured in Trap B1B. The ground skink ( Scincella lateralis ) and the gopher frog ( Rana capito ) were captured only in unburned areas (Table 4-1). The gopher tortoise ( Gopherus polyphemus ) was captured opportunistically next to trap arrays save one small juvenile . A plot of total species against total captures for burned and unburne d areas is shown in Figure 4-1. This rarefaction curve and consultation of the poten tial species list for Camp Blanding reveals that further sampling effort likely would have resulted in few, if any additional species sampled in the study areas. In addition, recaptures were far too low and sporadic with methods used (a non-transect and non-grid de sign) to enable any sort of mark-recapture analysis for population estimations. Results of two-factor ANOVA for herpet ofauna captures are discussed below (Table 4-2). Table 4-3 lists results of the same analysis performed on mean measurements of habitat variab les. Results of fire treat ment are those reached after allowing for differences in block and vice vers a. In some cases an interactive effect between treatment and block was also detected.
28 Table 4-1. Number of captures by species for burned and unburned areas at Camp Blanding Training Site. Species Common Name Unburned Burned Agkistrodon piscivorus conanti Florida Cottonmouth 1 1 Anolis carolinensis Green Anole 8 9 Cnemidophorus sexlineatus sexlineatus Six-lined racerunner 4 18 Coluber constrictor priapus Southern Black Racer 2 8 Drymarchon corais couperi Eastern Indigo Snake 0 2 Elaphe guttata Corn Snake 1 2 Eumeces egregius Mole Skink 0 1 Eumeces inexpectatus Southeastern five-lined skink 4 6 Gopherus polyphemus Gopher tortoise 1 4 Heterodon platyrhinos Eastern Hognose Snake 1 1 Heterodon simus Southern Hognose Snake 0 1 Masticophis flagellum Coachwhip 8 16 Sistrurus miliarus barbouri Dusky Pigmy Rattlesnake 0 2 Scincella lateralis Ground Skink 2 0 Sceloporus undulatus Fence Lizard 60 41 Ophisaurus ventralis Eastern Glass Lizard 0 1 Thamnophis sirtalis Garter Snake 1 1 Pituophis melanoleucus Florida Pine Snake 0 2 Bufo quercicus Oak Toad 1 1 Gastrophryne carolinensis Eastern Narrow-Mouthed Frog 0 1 Bufo terrestris Southern Toad 22 10 Rana utricularia Southern Leopard Frog 0 1 Scaphiopus holbrooki Eastern Spadefoot Toad 10 39 Rana capito Gopher Frog 1 0 Total reptiles 93 116 Total amphibians 34 52 Total captures 127 168 Number of species 16 21 Overall captures 295 Total snakes 14 36 Total frogs 34 52 Total turtles 1 4 Total lizards 78 76
29 0 5 10 15 20 25 050100150200250 Individuals capturedspecies #sppburned #sppnonburn Log. (#sppnonburn) Log. (#sppburned) Figure 4-1. Rarefaction curves of total spec ies against total capt ures in burned and unburned treatments at Camp Blanding Training Site. Effects of Fire Treatment on Herpetofauna Mean captures per trap of herpetofauna were higher (P = 0.026) in burned treatment areas than in unbur ned treatment areas (Table 4-2). Two factor ANOVA for herpetofauna richness and diversity also s howed significantly highe r number of species (P = 0.004) and larger ShannonÂ’s HÂ’ (P = 0.003) in burned areas over unburned areas. Total reptile captures followed the same trend in response to fire treatment as herpetofauna, with burned areas yielding greater numbers (P = 0.02). Interactions between treatment type and block were seen for this group. Effect of burning on reptile capture numbers was greatest in mixed flat woods blocks, which had more captures than either sandhill or logged areas (P <.005). Re ptiles also showed greater richness (P = 0.006) and diversity (P = 0.002) in burned areas than in unburned areas. Within reptiles, turtles had too few captures to enable separate statistical examination.
30Table 4-2. Mean, standard errors and P-va lues of captures per trap, richness and Sh annon-Weiner index (HÂ’) for herpetofauna, reptiles, anurans, snakes, and lizards in upland pine comm unities at Camp Blanding Traini ng Site. Different letters indicate a significant difference between block types. Significant P values are in italics. Treatment Habitat Burned Unburned F 1,12P Flatwoods Sandhill Logged F 2,12 P Total Herpetofauna Captures/ trap 19.67 Â± 1.5714.00 Â± 1.576.49 0.026 24.50 Â± 1.93A 15.17 Â± 1.93B10.83 Â± 1.93B13.14 >0.001 Richness 7.78 Â± 0.515.22 Â± 0.5112.56 0.004 8.67 Â± 0.62A 6.67 Â± 0.62B4.17 Â± 0.62C10.50 0.002 Shannon Index 1.73 Â± 0.081.28 Â± 0.0814.45 0.003 1.74 Â± 0.10A 1.60 Â± 0.10A1.17 Â± 0.10B8.67 0.005 Reptiles Captures/ trap 14.00 Â± 1.0010.22 Â± 1.007.14 0.020 15.33 Â± 1.23A 11.67 Â± 1.23 AB9.33 Â± 1.23B6.10 0.015 Richness 6.11 Â± 0.44 4.00 Â± 0.4411.28 0.006 6.83 Â± 0.58A 4.67 Â± 0.54B3.67 Â± 0.54B8.84 0.004 Shannon Index 1.56 Â± 0.100.99 Â± 0.1015.45 0.002 1.58 Â± 0.13A 1.26 Â± 0.13 AB1.0 Â± 0.13B5.48 0.020 Amphibians (anurans) Captures/ trap * 1.95 Â± 0.261.66 Â± 0.260.600.4542.85 Â± 0.32A 1.58 Â± 0.32B0.99 Â± 0.32B8.62 0.005 Richness 1.67 Â± 0.321.22 Â± 0.320.940.3511.83 Â± 0.40 1.50 Â± 0.401.00 Â± 0.401.120.359 Shannon Index 0.26 Â± 0.010.21 Â± 0.010.140.7120.27 Â± 0.12 0.31 Â± 0.120.12 Â± 0.120.750.492 Snakes Captures/ trap 4.0 Â± 0.511.56 Â± 0.5111.52 0.005 4.67 Â± 0.62A 2.50 Â± 0.62 AB1.17 Â± 0.62B8.02 0.006 Richness* 1.58 Â± 0.140.94 Â± 0.1410.89 0.006 1.78 Â± 0.17A 1.09 Â± 0.17B0.90 Â± 0.17B7.36 0.008 Shannon Index* 0.83 Â± 0.090.35 Â± 0.0914.45 0.003 0.95 Â± 0.11A 0.52 Â± 0.11B0.30 Â± 0.11B9.07 0.004 Lizards Captures/ trap 9.56 Â± 0.928.56 Â± 0.920.590.45810.17 Â± 1.13 9.17 Â± 1.137.83 Â± 1.131.070.373 Richness* 1.84 Â± 0.121.62 Â± 0.121.500.2441.83 Â± 0.15 1.77 Â± 0.151.60 Â± 0.150.640.515 Shannon Index 0.95 Â± 0.120.62 Â± 0.123.530.0850.86 Â± 0.15 0.89 Â± 0.150.61 Â± 0.151.040.382 * values were square root transformed
31Table 4-3. Means, standard errors and P-values for habitat struct ure variables measured at trap sites. Significant P-values ar e in italics. Treatment Habitat Burned Unburned F 1,12 P Flatwoods Sandhill Logged F 2,12 P Vegetation % Vegetation* 5.40 Â± 0.43 4.51 Â± 0.432.110.1725 .95 Â± 0.53 4.42 Â± 0.534.50 Â± 0.53 2.65 0.111 Vegetation Height 0.60 Â± 0.07 0.67 Â± 0.070.550. 4740.79 Â± 0.08 0.58 Â± 0.080.54 Â± 0.082.78 0.102 Groundcover % bare sand 26.51 Â± 3.08 10.08 Â± 3.0814.22 0.003 13.27 Â± 3.77 20.5 Â± 3.7721.13 Â± 3.771.34 0.300 Woody debris 9.13 Â± 0.93 8.20 Â± 0.930.510.4919.15 Â± 1.13 8.34 Â± 1.138.50 Â± 1.130.14 0.867 % LL cover 41.93 Â± 4.90 69.88 Â± 4.9016.24 0.002 50.33 Â± 6.01 59.06 Â± 6.0158.31 Â± 6.010.65 0.540 Depth LL (cm) 1.17 Â± 0.19 2.32 Â± 0.1919.12 <0.001 1.52 Â± 0.23 2.04 Â± 0.231.68 Â± 0.231.36 0.294 Volume leaf litter (m3/ha) 54.12 Â± 18.68 176.25 Â±18.6821.37 <0.001 87.15 Â± 22.88 137.17 Â±22.88121.25 Â±22.88 1.25 0.322 % Canopy cover 27.95 Â± 3.13 45.98 Â± 3.1316.64 0.002 34.88 Â± 3.83 40.05 Â± 3.8335.95 Â± 3.830.51 0.614 Differences with over 5% probabilit y of type one error were not reported * values were square root transformed
32 Snakes, however, were captured almost th ree times as often (P = .005) in traps placed in burned areas than those in unburned areas. Additionally, more snake species (P = 0.006) and higher diversity (P = 0.003) were recorded in burned areas. Unlike snakes, lizards showed no significant differences in any category between treatments or block pairings with the analysis used. Only dive rsity measures on treatment types were close enough to suggest a possible difference (P = 0.085) may exist. Only six-lined racerunners showed any difference with a higher total abundance in burned areas (P =.003). Anuran capture data produced no detect able differences in abundance between treatment types. Within levels of unburned treatment, a significant difference (P = .016) in amphibian abundance was detected, with sandhill levels exceeding those in logged sandhill. Burning also had a greater effect (P < 0.05) on reducing the number of anurans in sandhill and logged sandhill than in flatw oods. Data collected did not reveal any significant differences in amphibian ric hness or diversity between treatments. Effects of Habitat Type on Herpetofauna Flatwoods sites had higher (P < 0.001) herp etofauna captures per trap than either sandhill or logged sandhill. A significant (P < .05) interaction indicated that effect of burning on abundance of herpetofauna was greate r in flatwoods than in either type of sandhill treatment. Herpetofauna diversit y (Shannon HÂ’) was lower in logged sandhill than in the other two blocks (P = 0.005). A difference (P = 0.002) was found for species richness in all three habitat pairings with flatwoods having the highest number of species observed and logged sandhill the lowest (Table 4-2). Flatwoods exceeded logged sandhill (P = 0.015) in reptile abundance. Mean number of reptile species sampled (richness) was higher (P = 0.004) in flatwoods than in
33 sandhill or logged blocks. Howe ver, differences in diversit y were only detected between flatwoods and logged blocks (P = 0.020) (Table 4-2). Logged blocks had significantly fewer sn ake captures than flatwoods blocks, but neither was significantly different from unaltered sandhill when using a multiple comparison procedure (P = 0.006). Between bl ock types, snake richness and diversity were significantly higher (P= 0.008 and 0.004, respect ively) in the flatwoods type blocks than in either of the other two sandhill hab itat types. Lizards differed from snakes and total reptiles, showing no si gnificant differences between block type in richness or diversity. A difference was noted (P = .005) among bl ock types with significantly more anurans captured in flatwoods blocks than th e two other habitat bloc ks sampled (Table 42). Data collected did not reveal any si gnificant differences in anuran richness or diversity between blocks (Table 4-2). Effect of Treatment on Habitat Variables Differences in habitat struct ure were detected in four measured variables: percent bare sand, percent litter cover, depth of litter cover, and percent canopy cover. Mean vegetation cover and mean woody debris cover were not different between treatments or among block types (Table 3-3). Litter has a three-dimensional eff ect on habitat in the forest; depth and percent cover data were converted to a m easure of averag e litter volume at each trap site, in cubic meters per hectare. Predictably, this value also showed a highly significant difference between treatment and co ntrol sites. There were no differences between block types in any of the measured variables (Table 3-3). Canopy cover was significantly higher in unburned units than in burned units (P= 0.002). An interaction between block and treatment was detected for some units in percent canopy cover.
34 Sandhill and logged sandhill areas had signifi cantly more cover when unburned, whereas flatwoods did not (P< 0.05). Mean values for percent bare sand were very different in burned (26.5 %) and unburned (10.1%) units (Table 4-3). This va lue was inversely propor tional to percent litter cover (R2 = 0.50, P < 0.001), which was much highe r in unburned areas. Litter was nearly twice as deep in unburned areas, cont ributing to significantly higher volume in the unburned treatment (P <0.001). Percent woody de bris present was remarkably similar in every category, regardless of treatment or habita t block (Table 4-3). Habitat Variables and Individual Species Simple linear regression analysis (Table 4-4) of species with over 20 captures indicated a general positive correla tion of the six-li ned racerunner ( Cnemidophorus sexlineatus ) with open ground habitat features. Ther e was also a significant but weak correlation between abundance of this lizar d and open canopy, although not at the desired level (P = 0.054). Captures of this animal in burned areas we re significantly higher than for unburned areas (P = 0.003) Eastern fence lizards ( Sceloporus undulatus ) generally exhibited opposite trends from six-lined racerunners, s howing correlations with decrea sed bare sand and increased leaf litter volume across all treatment sites. In this case there was a positive but weak association between tota l captures of this lizard and in crease in canopy cover (P= 0.052) (Table 4-4). No significant difference in abundance of fence lizards was detected between treatments (P = 0.147). Upon individual analysis, differences in capture totals between treatments were insignificant for co achwhip (P = 0.094), southern toad (P = 0.111), and spadefoot toad (P = 0.232).
35 Table 4-4. Slope direction, R and P-value of si mple linear regression analyses with total captures per trap of six common speci es of herpetofauna as dependent variables and percent bare sand, canopy cover, volume of leaf litter and vegetation as independent variables. Association considered significant at P 0.05. % Bare Sand Leaf Litter Volume % Canopy Cover Vegetation density Species SlopeR P-valueSlopeR Pvalue SlopeR P-value SlopeR P-value Cnemidophorus sexlineatus* pos. 0.58 0.011 neg. 0.63 0.005 neg. 0.46 0.054 neg. 0.02 0.950 Masticophis flagellum pos. 0.37 0.126 neg. 0.410.095 neg. 0.21 0.404 neg. 0.04 0.875 Scaphiopus holbrooki pos. 0.05 0.853 neg. 0.270.285 neg. 0.18 0.484 pos. 0.35 0.154 Bufo terrestris neg. 0.02 0.951 pos. 0.130.597 neg. 0.10 0.710 pos. 0.38 0.119 Sceloporus undulatus neg. 0.49 0.039 pos. 0.48 0.043 pos. 0.47 0.052 neg. 0.01 0.973 * values were square root transformed
36 CHAPTER 5 DISCUSSION General Captures The composition and proporti on of reptile to amphibian captures reflects known species lists for sandhill in northeast Flor ida (Enge 1997, CBTS species list). The number of herpetofaunal species captured is comparable to those in other southeastern sandhill studies (Mushinsky 1985, Litt et al. 2001 ). Rare species captured include the eastern indigo snake, Florida pine snak e, and southern hognose snake. The proportionately higher reptile vs. amphibian abundance supports generalities about physiological characteristic s of these two groups. Low capture totals relative to other tra pping studies in upla nd Florida can be attributed to trap type. Us e of buckets would likely have increased total number of captures, especially for lizards. With the tr aps used, however, capture of even the largest snakes is possible and traps can be checked on alternating days with no mortality issues for trapped specimens. Effects of Fire Treatment on Herpetofauna Habitat conditions altered by prescribed fi re treatment appear to have positive effects on abundance, richness and diversity of snake, total reptile and other herpetofauna species in sandhill habitats of northeast Fl orida. The aforementioned behavioral and physiological adaptations of species in this sy stem to fire and its effects on habitat are largely responsible for their success in treated areas here. Changes induced by prescribed burning produced a more heterogeneous and ope n understory important to the success of
37 these animals and their prey. A seeming exception to this trend was the unforeseen neutral reaction of lizards to treatment. Ground dwelling anurans also appeared to be unaffected by the differences in habitat brought on by burn treatment. The higher herpetological abundance in burne d areas was similar to results of the sandhill trapping study of Mushinsky (1985). In that study, the hi ghest herpetological diversity and density occurred in a plot with a rather long sevenyear fire return interval. A fire occurred just prior to the sampling, cau sing destruction of mo st understory trees and exposing bare soil, which gradually became covered with herbaceous growth into the second year of sampling. Similar results were obtained post fire in the sandhill plots sampled here, although study blocks were much larger. Results here are different, however, from those of Moseley et al ( 2003), McLeod and Gates (1998) and Litt et al. (2001) who all detected no differences in he rpetological richness a nd diversity in their studies of fire treated areas. Many studies have reported significant e ffects of burning on certain herpetofaunal species, however. As with the results in this study, Moseley et al. (2003) found higher abundance and diversity of reptil es in the burned sections of bottomland forest studied. Greenberg et al. (1994) also showed the value of low, open structure promoted by fire for some reptiles of xeric Florida communities. The benefit of burning for snakes in all ca tegories was significant. Two species, southern black racer ( Coluber constrictor ) and coachwhip ( Masticophis flagellum ) represent 68% of snakes captu red. Both of these species are fast, wide foraging hunters that do well in more open habitats where it is easy to see and chase moving prey. Also, the relatively high body temperatur e preferred by racers and whip snakes is easier to attain
38 in these areas through basking (Herzog a nd Burghardt 1974, Secor and Nagy 1994). In addition, known prey items of these snakes (Camper and Dixon 2000) increase as fire dependent habitats recover in the months after a burn event. Typical prey includes orthoptera (Swengel 2000), lizards (Lillywh ite 1977, Mushinsky 1985, Litt et al. 2001, Cunningham et al. 2002) and common rodent species. Regarding the latter, condition, reproductive success, and abundance of small mammals are maximized in open canopied, heterogeneous habitats. These are maintained by disturbance such as fire, stochastic wind events and herbicides. Small patche s of microhabitat and an abundant mix of monocot and herbaceous dicot understory regrow th characterized favor able sites (Hanchy and Wilkins 1998, Davis et al . 2000, Schweiger et al. 2000). McLeod and Gates (1998) had comparable resu lts to this study in a burned area of pine forest in Maryland, finding more C. constrictor there than in an unburned area. Fire appears to be an important factor in su ccess of upland snake species where habitats experience some level of natural burning. Lizards as a group showed little diffe rence between burned and unburned study units at CBTS, with six-lined racerunners the sole exception. Li zard species in other studies showed a wide variety of responses to fire treatment depending on several factors including species studied, habita t and fire intensity. In s outhern California chaparral, lizards were more abundant in a recently burned open site than a 15-year-old unburned site with 90% brush cover (Lillywhite 1977). Cunningham et al. (2002) found similar benefits for many lizard species following a catastrophic fire that removed 90% of standing vegetation. It is very possible that habitat conditions prio r to the recent burn regime at CBTS had not deteriorated enough to significantly exclude many lizard species
39 or depress populations. Sim ilar results were reported in another Florida sandhill community by Litt et al. (2001), who found an in crease of six-lined racerunners and fence lizards in response to burning while abundan ce of four other lizard species remained largely unaffected by the treatment. Due to species-specific adaptations, the lizards in these communities appear to have a wide to lerance for a gradient of habitat conditions produced by different times since fire. Specifically, equal amount s of woody litter and perch sites in both treatments (Table 3C) may account for lack of a significant difference in abundance of fence lizards, which are depe ndent on perch site a nd litter availability (Allen and Neill 1953). Anuran species sampled seemed to be largely unaffected by the burn regime employed at CBTS. Several studies found a neutral effect of burning on terrestrial amphibians in a variety of situations (Kirkl and et al. 1995, Fredericksen and Fredericksen (2002) Moseley et al. 2003, Woinarski et al. 2004). Some deduced neutral effects of burning on amphibians. These were thought to be related to adaptation of the species involved to withstand open dry conditions (Kirkland et al. 1995, Fredericksen and Fredericksen 2002). Results of trapping in th is study could also be attributed to the adaptations of anurans capt ured. Southern toad ( Bufo terrestris ) and spadefoot toad ( Scaphiopus holbrooki ) comprised 94% of amphibian capt ures. Both of these animals have adaptations to prevent desiccation and facilitate living in fire prone communities. These adaptations include nocturnal activity cy cles, the ability to burrow down into moist subsurface soil or duff, use of other animal burrows to escape adverse conditions of moisture and temperature, and the ability to aestivate (Zug et al. 2001). These results do
40 differ, however, from the findings of Litt et al. (2001) who found fewer southern toads in burned sandhill plots. Effects of Habitat Type on Herpetofauna While fire did improve conditions for a vari ety of herpetofauna, especially reptiles and those other species adapted to more open conditions, an effect of habitat type was also seen in certain types of herpetofauna. The higher abundance and richness of reptiles and all herpetofauna in mixed flatwoods blocks can be attrib uted to the wider variety of habitats available there. Th e habitat-heterogeneity hypothesis states that an increase in habitat heterogeneity leads to an increase in species diversity (Ricklefs 1977). Many studies of animal conservation and diversity show the importance of heterogeneity of structure and vegetative species compositi on (McQuaid and Dower 1990, Tews et al. 2004) including effects produced by fire (Kut iel 1997). This effect is related to competitive exclusion and the availability of niches in heterogeneous landscapes for a greater number of species with different needs (MacArthur 1965). Animals captured in mixed blocks were thos e characteristic of flatwoods as well as sandhill habitats (Stout and Marion 1993). Th ese two were the only units with traps in mesic flatwoods or mixed areas of xeric sandhi ll near flatwoods. These areas were also wetter, frequently containing ephemeral stan ding water after rain events. Temporary standing water was only seen in other units af ter the intense hurricane rains of September 2004. Accordingly, two snakes, Florida cottonmouth ( Agkistrodon piscivorus ) and garter snake ( Thamnophis sirtalis) , and the southern leopard frog ( Rana utricularia ), typically associated with wet habitats, were found only in mixed flatwoods bloc ks. Differences in moisture level and water availability for breeding undoubtedly impacted anuran abundance among blocks, favoring we tter flatwoods (Table B3). Effects of block type
41 on other groups, especially where richness a nd diversity was higher, were likely related to the more productive and heterogeneous nature of mixed flatwoods over sandhill (Kirkman et al. 2001). Other species found only in mixed flatwoods blocks include eastern glass lizard ( Ophisaurus ventralis ), corn snake ( Elaphe guttata ), Florida pine snake ( Pituophis melanoleucas ), and hognose snakes ( Heterodon spp.). Corn snakes and eastern glass lizards are normally more common in me sic flatwoods communities than sandhill (Tennant 1997). Florida pine snakes and hognose snakes should be equally abundant in all blocks, but are also affected by conditions favoring their prey spec ies. Hognose prefer sandy uplands with sufficient moisture-retaini ng cover for their anuran prey (Tennant 1997), which were more abundant in flatwoods. Pine snakes are generally rare and spend up to 85% of time underground in the burrows of fossorial animals, many of which are prey items (Franz 1992, Tennant 1997). Burro ws of a common pine snake prey item, pocket gophers ( Geomys pinetus ) were more commonly seen in sandhill than flatwoods blocks. The absence of pine snakes from burned sandhill samples at CBTS cannot be readily explained except as an artifact of low capture rate or detection ability for this species. Capture sites in the block containing more than one major habitat type (higher heterogeneity) also demonstrated the most obs ervable treatment benefit (an interaction, P < 0.05) for herpetofauna studied. This effect is likely relate d to the more frequent fire requirements of flatwoods over sandhill type habitats due to higher rate of biomass production (Robbins and Myers 1992, Kirkman et al. 2001). Flatwoods sites had a greater proportion of woody understory shr ubs (e.g. wax myrtle, gallberry and saw
42 palmetto) that can quickly increase canopy cover and litter depth if unburned for more than a few years. In contrast to the diverse mixed flatwoods blocks, logged sandhill areas represent a more extreme habitat at the xeric end of the moisture gradient. The original tree harvest in this area left lower plant diversity a nd less heterogeneity. De spite a long recovery time, without much standing longleaf pine stoc k available, these units are still the least ecologically intact of the three categories sampled. Very little was done to restore longleaf pine and fire to these units until after 1997 (personal communication CBTS environmental staff). These areas have le ss to offer in terms of diversity and niche availability, conditions reflected in the depressed richness a nd diversity of herpetofauna. Effect of Treatment on Habitat Variables A lack of difference in vegetation volu me and woody debris structure shows the ability of upland pine systems to retain these characteristics w ith a typical prescribed fire program. As discussed in Chapter 2, the unde rstory plants here have the ability to resprout after a fire and quickly reesta blish the ground level structure and higher nutritional value necessary for many species, in cluding herpetofauna and their prey. Of variables measured, woody debris and canopy cover are the two that may require multiple burn rotations or a wildfire event for significan t changes to occur. Understory vegetation, leaf litter, and bare sa nd are more drastically altered with low-intensity prescribed fire. In contrast to high fire areas observed in other studies (e.g. Mushinsky 1985), management efforts here did not produce a significantly different landscap e with few trees and sparse groundcover. Effects here instead constituted a more intermediate le vel of disturbance, enhancing habitat and moving parameters measured to a more balanced state (Table 3C).
43 Habitat Variables and Individual Species In west-central Florida sa ndhill plots Mushinsky (1985) found that lizard species such as the six-lined racerunner favor open canopy stands and bare sand patches produced by frequent fires (<2 yr return interval). Litt et al. (2001) found more racerunner and fence lizards in burned areas. Cunningham et al. (2002) observed that the closely related Cnemidophorus tigris benefited most from conditions produced after a wildfire. Fence lizards also increased, but had wider preferen ce of habitat. In the upland pine areas here, the six-lined racerunner show ed the highest degree of preference for the open habitat features found in recent burn areas. The significant, but weak, negative correla tion of eastern fence lizards with open features (Table 4-4) seems curi ous in light of the results in Litt et al. (2001). This lizard depends on perch sites above th e ground, several inches to several feet, however, for hunting (Jackson 1973). Increased tree density w ould increase availability of perch sites on living or dead trees. In both treatment type s, however, this increase would also result in more leaf litter, less bare sand and mo re canopy cover, which would explain the negative response to these variables. Otherwise, enough woody debris and solar exposure was available near all trap sites to provide for the other needs of this animal.
44 CHAPTER 6 CONCLUSIONS Summary and Conclusions Examination of the interrelat ion of habitat heterogeneity, disturbance and diversity at CBTS revealed the importance of both fi re disturbance and community type to herpetofaunal communities in southeastern hi gh pine habitats. Recent fire events in upland pine communities were positively correl ated with greater measurable abundance, richness and diversity for herpetofauna, rept iles and snakes. Thes e observations support the original hypothesis concerning the benefit of prescribed fi re disturbance on habitat for herpetofauna and reptiles. Separate analys is of measurements on lizards and anurans revealed little treatment effect. Due to the la rge size of study areas a nd type of trap array used, larger snakes were captured and benef its of burning for these wide-ranging animals were also observed. It was found that differences in habita t type between blocks also affected herpetofauna, with the greatest benefit of burni ng appearing in flatwoods. Regardless of burn treatment, flatwoods blocks also had better habitat conditions over logged sandhill (and occasionally unlogged sandhill) for herpetofauna. Separa te analysis of reptiles and snakes also showed a favorable response to flatwoods blocks over sandhill and logged sandhill. This trend is most closely linked to the concept of overall habitat heterogeneity (increased in these mixed areas) and its pos itive relation to species productivity, richness and diversity.
45 Terrestrial anuran species sampled appeared to be unaffected ove rall by fire regime and only slightly affected by community type and water availability. Few species of anuran were captured, but those sampled had adaptations allowing the use of more open areas of decreased litter and lower moisture. Of the more commonly sampled species , only the six-lined racerunner had significantly higher captures in burned areas than in unburne d areas. Total captures of this animal were positively correlated with open sand and open canopy measurements, and negatively correlated w ith litter volume estimates. Conversely, the eastern fence lizard was negatively influenced by more ope n canopy and greater amounts of bare sand. This animal appears to thrive with increas ed litter volumes, but less than half of the variation in captures between treatments coul d be explained by any single factor. This species appears to have a wider tolerance of habitat conditions than the racerunner, supported by roughly equal capture s in both treatment types. The objective of knowing more about the shor t-term effects of prescribed fire on fauna in longleaf pine systems was also achieved. This study expands on findings of other studies in high pine (Mushinsky 1985, Litt et al. 2001) by examining the reactions of snakes to prescribed fire treatments in this community. Limitations of Study Unfortunately, capture methods were not successful at the cap ture of treefrogs (Hylidae) and caudates, which are typically present in sandh ill and flatwoods habitat of north Florida. Addition of ot her trap types may have enable d effective sampling of these other species, while doubling cost and effort required for the study. These exclusions could potentially have lowe red the total number of amphibians sampled, as well as altering diversity and richness measures for th is group. A complete picture of the effect
46 of fire on amphibians in this system cannot be gained from the results of this study. Also, as mentioned above, traps had no provision fo r the capture of turtles, including the terrestrial gopher tort oise and box turtles ( Terrapene carolina ) likely to be present in the study areas. All tortoises we re captured opportunist ically when observed near (<100 m from) the traps, save one small juvenile (42 mm TTL) captured inside trap B3B. This study is limited in that complete fi re histories are not known for the sites studied. While not an unusual situation, repe ated fire events prior to the known period could have potential additive effects on major habitat structure. Th erefore, preexisting differences between study units could not be factored out, but only minimized during selection. Though an ideal study would have complete control over length of sampling period, ignition type, timing, fr equency and intensity of burning for all plots, as mentioned in Chapter One, this project is limited to using recent burn history and community types to define and divide treatment areas. The project was also limited by the scope of time, funding and field assistan ce available. Despite efforts to improve experimental design, the comparative observati onal nature of this study combines with the factors mentioned above to reduce the streng th of causal inference of this study. Like many short-term ecological studies, control of many variables potenti ally affecting the sampling outcome was simply not possible. When combined with findings from many other studies, however, a stronger statement can be made about the benefit of prescribed fire in longleaf pine communities. Suggestions for Research and Management Results here are strongly applicable to a domain that includes many conservation areas in the southeastern US with upla nd pine communities and fire management
47 programs. Results are slightly less signifi cant and relevant when referenced under the context of fire dependent communities in other parts of the world. Further studies on herpetofauna in fire prone community types should be performed on areas where complete records of fire intens ity, frequency and seas on are available. In addition, use of a time series approach would drastically improve causal inference. This would require random selection of treatment areas after pre-sampli ng (observation) and prior to matching (intensity, season and freque ncy) fire treatments and re-sampling after each event. Level of disturba nce (fire intensity) should be given special attention during treatment. I would also suggest focus on more limite d taxonomic groups, as some researchers have done, examining only suborders or indivi dual species. This allows for a more thorough examination and testing of m echanistic sub-hypotheses and better understanding of specific needs of a species. In addition, examination of this question for certain types of herpetofauna, such as anur ans, may require the use of multiple trapping techniques to properly sample all species pres ent. If traps of the type used here are considered, a deeper trap that need not be as wide could be use d. This would further reduce stress to captured animals by taking ad vantage of thermally constant subsurface conditions. Based on findings of this study and other information available on this topic, I suggest that regular prescribed burning of southeastern upland pine communities be continued. Once areas have been returned to a maintenance state by initial cool weather burning followed within five years by a gr owing season fire, a random return interval such as that outlined in Robbins and Myers ( 1992) can be used to more closely replicate
48 natural fire frequency in any burn unit. Th is variation will al low open area and litter dwelling herpetofauna alike to find suitabl e habitat conditions throughout sandhill and flatwoods landscapes. In addition to maintenance of intact pine communities, I feel the results of this study have implications for altered sites as well. Reduced presence and diversity of herpetofauna in logged sandhill blocks points to reduced habita t value there. In order to maximize the quality and conn ection of larger sections of upland pine remaining on public and private conservation lands, it is important to restore all native component species and ecological processes. A restorat ion program that includes fire, Understory monitoring and replanting of longleaf pine in some sections of logged sandhill is currently occurring at CBTS. Programs lik e this are crucial to the responsible management of upland pine for the use and benefit of herpetofauna and other native wildlife. Due to their ectothermic and non-migrator y nature, many herpetofaunal species are highly sensitive to abse nce of natural fire events in l ongleaf pine sandhill and flatwoods. Even non-catastrophic burning events can signif icantly alter habitat structure at a scale important to reptiles and amphibians in this sy stem. With more info rmation on effects of prescribed burning, fire dependent ecosystems and their resident herpetofauna can be more effectively managed and maintained.
49 APPENDIX TRAP SITE CHARACTERISTICS Table A-1. Characteristics of individual trap sites at Camp Blanding Training Site Trap Notes Unit Size Trap Community Type Burn Dates UTM X -Coordinate UTM Y-Coordinate B1C RCW cluster 49.4 ha Flatwoods/ sandhill 1/16/02, 5/5/03 408621 3314192 B1A 49.4 ha Flatwoods 2/08/02, 5/5/03 408366 3313993 B1B RCW cluster 49.4 ha Sandhill/ flatwoods 1/16/02, 5/5/03 408168 3313544 B3C 55.5 ha Logged sandhill Summer 1997, 4/30/03 405369 3304351 B3B 55.5 ha Logged sandhill Summer 1997, 4/30/03 405538 3303636 B3A 55.5 ha Logged sandhill Summer 1997, 4/30/03 405738 3303633 B2A RCW cluster site, wetland area 61.7 ha Sandhill 1/24/02, 6/25/03 405011 3318799 B2B RCW cluster site,wetland area 61.7 ha Sandhill 1/24/02, 6/25/03 404175 3318778 B2C RCW cluster site, wetland area 61.7 ha Sandhill 1/24/02, 6/25/03 404351 3318852 U1A RCW cluster site 48.6 ha Flatwoods No burn record 403934 3318023 U1B RCW cluster site 48.6 ha Flatwoods/ sandhill No burn record 403807 3317750 U1C RCW cluster site 48.6 ha Sandhill/ flatwoods No burn record 404609 3318197 U2A Small wetland area 56.1 ha Sandhill No burn record 407702 3304756 U2B Small wetland area 56.1 ha Sandhill No burn record 407461 3304712 U2C Small wetland area 56.1 ha Sandhill No burn record 407406 3304323
50 Table A-1 Continued Trap Notes Unit Size Trap Community Type Burn Dates UTM X -coordinate UTM Y-Coordinate U3A 78.2 ha Logged sandhill No burn record 407471 3303346 U3B 78.2 ha Logged sandhill No burn record 407360 3303512 U3C 78.2 ha Logged sandhill No burn record 407186 3303259
51 A B Figure A-1.Typical unburned (A) and burne d (B) logged sandhill at Camp Blanding Training Site.
52 A B Figure A-2. Typical unburned (A) and burned (B) sandhill community at Camp Blanding Training Site.
53 A B Figure A-3. Typical unburned (A) and burne d (B) flatwoods community at Camp Blanding Training Site.
54 Figure. A-4. Map of north study units B2 and U1 at Camp Blanding Training Site.
55 Figure. A-5. Map of Study unit B1 at Camp Blanding Training Site.
56 Figure A-6. South study units B3, U2 and U3 at Camp Blanding Training Site.
57 Figure A-7. Relative positi on of study units at Camp Blanding Training Site.
58 LIST OF REFERENCES Allen, R. and W.T. Neill. 1953. The race -runner lizard. Florida Wildlife 6: 46-47. Anderson R.C. and E.S. Menges. 1997. Effects of fire on sandhill herbs: Nutrients, mycorrhizae, and biomass allocation. Am erican Journal of Botany 84(7): 938-948 Aresco, M.J. and C. Guyer. 1999. Burrow abandonment by gopher tortoises in slash pine plantations of the Conecuh National Fore st. Journal of Wildlife Management 63(1): 26-35. Barbault, R. 1976. Structure et dynamique dÂ’un peuplement dÂ’Amphibiens en savane protÃ©gÃ©e du feu (Lamto, cote-d Â’Ivoire). Terre Vie 30: 246-263 Biswell, H.H. 1974. Effects of fire on chaparral. Pp. 321-364 in Kozlowski, T.T. and C.E. Ahlgren (eds.). Fire and Ecosystems . Academic Press, New York, NY. Brattstrom, B.H. 1979. Amphibian temperat ure regulation studies in the field and laboratory. American Zool ogist 19(1): 345-356. Brisson, J.A., J.L.Strassburg and A.R.Temp leton. 2003. Impact of fire management on the ecology of the collared lizard ( Crotaphytus collaris ) populations living on the Ozark Plateau. Animal C onservation 6: 247-254 Bury, R.B. 1983. Differences in amphibian populations in logged and old growth redwood forests. Northwest Scientist 57: 167-178 Campbell, H.W. and S.P. Christman. 1982. Th e herpetological components of Florida sandhill and sand pine scr ub associations. Pp. 163-171 in N. J. Scott, Jr., (ed.). Herpetological communities. Wildlife Research Report No.13. U.S. Fish and Wildlife Service, Washington, D.C. Camper, J.D. and J.R. Dixon. 2000. Food habits of three species of striped whipsnakes, Masticophis spp. (Serpentes: Colubridae). Texas Journal of Science 52(2): 83-92 Cavitt, J.F. 2000. Fire and tallgrass prairi e reptile community: effects on relative abundance and seasonal activity. Jour nal of Herpetology 34(1): 12-20 Connell, J.H. 1978. Disturbance in tropical rain forests and coral reef s. Science 199(24): 1302-1309.
59 Cook, G.D. 1994. The fate of nutrients during fires in a tropical savanna. Australian Journal of Ecology 19: 359-365 Corn, P. S. 1994. Straight-line drift fences and pitfall traps. Pp. 109-117 in W. R. Heyer, M. A. Donnelly, R. W. McDiarmid, L. C. Hayek, and M. S. Foster (eds.). Measuring and Monitoring Biological Diversity: Standard Methods for Amphibians. Smithsonian Instit ution, Washington, D.C. Cox, J., D. Inkley and R. Kautz. 1987. Ecology and Habitat Protection Needs of Gopher Tortoise (Gopherus polyphemus) Populati ons Found on Lands Slated for Large Scale Development in Florida. N ongame Wildlife Program Technical Report No. 4. Florida Game and Fresh Water Fish Commission, Tallahassee, FL. Cunningham, S.C., R.D. Babb, T. R. Jones, B.D. Taubert and R.Vega. 2002. Reaction of lizard populations to a catastrophic wildfire in a central Arizona mountain range. Biological Conservation 107: 193-201. Davis, S.S., R.B. Mitchell and S. Demara is. 2000. Trap-revealed microhabitat use by small mammals in monoculture grasslands. Texas Journal of Science 52(3): 195200 Diemer, J.E. 1986. The ecology and management of the gopher tortoise in the Southeastern United States. Herpetologica 42(1): 125-132 Driscoll, D.A. and J.D. Roberts. 1997. Im pact of fuel-reduction burning on the frog Geocrinia lutea in southwest Western Australia. Australian Journal of Ecology 22(3): 334-339 Enge, K.M. 1997. Habitat occurrence of Florid aÂ’s native amphibians and reptiles. Tech. Rep. No. 16. Florida Game and Freshwat er Fish Commission, Tallahassee, FL. 44 pp + vi. Enge, K.M. 2001. The pitfalls of pitfall traps. Journal of Herpetology 3: 467-478. Enge, K.M. and W.R. Marion. 1986. Effects of clearcutting and site preparation on herpetofauna of a north Florida flat woods. Forest Ecology and Management 14:177-192. Environmental Systems Research Institute. 1999. Arcview Version 3.2. Environmental Systems Research Institute, Inc., Redlands, CA. Faria, A.S., A.P. Lima and W.E. Magnusson. 2004. The effects of fire on behavior and relative abundance of three lizard species in an Amazonian savanna. Journal of Tropical Ecology 20: 591-594 Florida Natural Areas Inventory (FNAI). 1990. Guide to the natural communities of Florida. Florida Department of Natu ral Resources. Tallahassee, FL. 111 pp.
60 Franz, R. 1992. Florida pine snake. Pp. 254-258 in Moler, P.E. (ed.). Rare and Endangered Biota of Florida, Volume III. Amphibians and Reptiles. University Press of Florida, Gainesville, FL. Fredericksen, N.J. and T.S. Fredericksen. 2002. Terrestrial wildlife responses to logging and fire in a Bolivian tropical humid fore st. Biodiversity and Conservation 11: 2738 Gillison, A.N. 1983. Tropical savannas of Au stralia and the southwest Pacific. in BourliÃ¨re, F. (ed.). Ecosystems of the World: Tropical Savannas. Elsevier Scientific Publishing Co., Amsterdam, Netherlands. Gillon, D. 1983. The fire problem in tropical savannas. Pp. 617-641 in F. BourliÃ¨re, (ed.). Ecosystems of the World: Tropical Savanna s. Elsevier Scientific Publishing Co., Amsterdam, Netherlands. Glitzenstein, J.S., W. J. Pl att and D. Streng. 1995. Effects of fire regime and habitat on tree dynamics in north Florida longleaf pine savannas. Ecological Monographs 65(4): 441476 Glitzenstein, J.S., D.R. Stre ng and D.D. Wade. 2003. Fire frequency effects on longleaf pine ( Pinus palustris , P.Miller) vegetation in South Ca rolina and Northeast Florida, USA. Natural Areas Journal 23:22-37 Greenberg, C.H. 2001. Response of reptile and amphibian communities to canopy gaps created by wind disturbance in the southe rn Appalachians. Forest Ecology and Management 148: 135-144 Greenberg, C.H., D.G. Neary, a nd L.D.Harris. 1994. Effect of high-intensity wildlfire and silvicultural treatments on reptile communities in sand-pine scrub. Conservation Biology 8(4): 1047-1057 Hanchey, M.F. and K.T. Wilkins. 1998. Hab itat associations of the small mammal community in the Grand Prairie of north-cen tral Texas. Texas Journal of Science 50(2):107-122 Herzog, H.A., Jr. and G.M. Burghardt. 1974. Pr ey movement and pred atory behavior of juvenile western yellow-bellied racers Coluber constrictor mormon. Herpetologica 1974 30(3): 285-289 Hoffman, W. A. 1999. Fire and population dyna mics of woody plants in a neotropical savanna: matrix model projec tions. Ecology 4: 1354(1) Hooge, P.N., W. Eichenlaub, and E. Solom on. 1999. The animal movement program. USGS, Alaska Biological Scie nce Center, Glacier Bay, AK. Jackson, J.F. 1973. The phenetics and ecology of a narrow hybrid zone. Evolution 27(1): 58-68
61 James, S.E. and R.T. MÂ’Closkey. 2003. Lizard microhabitat and fine fuel management. Biological Conservation 114: 293-297 Jandel Corporation. 1995. Sigmastat for Wi ndows Version 2.0. Jandel Scientific Software Corporation. San Rafael, CA. Karr, J. R. and K.E. Freemark. 1985. Distur bance and vertebrates: an integrated perspective. Pp. 153-168 in S.T.A. Pickett and P.S. White (eds.). The Ecology of Natural Disturbance and Patch Dynamics. Academic Press, Inc. Orlando, FL. Kemp, E.M. 1981. Pre-quarternary fire in Australia. Pp.3-21 in A.M. Gill, R.H. Groves and I.R. Noble (eds.). Fire and the Aust ralian Biota. Australian Academy of Science, Canberra, Australia. King, F.W. 1998. Florida Army Na tional Guard, Camp Blanding Tr aining Site Integrated Natural Resources Management Plan. Fl orida Museum of Natural History, Gainesville, FL. Kirkland, G.L. Jr., H.W. Snoddy and T.L. Amsl er. Impact of fire on small mammals and amphibians in a central Appalachian deciduous forest. American Midland Naturalist 135: 253-260. Kirkman, L.K., R.J. Mitchell, R.C. He lton and M.B. Drew. 2001. Productivity and species richness across an environmental gr adient in a fire-dependent ecosystem. American Journal of Botany 88(11): 2119-2128 Komarek, E.V. 1964. The natural history of lightning. Proc. Tall Timbers Fire Ecology Conference 3: 139-183. Tall Timbers Re search Station, Tallahassee, FL. Komarek, E.V. 1969. Fire and animal be havior. Proc. Tall Timbers Fire Ecology Conference 9:161-207. Tall Timbers Research Station, Tallahassee, FL. Komarek, E.V. 1972. Lightning a nd fire ecology in Africa. Proc. Tall Timbers Fire Ecology Conference, 11: 473-512. Tall Timber s Research Station, Tallahassee, FL. Kutiel, P. 1997. Spatial and temporal he terogeneity of species diversity in a Mediterannean ecosystem following fire. International Journa l of Wildland Fire 7(4): 307-315 Labisky, R.F. and J.A. Hovis. 1987. Comparis on of vertebrate wildlife communities in longleaf pine and slash pine habita ts in North Florida. Pp. 201-228 in H.A. Pearson, F. Smeins and R. Thill (eds.). Ecological, Physical, and Socioeconomic Relationships within Southern National Fo rests. General Tec hnical Report SO-68. Southern Forest Experiment Station, New Orleans, LA. Letnic, M, C. R. Dickman, M. K. Ti schler, B. Tamayo and C.L. Beh. 2004. The responses of small mammals and lizards to post-fire succession a nd rainfall in arid Australia, Journal of Arid Environments 59(1): 85-114
62 Lillywhite, H.B. 1977. Animal responses to fire and fuel management in chaparral. Pp. 368-373 in H. A. Mooney and C.E. Conrad (e ds.). Proceedings of the Symposium on the Environmental Consequences of Fire and Fuel Management in Mediterranean Ecosystems. Forest Service, USDA. Washington, DC. Lillywhite, H.B. and F. North. 1974. Perching behavior of Sceloporus occidentalis in recently burned chaparral. Copeia 1974: 256-257 Lips, K.R. 1991. Vertebrates associated with tortoise ( Gopherus polyphemus ) burrows in four habitats in south-ce ntral Florida. Journal of Herpetology 25(4): 477-481 Litt, A.R., L. Provencher, G.W. Tanner a nd R. Franz. 2001. Herpetofaunal responses to restoration treatments on longleaf pine sa ndhills in Florida. Restoration Ecology 9(4): 462-474. MacArthur R.H. 1965. Patterns of species diversity. Biological Review 40: 510-573 Macdonald, L.A., and H.R. Mushinsky. 1988. Foraging ecology of the gopher tortoise, Gopherus polyphemus , in a sandhill habitat. He rpetologica 44(3): 345-353. Mackey, R.L. and D.J. Currie. 2001. The di versity-disturbance relationship. Is it generally strong and peaked? Ecology 82(12): 3479-3482 McLeod, R.F. and J.E. Gates. 1998. Response of herpetofaunal communities to forest cutting and burning at Chesapeake Farms, Maryland. The American Midland Naturalist 139(1): 164-173 McQuaid, C.D. and K.M. Dower. 1990. Enhancement of habitat heterogeneity and species richness on rocky shores inu ndated by sand. Oecologica 84(1): 142-144. Means, D. B. 1985. Radio-tracking the easte rn diamondback rattlesnake. National Geographic Society Resear ch Report, 18: 529-536. Means, D. B. and H.W. Campbell. 1982. Eff ects of prescribed burning on amphibians and reptiles. Pp. 89-97 in Ge ne W. Wood, (ed.). Prescrib ed fire and wildlife in Southern forests. Belle W. Baruch Fo rest Science Institu te, Georgetown, SC. Means, D.B., C.K. Dodd, S.A. Johnson and J.G. Palis. 2004. Amphibians and fire in longleaf pine ecosystems: Response to Schurbon and Fauth. C onservation Biology 18(4): 1149-1153 Mistry, J. 2000. World Savannas: Ecology a nd Human Use. Pearson Education, Ltd. Harlow, England. Moseley, K.R., S.B. Castleberry, and S.H. Sc hweitzer. 2003. Effects of prescribed fire on herpetofauna in bottomland hardwood forest s. Southeastern Na turalist 2(4): 475486
63 Mushinsky, H.R. 1985. Fire and the Florida sandhill herpetofaunal community: With special attention to the responses of Cnemidophorus sexlineatus . Herpetologica 41: 333-342 Mushinsky, H.R. 1992. Natural history and abu ndance of southeastern five-lined skinks, Eumeces inexpectatus on a periodically burned sandhill in Florida. Herpetologica 48: 307-312. Norman, M.J.T. and Wetselaar, R. 1960. Losse s of nitrogen on burning native pasture at Katherine, NT. Journal of Australian In stitute of Agricultural Science 26: 272-273. Odum, E.P. 1993. Ecology and Our Enda ngered Life-support Systems. Sinauer Associates, Inc. Sunderland, MA. Outcalt, K.W. 2000. The longleaf pine ecosyst em of the south. Native Plants Journal 1(1): 43-51. Outcalt, K.W. and R.M. Sheffield. 1996. The longleaf pine forest: Trends and current conditions. Resource Bulletin SRS-9, 23 pp. USDA Forest Service, Southern Research Station, Asheville, NC. Parr, C.L. and S.L. Chown. 2003. Burning issues for conservation: A critique of faunal fire research in southern Afri ca. Austral Ecology 28: 384-395 Pielou, E.C. 1977. Mathematical Ecology. John Wiley and Sons, New York, NY. 385 pp. Pough, F.H., R. M. Andrews, J. E. Cadle, and M. Crump. 2000. Herpetology. Prentice Hall, NJ. 577 pp. Pounds, J.A. and J.F. Jackson. 1983. Utilization of perch sites by sex and size classes of Sceloporus undulatus undulatus. Journal of Herpetology 17: 287-289 Reice, S.R. 2001. The Silver Lining: The Be nefits of Natural Disasters. Princeton University Press. Princeton, NJ. Ricklefs, R.E. 1977. Environmental hetero geneity and plant species diversity: A hypothesis. American Naturalist 111: 376-381. Robbins, L.E. and R.L. Myers. 1992. Seasonal ef fects of prescribed bu rning in Florida: a review. Misc. Pub. No. 8, Tall Timber s Research, Inc. Tallahassee, FL. Rudolph, C., S. Burgdorf, R. Conner, and R. Schaefer. 1999. Preliminary evaluation of the impact of roads and associated vehicu lar traffic on snake popul ations in eastern Texas. Pp. 129-136 in G. L. Evink, P. Garrett, and D. Ziegler (eds.). Proceedings of the third international conference on wildlife ecology and transportation. FLER-73-99. Florida Department of Tr ansportation, Tallahassee, Florida.
64 Russell, K.R., D.H. Van Lear, and D.C. Guynn, Jr. 1999. Prescribed fire effects on herpetofauna: review and management im plications. Wildlife Society Bulletin 27(2): 374-384. Russell-Smith, J. 2002. Australian fire regi mes: Contemporary patterns (April 1998March 2000) and changes since European se ttlement. Australia: State of the Environment Second Technical Paper Series (Biodiversity), Series 2. Department of the Environment and Heritage, Australia. San Jose, J.J. and M.R. Farinas. 1983. Changes in tree density and species composition in a protected Trachypogon savanna, Ve nezuela. Ecology 64(3): 447-453 SAS Institute. 1997. SAS/STAT software: changes and enhancements through release 6.12. SAS Institute, Carey, NC. Schweiger, E.W., J.E. Diffendor fer, R.D. Holt, R. Pierotti, and M.S. Gaines. 2000. The interaction of habitat fragmentation plant, and small mammal succession in an old field. Ecological Monographs 70(3): 383-400 Secor, S.M. and K.A. Nagy. 1994. Bioenerge tics correlates of foraging mode for the snakes Crotalus cerastes and Masticophis flagellum . Ecology 75(6): 1600-1614 Silva, J.F. 1996. Biodiversity and stability in tropical savannas. Pp. 161-171 in Solbring, Medina and Silva (eds.). Biodiversity a nd Savanna Ecosystem Processes. SpringerVerlag, Berlin, Germany. Singh, S., A.K. Smyth and S.P. Blomberg. 2002. Effect of a control burn on lizards and their structural environment in a eucalypt open-forest. Wildlife Research 29(5): 447-454 Smith, L.J., A.T. Holycross, C.W. Painte r and M.E. Douglas. 2001. Montane rattlesnakes and prescribed fire. Southw estern Naturalist 46(1): 54-61 Snyder, J.R. 1986. The Impact of Wet Season and Dry Season Prescribed Fires on Miami Rock Ridge Pineland, Everglades National Park. South Florid a Research Center Report SFRC-86/06. Everglades National Park, Homestead, FL. Sousa, W.P. 1984. The role of disturbance in natural communities. Annual Review of Ecological Systems 15: 353-391 Sparks, J.C., R.E. Masters, D.M. Engle, M.E. Payton and G.A. Bukenhofer. 1999. Influence of fire season and fire beha vior on woody plants in red-cockaded woodpecker clusters. Wildlife So ciety Bulletin 27(1): 124-133. Stout, I.J. and W.R. Marion. 1993. Pine flatwoods and xeric pine forest of the Southern (lower) Coastal Plain. Pp. 373-446 in W.H. Martin, S.G. Boyce, and A.C. Echternacht (eds.). Biodiv ersity of the southeaste rn United States, lowland terrestrial communities. John Wiley and Sons, New York, NY.
65 Swaine, M.D., W.D. Hawthorne, T.K. Orgl e. 1992. The effects of fire exclusion on savanna vegetation at Kpong, Ghana. Biotropica 24(2) A: 166-172 Swengel, A.B. 2001. A literature review of ins ect responses to fire, compared to other conservation managements of open habitat. Biodiversity and Conservation 10(7): 1141-1169. Tennant, A. 1997. A Field Guide to Snakes of Florida. Gulf Publishing Co. Houston, TX. 257 pp. Tews, J., U. Brose, V. Grimm, K. TielbÃ¶r ger, M.C. Wichmann, M. Schwager and F. Jeltsch. 2004. Animal species diversity driv en by habitat heterogeneity/diversity: the importance of keystone structures . Journal of Biogeography 31(1): 79-92 Turney, C.S.M., A.P. Kershaw, P. Moss, M.I. Bird, L.K. Fifield, R.G. Cresswell, G.M. Santos, M.L. Di Tada, P.A. Hausladen, and Y. Zhou. 2001. Redating the onset of burning at LynchÂ’s Crater (North Queenslan d): Implications fo r human settlement in Australia. Journal of Quar ternary Science 16(8): 767-771 Varner, J.M. III., J.M. Kush and R.S. Meldah l. 2000. Ecological restoration of an oldgrowth longleaf pine stand util izing prescribed fire. Pp. 216-219 in W. K. Moser and C. F. Moser (eds.). Fire and Fore st Ecology: Innovative Silviculture and Vegetation Management. Tall Timbers Fi re Ecology Conference Proceedings, No. 21. Tall Timbers Research Station, Tallahassee, FL. Wade, D. 1989. A guide for prescribed fire in southern forests. U.S. Deptartment of Agriculture, Forest Service, Southern Region, Technical Publication R8-TP 11. Atlanta, GA. Williams, K.L. and K. Mullin. 1987. Amphibian s and reptiles of longleaf-slash pine stands in Central Louisiana. Pp. 116-120 in H.A. Pearson, F. Smeins and R. Thill (eds.). Ecological, Physical, and Socio economic Relationships Within Southern National Forests. General Technical Re port SO-68. Southern Forest Experiment Station, New Orleans, LA. Williams, R.J., A.D. Griffiths and G.E. Alla n. 2002. Fire regimes and biodiversity in the savannas of northern Australia. Pp. 281-304 in Bradstock, R.A., J.E. Williams, and M.A. Gill (eds.). Flammable Australia: The Fire Regimes and Biodiversity of a Continent. Cambridge University Press, Cambridge, UK. Woinarski, J.C., J. Risler and L. Kean. 2004. Response of vegetation and vertebrate fauna to 23 years of fire exclusion in a trop ical Eucalyptus open forest, Northern Territory, Australia. Aust ral Ecology 29(2): 156-176 Wootton, J. T. 1998. Effects of distur bance on species diversity: a multitrophic perspective. The American Naturalist 152(6): 803-825.
66 Zimmerman, B.L. 1994. Inventor y and monitoring: audio st rip transects. Pp. 92-97 in Heyer, W. R., M.A. Donnelly, R.W. Mc Diarmid, L.C. Hayek and M.S. Foster (eds.). Measuring and Monitoring Biologi cal Diversity: Standard Methods for Amphibians. Smithsonian Inst itution, Washington, DC. Zug, G.R., L.J. Vitt and J.P. Caldwell. 2001. Herpetology, 2nd Edition. Academic Press, San Diego, CA.
67 BIOGRAPHICAL SKETCH Keith Charles Morin was born and grew up in southeastern Massachusetts. He received a Bachelor of Science degree in biology and environmental science from Jacksonville University in 1995. Between 1996 and 1999 he successfully completed three full time volunteer and National Service positions in natural resource management in Maryland, Virginia, and Geor gia. After working for the Department of the Air Force as a wildlife technician for over three years, Ke ith entered the University of Florida in the Department of Wildlife Ecology and Conserva tion to continue his education. He currently works for the Florida Department of Environmental Protection as a park biologist in Crysta l River, Florida.