1 HOW AQUATIC FAUNA RESPO NDED TO LARGE SCALE MANAGEMENT ON LAKE TOHOPEKALIGA, FLORIDA By MELISSA ANN DESA 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 2008
2 2008 Melissa Ann DeSa
3 To my parents, whose love, encouragement and s upport got me this far; and to my husband, who was supportive and kept me sane throughout the process.
4 ACKNOWLEDGMENTS I would lik e to thank my advisor, Dr. Wile y Kitchens, who accepted me as a graduate student. His support throughout th is project has been helpfu l and encouraging. I feel particularly privileged to have worked for Dr. Kitchens and the wonderful people who comprise the Florida Fish and Wildlife Cooperative Resear ch Unit in Gainesville, FL. Without many of these people this thesis could not have happe ned. Many long hours in th e field were spent on this project, I was lucky to have great companionship. Thanks go to the following people for their hard work. Adam Betuel, Andrea Bowlin g, Janell Brush Melinda Connors, Carolyn Enloe, Jonathan Felix, Edward Larivee, Kate Leonard, Ashley Peele, Alison Pevlar, Derek Piotrowicz, Jono Saunders, Amy Sc hwarzer, Bradley Shoger, Taylor Tidwell, Zach Welch and Christa Zweig. Special thanks go to Julien Martin and the PATUXENT folks who offered advice and insight during the analysis phase. Additionally I would like to acknowledge the contributions of Hardin Waddle and Christopher Rota, who helped get me started on learning and using PRESENCE. Finally, and perhaps most importantly, I thank Ann Marie Muench, for whom this thesis was originally intended! A nn Marie conducted the or iginal pre-management studies, but due to changes in the scheduled pr ogram, was unable to document both the before and after responses. Her work was exemplary and paved the way for me to continue the research. She was always helpful in answering my questions, and her hard work establishing the initial project was greatly appreciated.
5 TABLE OF CONTENTS page ACKNOWLEDGMENTS...............................................................................................................4 LIST OF TABLES................................................................................................................. ..........7 LIST OF FIGURES.........................................................................................................................8 ABSTRACT...................................................................................................................................10 CHAP TER 1 INTRODUCTION..................................................................................................................12 Brief History of Floridas Wetland Landscape....................................................................... 12 Florida Fisheries and Lake Management................................................................................14 Lake Tohopekaliga.................................................................................................................16 Need for New Sampling and Statistical Techniques..............................................................17 Study Objectives.....................................................................................................................21 Overview of Terms Used........................................................................................................22 Habitat.............................................................................................................................22 Location...........................................................................................................................22 Statistical Parameters and Covariates..............................................................................23 2 METHODS AND STATIS TICAL ANALYSES ................................................................... 26 Trapping..................................................................................................................................26 Trap Descriptions....................................................................................................................27 Sampling Methods..................................................................................................................27 Post-Management Field Methodology Adjustments.............................................................. 28 Hurricane Complications........................................................................................................30 Data Analyses.........................................................................................................................31 Assumption Violations.................................................................................................... 32 Model Selection Criteria.................................................................................................. 33 3 RESPONSE OF HERPETOFAUNAL COMMUNITIES......................................................37 Introduction................................................................................................................... ..........37 Species Accounts....................................................................................................................38 Results.....................................................................................................................................40 Discussion...............................................................................................................................41 Amphibians..................................................................................................................... .42 Reptiles............................................................................................................................44 The Bigger Picture........................................................................................................... 45
6 4 RESPONSE OF FISH COMMUNITIES............................................................................... 53 Introduction................................................................................................................... ..........53 Results.....................................................................................................................................55 Effects of Habitat....................................................................................................................57 Discussion...............................................................................................................................78 Largemouth Bass............................................................................................................. 79 Bluegill............................................................................................................................81 Other Fish Species........................................................................................................... 81 Summary..........................................................................................................................84 5 RESPONSE OF NATIVE AND EXOTIC APPLE SNAIL COMMUNITIES...................... 86 Introduction................................................................................................................... ..........86 Results.....................................................................................................................................87 Discussion...............................................................................................................................93 6 FINAL CONCLUSION AND IMPLICATI ONS FOR FUTURE MANAGEMENT ............ 97 Landscape Level Considerations............................................................................................ 98 Water Level Considerations...................................................................................................99 Vegetation Management Considerations.............................................................................. 100 Final Thoughts and Summary...............................................................................................102 APPENDIX A HERPETOFAUNA SPECIES LIST AND OCCUPANCY MODEL OUTPUT S............... 104 B FISH SPECIES LIST AND O CCUPANCY MODEL OUTPUTS ...................................... 106 C APPLE SNAIL OCCUPANCY MODEL OUTPUTS......................................................... 110 D FUTURE MONITORING RECOMMENDATIONS.......................................................... 111 E IMPORTANCE OF DETECTION PROB ABILITY: SOME EXAMPLES ........................ 112 LIST OF REFERENCES.............................................................................................................118 BIOGRAPHICAL SKETCH.......................................................................................................129
7 LIST OF TABLES Table page A-1 Herpetofauna species list.................................................................................................104 A-2 Occupancy models within 3 AIC for each herp etofaunal species................................ 105 B-1 Fish species list.......................................................................................................... ......106 B-2 Occupancy models within 3 AIC for each fish species ................................................ 107 C-1 Occupancy models within 2 AIC for both exotic and native apple snail species.........110
8 LIST OF FIGURES Figure page 1-1 Map of Lake Tohopekaliga within greater Everglades watershed. ....................................24 1-2 Daily mean lake stage for Lake Tohopekaliga from January 1942 until June 2008.......... 25 2-1 Trap set-up................................................................................................................ .........35 2-2 Lake Toho transect locations.............................................................................................36 3-1 Occupancy estimates with standard error for amphiuma................................................... 46 3-2 Occupancy estimates with standard error for siren............................................................ 47 3-3 Occupancy estimates with standard error for pig frog....................................................... 48 3-4 Occupancy estimates with standard error for Florida banded water snake. ...................... 49 3-5 Occupancy estimates with standard error for Florida green water snake. ......................... 50 3-6 Occupancy estimates with st andard error for stinkpot. ......................................................51 3-7 Occupancy estimates with standa rd error for striped m ud turtle....................................... 52 4-1 Occupancy estimates with stand ard error for bluegill....................................................... 59 4-2 Habitat occupancy estimates with standard error for bluegill 2006-2007......................... 60 4-3 Occupancy estimates with standard error for dollar sunfish.............................................. 61 4-4 Habitat occupancy estimates with st andard error for dollar sunfish 2006-2007. .............. 62 4-5 Occupancy estimates with stan dard error for largem outh bass......................................... 63 4-6 Habitat occupancy estimates with sta ndard error for largem outh bass 2005-2006........... 64 4-7 Occupancy estimates with standa rd error for Seminole killifish. ...................................... 65 4-8 Habitat occupancy estimates with sta ndard error for Sem inole killifish 2005-2007......... 66 4-9 Occupancy estimates with standard error for sailfin catfish.............................................. 67 4-10 Occupancy estimates with standard error for redear.......................................................... 68 4-11 Occupancy estimates with stan dard error for armored catfish. ..........................................69 4-12 Occupancy estimates with standa rd error for blue-spotted sunfish. ..................................70
9 4-13 Occupancy estimates with st andard error for chubsucker ................................................. 71 4-14 Occupancy estimates with standard error for gar.............................................................. 72 4-15 Habitat occupancy estimates with standard error for gar 2006-2007. ............................... 73 4-16 Occupancy estimates with stan dard error for sailf in molly............................................... 74 4-17 Occupancy estimates with standard error for spotted sunfish........................................... 75 4-18 Occupancy estimates with standard error for warmouth................................................... 76 4-19 Habitat occupancy estimates with standard error for warm outh 2005-2007..................... 77 5-1 Occupancy estimates with standard error for both apple snail species. ............................. 88 5-2 Habitat occupancy site es tim ates with standard error for exotic apple snails 20052006....................................................................................................................................89 5-3 Habitat occupancy estimates with standard error for native apple snails 2006-2007. ....... 90 5-4 Colonization estimates with standard error for both apple snail species. .......................... 91 5-5 Habitat colonization estima tes with standard error for exotic apple snails 2006-2007. .... 92 E-1 Fish detection probabilities from 2002 that were influenced by tem perature................. 113 E-2 Fish detection probabilities from 2002 th at were influenced by lake stage and tem perature.................................................................................................................... ..114 E-3 Bass detection probabilities from 2005-2006 that were influenced by tem perature....... 115 E-4 Fish detection probabilities from 2006-2007 that were influenced by lake stage and tem perature.................................................................................................................... ..116 E-5 Herpetofauna detection probabilities from 2002 that were influenced by lake stage ...... 117
10 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 HOW AQUATIC FAUNA RESPO NDED TO LARGE SCALE MANAGEMENT ON LAKE TOHOPEKALIGA, FLORIDA By Melissa Ann DeSa December 2008 Chair: Wiley M. Kitchens Major: Interdisciplinary Ecology Many Florida lakes have undergone dramatic ch anges in hydrology that have restricted water level fluctuation, reduced floodplai n area and triggered accumulation of dense monospecific stands of vegetation and muck (tussocks) in narrowed littoral zones. Dense plant communities and the conditions they impart on the environment are considered unsuitable habitat for aquatic fauna, causing lake managers to employ draw downs and sediment removal as a means of offsetting these conditions. Multiple studies document the benefit conveyed to economically important sportfish, primarily largemouth bass ( Micropterus salmoides ). However, most fish studies fail to meet two major criteria, including ability to actually sample fishes within dense tussock formations and the failure to consider de tection probabilities. Surprisingly scant data exists documenting the impacts of lake management on other aquatic inhabitants. Our study sampled a variety of littoral fauna within dense pickerelweed (Pontederia cordata ) tussocks, including sport and forage fishes reptiles, amphibians and apple snails prior to, and after a major lake management project on Lake Tohopekaliga, FL. Unfortunately an active hurricane season followed on the heels of management actions, complicating the study somewhat. However, we still gained valuable insight into faunal respon se to extreme habitat management.
11 Results for fish were mixed. Of the thirteen species captured enough to obtain statistical estimates, five including bluegill and largemouth bass increased occupancy after the management. One sportfish, the redear sunfis h barely responded. The other seven species including two sportfish, the warmouth and spotted sunfish, as well as some forage fish species declined in occupancy. All herpetofaunal specie s declined dramatically after the management. The large-scale extent of this project probably triggered migrations, survival strategies and caused direct mortality for many caught in the pr ocess of the dry down and sediment removal. Some species commonly associated with vegetated and murky envi ronments began to recover by 2006-2007, indicating that pre-management conditions were returning. As vegetation continues to recover across the lake, animals will continue to respond. This study documents the immediate responses to management, which ought to be supplemented with further monitoring to understand long term impacts of lake management.
12 CHAPTER 1 INTRODUCTION Brief History of Floridas Wetland Landscape The state of Florida is literally awash in water with over 7700 lakes dotting the landscape, the enormous swath of Everglades swamp, m eandering rivers, floodplains, springheads, and numerous urban retention ponds. Unlike many lake s around the world, most of Florida lakes are very shallow, with large surface area to volume ratios. Even the largest lake, Lake Okeechobee only has a mean depth of less than 3m (Havens et al 1996). Most naturall y formed lakes are the result of either solution processe s or relict sea-bottom depressions that filled with freshwater as the oceans receded (Edmiston and Myers 1983). Before human settlement and manipulation of this vast wetland system, much of Florida was essentially a large floodplain, receiving and storing rainfall by flooding the extensive littoral zones of lakes and swelling and spilling over river banks. The Kissimmee Chain of Lakes drained through the Kissimmee River which then emptied into Lake Okeechobee. During high water events, the giant lakes southern rim would spill over into the vast Everglades, a wetland ecosystem comprised of sawgrass ridges, sloughs and tree islands. This giant freshwater wetland expanded to the east and west coasts, transiti oning to brackish marsh and mangrove communities that greeted the ocean. This vast expanse of the greater Everglades watershed beginning in central Florida (Figure 1-1), would ebb and fl ow depending on weather patterns, shaping the Floridian landscape and its native flora and fauna. These seasonally inundated and fluctuating we tlands provide habitat for many animals. Local flora and fauna are adapted to this dyna mic system, and comprise a diverse food web. Aquatic plants support epiphytic algae which feed s a macroinvertebrate community that in turn nourishes small vertebrates, consumed by larger animals such as bass, egrets, otters and
13 alligators to name a few. The structural co mplexity provided by aquatic vegetation offers protection from predators for many animals in cluding young fish species and herpetofauna (Savino and Stein 1982, Rozas and Odum 1988). Wet seasons tend to tr igger the onset of breeding for many animals, and the degree of wate r fluctuation influences nesting success of many aquatic and semi-aquatic vertebrates from turtles to snail kites ( Rostrhamus sociabilis ). The dry season will send some animals into aestivation or force migration, until suitable conditions return. In addition to providing habitat to many anim als, wetlands provide valuable ecosystem services and functions including storage of exces s runoff, flood and erosion control, groundwater recharge and discharge and improved water qual ity through filtration pro cesses (SFWMD 2003). As water spills into floodplains moving across the landscape it carr ies with it an assortment of sediments that disperse and settle out. This is accomplished through many mechanisms including stabilizing and trapping sediments thereby reducing turbidity (Schriver et al. 1995; Vermaat et al. 2000), uptake of nutrients through pl ant tissues and their ox idation of sediments that reduce phosphorous fluxes in the wa ter column (Wigand et al. 1997). The natural functions of wetlands have been degraded over the years, due mostly to human settlement. Flooded shorelines impacted agricu lture and other human uses, often causing great catastrophe and loss of lives, particularly after large tropical storm even ts. In response, many Florida lakes were dyked, dredged, and lowered, canals were dug and exis ting outflow channels were straightened (Schiffer 1998). The altera tion of this landscape for the purpose of flood control and use of adjacent floodplain began in the 1800s and continues today. The result of less flooding and stabilized water conditions for hu mans has resulted in a slew of unintended negative consequences. Increased eutrophy a nd lake succession, decr eased water quality,
14 decreased plant diversity, dense vegetation growt h, loss of wildlife habitat, reduced recreational use and poor aesthetic qual ity are common symptoms of highly regulated lakes. Some milestone projects are unde rway to reverse years of wetl and habitat degradation such as the Kissimmee River Restorati on Project and the Comprehensive Everglades Restoration Plan. However, many Florida lakes face continued pressure as the human population continues to increase and demands for land and water increase. Lake management has evolved in response to the problems afflicting lakes, but also to en sure continued protecti on against catastrophic flooding and destruction of human property and lives. Because of this later concern, management actions typically cannot fix the sour ce problem, but can be effective at delaying continued degradation. Florida Fisheries and Lake Management Restoring historical flow to Floridas floodplains is rarely a viable management option. Although natural water variation ha s been considerably stifled (Figure 1-2), lake mangers still allow for some variation throughout the year. A standard management response to address the accelerated lake succession and degradation is ro utine draw downs sometimes accompanied with plant and mud removal. These lake management projects are regularly used on many Florida lakes including Lake Tohopekaliga (hereafter referred to as Lake Toho) to offset negative (or degraded) conditions. Draw-downs work by c onsolidating and oxidizing organic sediments, stimulating germination of native plants and increasing macroinvertebra te and forage fish populations (Moyer et al. 1995). Coupled with sedi ment removal, the effects of draw-downs are more powerful and long-lasting. Dense stands of vegetation including offshore floating mats and accumulated organic sediments are undesirable consequences of restricted flow and cultura l eutrophication. These conditions have been documented to alter the abio tic properties of water, resulting in poor habitat
15 quality for aquatic organisms (Crowder a nd Cooper 1982, Savino and Stein 1982, Sculthorpe 1985, Frodge et al. 1990, Moyer et al. 1995, Mi randa et al. 2000, Miranda and Hodges 2000, Allen and Tugend 2002). Such conditions are th ought to be detrimental to the thriving freshwater fishing industry in Florida, which brin gs in an estimated two billion dollars to the state each year. Thus improving sportfish populat ions and maintaining boater access is a major objective of many management practices in addition to offsetting plant and muck accumulation (Moyer et al. 1995, Olson et al. 1998, Tugend 2001, Allen and Tugend 2002, Allen et al. 2003). For years, fisheries science and management have subscribed to the notion that dense vegetation is unsupportive of healthy and abundant sport fish populati ons. They attempt to offset conditions that have resulted fr om years of impoundment and cultural eutrophica tion. While most managers can probably agree that dense ve getation is undesirable, they also recognize the importance of vegetation in supporting fish populations as well as other aquatic fauna. Vegetation provides structural complexity, substrate for macro and microinvertebrates, egg attachment for various species, protection from predation and weather and an increased prey base. Littoral food webs are co mplex, containing more trophic guilds, species and links than pelagic zones (Havens et al. 1996) Although pelagic food webs can also be complicated, they are usually simpler and support lower diversit y. However, when plant densities reach exceedingly high levels in littoral zones, can the animals that rely on vegetated habitats survive the ensuing conditions? Determini ng how much plant biomass is detr imental versus beneficial is no easy task. This study presented a unique opportunity to examine the habitability of dense pickerelweed ( Pontederia cordata ) to reptiles, amphibians, fish and apple snails inhabiting a eutrophic, degraded lake. An unprecedented management project was planned for Lake Toho,
16 involving dry down and removal of an estimated 6.9 million m3 or 1,351 ha of the organic plant and muck accumulated in most of the littoral zone What follows is a description of the lake, a review of past research a nd the objectives of this study. Lake Tohopekaliga Lake Toho is a large shallow lake, about 9800 hectares with an average depth of 2.1 m eters. It is a naturally formed lake located in Kissimmee, Florida (O sceola County), a highly developed and populated area. Li ke many Florida wetlands, it has suffered the consequences of restricted water flow, cultural eutr ophication and intense human use. Between 1942 and 1964, the lake fluctuated w ith a range of about 3.2 meters (USGS, unpublished data). After a lock and spillway structure was completed, the time between 1964 and1970 experienced a range of only 1.44 meters (Wegener and Williams 1974). Further regulation reduced the range even further to 0.91 meters, with a one in three year drop of an extra 0.15 meters (Hoyer et al. 2008). T hus the historical recorded fl uctuation dropped from about 3.2 meters to 0.91 meters (Figure 1-2). Beginning in the 1950s, the first municipal wastewater discharge entered the lake (Williams 2001), deteriorating water quality and aquatic habitat. Phosphorous and nitrogen loading was enormous, causing managers to st ep in to address the problem. By 1987 the wastewater discharge was almost completely eliminated and there has been substantial improvement since. However the damage done during this time period was exacerbated by water level restrictions that led to algal pr oduction and accelerated lake succession (Hoyer et al. 2008). Draw downs and occasional muck removal became standard practice, with the first draw down on Lake Toho occurring in 1971, and anot her in 1979. In 1987 after the wastewater discharge was nearly eliminated, managers drew the lake down again, this time accompanying it
17 with mechanical removal of an estimated 172,000 m3 vegetation and muck. Within two years however, there was almost complete recovery of Pontederia cordata the vegetation targeted for removal, though several grass species increased in frequency as well (Moy er et al. 1989). An unprecedented initiative in 2004 to dr aw down and remove 6.9 million m3 of sediment had three main objectives: offset lake succession, improve la ke access and aesthetics and restore fish and wildlife habitat (Hoyer et al. 2008). The lake was lowered from 16.8 m to 14.9 m NGVD exposing and drying the entire emergent littoral zone, and permitting access and use of heavy equipment to remove materials. Twenty nine in-lake spoil islands were crea ted, in addition to material placement upland (FFWCC 2004). Upon completion, water stages returned to normal levels during late summer/early fall of 2004. Need for New Sampling and Statistical Techniques Degraded habitats are co ntinually targeted for removal, yet there has been inadequate evidence for its unsuitability to any guild of aquati c vertebrate other than sportfish. Further, the majority of fish studies were unsuccessful at sampling in dense vegetation, rendering dubious results. Simply put, traditional sampling techni ques perform poorly in dense vegetation. The same might be said for herpetofaunal sampling methods. While some of these methods might function well under different ci rcumstances, none are particularly useful or practical for sampling in dense aquatic vegetation. Fish sampling methods typically include electrofishing, rotenone/blocknet and Wegener rings (Moyer et al. 1995, Hoyer and Canfie ld 1996, Tugend 2001, Allen and Tugend 2002, Allen et al. 2003, Bonvechio and Bonvechio 2006). Th ese methods only perform well in sparsely vegetated and open water habitats due to limitations of equipment and boat access. Many studies attest to have employed these methods in degrad ed habitats when they are probably sampling at
18 the edges of these inaccessible areas. Moyer et al. (1995) admitted this, and also that far fewer control sites were sampled because of dens e vegetation. Allen and Tugend (2002) found it nearly impossible to sample with blocknets in the dense vegetation and thus used only rotenone, whilst using blocknets in enhanced sites. Bonvechio and Bonvechio (2006) found discre pancies between angler catch rates and electrofishing, suggesting that either one or both methods were poor indicators of largemouth bass abundance. They noted that envi ronmental factors such as hydrilla ( Hydrilla verticillata ) coverage, trophic state and water quality can limit the ability of electrofishing methods to estimate abundance. Furthermore, large adult fish are more readily captured during electrofishing (Rey 1996, Bayley and Austen 2002) which neglects smaller fish representative littoral zone inhabitants. Creel surveys are another method used in fisherie s research, used to attain fishing pressure estimates from angler counts. Various methods are available, and the estimates can be very inconsistent, affecting virtually every aspect of creel survey results (S oupir et al. 2006). For example, aerial estimation methods have been referred to as fair-weather fishing estimates because of their reliability on good flying weathe r (Soupir et al. 2006). Additionally, shoreline anglers can be difficult to detect and the high costs that typicall y limit the number of surveys, reduces the accuracy and precision of the study (Soupir et al. 2006). Importantly, angler success might be the result of factors ot her than actual fish abundance. Ch anges in fish behavior, fishing technique, effort, access, environment, and reliability of angler information can influence these surveys (Bonvechio and Bonvechio 2006). While creel surveys provi de useful information about harvest rates, fishing pressure and angler effort they should not be used to infer habitat or
19 population changes. These types of surveys should be supplemental to research that investigates the effectiveness of management on habitat qua lity and its effects on species of interest. Trapping has been used intermittently to sample littoral fishes (Bendell and McNichol 1987, Conrow et al. 1990, Chick and McIvor 1994, Jackson and Harvey 1997, Whittier and Hughes 1998, MacRae and Jackson 2006, Bunch et al. 2008). Various trapping methods target an array of fish sizes and speci es and have been successful in vegetated habitats. However, Conrow et al. (1990) struggled with vegeta tion entanglement, compromising their method somewhat. Bunch et al. (2008) used active tra pping methods which were successful in various types of dense vegetation. Other than this, there has been little use of traps probably because of the time investment required to use them. St udies of lake management effects in Florida continue to rely primarily on the traditional fishing methods referred to above. If a densely vegetated habita t is not accessible prior to management but then becomes available afterwards, any perceived changes in habitat quality or abunda nce of fish could be misleading. What is likely provided is a de scription of the surveyors ability to find the species on the landscape, not where the species is on the landscape (MacK enzie 2005). Although improved boater access is an important and comm only stated goal for management projects, whether or not it is improving habitat for target species is questionabl e based on these methods. Studies have shown that aquatic macrophytes are commonly used by larval and juvenile fishes as well as small forage fishes (Shi reman et al. 1981, Killgore et al. 1989, Chick and McIvor 1994, Miranda et al 2000, Miranda and Hodges 2000, A llen and Tugend 2002, Havens et al. 2005, Pelicice et al. 2005, MacRae and Ja ckson 2006, Johnson et al. 2007, Bunch et al. 2008). Following the removal of this habitat, so me studies document higher juvenile catch rates and abundances (Moyer et al. 1995, Allen and Tugend 2002, Allen et al. 2003). This might be a
20 result of the previously inaccessible fishes now being more exposed and readily captured in open water where the traditional fisheries technique s excel. Allen et al. (2003) explain their overestimation of age-1 largemouth bass abundance as a likely result of improved catchability after the management. Allen et al. (2003) also noted no difference between pre and post management catch of harvestable-sized fish ( 356mm TL) sampled via electrofishing and creel surveys. Since adult bass do not occupy littoral vegetation to the extent that juvenile s do, their undetected change in abundance is not surprising. Hoyer and Canfield (1996) confirmed that in some small Florida lakes, adult largemouth bass have been known to exist without any macrophyte coverage. They also recognized positive correlations between juvenile bass abundance and aquatic macrophyte abundance, but no such relationship with adult bass. They acknowle dge that large lakes such as Toho, Kissimmee and Okeechobee have significantly less shoreline to surf ace area ratios and might require macrophytes as refuge for young bass, and call for more research on larger lake systems (>300ha). All of the studies focusing on fish response to management have failed to incorporate detection probability into their analyses. This weakness is also common in many herp studies (Mazerolle et al. 2007). Such approaches assu me perfect detection ( 100% chance of detecting the animal in each sample), which is seldom real istic. Time of day, season, temperature, water level, vegetation cover, observer technique, trap type and boat noise are the sorts of things that can influence detection probabilities and are aff ecting how scientists and managers interpret faunal responses to management. For example, a simple observed change in count might be the result of random variations or changes in detection (MacKenzie et al. 2002), rather than an actual change in abundance, density or occupancy. A dditionally, non-detection of a species does not
21 necessarily imply absence, but al so the possibility that it simply went unnoticed. Statisticians have devised sophisticated approaches to th is problem (Mazerolle et al. 2007), which are explored in this study. Although counts uncorrect ed for detection (ad hoc) perform poorly, they continue to be used extensively (MacKenzie 200 5, Mazerolle et al. 2007). In fact, a supposedly famous fisheries saying is, st udying fish is like studying trees, except that they move around and are invisible (She pherd in Hillborn 2002). Study Objectives Because of the paucity o f research on faunal re sponses other than targeted sport fishes, and lack of effective management analysis (OPPAGA 2001, 2003), the FWC proposed a study to document how herpetofauna and fishes would respond to the Lake Tohopekaliga Management Project. This unprecedented largescale project provided a unique opportunity to study lake-wide impacts of the management project with both before and after lake-wide comparisons as well as smaller scale resolution c ontrol and treatment habitats. Armed with an appropriate sampling and statisti cal analysis approach, we were able to study changes in occupancy of littoral species after the management project. This study was intended to answer the following research questions and hypotheses. 1) What faunal species occupied the dense P. cordata littoral habitat of Lake Tohopekaliga prior to management? 2) Does the estimated proportion of area occupied for each species change after management? Hypothesis 1 : species that have high occupancy estimates pre-management will decrease after management operations. Hypothesis 2 : Mud-burrowing species such as amphiumas and sirens will decrease in occupancy post-management. Hypothesis 3 : Largemouth bass, bluegill and Seminole killifish will increase occupancy post-management. 3) Does the estimated proportion of area occupied for each species differ between control and treatment habitats? Hypothesis 4 : species with high occupancy pre-management will have the highest occupancy in control habitats.
22 Hypothesis 5 : largemouth bass, bluegill and Seminole killifish will have the highest occupancy in treated habitats. 4) What environmental factors influence dete ction probabilities and other parameters, particularly occupancy for each species? Hypothesis 6 : lake stage and temperat ure are strong determinants of detection probability; habitat is a strong determinant of occupancy. Overview of Terms Used Habitat In this study we identify three habitat type s after the m anagement project: control, treatment and all-lake. The control and treatm ents that comprised the small scale resolution focus studies, allowed for post-management treatm ent and control comparisons. Each study area consisted of 1600m of shoreline, with two cont rol and two treatment blocks each 400m. One study area was located in Goblets Cove (Area 3) another along South St eer Beach (Area 2) and the third south of Browns Point (A rea 1) refer to Figure 2-2. The third habitat type identified is all-lake, which can simply be thought of as an alternative treatment type. These areas were located where extensive shoreline scraping had occurred, primarily along the southeastern shoreline and entire western shoreline up to Browns Point. This treatment type might differ from the study area treatments because of lack of proximity to vegetation, and large tracts of exposed open water habitat which might influence its habitability. Each of the three habitats control, treatment and all-lake were intentionall y surveyed in order to detect any differences among them. Location Five distinct locations around the perim eter of the lake were identified (Figure 2-2). These are study areas 1, 2, 3, all-lake east, and al l-lake west. Lake T oho is large enough that differences might exist based on location due to various factors such as fetch, natural sedimentation and human disturbances that might influence our statistical parameter estimations.
23 Statistical Paramet ers and Covariates Param eters for this study are occupanc y, detection probability and colonization. Environmental covariates that may influence thes e parameters are habitat, location, lake stage, temperature, and season. The software program PRESENCE (Hines J.E. 2006) was used to calculate these estimates. Patch occupancy analysis can be thought of as proportion of area occupied (PAO). This estimate of PAO obtained from repeated sampling, can be used to infer the extent to which an animal occupies the system in general. For exam ple, if a particular species obtains an occupancy estimate of 0.40, this means that the animal occupi es approximately 40% of the areathe PAO is 40%. A colonization value of 0.30 means that an unoccupied site has a 30% chance of becoming colonized. A detection probability of 0.50 means that if an animal is present in the habitat, there is a 50% chance of detecting it. In occupancy analyses, the term season does not imply an actual season (breeding season, winter, summer, dry, wet etc.). Instead, it is defined as some period of time in which the observer considers an area closed to localized exti nctions and colonizations of a species. In this study two consecutive sample occasions that sh are similar lake stage and temperature were considered closed seasons. The reason for select ing these discrete time pe riods is explained later in the methods section unde r Assumption Violations.
24 Figure 1-1: Map of Lake Tohope kaliga within greater Evergl ades watershed (source: Brush 2006)
25 Figure 1-2: Daily mean lake stage in feet (NGVD) from January 1942 until June 2008. The dotted red lines indicate managed drawdow ns. Impoundment occurred in 1964 and lake stage was regulated between 52-55 ft NGVD (Source: Brush et al. 2008)
26 CHAPTER 2 METHODS AND STATISTICAL ANALYSES Trapping Traditional m ethods of sampling fish and herp etofauna would have been ineffective and restrictive in the observed species Densely vegetated li ttoral zones require a different technique capable of sampling a variety of littoral fauna. Some researchers have explored funnel traps (Darby et al. 2001, Casazza 2000, Sorensen 2003), leading to a pilot study by Muench (2004) who tested a few trap designs. Muench discovere d that 1.3cm mesh size m odified crayfish traps and floating minnow traps worked we ll at capturing herpetofauna, and a variety of fish species in dense P. cordata. Traps retained mostly small adult species and j uveniles of larger fish. For example, small forage fish species included blue-spotted sunfish ( Enneacanthus gloriosus, ) Seminole killifish ( Fundulus seminolis) and sailfin mollies (Poecilia latipinna) Juvenile fish included many sportfishes such as bluegill (Lepomis macrochirus), largemouth bass ( Micropterus salmoides), redear sunfish( Lepomis microlophus ), spotted sunfish ( Lepomis punctatus). and warmouth ( Lepomis gulosus ). Table B-1 lists all species captured. Herpetofaunal species were mostly adults as young were able to escape through the mesh. Stinkpots ( Sternotherus odoratus) and striped mud turtles (Kinosternon baurii ) are very small adults that were successfully captured. Large adults such as softshell turtles (Apalone ferox ), snapping turtles (Chelydra serpentina osceola) and peninsula cooters (Pseudemys floridana peninsularis ) were excluded from the small trap size and only a few young were ever encountered. Despite the ability to retain very small or very large individuals, overall the traps captured a variety of fauna representative of th e littoral zone. For more details on the trap selection process see Muench (2004).
27 Trap Descriptions Both trap types were m ade from green vinyl-c oated hardware cloth. Crayfish traps were about 80cm tall including an extension chimney which allowed additional space above water for breathing of trapped animals (Figure 2-1). Three entry funnels at the base were each approximately 6cm in diameter. The tops of the traps were fitted with lids made from similar hardware cloth and small bungee cords. Rect angular minnow traps measured 60 x 25 x 18cm with an entry funnel at opposite ends approximately 9cm wide and 6cm tall. Minnow traps were fitted with floating devices to keep the funnels ab out even with the water surface (Figure 2-1). Prior to management styrofoam pool toys (w acky noodles) were used, but these did not hold up in the years following management and were switched to industria l marine floats from Sterling Net and Twine. Two float sizes per trap were used (#SB10 x 31/2in and SB11-4 x 6in). They were affixed through their central op ening with zip-ties to the wire mesh. These floats resisted sun, desiccation, water-logging and atta ck by vultures (which regularly destroyed the pool toys when they were more visible af ter the management). Traps were not baited, although trapped animals may have encouraged others to enter. Sampling Methods The following m ethodology was developed and implemented by Muench (2004) and later adapted and modified for the entire study. Sa mpling was conducted around the periphery of the southern two-thirds of the lake, in order to und erstand occupancy at the lake-wide scale (Figure 2-2). The northern third was not scraped, thus it was not sample d. Eighteen sites were randomly selected and at each, a transect was set-up comprising 3 trap point s about 10m apart, perpendicular to the shoreline. One crayfish an d one minnow trap were attached to a pole. The crayfish trap rested on the bo ttom, with the floating minnow tr ap beside it (Figure 2-1), both partially above water to permit surfacing of tr apped animals. Where dense tussocks existed,
28 traps were punched through and rested among the organic sediment and vegetation. In 2002, the traps were placed in the most lakeward portion of the pickerelweed zone when possible. This was approximately 0.6-0.9m deep, with the la ke stage holding at 13.8m NGVD. Trapping proceeded throughout the entire year, pending suitable water levels (greater than 16m NGVD). Traps remained in place and were checked once per week throughout the year. Efforts were made to space sampling seven days apart but this was not always logistically possible. Traps were approached by airboat and observers would walk to th e three trap points, retrieving the animals. All animals were identified and weighed with only herpetofauna receiving length measurements. Herpetofauna were weighed indi vidually, while fish species were grouped and weighed together. All animals were then released on site. Trapping began on January 24, 2002. From May 2-June 12 traps were removed to due low water levels that stranded traps on dry ground. Trapping resumed entirely by July 24 (traps were replaced slowly as water and time permitted). By December 3, traps were again removed to due low water levels associated with the attempted 2002 drawdown. Post-Management Field Methodology Adjustments After the m anagement project was comple ted, trapping was continued in a similar manner. Due to differences in habitat structure, some modifications had to be made in order for trapping to function properly. Fi rst, exact 2002 locations were not possible to re-sample due to deeper water from scraping and associated loss of organic sediment/vegetated bottom. Transects were moved shoreward from original GPS loca tions to where they we re comfortably above water. The chimneys of the crayfish traps were extended even further to approximately 125cm to allow for more fluctuation in lake stages without swamping the trap. Wind and waves became a serious issue, causing loss and damage of traps. The flimsy 1 inch PVC poles originally used were replaced wi th sturdier 2 inch poles Furthermore, these
29 poles had to be pounded into the lake substrat e, as there was no longer vegetation to support either the traps or poles. Coated electrical wire was used to affix traps to the poles, as the original zip tie method only resulte d in snapped ties and lost trap s. Every effort was made to make sure transects were properly functioning, but environmental conditions often resulted in missed samples due to trap disrepair. These logis tical issues as well as a change in technicians, prevented trapping from fully resuming until May 3, 2005. It took a few months of revisions to find an adequate method and familiarize staff. Unfortunately, Hurricane Wilma struck the Kissimmee area in late October, wh ich destroyed and displaced most traps. Thus from October 25, 2005 through to the end of November, trappi ng was sparse and intermittent as time and resources slowly allowed re placement of all traps. Post-management trapping also included 12 additional transects coinciding with the newly created study areas. One transect per cont rol and treatment block was randomly selected and transects were established in the same ma nner as 2002. There were two controls and two treatments per study area, thus 4 blocks in 3 study areas provided 12 additi onal transects. This brought the total number of transects to 30 (90 tr ap points and 180 traps). After 7 months of checking these 30 transects, trappi ng effort was reduced by elimina ting the original 18 all-lake transects, in order to focus on control and tr eatment habitats. Beginning December 1, 2005 only the 12 study areas were sampled. This method continued until October 17, 2006. Up to this point, all transects were in open water, even control plots presumably due to 2004-2005 hurricane damage. Vegetation was rec overing shoreward, but the traps were not sampling within this community. Thus an adju stment to the protocol was implemented, with longer transects spaced out accordin g to water depth. The first trap point began close to shore at 15cm deep, then 25cm, 35cm, 45cm, 55cm (at 16.20 m NGVD). This allowed sampling in the
30 recovering littoral zone and out in to the deeper reaches as well. With this protocol we were better reflecting the recovering littoral zone fa unal communities. The 12 study area transects were used, in addition to 6 alllake transects randomly selected from the original 18 (excluding locations that fell within proxim ity of a study area, in order to adequately sample expansive scraped shoreline habitat; Figure 2-2). This pr ovided even sampling of control, treatment and all-lake areas, with 6 transects per habitat type Trapping resumed in the same manner as in previous years with no other revisions until October 26, 2007. Hurricane Complications The 2005 hurricane season was very active in the state of Florida. Shortly after m anagement had been completed and water was steadily returning to the lake, three major hurricanes swept through. Hurricanes Charley, Fran ces and Jeanne all stru ck within 40 miles of Lake Toho. Not surprisingly, the copious rainfall resulted in a rapid and dramatic rise in lake level. Water levels were at their highest si nce the 1960s, approximately 2 feet above maximum regulated levels (Brush et al. 2008). Basically, the water rose over 8 feet in a 4 month period. The timing couldnt have been worse, followi ng so closely behind management activities. Based on hurricane trajectories, Lake Toho s geography and direct observation of the aftermath, it was surmised that South Steer Beac h and Goblets Cove were somewhat sheltered from battering wind and waves, while the southern and south western parts of the lake were highly impacted (Brush et al. 2008). This is reflected by later vegetation samples showing slower recovery than might be expected, particular ly in the south part of the lake (Brush et al. 2008). Unfortunately the hurricanes rendered control and treatment sites to look very similar, especially in the southern areas. It is difficult to tease out the effects of th e hurricanes versus the treatment and to compare control and treatment especially in the im mediate time period after both management and
31 hurricanes. Possible confounding hurricane effects are noted throughout this thesis where appropriate. Data Analyses The statistical m odeling program PRESENCE (Hines, J.E. 2006) was used to obtain patch occupancy estimates. If an animal was de tected, a 1 was recorded, a 0 if it was not, and a if the sample was missed. Missed occasions do not affect the data, they simply contribute nothing. Over time, a detection history is obtaine d comprising of a series of 1s, 0s and dashes for each transect. For example, a particular tran sect might have a warmouth detection history of 10 00 01 which indicates two consecutive sample o ccasions that comprise a season for a total of three seasons. Each species at each transect was recorded in this way, and the detection histories were entered into Microsoft Excel and c opied directly into PRESENCE software. Covariates are variables thought to potent ially influence dete ction probabilities, occupancy, colonization and/or ex tinction rates. Categorical s ite covariates were location on lake (Figure 2-2) and habitat (post-management only: treatment, control, and all-lake). Continuous sample covariates (environmental vari ables that change for each sample occasion) were average air temperature, lake stage and season. Lake stage for each sample occasion was obtained from the South Florida Water manage ment Districts DBH YDRO browser website (http ://my.sfwmd.gov/dbhydroplsql/ show_dbkey_info.m ain_menu ). The mean daily average lake level at headwater station S61_H was used. This is the water contro l structure in the south part of Lake Toho leading to Lake Cypress via the South Port Canal). Temperature was obtained from http://www7.ncdc.noaa.gov/CDO/cdodateoutmod.cmd Average daily temperatures from the day of the previous sample occasion until th e day before the current sample occasion were averaged.
32 These covariates were standardized so that each value was between 0-1, which performs more efficiently in PRESENCE, compared to entering the actual number. For example 26 degrees Celsius was standardized to 0.26 instead. The categorical covariates were coded with ones and zeroes so that each had a unique code. An example of a model name might be psi (loc ation), gam (.), p (lake stage). What this means is occupancy (psi) differs by location on the lake, colonization (gam) is constant across all sites at all times, and detection probability (p) changes with lake stage. All model outputs are reported in Appendix A and B. The second model parameterization type in PRESENCE labeled seasonal occupancy and colonization, detection was selected as the standa rd model to run for all species. This way any seasonal occupancy differen ces could be estimated. Assumption Violations Occupancy analyses entail a set of assum pti ons that should be addressed by study design. One major assumption is that occupancy status at each site does not change during the season, in other words the sites are closed to changes in occupancy (MacK enzie et al. 2006). This does not refer to individuals, but to a species. Unfortunately the original design does not accommodate this assumption, as animals are tra pped year round, week after week. Surely in this time period with fluctuating water levels, temperature and vegetation changes, species are moving into and out of the sample areas. In order to satisfy this assumption, two subsequent sample occasions that remained relatively equal in lake stage and temperature were grouped as a single closed season. Thus, only one week elapsed per season. The next two occasions were omitted, the next two used etc. In some instances where a dramatic change occurred (ie: lake stage or temperature dropped or rose sharply), that sample would be discarded if it could not be groupe d with the subsequent
33 sample occasion of similar lake stage and temperat ure. This was the only sensible way to ensure that closure assumptions for the statistical model were satisfied. This technique was applied to the 2002 data which had been analyzed differently by Muench (2004). Four years la ter, the program PRESENCE is more powerful, and with the addition of the post-management data we were able to run several multi-season models. Muench (2004) was unable to successfully enter covariates into the progr am, but this feature was now possible. This method is preferable to using NMS correlation data that does not account for detection probabilities and is not generally advisable for animal studies. In light of these changes, any future sampling should be adjust ed to avoid data disc arding, see Appendix D. Model Selection Criteria Multi-seaso n models for each year were r un for every species. First, detection probabilities were modeled with a mixture of covariates until a reasonable model was obtained. This best model was added to by incorpora ting covariates into th e occupancy (psi) and colonization (gamma) parameters. In many cases, these models were unable to explain the data adequately, and thus a constant psi model was selected. Seas onal occupancy estimations would not run, returning spurious results resulting in cons training psi as a constant throughout the year. Thus, one occupancy estimate was obtained for each sample year. Model selection was accomplished using the AIC (Akaike Information Criterion) method. It is a method based on likelihood, which applies penalties for incorporating too may parameters into the model (MacKenzie et al. 2006). The simpler, more parsimonious the model, the better. The actual AIC value reported is not of importance, but rather the differences in AIC between competing models that a llows one to select the best model. Burnham and Anderson (2002) provide a rule of thumb that models with AIC differences ( AIC) of 2 or less have a
34 substantial level of support. As it turns out, there ar e often several contending models, with no best model. Before deciding on the model(s) to draw infere nces from, output results were examined. Several issues may appear in the results output, that were not evid ent in the model list. If an animal was captured very infrequently resulting in low detection probabilities (<0.3), this tended to drive occupancy estimates towards 1.0. Thus if a psi value close to or equal to 1.0 was reported, the p value and detection history was scanned to see if psi was a reasonable estimate. Many species with infrequent captu res returned spurious results that were ranked among the top models, so these models had to be discarded. If the species was detected so infrequently that results appeared suspicious, the species was determined to be unanalyzed. Some model outputs contained the warning N umerical convergence was not reached. Parameter estimates converged to approximately x significant digits. However, greater than 3 significant digits reported was okay, and the warning could be safely ignored. Anything less and the model was suspicious. The standard errors and confidence intervals for each parameter estimate were examined. If they were very large relative to the estimate itself, particularly for more than one parameter (ie: psi and p were both outrageous) then the m odel was deleted. This technique was slightly problematic, because the relatively small number of transects as well as spar se detection histories of some species would necessarily result in higher errors, alth ough the general pattern was still evident. Thus discretion on a case by case basis was used to decide if a model was appropriate or not. To summarize, detection histor ies for individual species were inputted along with site and sample covariates into the progr am PRESENCE. Models were firs t run to find which covariates
35 best explained detection probabilit y, p. Once settled, that model was tested with covariates for psi and gamma to explore which models fitted th e data best. Using th e AIC method, the highest ranked models (anything within 2 AIC) were carefully examined for any dubious results including inflated psi values, very low p values, no numerica l convergence and high standard errors. Once satisfied with the model selecti on outputs, the best model(s) were used for inference about the system. Figure 2-1: Trap set up showing both crayfish and floating minnow trap fitted with new float design
36 Figure 2-2: Lake Toho tr ansect locations. Inset A shows all transect locations for each year. Note than in 2005-2006, all the yellow 2002 dots were also sampled for the first part of the sampling, but were scaled back to onl y the study area (blue dots). See text for details. Insert B shows transect set-up post-management in open water. C shows control and treatment plots just prior to re -flooding of newly scraped study areas. The bottom right corner shows a spoil island.
37 CHAPTER 3 RESPONSE OF HERPETOFAUNAL COMMUNITIES Introduction Wetland ecosystem s can support a diversity of he rpetofaunal species (hereafter referred to as herps) including snakes, frogs, salamanders and turtles. The shallow vegetated littoral zone offers critical habitat used for refuge, feed ing and reproduction. Some are considered semiaquatic, intermittently using terrestrial and littoral habitats. For example, some turtles spend much of their life in the water, but emerge dur ing breeding season to nest in upland habitat. Others are entirely aquatic, carrying out al l of their life functions in the water. Most Florida wetlands are dynamic ecosystems, with fluctuating wate r levels throughout the year that consequently alters the physical and chemical environment. Native fauna have adapted to these changing conditions, even developing means by which to survive extreme conditions. Amphibians tend to be less mobile occurring in metapopulations and exhibiting high site fidelity (Harrison and Taylor 1997, Hanski 1999, Sinsch 1990, Berry 2001). Thus, a typical response to drought conditions is to burrow into mud, where some species (ie: sirens and amphiumas) can survive without food and water for prolonged periods of time. More mobile species can either follow the receding water or mi grate in search for suitable habitat. The extremes to which these characteristics limit ad aptation and survival in a changing environment will vary greatly depending on the species and surrounding landscape (Smith and Green 2005). The littoral zones of many Florid a lakes have dramatically ch anged since the onset of flood control measures and other human disturbances. On Lake Toho, years of restricted water levels coupled with both natural and anthropogenic eu trophication has resulted in dense monospecific stands of P. cordata in the littoral zone. As this conti nually grows and dies back, a layer of organic sediments (muck) accumulates. This layer is normally considered detrimental to wildlife
38 and leaves many lake-users disappointed with th e aesthetics and limited boat access. This muck and dense vegetation is continua lly targeted for removal by lake managers, to offset succession of lake degradation as well as to promote recreational use. Aresco and Gunzburger (2004) comment on th e misconception that this organic sedimentation is unnatural and harmful to wildlif e, water quality and recr eation. In fact, some herps are actually associated w ith this muck such as sirens amphiumas and snapping turtles (Carr 1940, Bancroft et al. 1983, Aresco and Gunzbur ger 2004). Macrophytes and the flocculent layer of organic detritus provide s a rich source of invertebrate prey (Butler et. al 1992, Schramm et al. 1987, Schramm and Jirka 1989) and al so permits burrowing to escape drought. Amidst the plethora of studi es documenting the benefits of management projects to sportfishes and overall habitat improvements, th ere is scant information regarding the impacts on herp species. One study, by Aresco and Gunzbur ger (2004) noted dramatically high mortality rates of several species during sediment remova l operations. Johnson (2005) identify major gaps in the understanding of herps in Lake Okeechobee, the most well-studied lake in Florida. Their review discovered little informa tion other than species lists. In fact, the only relevant literature available was regarding alligators ( Alligator mississippiensis ). If draw downs and scraping continue to be used, then it is imperative to und erstand how this affects native herp residents. The large scale management that took pl ace at Lake Toho in 2003-2004 provided a unique opportunity to address these knowledge gaps. Species Accounts Sirens ( S iren spp.) and amphiumas ( Amphiuma means ) are large aquatic salamanders commonly associated with muddy and weed-cho ked ditches, heavily vegetated ponds and lakes, or sluggish streams (Funderburg and Lee 1967, Bancroft et al. 1983 Franz 1995). Both are capable of surviving droughts by burrowing into the mud where they can survive for about 2-5
39 years without food or water (Martof 1969, Ethe ridge 1990). Bancroft et al. (1983) found increased amphiuma and siren densities with incr easing sediment depth, and absence from bare sandy shorelines. Both are mainly nocturnal, be nthic creatures feeding on a diversity of prey items such as crayfish, snails, adult and larv al insects, tadpoles, salamander larvae and fish (Hanlin 1978, Martof et al. 1980). They are cons idered top predators in wetlands and because they are so similar in diet and life histories, they may be restricted by competition or predatorprey interactions (Snodgrass et al. 1999). Hypothesis 2 stated that these species, because of their mud-burrowing habits will decline in o ccupancy after management operations. Pig frogs ( Rana grylio ) are highly aquatic animals, known to occupy marshes containing emergent vegetation, which they use for cover, foraging and substrate for egg-laying (Carr 1940, Wright and Wright 1949, Dundee and Rossman 1989, Conant and Collins 1991). They survive droughts by burrowing into mud and peat (Lig as 1960, Wood et al. 1998), however drought conditions can be detrimental. Ugarte (2004) documented an Everglades drought event in which all juveniles died or immigrated. Many aquatic turtles and water snakes are al so known to inhabit shallow wetlands with emergent vegetation. They tend to be better dispersers than amphibians, although movements are infrequent, occurring only out of necessity such as changing environmental conditions or mate searching (Roe et al. 2003). The stinkpot ( Sternotherus odoratus ) is highly aquatic and rarely ventures onto land except to lay eggs. They are almost always found in the water, and their drought response is to follow the receding wate r. Should all the wate r disappear, they will burrow into the mud. Extreme conditions result in temporary cessation of reproduction (Gibbons et al. 1983). The st riped mud turtle ( Kinosternon baurii) is more terrestrial, foraging through murky shallow water looking for seeds, mollusks, fish and algae (VGDIF 2008). They do spend
40 more time traversing land, and so drying events would probably result in migration from the area to somewhere more suitable for the time being. Both the Florida banded water snake ( Nerodia fasciata pictiventris ) and Florida green water snake ( Nerodia floridana ) prefer shallow and still wetlands (Ernst and Ernst 2003). Florida banded water snakes less than 50cm are thought to prey on fish, but then switch mainly to frogs once past this size class (Mushinsky et al. 1982). The Florida green water snake feeds primarily on fish but also opportunistically on frogs, salamanders, tadpoles, small turtles and invertebrates (Ernst and Ernst 2003). As with other herpetofauna, the extent to which these animals will travel and utilize terrestrial, littoral and pelagic habitats depends on specific life history traits. Their movements are probably influenced by the distri bution of wetlands and preferred prey in the land scape (Roe et al. 2004). Results Before m anagement in 2002, all species were es timated to occupy a high percentage of the littoral zone, ranging from approximately 60 -90%. Following management all declined significantly. In fact, they were captured in such low numbers that they were not even statistically useful in the firs t year post-management, with the exception of the stinkpot. Amphiumas were not encountered at all in this time, and so rarely in the second year that analysis was not possible. Throughout the postmanagement sampling, both water snake species had so few captures that they were inestimable. By the second sampling year, all other species showed slightly increa sed occupancy, with sirens and st riped mud turtles showing a marked increase. The following occupancy estimates for each spec ies are reported as proportion of area occupied with the standard error following in br ackets. The first percentage listed is from 2002,
41 the next from 2006-2007 for all species, since 2005-2006 estimates were not possible. The only exception was the stinkpot. Figures 3-1 through 3-7 display these results. Amphiuma occupancy was 0.84(0.14), afte rwards inestimable. Figure 3-1. Siren occupancy was 0.74(0.07), then 0.48(0.13). Figure 3-2. Pig frog occupancy was 0.88(0.10) then 0.29(0.21). Figure 3-3. Banded water snake occupancy was 0.84(0.12), afterwards inestimable. Figure 3-4. Green water snake occupancy was 0.59(0.22), afterwards inestimable. Figure 3-5. Stinkpot occupancy was 0.69(0.25) then 0.21(0.08), then 0.35(0.09). Figure 3.6. Striped mud turtle occupancy was 0.91(0.08) then 0.74(0.15). Figure 3-7. Although some animals did not appear in the estimates due to low detection, they are worth briefly noting. Prior to management mud snakes, peninsula cooter s and striped crayfish snakes were captured, but were never encount ered afterwards. Other species such as cottonmouth, snapping turtle, Florida softshell turtle and leopard frog were also captured prior to management, but were only encountered sparingl y afterwards. Leopard frogs were captured 15 separate times in 2002, not once in 2005-2006, a nd only 6 times in 2006-2007.Refer to Table A1in Appendix A for species list. Discussion The response of the herpetofa una was clear-all species occupa ncy declined dram atically after management. Except for the stinkpot, all speci es were caught in so fe w numbers in the first sample year that estimates were not even possible, with amphiumas going completely undetected the entire first year. The second year post mana gement showed some improvement for pig frogs, stinkpots, striped mud turtles and sirens, particularly the later two. Howeve r, both water snake species as well as amphiumas were still inestimab le due to low detection. Many of the scraped areas were essentially barren for the first three years (Brush et al. 2008), which probably relates
42 to the lag in recovery for most species. All species had relatively high occupancy premanagement thus our hypothesi s stating that animals with high pre-management occupancy would decline afterwards is strengthened. We c ould not test for the hypothesis that these animals would also occupy control habitats, as data were too sparse and hurricanes effectively reduced control habitats in the first year.. Amphibians The increase in siren occupancy im plies th at organic sedimenta tion and vegetation had attained suitable levels, as this species is highly dependent on both. Oddly, amphiumas did not show increased occupancy by the second year desp ite their similar habits to the siren. Both salamander species burrowed into the muck dur ing drawdown and had the scraping not occurred, would probably have survived, as they are capable of doing so for long periods of time. The scraping process unearths aestivating animals a nd crushes them with heavy machinery (Aresco and Gunzberger 2004, Muench 2004). Potentially large numbers were uncovered and killed, leaving behind a depleted population left to recover in extremely different habitat. Surmising reasons for the discrepancy between siren and amphiuma recovery is purely speculative since little is known about either species. Thei r similar morphologies, life histories and diets might lead to competitive exclusion. S nodgrass et al. (1999) suggest that these two can coexist under favorable conditions such as long hydroperiods absent from complete drying, less wetland isolation, low disturbance frequency and high immigration rates. Prior to management they did coexist, occupying a la rge proportion of the lake. The ample habitat on such a large lake probably allowed for mutual coexisten ce. There are many possibilities as to why amphiumas might have failed to respond similar to sirens. Perhaps they are less tolerant of adverse conditions, have lower foraging effici ency, lower reproductive capacity, increased susceptibility to predation or vulnerability to disease. Johnson (2005) cites amphibian disease as
43 a major stressor for animal populations. Any of these factors might have been exacerbated by stressful conditions, contributing to their failure to recover. Pig frogs declined significantly after manage ment, plummeting from very high occupancy with only a moderate recovery in 2006-2007. It is likely that lack of vegetative substrate protection, reproduction and resultin g diminished prey base were to blame for the decreased occupancy of these frogs. On a positive note, evidence from one study suggests that these frogs (and perhaps leopard frogs which were captured infrequently throughout the entirety of the study) might find permanent retention ponds in re sidential developments to be satisfactory habitat (Delis et al. 1996). It is possible that during the management project, they escaped to and found refuge in nearby urban ponds. If they can avoid the other ca lamities afflicting amphibians, they might eventually rebound at Lake Tohopekalig a and even flourish in adjacent wetlands. Amphibians are particularly susceptible in todays world of climate change and rapid development. Globally, they are experiencing declining populations and extinctions (Houlahan et al. 2000, Stuart et al 2004). There are several possible re asons including low vagility, narrow habitat tolerance, habitat loss and degradation, edge effects, high vulnerability to pathogens, invasive species, climate change, UV-B expos ure and pollution (Stuar t et al. 2004, Cushman 2006). Furthermore, many life history characteri stics such as delayed sexual maturity, low reproductive rates and high juvenile mortality render many herpet ofaunal species even more vulnerable to environmental changes. These trai ts can exacerbate an already stressed population that has lost many adults, making recovery ve ry difficult (Congdon et al 1994, Snodgrass et al. 1999, Klemens 2000). Given that am phibians are considered important components of aquatic food webs and ecosystems, are globally considered an at risk animal group, and are sometimes
44 commercially and recreationally important, it is crucial that management decisions consider these animals. Reptiles Reptile s were similarly affected by the management. However, striped mud turtles made a significant recovery, and perhaps will completely recover in the years to come. The reason for such resilience is probably reflected in their more terrestrial nature. Striped mud turtles are semiaquatic, but spend much of their time on land. Unlike some animals that burrowed into the mud to survive the dry-down, and were then subs equently removed during sediment removal operations, the striped mud turtles might have been able to leave during the process. Perhaps they were able to survive in the first year desp ite the bare littoral zone, opportunistically foraging in both water and on land. By the second year, substantial vegetative re-growth might have provided better cover and food resources to encourage re-colonizati on, boosting occupancy levels close to that which they once were Stinkpots were probably still estimable in the first year after management because of their highly aquatic nature. Although they are also kno wn to inhabit shallow muddy water, they tend to occur more in permanent aquatic habitats a nd rarely leave the water except to lay eggs (Gibbons et al. 1983, VDGIF 2008). Although they t oo are mud burrowers, they tend to follow the receding water (Gibbons et al. 1983) before burrowing and thus some might have remained in the wet portions of the othe rwise dry lake, escaping excavation. They were still impacted by the process though, as indicated by their low occupancy estimates in the two years after management. Although only half what they were in 2006-2007, it is still a substantial improvement in a short time. Marchand and Litvaitis (2004) found more turtles in ponds w ith organic substrate and shoreline vegetation. They sugge st that organic subs trate provides bette r foraging opportunities
45 compared to hard substrates. However they also note that dense vegetation might reduce the suitability of the habitat by restri cting turtle movement. This is probably more relevant for larger bodied turtles not studied in this project. Smalle r turtles might navigate dense vegetation easier indeed they had high occupancy estimates prior to management. The Bigger Picture There are a wide range of issues affecti ng herp populations that extend beyond the lake ecosystem itself, and also beyond th e scope of this study. Some were mentioned earlier, such as sensitivity to climate, UV, pollution and dis ease. Additionally, the surrounding terrestrial landscape will have huge impacts on the lake and its resident fauna. Lake Toho has increasingly become isolated from other wetlands due to rapid urbanization includ ing road networks and buildings. In addition to how asphalt jungles can negatively impact water quality, they also restrict animal movement between wetlands, lakes and retention ponds nearby. Many aquatic animals emerge from the water to migrate over land, in response to drought or other disturbances, or for reproductive purposes. The extent to which reptiles and amphibians will migrate depends on the species and surrounding landscape (Semlitsch and Bodie 2003). Some snake and turtle species rarely leave their aquati c habitats, whereas others will routinely travel great distances. Most amphibians will not trav erse land, except for very short distances to lay eggs. Upland habitat that surrounds Lake Toho is risky business for many animals that will encounter automobiles (Aresco 2002, 2005). Female turtles are most vulnerable because they routinely make this journey, which in the long te rm can skew sex ratios (Marchand and Litvaitis 2004). Johnson (2005) suggests that development and roads within several hundred meters of lake edges can have long term e ffects on aquatic turtle populations.
46 Amphiuma Year 2002 2005 20062006 2007 Occupancy (psi) 0.0 0.2 0.4 0.6 0.8 1.0 ** **not captured*inestimable* Amphiuma Year 2002 2005 20062006 2007 Occupancy (psi) 0.0 0.2 0.4 0.6 0.8 1.0 ** **not captured*inestimable* Amphiuma Year 2002 2005 20062006 2007 Occupancy (psi) 0.0 0.2 0.4 0.6 0.8 1.0 ** **not captured*inestimable* Figure 3-1: Occupancy estimates with standard error for amphiuma, one year pre-management and two years post-management
47 Figure 3-2: Occupancy estimates with standard error for siren, one year pre-management and two years post-management
48 Figure 3-3: Occupancy estimates with standard error for pig frog, one year pre-management and two years post-management
49 Figure 3-4: Occupancy estimates with standard error for Florida banded water snake, one year pre-management and two years post-management
50 Figure 3-5: Occupancy estimates with standard error for Florida green water snake, one year pre-management and two years post-management
51 Figure 3-6: Occupancy estimates with standard error for stinkpot, one year pre-management and two years post-management
52 Figure 3-7: Occupancy estimates with standard error for striped mud turtle, one year premanagement and two years post-management
53 CHAPTER 4 RESPONSE OF FISH COMMUNITIES Introduction Years of water restriction s coupled with cultural eutrophication and rapid succession of lake ecosystems have promoted conditions fa vorable to dense monospecific stands of P. cordata and Typha spp. Dense vegetation can alter the abio tic properties of the surrounding water, resulting in sub-optimal environmen t for fauna. This can include lethally low levels of dissolved oxygen, temperatures beyond survival thresholds, re duced foraging efficiency due to visual and swimming barriers, and damagingl y low pH from plant respir ation (Crowder and Cooper 1982, Savino and Stein 1982, Sculthorpe 1985, Frodge et al. 1990, Moyer et al. 1995, Miranda et al. 2000, Miranda and Hodges 2000, Allen and Tugend 2002). These conditions have been linked to stunted fish populations exhi biting reduced growth and overall condition (Colle and Shireman 1980, Crowder and Cooper 1982, Savino and Stein 1982, Bettoli et al. 1992). Offshore berm formation also occurs, which not only imparts the same adverse environment as dense vegetation, but is also thought to restrict fish acce ss to shoreline spawning habitat (Moyer et al. 1995). These negative attributes have drawn much attention from fisheries managers and scientists, because freshwater recreational fishing is a boomi ng business in Florida. A 2003 report indicated that it has an economic impact of nearly $2billi on per year to the state (Harding 2003). Not surprisingly, maintainin g a habitat that supports both juvenile and adult stages of sportfish, and facilitates boater access is an impor tant goal for many lake management projects. Common methods used to achieve such conditions are routine lake draw downs, mechanical sediment removal and herbicide application. Draw downs are in tended to mimic natural drought conditions that expose the littoral zone, kil ling off excessive plan ts, promoting oxidation,
54 consolidation and germination of desirable native species. Mech anical removal of plants and sediment provides open, sandy areas for adult fish and appropriate spawning substrate. It also immediately opens up boater access and prolongs the effects of dr aw downs. Numerous studies document the positive effects of such management activities for important sport fish in Florida, primarily largemouth bass ( Micropterus salmoides) and bluegill ( Lepomis macrochirus ) (Moyer et al. 1995, Olson et al. 1998, FFWCC 2001, Tugend 2001, Allen a nd Tugend 2002, Allen et al. 2003). The unsuitable nature of dense aquatic vegetation has more or less been considered fact for a long time. However, plants ar e also recognized as providing valu able habitat to juvenile fish and small forage species (Shireman et al 1981, Crowder and Cooper 1982, Wiley et al. 1984, Killgore et al. 1989, Chick and McIvor 1994, Dewey et al. 1997, Miranda and Pugh 1997, Miranda et al. 2000, Miranda and Hodges 2000, Pe licice et al. 2005, MacRae and Jackson 2006, Bunch et al. 2008). Largemouth bass anglers, fisheries biologists a nd other professionals involved in lake management often strongly supp ort the statement that aquatic macrophytes are essential for largemouth bass populations (Hoyer and Canfield 1996). The precise threshold beyond which aquatic macrophytes are considered pr oblematic is highly controversial, with reference only to an intermediate macrophyt e density (Crowder and Cooper 1982, Wiley et al. 1984, Olson et al. 1998, Miranda an d Hodges 2000). Trebitz et al (1997) suggest about 20-30 percent removal and not over 50 percent in or der to promote bluegill and largemouth bass populations based on computer simulation models Hoyer and Canfield (1996) refer to a ballpark figure of 20-30 percent of lake area cove red (PAC), which is thought to strike a healthy balance for fish in small water bodies (Hoyer a nd Canfield 1996). But they maintain that any justification for a specific leve l will remain controversial.
55 Most studies have been supportive of lake ma nagement and its benefit for sportfish (Moyer et al. 1995, Olson et al. 1998, FFWCC 2001, Tugend 2001, Allen a nd Tugend 2002, Allen et al. 2003). However, Allen et al. (2003) report that while draw dow ns and muck removals support recreational activities and improve habitat, it is difficult to clearly decipher the impacts on adult largemouth bass. They suggest instead that effo rts be focused on recreational benefits rather than towards a single species. To complicate th e matter further, many traditional fish sampling techniques fail to adequately sample within de nse vegetation and ignore de tection probabilities, a critical factor in animal studi es (see Introductiona need for a new approach to sampling and statistical analyses). Given that recreational fishing is such an important industry, and that there is some uncertainty as to how fishes, even largemout h bass respond to management, some work was needed to address the issue. This study was an important first step in understanding how littoral zone fish communities responded to a large-scale lake management program. Focus needed to extend beyond bass and bluegill in order to underst and how other species th at comprise littoral food webs are similarly affecte d. It was also an opportunity to utilize new methods more appropriate for sampling small fish in dense vegetation, as well as to employ some modern statistical analyses. Results The respons e of fishes was variable. Of the 13 estimable species, 5 (bluegill, dollar sunfish, largemouth bass, Seminole killifish and sailfin catfish) showed increased occupancy post-management. One, the redear sunfish did not show much appreciabl e change. The other 7 species showed notably lower occupancy estimates, in some cases having such low captures that estimates were not possible. The hypothesis st ating that animals with high occupancy would decline after management was generally true, except for bluegill and largemouth bass which
56 were expected to increase, regardless of pre-ma nagement occupancy estimates. Both bass and Seminole killifish had higher occupancy in treatme nt habitats which was also expected, but the bluegill did not, thus the hypothe sis was not entirely supported. The following occupancy estimates for each spec ies are reported as proportion of area occupied with the standard error following in br ackets. The first percentage listed is from 2002, the next from 2005-2006 and the last from 2006-2007. Bluegill occupancy was 0.38(0.18), 0. 85(0.07) then 0.75(0.09). Figure 4-1. Dollar sunfish occupancy was 0.18(0.05), 0.47(0.06) then 0.56(0.07). Figure 4-3. Largemouth bass occupancy was 0. 40(0.27), 0.71(0.17) then 0.64(0.30). Figure 4-5. Seminole killifish occupancy prior to management was in estimable, then 0.70(0.20) and 0.73(0.11). Figure 4-7. Sailfin catfish occupancy prior to management and in the first year after was not estimable, then rose 0.26(0.11). Figure 4-9. Redear sunfish occupancy was 0.59(0.12) 0.49(0.23), then 0.54(0.13). Figure 4-10. Armored catfish occupancy was 0.83(0.09), af terwards inestimable. Figure 4-11. Blue-spotted sunfish occupancy was 0.77( 0.08), 0.28(0.07), then 0.39(0.12). Figure 4-12. Chubsucker occupancy was 0. 43(0.21), af terwards inestimable. Figure 4-13. Gar occupancy was 0.67(0.09), inestima ble, then 0.41(0.13). Figure 4-14. Sailfin molly occupancy was 0.36(0.12), af terwards inestimable. Figure 4-16. Spotted sunfish occupancy was 0.31(0.15), af terwards inestimable. Figure 4-17. Warmouth occupancy was 0.84(0.08), 0.53(0.07), then 0.66(0.08). Figure 4-18. Although many species were not captured enough to allow statistical estimates throughout the study they are worth mentioning. Bowfin, br own bullhead, chain pickerel, flagfish, golden shiner, golden topminnow, redfin pickerel, tadpole madtom and th readfin shad were all captured at some point throughout the study but not cons istently enough to allow for estimates.
57 Interestingly, two exotic species not detected pr ior to management that appeared sporadically afterwards were blue tilapia and Mayan cichlid. The Mayan ci chlid had not been documented here before and a specimen was donated to the Fl orida Museum of Natural Historys Ichthyology Department. Effects of Habitat Six out of 13 fish species indicated that habitat influen ced site occupan cy estimates, including bluegill, gar, warmouth, dollar sunfish, largemouth bass and Seminole killifish. Of the 6, 4 had higher occupancy estimates in control sites. Both largemouth bass and Seminole killifish had higher estimates in all-lake sites. The following occupancy estimates for each species are reported as proportion of area occupied with the standard error following in brackets for each habitat type. These values are also reflected in the figures at the end of the chapter. Bluegill only indicated influence of ha bitat on occupancy in 2006-2007, with 0.72(0.10) in control, 0.57(0.10) in treatment and 0.59(0.11) in all-lake habitats. This was the third possible model with a AIC of 1.03 and model weight of 17%, thus it is a plausi ble model, but it is also likely that occupancy was constant Either way habitat was influential on this species to some degree because there was also a 20% chance that habitat influen ced colonization, this model had AIC of 0.62, with a colonization probability of 0.28(0.13) in control, 0.54(0.21) in treatment and 0.55(0.18) in all-lake ha bitats. See Figure 4-2. Gar only indicated influence of habitat in 2006-2007, with 1.0(0.0) in control, 0.32(0.13) in treatment and 0.37(0.15) in all-lake habitats. There is gr eat support for habitat occupancy models, with 2 habitat models ra nking as the first two models w ith a combined weight of 73%. Although the estimate of 1.0 with 0 standard e rror would initially ca use concern for model estimates, upon examining capture data for control habitats, it seems reasona ble as captures were
58 very frequent. Sometimes inflated estimates with low error are reported in cases where detection histories are sparse, but this does not seem to be the case. See Figure 4-15. Warmouth indicated influence of habitat in both post management years. In 2005-2006, occupancy was estimated at 0.79(0.12) in control, 0.36(0.12) in treatment and 0.46(0.10) in alllake habitats. There was approximately 75% suppor t for habitat occupancy m odels in this year. In 2006-2007 occupancy rose in all habitat types, with 0.98(0.18) in control, 0.52(0.09) in treatment and 0.55(0.10) in all lake habitats. Habitat occupancy differences are even more supported in this year at nearly 99% weight. See Figure 4-19. Dollar sunfish only indicated influence of habitat in 2006-2007, with 0.66(0.12) in control, 0.60(0.13) in treatment and 0.37(0.10) in all-lake habitats. This was the second of two possible models, with a AIC of 0.66 and a model weight of 38%. Thus there is more support for the model indicating constant occupancy acr oss habitats, but the habitat model is still plausible. See Figure 4-4. Largemouth bass only indicated influence of habitat in 2005-2006, with 0.30(0.26) in control, 0.41(0.28) in treatment and 0.72(0.26) in al l-lake habitats. This was the third ranked model with a AIC of 0.41 and model weight of 24%. There is more support (57%) for constant occupancy across habitats, but it is still plausible that habitat influenced occupancy. See Figure 4-6. Seminole killifish indicated influence of hab itat in both post management years. In 20052006, occupancy was estimated at 0.41(0.13) in c ontrol, 0.56(0.17) in treatment and 0.84(0.22) in all-lake habitats. There was the top ra nked model with a weight of 43%. In 2006-2007 occupancy rose in all habitat types, with 0.45(0.15) in cont rol, 1.0(0.0) in treatment and
59 0.89(0.12) in all lake habitats. Habitat occupancy differences ar e even more supported in this year with approximately 77% weight, ranked again as the top model. See Figure 4-8. Figure 4-1: Occupancy estimates with standard error for bluegill, one year pre-management and two years post-management
60 Figure 4-2: Habitat site occupancy estimates with standard error for bluegill 2006-2007
61 Figure 4-3: Occupancy estimates with stan dard error for dollar sunfish, one year premanagement and two years post-management
62 Figure 4-4: Habitat site occ upancy estimates with standard error for dollar sunfish 2006-2007
63 Figure 4-5: Occupancy estimat es with standard error for largemouth bass, one year premanagement and two years post-management
64 Figure 4-6: Habitat site occ upancy estimates with standard error for largemouth bass for 20052006
65 Figure 4-7: Occupancy estimates with standard error for Seminole killifish, one year premanagement and two years post-management
66 Figure 4-8: Habitat site occupa ncy estimates with standard e rror for Seminole killifish for 20052007
67 Figure 4-9: Occupancy estimat es with standard error for sailfin catfish, one year premanagement and two years post-management
68 Figure 4-10: Occupancy estimates with standard error for redear, one year pre-management and two years post-management
69 Figure 4-11: Occupancy estimates with standa rd error for armored catfish, one year premanagement and two years post-management
70 Figure 4-12: Occupancy estimates with standard error for blue-spotted sunfish, one year premanagement and two years post-management
71 Figure 4-13: Occupancy estimates with standard error for chubsucker, one year pre-management and two years post-management
72 Figure 4-14: Occupancy estimates with standard error for gar, one year pre-management and two years post-management
73 Figure 4-15: Habitat site occupancy estim ates with standard error for gar 2006-2007
74 Figure 4-16: Occupancy estimates with standa rd error for sailfin molly, one year premanagement and two years post-management
75 Figure 4-17: Occupancy estimates with standa rd error for spotted s unfish, one year premanagement and two years post-management
76 Figure 4-18: Occupancy estimates with standard error for warmouth, one year pre-management and two years post-management
77 Figure 4-19: Habitat site occupancy estimat es with standard error for warmouth 2005-2007
78 Discussion Previous fisheries studies linki ng the effects of lake m anageme nt to fish populations had three major short comings. First, traditional fish and herpetofaunal sampling methods could not sufficiently characterize occurrence of species in dense vegetation. Second, all studies have either employed raw count data to infer abunda nce and population estimat es, or creel surveys which provide only angler access information, rather than any real clues as to habitat quality or use. Finally, no studies have ever examined the impacts of management on species other than sport fishes. This study has addressed all three of these concerns. The trapping protocol effectively sampled within dense vegetation stands including dense floating mats, and it also performed well in open water habitat in the following manageme nt. Raw counts, abundances and angler surveys were abandoned for modern statistical and met hodological approaches using site occupancy and detection probability estimations. Perhaps most importantly, we were able to document changes in occupancy of several species including s port fishes, before and after management. The response of littoral fish communities vari ed by species. Of the thirteen estimable species, five showed a positive response after management including two important sportfish, the largemouth bass and bluegill as well as dolla r sunfish, Seminole killifish and sailfin catfish. The three other sportfish, redear, warm outh and spotted sunfish declined after management. Redear declined only slightly followed by a slight increase, the change was minimal compared most other species. Three fora ge species also declined including blue-spotted sunfish, chubsucker and sailfin molly. Unfortunately the hurricane effects from 2005 complicated our desire to understand postmanagement treatment effects. Control and treatme nt habitats were similar at least in the first year afterwards, with controls slowly recovering by the second year and treatments remaining
79 relatively bare (Brush et al. 2008). In the abse nce of hurricanes, we would have expected more differences in habitat occupancy than the mere three species observed in the first year postmanagement (bass, warmouth, Seminole killifish) By the second year, as habitats were becoming more distinguishable, f our species showed a difference. It wasnt until the final June 2008 vegetation sample that relativel y stable control plots were ev ident as well as treated areas colonized with eelgrass comm unities (Brush et al. 2008). Hurricane effects were probably not the onl y limiting factor affecting estimation possibilities. Low captures of many animals coupled with relativ ely few site replicates limits complex covariate modeling. In most cases, only constant occupancy models were possible. Despite not being able to tease out the finer de tails of habitat differences, overall occupancy changes are still very tell ing of how a species responded to management, confounded by hurricanes at the lake wide scale. Largemouth Bass The long studied and valuable largemouth ba ss have been reported to benefit from management activities (Moyer et al. 1995, Ol son et al. 1998, FFWCC 2001, Tugend 2001, Allen and Tugend 2002, Allen et al. 2003). This provides incentive for managers to continue with draw downs and muck removals as they seem to improve habitat. In this study, juvenile bass increased occupancy immediately following mana gement. There is evidence that habitat influenced either detection probability, occupanc y, or perhaps both to some degree in the first year following management. The most likely model indicates that habitat influenced detection, and occupancy remained constant. Detection probability was lowest in control, highest in treatment then all-lake sites, reiterating why det ection probabilities are so important to consider. Count statistics would have implied that abundance or density estimates were different across habitats, when in fact they were not.
80 An alternative but less likely possibility is that occupancy was in fact influenced by habitat. Occupancy was similarly affected as dete ction probability. That is, lowest in control, then treatment and all-lake. Interestingly, both overall occupancy estimates for each year as well as each habitat, had fairly large standard erro rs, just under 30%. These less precise estimates potentially make differences between years and habitats less striking th an they first appear. Either way, if abundance is th e underlying cause of higher occ upancy and/or detections in treated areas, then we might infer that more juvenile bass occupied the lake after management, particularly in treated areas in the first year. This would co rroborate earlier studies that found increased bass abundance in enhanced lake s (Moyer et al. 19 95, Allen et al. 2003). Juvenile fish are typically associated with shallow vegetated habitat that they use for forage and cover (Hoyer and Canfield 1996, Miranda and Pugh 1997). An apparent increase in occupancy after scraping seems perplexing, b ecause open water provides neither forage nor cover. The explanation might lie in the ava ilability of sandy substrate for spawning adults, resulting in more opportunity to re produce. Less heavily vegetated habitats have been associated with increased recruitment and growth rates, as this type of habitat is thou ght to strike a balance between protection and prey ava ilability (Miranda and Pugh 1997). The immediate increase in occupancy seen in 2005-2006, was not sustained the following year, nor the effect of habitat on occupancy and detection probability. Occupancy may continue to decline approaching pre-management levels as control plots stabilize and P. cordata threatens to dominate again. However, if the V. americana community currently es tablished in scraped areas remains competitive, it might support both j uvenile and adult bass populations. Further monitoring will help us to unde rstand the long term management effects, including how recovering plant communities impact resident fishes.
81 There was no concurrent study looking at th e effects of management on adult bass. However, a similar management activity conducted on Lake Kissimmee in 1995-1996, was unable to detect any increase in adult bass (Allen et al. 2003). The goal of the project was to improve largemouth bass fishing, but electrofishing a nd angler catch rates c ould not confirm this. In fact, fishing effort actually declined. Bluegill Bluegill also increased in occupancy quite dram atically following management, perhaps for similar reasons as largemouth bass. Although habitat differences were not readily evident until the second year, the species clearly benefited from enhanced habitat. Highest occupancy was in control, then all-lake and treatment habita ts (treatment and all-lake were very similar). Interestingly, there was also support for habitat colonization differences which did not occur for any other species. In this scenario, coloniza tion probability was highest in all-lake then treatment and control. At first, this seems to contradict the habitat o ccupancy model. But the contending models might not be mutually exclus ive. Perhaps preference for control sites reflected by high occupancy, results in lower colonization rates due to saturation of the habitat. As vegetation recovers in scraped areas, mo re vegetated habitat becomes available and individuals rapidly colonize. Other Fish Species Besides blu egill and largemouth bass, four other species including dollar sunfish, gar, Seminole killifish and warmouth showed habitat occupancy differences. All but the Seminole killifish had highest occu pancy in control sites. The other three sport fish encountered in this study, the spotted sunfish, redear and warmouth did not show any positive response to management. While the spotted sunfish and warmouth declined, redear barely changed, rema ining relatively stable prior to and after
82 management. The main food sources of redear are small snails and mussels, supplemented also with aquatic insects (Carlander 1977, Lee 1980), thus the common name shell-cracker. Perhaps these food sources were still plentifu l post-management, such that both adults and juveniles were able to cope with the changed environment. Also, they do not seem particular as to the type of substrate they nest in, having been documente d in both sandy and soft muddy bottoms as well as in aquatic vegetation (Wilbur 1969). Thus they were not limited in nesting availability, nor benefited from more open sandy bottoms, perhaps resultin g in their indifference to the management. The apparent partiality for control habitats in both post-management years exhibited by warmouths, in addition to their overall decline is not surprisi ng given their known preference and/or tolerance for sluggish swampy conditions (Lee 1980, FFWCC 2008). Males are thought to construct nests near clumps of vegetati on (Larimore 1957), which was lacking immediately after management. This may have reduced spawning or maybe just increased juvenile susceptibility, as they are not guarded past 5 or 6 days after spawning (Carlander 1977). Their increase by the second year suggests the return of pre-management vegetation conditions, more suitable to their habitat prefer ences and nesting requirements. Spotted sunfish are thought to inhabit sluggish, heavily vegetated environments with sandy gravel bottoms (FFWCC 2008). They are a ubiquitous species, however their relatively low occupancy on Lake Toho prior to management as well as continually low captures afterwards, will likely prevent them from becoming a common species on this lake. Similar to the warmouth with preference fo r swampy, muddy, heavily vegetated habitats, gar also declined after management and indicate d a preference for contro l habitat in the second year post-management. Gar are particularly adapta ble to anoxic conditions, with their ability to
83 take in oxygen through their air bladder (FFWCC 2008). This spec ies is also showing recovery which indicates return of their preferred habitat conditions. Seminole killifish showed higher occupancy in all-lake and treated sites for both postmanagement years. This is not surprising, as this species has commonly been associated with open sandy habitats, indicative of successful rest oration (Wegener and Williams 1975, Moyer et al. 1995, Tugend 2001). Dollar sunfish would be expected to pref er vegetated habita t (Bauer 1980, HassanWilliams and Bonner 2007), which they do indicate, but their overall occupancy rates increased after the management. They seem to have varied diets and are both benthic and surface feeders (Goldstein and Simon 1999). This versatility might enable them to be opportunistic, able to exploit both open and vegetated areas. Due to th eir small size, they might have preference for vegetated habitats, but like wise capitalize on open water environments as well. Small forage fish that are typically associated with shallow, vegetated still waters, such as the lake chubsucker, sailfin molly and blue -spotted sunfish all responded negatively to management. The chubsucker and sailfin molly ha ve not yet shown any signs of recovery, but it would be anticipated given the in crease in vegetation biomass retu rning to the lake. Some fish and herp species are also signaling that pre-mana gement habitat conditions are returning, which should eventually benefit these small forage species. The armored catfish (brown hoplo) declined in occupancy after management, while the sailfin catfish increased. The armored catfish is a highly adaptable cr eature, surviving hypoxic conditions with their ability to breathe air (Brauner et al. 1995, Affonso and Rantin 2005). They are also aggressive nest guarders (Nico et al. 1996). These traits likely allowed its persistence in the pre-management habitats, and we would exp ect such an aggressive species to continue
84 flourishing. Maybe they were out-competed by ot her species afterwards or were directly affected by the scraping process. We would exp ect this species to recover based on its highly competitive nature, especially if the origin al vegetation communities begin to establish. Sailfin catfish only showed an increase in the second year, and were captured so infrequently prior to management and in the first year after, that it is impossible to conclude whether or not their increase was a direct resu lt of the management. Similar to many exotic species, they exhibit hard y traits allowing them to survive and compete with natives. They are known to construct burrows into mud, which are then used for reproduction but also allow them to survive drought conditions (Hoover et al. 2004). They have been found in dried-out burrows appearing dead, but are in fact al ive and once returned to the wate r, revive quickly (Hoover et al. 2004). This adaptation to extreme conditions has lik ely led to their spread and ability to survive and compete. Two other exotic species, the blue tilapia ( Oreochromis aureus ) and Mayan cichlid ( Cichlasoma urophthalmus ) were not encountered prior to management, but were found in low numbers afterward. Blue tilapias have been es tablished in southern and central Florida for sometime, and are known to inhabit Lake Toho. The Mayan cichlid had not been documented at Lake Toho and specimens were submitted to the Florida Museum of Natural History in Gainesville, FL for cataloging. Whether or not the appearance of the fish are related to management is unclear. It is possible that the disturbance allowed them to compete with natives, but it could also be a coincidence in that they first appeared post-management. Summary Overall it ap pears that only a fraction of fish es benefited from the management, while the rest declined. Two important spor t fishes (bass and bluegill) that showed increased occupancy in the first year, declined somewhat in the second year. Three other sport fish (redear, spotted
85 sunfish and warmouth) did not respond positively. It would be useful and worthwhile to conduct further studies to assess the longe vity of management effects on a ll fishes. The short term study as well as lack of adult sampling leaves some am biguity as to the real and long-term impacts of management on various fish species. Thus, c oncurrent with Allen et al. (2003), it is very difficult to detect the true response of bass to management, as well as other species. Results would have been more comprehensive if we knew whether or not the increase in juveniles actually resulted in succe ssful recruitment into the adult population. Of course the hurricanes also complicate the results. According to vegetation studies (Brush et al 2008), control areas have stabilized with recovery of P. cordata and in scraped areas, V. americana has become quite abundant. This has been accompanied by the recovery of warmouth and gar, two species that naturally tolerant of densely vegetated, soft-bottom environments. As noted in chapter 3, sirens also recovered substantially by the second year, another spec ies indicative of pre-management conditions. Although focus is typically directed towards re creational sportfish for management studies, also understanding how other species are affected by management is important. Small forage species provide vital links between smaller epiphy tic fauna and larger predatory fish, herps and birds. Although single species management for largemouth bass will probably remain the underlying motivation for many future activities, it is important to have a holistic ecosystem approach. Species diversity is often touted as th e key to healthy ecosystems, which in the end is a net benefit to aquatic fauna, managers and lake users. Continued monitoring should be implemented, to track how the succession of lake vegetation continues to shape faunal communities.
86 CHAPTER 5 RESPONSE OF NATIVE AND EXOTIC APP LE SNAIL COMMUNITIES Introduction Although apple snails w ere not intended to be studied, they were captured and recorded throughout our sampling. The exotic island apple snail ( Pomacea insularum ) appeared on the lake shortly before management, providing an opportunity to document the response of both the native Florida apple snail ( Pomacea paludosa ) and the exotic to the management. The presence of the exotic apple snail from South America presents stiff competition for the native species. They have a voracious appetite, consuming al most any aquatic plant in sight (Power 2007). Egg masses are large with the average clutch containing over 2000 eggs, and a field hatching success estimated at 70% (Barnes et al. 2008). In the warm months typical of the southeastern US, a new clutch ev ery week is possible (Barnes et al. 2008). Not surprisingly, it is considered a major pest in many parts of the worl d where it has been introduced. Its threats to Floridas wetland ecosystems are serious. This aggressive and rapid reproducing species will likely compete with the native apple snail, a critical prey item for the snail kite ( Rostrhamus sociabilis ). It might also incur damage to wetland plant communities. Although the native Florida appl e snail is well adapted to the dynamic Florida ecosystem in which it evolved, competition with a much la rger and more actively reproducing species could be a serious problem. Natural drying events are survived for weeks or months at a time by aestivating, since they are not capable of moving far. Snails are stranded on dry ground and will wait for the water to return (D arby et al. 2002). The chances of surviving such conditions are much higher for adults than juveniles.
87 Individual eggs are much larger than that of the exotic species, but clutches contain far fewer eggs, about 10-80 on average (Brown 2005). While drying processes are a natural part of the Florida ecosystem and are beneficial to apple snail habitat and reproduction, prolonged drying or flooding events and the timing in which th ey occur is critical to this species. This species is considered an annual br eeder, with the peak from Apr il-June. Essentially they have one opportunity to successfully reproduce (Darby et al. 2008). If conditions are excessively dry or wet, reproduction will fail that year. Although the native Florida apple snail is well-adapted to its environment, the presence of a potentially aggressive competitor could cause problems. It is possible that the management created a disturbance that might favor the esta blishment of the exotic Here we present occupancy trends over time for both speci es before and after the management. Results The following occupancy and colonization resu lts are ind icated in percentages, with standard error following in brackets. Exotic apple snails were not detected prior to management, but in the first year after, were estimated to occupy 45 %(0.10) of the la ke, further increasing the following year to 83%(0.06) (Figur e 5-1). In the first year control and treatments had nearly equal occupancies, 53 %(0.15) and 52%(0.15) respectively, with lower o ccupancy occurring in the all-lake sites at 20%(0.12) (F igure 5-2). In the second year, habitat occupancy models were not successful, instead a colonization model i ndicated differences between habitats. The probability of colonization was much higher in control, at 76 %(0.21) compared to both treatment and all-lake areas, 3% (0.10) and 6% (0.06) respectively (Figure 5-5). Native apple snails declined dramatically in th e first year after the scraping, but increased beyond pre-management levels by 2006-2007 (Figure 5-1). In 2002 occupancy was estimated at 66 %(0.11), compared to 13%(0.0427) then 80% (0.08) in the succeeding post-management
88 years. Differences in occupancy between hab itats was only evident in the second year, with highest occupancy estimated in c ontrol sites at 90 %(0.08), then treatment at 69%(0.12) and alllake at 53%(0.23) (Figure 5-3). In 2002 coloni zation probability was fairly low at 15 %(0.09), dropped in 2005-2006 to 7%(0.03), and then increased dramatically in the final sample year to 43%(0.14) (Figure 5-4). In the second year, overall occupancy estim ates for both species were quite high and close to one and other, with exotics indicating only a slightly higher occupancy. Native occupancy was estimated at 80%(.08) a nd 83% (0.06) for exotics (Figure 5-1). Figure 5-1: Occupancy estimates with standard error for both snail species, one year premanagement and two years post-management
89 Figure 5-2: Habitat site occupancy site estimates with standard e rror for exotic apple snails 20052006
90 Figure 5-3: Habitat site occupa ncy estimates with standard e rror for native apple snails 20062007
91 Figure 5-4: Colonization estimates with standard error for both apple snail species, one year premanagement and two years post-management
92 Figure 5-5: Habitat colonizati on estimates with standard erro r for exotic apple snails 2006-2007
93 Discussion It is not certain when exactly the exotic apple snail arrive d at Lake Toho, only that it was observed in 2003 in Goblets cove (field staff, personal communication). Its establishm ent has been remarkable, with a spike from no detecti on at all in 2002 to an estimated 45% lake wide occupancy in 2005-2006, then nearly doubling the ne xt year to 83%. This rapid increase confirms their ability to rapi dly reproduce. One individual female can produce thousands of eggs per season (Kolar and Lodge 2001, Barnes et al. 2008), with an estimated successful hatching rate of over 70%. Their successful spread within a shor t time period will probably lead to competitive interaction with the native apple snail (Halwart 1994), and there may also be the potential that the two will hybr idize (Rhymer and Simberloff 1996). As of this writing, both species have rapi dly colonized since the management and are estimated to occupy just over 80% of the lake. We have no idea how this will play out, whether or not the exotic will con tinue to occupy the entire lake out-co mpeting the native, or if the highly resilient native will still be able to successfully compete. Rawli ngs et al. (2007) noted a decline of the Florida apple snail in the presence of this exotic species. In the first post-management sampling year, co ntrol and treatment habitats were occupied roughly the same, with lower occupa ncy in the all-lake habitats. This is probably an artifact of the active 2004 hurricane season th at essentially rendered control and treatments areas the same, especially in the south end of the lake. The all-lake ar eas were completely devoid of any vegetation or organic sediments a nd thus would have been unsuitabl e habitat for either species, with no means by which to emerge from the water to breathe or lay eggs. The fact that they were captured at all in such barren areas might be the result of the traps presence, offering a substrate for both of these important activities.
94 By the second year, colonization of the exotic species was extr emely high in control plots at 76%, with a mere 3% and 6% for treatment and all-lake respectivel y. Although this species was capable of early establishmen t despite sparse plant communities it is evident that vegetated areas are preferred. We have three pieces of evidence to support this. First, near equal occupancy in the first year be tween control and treatment hab itats, compared to much lower occupancy in all-lake areas. Second, their high colonization probability in control sites in the second year. Third, the near doubling of occupa ncy by the second year which was probably at least in part due to recovering vegetation communities. The island apple snails persistence in the firs t year post-management is testament to their ability to survive adverse conditions. Not only did they endure the management operations, but they were able to establish in a lake ecosys tem almost completely devoid of vegetation, an important component of apple snail habitat. The native Florida apple snail appears highly re silient. Prior to management it was fairly widespread on Lake Toho estimated to occupy about 66% of the lake. The management greatly affected this species, as reflected by a sharp decline in occupancy to 13%, coincidentally synchronous with the population explosion of ex otic snails. The natives were captured too infrequently to offer any clues as to habitat effects. However th eir dramatic drop in occupancy is telling of their vulnerability to the management ope rations and/or resultant habitat. Despite this, they appear highly resilient, showing an even la rger site occupancy estimate by the second year than they did prior to management (80% compar ed to 66%). Although they still lagged slightly behind the exotics, this was a signi ficant recovery in a short time period. This resilience is also reflected in the much higher colonization in the second year (43%), compared both to premanagement (15%) and the first year after (3%).
95 By the second year we were able to detect differences in occupancy between habitats, with control being the most highly occupied, th en treatment and all-lake. This apparent preference for vegetated habitat is concurrent wi th Karunaratne et al. (2006) who found that native apple snail densities were much greater in wet prairie habitats than sloughs, the main difference being the presence of emergent vegetati on in the wet prairies and none in the sloughs. The authors attribute this pattern to the snails use of emergent vegetation for oviposition and for climbing out of the water for aerial respiration. Eggs must be laid several centimeters above the water (Turner 1996), thus requiring tall emergent vegetation and precluding floating leaf communities as suitable egg deposition sites. However, the density of emergent plants is likely to influence habitat usage. Karunaratne et al. (2006) noted that dense stands of Eleochari s had lower snail densities and the authors surmise that dense vegetation might hinder horizontal move ments of snails as well as their ability to climb vegetation. They noticed less peri phyton (a major food source) in dense Eleocharis stands as well. This might explain the higher occupancy noted in 2006-2007 compared to premanagement. Pickerelweed communities were ve ry dense prior to the management, but had not yet reached such intensity after management. Peak egg laying and hatching occurs in the dry season in April-May with hatchlings reaching adult size by June. With the draw down on Lake Toho beginning in November 2003, peak reproduction had already occurred and most individuals were adults, better able to survive drought conditions (Darby et al. 2002). Thus the native snail was likely spared imminent danger by having many adults in the population, that were either able to follow the receding water or perhaps evade removal by remaining stranded far enough on shore to miss the bulldozers.
96 This corroborates accumulated evidence that the Florida apple snail is tolerant and adaptable to periodic drying events Not only that, they were able to make a significant recovery in a short time, despite a sudden decline in o ccupancy immediately after the management. The reduction of high density plant stands may also have contributed to their success. The exotic island apple snail appears a very ag gressive competitor, ab le to spread quickly and occupy nearly half the lake within a ve ry short time. This was accomplished in an environment containing very spar se emergent vegetation and expos ed shorelines subject to wind and wave action. During this time the natives were captured infrequently, as they were attempting to recover from the management and perhaps also competing with the exotic. The native was able to make a strong recovery, and by the last year was essen tially at par with the exotic that had nearly doubled its occupancy w ithin a year. Perhaps the differences in diet (exotics consume vegetation, natives periphyton) will allow the two to co-exist even if the exotic remains widespread. Knowing how the nativ e will respond is unclear without further monitoring. As vegetation continues to recover, the sn ail populations will probabl y react. Continued monitoring would be useful to reveal how the na tive versus exotic species saga plays out, and how the long-term effects of the management pr oject continue to infl uence snail occupancy.
97 CHAPTER 6 FINAL CONCLUSION AND IMPLICATI ONS FOR F UTURE MANAGEMENT As Floridas wetlands continue to di sappear and become impacted by human development, it becomes even more importa nt to understand how local fauna respond to management actions. Most remaining wetlands including large lakes such as Lake Toho are becoming isolated in the landscape, fragmented by road networks and development, effectively reducing habitat availability for many animals. Ironically the purposef ul human alteration of water bodies has had such major unintended cons equences, that even more human involvement is required to ameliorate the problems. Management actions act only as band aids temporarily offsetting the problems, because the actual cure is not possible. Intelligent decisions are needed that try to strike a balance be tween the needs of several spec ies, maintenance of a viable sportfish industry and other recr eational uses, and flood control. This study was a first step in understanding the impacts of management on vege tation, birds, herps, fish, and apple snails (Muench 2004, Welch 2004, Brush 2006, Brush et al. 2008?). The main goals of Florida lake management are typically to offset or stall succession, remove undesirable accumulations of plants and muck, prom ote quality habitat and open up boater access. Arguably, the main impetus of all these is to preserve sport fish populations. However the ambiguity involved in whether or not important species such as largemouth bass are actually benefiting from management practices has led to the idea that management goals really ought to be more about the habitat quality and re creational value rather than targeted toward a single species (Allen et al. 2003). This study documented an immediate increase in occupancy of juvenile largemouth bass and bluegill, but it also s howed that many other species of fish as well as all resident reptiles and amphibians were negatively affected. Without long te rm studies to follow this trend, we are left
98 again with an unclear perspect ive on how management actions ar e performing. For example, it is unclear whether the increased occupancy of some fishes actual ly benefited the population as a whole. We have no idea if increased occupa ncy of young largemouth bass and bluegill actually recruited into the adult populat ion. We are also unsure how l ong it will take for some of the reptiles and amphibians to rebound because of li miting life history characteristics such as delayed sexual maturity, high juvenile mortality a nd/or low vagility. These traits are likely to prohibit immediate response to habitat change, it will take time to re-establish some populations but without long term data we have no idea how long that will take, if it happens at all. This study was a first step towards adopting a more holis tic view to lake management, but still leaves questions and further work. Regardless of the direction in which management moves, it is realistic to expect the preserva tion of a reputable sport fishing industry as an important goal, because of the huge economic stimulus it brings to the state of Florida. Landscape Level Considerations Managing a lake shou ld naturally extend beyond the lake itself, because neighboring environments influence the lake ecosystem and many animals utilize both habitats. The more wetlands become fragmented in the landscape, the more vulnerable aquatic and semi-aquatic animals become. Depending on the severity of the situation, if localized extinctions occur, recolonization might not be possible (Brown a nd Kodric-Brown 1977, Semlitsch and Bodie 1998; Cushman 2006). Part of this solution might be the maintenan ce of substantial wetland buffers that not only improve water quality, but also provide valuable ha bitats critical for life-history functions such as nesting, basking, foraging and refuge (Semlitsch and Bodie 2003). If development is continually permitted in buffer zones surrounding lakes, the success of many species will likely decline.
99 Water Level Considerations W ater level fluctuations have powerful imp acts on vegetation and animal communities. Various studies have noted that water level fluctuations during critical bass spawning times can reduce hatching success and year -class strength (Mitchell 1982, Kohler et al. 1993, Waters and Noble 2004). Lake stage determines aquatic plan t communities, which in turn influences fish, herp, bird and invertebrate community dynamics via changes in habitat structure, predator protection, prey availability and nesting and spawning substrate. Havens et al. (2005) documented bass recruitment changes resulti ng from lake stage and vegetation structure variations. Prolonged high wate r reduced plant biomass and c overage, coinciding with failed bass recruitment. Return of moderate water le vels triggered a structur ally diverse vegetation community that coincided with years of str ong bass recruitment. Simply put, water level regulations that permit a diverse plant commun ity, will promote a diverse fish community as well (Johnson et al. 2007). In their review paper, Johnson et al. (2007), not e that many lake and reservoir studies cite high water levels as drivers of increased year-class strength of largemouth bass. High water levels are associated with increased spawning substrate, protective c over and invertebrate production. In this case high water resulted in more inundated area, thus more available wetland habitat. Additionally, fluctuating water levels have been implicated in failed recruitment. If fluctuation occurs during spaw ning season, this can reduce hatc hing success. Perhaps high and stable water levels during important stages such as spawning and nursery seasons are important to largemouth bass. Too much fluctuation in this time might impact successful reproduction, while maintaining high water too long can adversely affect plant communities. The effects of
100 plant communities cascade thr ough the ecosystem affecting ev erything from water quality, dissolved oxygen, epiphytic and invertebrate communities and all higher fauna. Many herp species would also be influenced by water stage as it relates to critical life stages. Johnson (2005) recommends that a slow decline in lake stag e during the spring and summer followed by a reversal in the fall would most benefit amphiuma nesting success. For aquatic turtles Johnson suggests stable or slowly declining levels in the spring and summer, continuing into the fall. Any lake stage regime th at results in significant increase in water levels over a short period of time prior to Septembe r would likely drown some nests. However variation in upland habitat and el evation of nesting s ites are important to protecting some nests from inundation. Thus lake hydrology and upland habitat quality are two important stressors influencing turtle population, and most lik ely many other herpetofauna as well. When necessary draw downs are performed, th e welfare of several species ought to be considered. If draw downs happen when they are convenient for people, but at a time when animals are sensitive to change, unintended cons equences such as failed reproduction are likely to occur. Considering the powerful impact that water levels have on plant growth and animal communities, it should be more carefully considered. A comprehensive review of critical nesting, spawning, migration s easons of all aquatic vertebrates should be compiled and examined. Perhaps a suitable water schedule that would encourage favorable plant growth and animal use that is still compatible with flood control needs. Vegetation Management Considerations Prior to m anagement, the dense stands of ve getation consisted mostly of pickerelweed and cattail, and were targeted for removal because past research indicated that it was poor quality habitat. While there is no doubt that extremely dense plant stands create difficult conditions such
101 as hypoxia and limited access to various faunal speci es, dense plants might have a place in the ecosystem. This study documented high occupancy of many reptiles and amphibians as well as a host of fish species. Other studies also provide evidence that dense plant stands provide pockets of refugia to small fishes (Miranda et al. 2000, Bunch et al. 2008). Differences among plant species and their structural complexity will affect the surrounding environment and habitat suitability. For example, Bunch et al. (2008) found highest species richness in cattail, pickerelweed and torpedograss when compared to smartweed and water primrose. But even high density torpedograss and primrose communities pr ovided habitat for high abundances of stresstolerant fishes. The authors suggest that in th e interest of maintaining high fish diversity, managers should prevent large areas of dens e plant coverage, but recognize that these communities to provide important habitat for fishes as well as other wildlife such as birds, reptiles and amphibians. Pelicice et al. (2005) found that in a tropi cal reservoir, high macrophyte density was beneficial to fish assemblages. They conclude that routine vegetation removal for multiple use purposes would be deleterious to the fishes dependent on dense macr ophytes, reducing littoral fish density, biomass and species richness. Some literature has referred to the useful ness of vegetation edges (Trebitz et al. 1997, Miranda and Hodges 2000). The mowing or scra ping of narrow channels through vegetation could permit edge use by species that benefit from both open water habi tat and vegetation. It might strike a balance between predator a nd prey interactions and provide oxygenated microhabitat amidst dense vegetation. Mowed channels would also benefit boater access. Experimenting with minimal plant removal techni ques that achieve goals of boater access, while opening up some edge habitat might be useful.
102 Final Thoughts and Summary The results from this study document only th e immediate response of animals to habitat change. This is hardly conclusive informati on with which to recommend specific management actions. Lake managers and scientists working on these systems understand them best and the realities of what is possible, given the highly interdisciplinary nature of lake management. Continuing with studies that document faunal response to management will help shape our understanding of lake ecosystems, and support a adaptive and flexible ou tlook in light of new information. Although the immediate effects documented he re are important to understand, it does not paint a complete picture of long term response to management. As the vegetation continues to recover, faunal communities will change. Curre ntly, most control sites have stabilized to communities similar to pre-management conditi ons, while scraped areas are still changing. Currently in these areas, carpets of V. americana) are dominating scraped areas(Brush et al. 2008). But it is unknown whether this community will continue to flourish or eventually be outcompeted by P. cordata Although such a large scale mana gement project is not likely to take place again (FWC, personal communication), there have been valuable lessons learned. Management will always be needed in lakes because they have been so far removed from their natural state. This precludes restoration to historical condi tions, and these highly urbanize d and regulated lakes must be managed realistically. It might not be necessary to spend too much time, effort and money eradicating plant communities that are considered unacceptable and tend to return quickly after removal. Plus, in this study these communities were used by many animals that subsequently lost ground after the management. Probably no one would argue in favor of dense plants and
103 floating mats to dominate an entire lake, especi ally in small lakes where this could rapidly progress into a dried out depression. By using drawdowns and minimal plant remova l at appropriate times, managers might be able to establish lake habitats that are supportive of many specie s and also permit boater access. The realistic situation is that no one will ever be happy. Lakefront property owners will always despise dense plants and tall cattail stands, because it isnt what a lake is supposed to look like. Many Florida lakes are naturally eutrophic which promotes thes e conditions and the continual influence of humans will continue to drive lakes into undesirable states. Fisherman do not appreciate dense hydrilla, but are tolerant of some because of its purported benefit to fishes. Nor do they like such dense emergent plants that they cannot access the shorelines. They do want some plants though, as most fishermen realize th at where there are plants, there are fish. Herpetophiles will want weedy mucky habitats th at support these critters. Birders who enjoy limpkins, rails, herons and gallinules will similarl y want these vegetated lakes with some floating tussocks. Lake managers are challenged to maintain healthy ecosystems that promote economically viable sportfishing, as many other species as possible and satisfying lake residents. All this in the face of climate change, continued urbanization and heightened demands on wetland ecosystems.
104 APPENDIX A HERPETOFAUNA SPECIES LIST AND OCC UPANCY MODEL OUTPUTS Table A-1: Herpetofauna species list Common Name Scientific Name 2002 20052006 20062007 Amphiuma Amphiuma means Cottonmouth Agkistrodon piscivorous conanti Fl. banded water snake Nerodia fasciata pictiventris Fl. green water snake Nerodia floridana Fl. snapping turtle Chelydra serpentina osceola Fl. softshell turtle Apalone ferox Leopard frog Rana sphenocephala Mud snake Farancia abacura abacura Peninsula cooter Pseudemys floridana peninsularis Pig frog Rana grylio Siren Siren spp. Stinkpot Sternotherus odoratus Striped crayfish snake Regina alleni Striped mud turtle Kinosternon baurii A check indicates the species was captured enoug h to provide occupancy analyses, an x it was captured but not enough to provide analyses, no mark indicates the species was not captured at all in that year.
105 Table A-2: All occupancy models within 3 AIC for each herpetofaunal species Year Model AIC wi Amphiuma 2002 (.), (.),p(temperature) 0 0.70 2005-2006 Not captured 2006-2007 Inestimable Florida Banded Water Snake 2002 (.), (.),p(season) 0 0.52 (temperature), (.),p(season) 1.53 0.24 (lake stage), (.),p(season) 1.58 0.24 2005-2006 Inestimable 2006-2007 Inestimable Florida Green Water Snake 2002 (.), (.),p(lake stage) 0 0.84 2005-2006 Inestimable 2006-2007 Inestimable Pig Frog 2002 (.), (.),p(temperature) 0 0.46 (.), (.),p(.) 0.32 0.39 (.), (.),p(lake stage) 2.16 0.15 2005-2006 Inestimable 2006-2007 (.), (.),p(temperature) 0 0.79 Siren 2002 (lake stage), (.),p(lake stage) 0 0.22 (.), (.),p(lake stage) 0.33 0.18 (.), (.),p(temperature) (lake stage & temperature), (.),p(lake stage) 0.55 0.17 (location), (.),p(lake stage) 0.85 0.14 (temperature), (.),p(lake stage) 1.06 0.13 (.), (.),p(lake stage & temperature) 2.19 0.07 2005-2006 Inestimable 2006-2007 (.), (.),p(habitat) 0 0.85 Stinkpot 2002 (.), (.),p(lake stage) 0 0.67 (lake stage), (.),p(lake stage) 1.99 0.25 2005-2006 (.), (.),p(habitat) 0 0.45 (lake stage), (.),p(habitat) 0.21 0.40 (.), (.),p(lake stage) 2.52 0.13 2006-2007 (.), (.),p(lake stage) 0 0.58 (lake stage), (.),p(lake stage) 1.69 0.25 Striped Mud Turtle 2002 (location), (.),p(.) 0 0.68 (.), (.),p(season) 2.60 0.18 2005-2006 Inestimable 2006-2007 (.), (.),p(.) 0 0.51 (.), (.),p(temperature) 0.48 0.40 Also indicated are the model weights (wi). Where there were not enough captures for estimates, it is indicated inestimable. Where the species was not captured at all that sample year, it is indicated not captured.
106 APPENDIX B FISH SPECIES LIST AND OCCUPANCY MODEL OUTPUTS Table B-1: F ish species list Common Name Scientific Name 2002 20052006 20062007 Armored catfish Hoplosternum littorale Black crappie Pomoxis nigromaculatus Bluegill Lepomis macrochirus Blue-spotted sunfish Enneacanthus gloriosus Blue tilapia Oreochromis aureus Bowfin Amia calva Chain pickerel Esox niger Chubsucker Erimyzon spp. Dollar sunfish Lepomis marginatus Flagfish Jordanella floridae Gar Lepisosteus spp. Golden shiner Notemigonus crysoleucas Golden topminnow Fundulus chrysotus Largemouth bass Micropterus salmoides Mayan cichlid Cichlasoma urophthalmus Sailfin catfish Pterygoplichthys spp. Redear sunfish Lepomis microlophus Redfin pickerel Esox americanus Sailfin molly Poecilia latipinna Seminole killifish Fundulus seminolis Spotted sunfish Lepomis punctatus Threadfin shad Dorosoma petenense Warmouth Lepomis gulosus A check indicates the species was captured enoug h to provide occupancy analyses, an x it was captured but not enough to provide analyses, no mark indicates the species was not captured at all in that year.
107 Table B-2: All occupancy models within 3 AIC for each fish species Year Model AIC wi Armored Catfish 2002 (.), (.), p(season) 0 0.47 2005-2006 Inestimable 2006-2007 Inestimable Bluegill 2002 (.), (.),p(lake stage & temperature) 0 0.66 (lake stage), (.),p(lake stage & temperature) 1.93 0.25 2005-2006 (.), (.),p(location) 0 0.33 (.), (.),p(habitat) 1.67 0.14 (.), (.),p(habitat & location) 2.30 0.10 (habitat), (.),p(location) 2.31 0.10 2006-2007 (.), (.),p(temperature & time) 0 0.29 (.), (habitat),p(temperature) 0.62 0.21 (habitat), (.),p(temperature) 1.03 0.17 (.), (.),p(habitat) 2.00 0.11 Blue-spotted Sunfish 2002 (location), (.),p(temperature) 0 0.80 2005-2006 (.), (.),p(lake stage) 0 0.75 (.), (habitat),p(lake stage) 2.22 0.25 2006-2007 (.), (.),p(lake stage) 0 0.59 (habitat), (.),p(lake stage) 2.72 0.15 Chubsucker 2002 (.), (.),p(temperature) 0 0.34 (.), (.),p(.) 1.43 0.17 (lake stage), (.),p(temperature) 1.75 0.14 (.), (.),p(lake stage & temperature) 1.81 0.14 (temperature), (.),p(temperature) 1.81 0.14 2005-2006 Not captured 2006-2007 Not captured Dollar Sunfish 2002 (.), (.),p(.) 0 0.2 (temperature), (.),p(temperature) 0.08 0.19 (.), (.),p(temperature) 0.33 0.17 (lake stage & temperature), (.),p(temperature) 0.62 0.14 (lake stage), (.),p(temperature) 1.4 0.10 (.), (.),p(lake stage) 1.59 0.09 (.), (.),p(lake stage & temperature) 2.15 0.07 (lake stage & temp), (.),p(lake stage & temp) 2.61 0.05 2005-2006 (.), (.),p(season) 0 0.63 2006-2007 (.), (.),p(temperature) 0 0.52 (habitat), (.),p(temperature) 0.66 0.38 Also indicated are the model weights (wi). Where there were not enough captures for estimates, it is indicated inestimable. Where the species was not captured at all that sample year, it is indicated not captured.
108 Table B-2: Continued Year Model AIC wi Gar 2002 (.), (.),p(lake stage & temperature) 0 0.69 2005-2006 Inestimable 2006-2007 (habitat), (.),p(temperature) 0 0.53 (habitat), (.),p(lake stage & temperature) 1.99 0.20 (.), (.),p(temperature) 2.72 0.14 Largemouth Bass 2002 (.), (.),p(temperature) 0 0.36 (lake stage), (.),p(temperature) 1.09 0.21 (.), (.),p(lake stage & temperature) 1.59 0.16 (.), (.),p(.) 1.98 0.13 2005-2006 (.), (.),p(temperature & habitat) 0 0.30 (.), (.),p(habitat) 0.16 0.27 (habitat), (.),p(temperature) 0.41 0.24 2006-2007 (.), (.),p(lake stage & temperature) 0 0.72 (.), (.),p(temperature) 2.99 0.16 Redear Sunfish 2002 (.), (.),p(season) 0 0.68 2005-2006 (.), (.),p(temperature) 0 0.36 (.), (.),p(.) 0.78 0.25 (.), (.),p(lake stage & temperature) 1.95 0.14 (.), (.),p(lake stage) 2.37 0.11 (location), (.),p(temperature) 2.72 0.09 2006-2007 (.), (.),p(lake stage) 0 0.34 (.), (.),p(lake stage & temperature) 1.6 0.15 (.), (.),p(.) 1.66 0.15 (.), (habitat),p(lake stage) 2.30 0.11 (habitat), (.),p(lake stage) 2.91 0.08 Sailfin Catfish 2002 Not captured 2005-2006 Inestimable 2006-2007 (.), (.),p(.) 0 0.36 (.), (.),p(temperature) 0.12 0.34 (.), (.),p(lake stage) 1.99 0.13 (.), (.),p(lake stage & temperature) 2.04 0.13 Sailfin Molly 2002 (.), (.)p(lake stage) 0 0.53 (temperature), (.),p(lake stage) 1.9 0.21
109 Table B-2: Continued Year Model AIC wi (lake stage), (.),p(lake stage) 1.97 0.20 2005-2006 Inestimable 2006-2007 Inestimable Seminole Killifish 2002 Inestimable 2005-2006 (habitat), (.),p(season) 0 0.43 (.), (.),p(season) 0.72 0.30 (.), (.),p(lake stage & habitat) 2.01 0.16 2006-2007 (habitat), (.),p(season) 0 0.77 (.), (.),p(season) 2.47 0.22 Spotted Sunfish 2002 (.), (.),p(lake stage) 0 0.53 (lake stage), (.),p(lake stage) 1.84 0.21 (.), (.),p(.) 2.75 0.13 2005-2006 Inestimable 2006-2007 Inestimable Warmouth 2002 (temp), (.),p(temperature) 0 0.41 (.), (.),p(temperature) 0.74 0.28 (.), (.),p(lake stage & temperature) 2.69 0.11 (lake stage), (.),p(temperature) 2.72 0.11 2005-2006 (habitat), (.),p(lake stage) 0 0.61 (.), (.),p(lake stage) 2.06 0.22 (habitat), (habitat),p(lake stage) 2.92 0.14 2006-2007 (habitat), (.),p(temperature) 0 0.99
110 APPENDIX C APPLE SNAIL OCCUPANCY MODEL OUTPUTS Table C-1: All occupancy m odels within 2 AIC for both exotic and native species Year Model AIC wi Exotic Apple Snail 2002 Not applicable 2005-2006 (.), (.),p(location) 0 0.64 (habitat), (.),p(location) 1.18 0.36 2006-2007 (.), (habitat),p(temperat ure & location) 0 0.89 Native Apple Snail 2002 (.), (.),p(location) 0 0.55 (.), (.),p(location & temperature) 1.07 0.32 2005-2006 (.), (.),p(lake stage) 0 0.52 (.), (.),p(lake stage & temperature) 0 0.40 2006-2007 (.), (.),p(location) 0 0.58 (habitat), (.),p(location) 1.29 0.31 Also indicated are the model weights (wi)
111 APPENDIX D FUTURE MONITORING RECOMMENDATIONS For continued m onitoring of the herpetofauna and fish communities, sampling in the same manner as in the last year (2006-2007), with 18 transects evenly distri buted in the control, treatment and all-lake sites is recommended. The sampling effort should be adjusted, such that traps are in place only at specific times throughout the year, rather than being deployed yearround. This requires too much time and investment in trap maintenance and field personnel. If traps are active for about 7 days, this would be the approximate time period elapsed in one season for this study. But this is not too important, as long as there are two or more consecutive trap checks (the more the better for increased accuracy). Traps should be checked every day, to reduce stress and mortality of trapped animals. Perhaps once per month or as little as 6 months spread throughout th e year would be sufficient to encompass the extremes of the lake environment that affect de tection probabilities. For exampl e, trapping would ideally take place during cold winter weather with high lake stage, moderate weather and lake stage and very hot summer temperatures with lower lake stages. Although the protocol would be slightly different, it would still provide re liable occupancy estimates for co mparison. In fact, many of the species will almost certainly be more abundant than they were in the two years following management, which increases detections providing better precision estimates.
112 APPENDIX E IMPORTANCE OF DETECTION PROBABILITY: SOME EXAMPLES A fa mous fisheries saying is something to eff ect of, studying fish is like studying trees, except that they move around a nd are invisible (M. Rogers personal communication). This emphasizes why studying animal populations without considering detect ion probabilities is unwise (Mackenzie 2005, Mazerolle et al. 2007). Studies that report raw counts without considering detection probability are likely reporting results that reflect some random environmental or observer variable. This wa s part of the reasoning for using the program PRESENCE. In this study we chose five covariates suspected of influencing the study animals. The three (continuous) envi ronmental variables lake stage, temp erature, and time were incorporated as well as two (categorical) habitat variables loca tion on lake and habitat. As it turns out, most animals were influenced by temperature and/or la ke stage. In fact when plotting detection probabilities against these two covariates, there we re clear patterns. Such variation underscores their importance in animal studies. The following are just a few examples.
113 Figure E-1: Fish detection pr obabilities from 2002 that were influenced by temperature
114 Figure E-2: Fish detection pr obabilities from 2002 that were influenced by lake stage and temperature. The temperature curve does not correlated directly to the y-axis, it is superimposed onto the figure for reference.
115 Figure E-3: Bass detection proba bilities from 2005-2006 that were influenced by temperature
116 Figure E-4: Fish detection proba bilities from 2006-2007 that were influenced by lake stage and temperature
117 Figure E-5: Herpetofauna detec tion probabilities from 2002 that we re influenced by lake stage
118 LIST OF REFERENCES Affonso, E. G. and F. T. Rantin. 2005. Resp iratory responses of the air-breathing fish Hoplosternu m littorale to hypoxia and hydrogen-sulfide. Comparative Biochemistry and Physiology: Toxicology and Pharmacology 141:275-280. Allen, M., K. I. Tugend, M. J. Mann. 2003. Larg emouth bass abundance and angler catch rates following a habitat management project at La ke Kissimmee, FL. North American Journal of Fisheries Management 23:845-855. Allen, M. S., K. I. Tugend. 2002. Effects of a larg e-scale habitat management project on habitat quality for age-0 largemouth bass at Lake Kissimmee, FL. p. 1-11. Proceedings of the International Black Bass Symposium, Ameri can Fisheries Society. Bethesda, Maryland. Aresco, M. J. 2002. Surviving drought: Lake Ja ckson's turtles. Florida Wildlife 56(2):26. Aresco, M. J., M. S. Gunzburger. 2004. Effects of large-scale sediment removal on herpetofauna in Florida wetlands. Jour nal of Herpetology 38:275-279. Aresco, M. J. 2005. Mitigation measures to re duce highway mortality of turtles and other herpetofauna at a north Fl orida lake. The Journal of Wildlife Management 69:549-560. Bancroft, G. T., J. S. Godley, D. T. Gross, N. N. Rojas, D. A. Sutphen and R.W. McDiarmid. 1983. Large-scale operations management test of use of the white amur for control of problem plants. The herpetofauna of Lake Conway: species accounts. US Army Engineer Waterways Experiment Station, Aquatic Plan t Control Research Program, Vicksburg, Mississippi. A-83-5 Barnes, M. A., R. K. Fordham and R. L. Burk s. 2008. Fecundity of the exotic applesnail, Pomacea insularum Journal of North American B Benthological Society 27(3):738 745. Bauer, B. H. 1980. Lepomis marginatus (Holbrook), Dollar sunfish. pp. 599 In D. S. Lee, et al. Atlas of North American Freshwater Fish es. NC. State Museum of Natural History, Raleigh, NC, USA. Bayley, P. B. and D. J. Austen. 2002. Capture effi ciency of a boat electrofi sher. Transactions of the American Fisheries Society 131:435-451. Bendell, B. E. and D. K. McNi chol. 1987. Fish predation, lake acidity and the composition of aquatic insect assembla ges. Hydrobiologia 150:193-202. Berry, O. 2001. Genetic evidence for wide dispersal by the sand frog, Heleioporus psammophilus (Anura: Myobatrachidae) in western Australia. Journal of Herpetology 35:136-141.
119 Bettoli, P. W., M. J. Maceina, R. L. Noble, R. K. Betsill. 1992. Piscivory in largemouth bass as a function of aquatic vegetation abundance. North American Journal of Fisheries Management 12:509-516. Bonvechio, K. I. and T. F. Bonvechio. 2006. Re lationship between habitat and sport fish populations over a 20-year peri od at West Lake Tohopekaliga, Florida. North American Journal of Fisheries Management 26:124-133. Brauner, C. J., C. L. Ballantyne, D. J. Randall, A. L. Val. 1995. Air breathing in the armoured catfish ( Hoplosternum littorale ) as an adaptation to hypoxic, acidic and hydrogen sulphide rich waters. Canadi an Journal of Zoology 73:739-744. Brown, J. H. and A. Kodrick-Brown. 1977. Turnove r rates in insular biogeogrpahy: effect of immigration on extinction. Ecology 58:445-449. Brown, K.2005. Pomacea insularum island apple Accessed September 15, 2008. http://aquat1.ifas.ufl.edu/guide/pomcan.html. Bunch, A., M. Allen an d D. Gwinn. 2008, unpublished data. Evaluation of littoral fish communities in dense emergent plants at Lakes Kissimmee and Istokpoga. University of Florida School of Forest Resources and Conservation, Gainesville, FL. Brush, J. M. 2006. Wetland avifauna usage of littoral habitat prior to extreme habitat management in Lake Tohopekaliga, Florida. Masters Thesis. University of Florida, Gainesville, Florida, USA. Brush, J. M., M. DeSa, Z. Welch and W. Kitc hens. 2008 Final Comprehensive Report Florida Fish and Wildlife Conservation Commission. Monitoring floral and faunal succession following lake management in the litto ral reaches of Lake Tohopekaliga, FL. Butler, R. S., E. J. Moyer, M. W. Hulon, V. P. Williams. 1992. Littoral zone invertebrate communities as affected by a habitat restorat ion project on Lake Tohopekaliga, Florida. Journal of Freshwater Ecology 7:317-328. Carr, A. F. Jr. 1940. A Contribution to the Herpetol ogy of Florida. University of Florida Press, Gainesville, FL. Carlander, K. D. 1977. Handbook of Freshwater Fishery Biology. The Iowa State University Press, Ames, IA, USA. Casazza, M. L., G. D. White, C. J. Gre gory 2000. A funnel trap modification for surface collection of aquatic amphibians and re ptiles. Herpetologi cal Review 31:91-92. Chick, J. H. and C. C. McIvor. 1994. Patterns in the abundance and composition of fishes among beds of different macrophytes: viewing a littoral zone as a landscape, Canadian Journal of Fisheries and Aquatic Science 51(1994):2873.
120 Colle, D. E. and J. V. Shireman. 1980. Weight-le ngth relationships and coefficient of condition of largemouth bass, bluegill and redear sunfish in hydrilla infested lakes. Transactions of the American Fisheries Society 109:521-531. Conant, R. and J. T. Collins. 1991. A Field Gu ide to Reptiles and Am phibians: Eastern and Central North America. Third edition. Houghton Mifflin Company, Boston, MA, USA. Congdon, J. D., A. E. Dunham, R. C. Van Loben Sels. 1994. Demographics of common snapping turtles ( Chelydra serpentina ): implications for conservation and management of long-lived organisms. Amer ican Zoologist 34:397-408. Conrow, R., A. V. Zale, R. W. Gregory. 1990. Dist ributions and abundances of early life stages of fishes in a Florida lake dominated by aquatic macrophytes. Tr ansactions of the American Fisheries Society 119:521-528. Crowder, L. B. and W. E. Cooper. 1982. Habita t structural complexity and the interaction between bluegills and their prey. Ecology 63:1802-1813. Cushman, S. A. 2006. Effects of habitat loss and fragmentation on amphibians: a review and prospectus. Biological Conservation 128:231-240. Darby, P. C., P. L. Valentine-Darby, H. F. Pe rcival, W. M. Kitchens. 2001. Collecting Florida applesnails ( Pomacea paludosa ) from wetland habitats using funnel traps. Wetlands 21:308-311. Darby, P. C., R. E. Bennetts, S. J. Miller and H. F. Percival. 2002. Movements of Florida apple snails in relation to water levels and drying events. Wetlands 22:489-498. Darby, P. C., R. E. Bennetts and H. F. Percival. 2008. Dry down impacts on apple snail ( Pomacea paludosa ) demography: implications for we tland water management. Wetlands 28:204-214. Delis, P. R., H. R. Mushinsky and E. D. Mc Coy. 1996. Decline of some west-central Florida anuran populations in respons e to habitat degradation. Bi odiversity and Conservation 5:1579-1595. Dewey, M. R., W. B. Richardson, S. J. Zigler. 1997. Patterns of foraging and distribution of bluegill sunfish in a Mississippi River backwater: influence of macrophytes and predation. Ecology of Freshwater Fish 6. Dundee, H. A. and D. A. Rossman. 1989. The Amphibians and Reptiles of Louisiana. Louisiana State University Press, Baton Rouge, Louisiana.
121 Edmiston, H. L., and V. B. Myers. 1983. Florida Lakes: a description of lakes, their processes, and means of protection. Wild erness Graphics, Tallahassee, FL Ernst, C.H. and E. Ernst. 2003. Snakes of the United States and Canada. Smithsonian Institute Press, Washington, D.C. Ernst, C. H. and Ernst, E. M. 2003. Snakes of the United States and Canada. Smithsonian Books, Washington, D.C. Etheridge, K. 1990. The energetics of estivating sirenid salamanders ( Siren lacertina and Pseudobranchus striatus ). Herpetologica 46:407-414. Florida Fish and Wildlife Conservation Commission. 2001. Accessed October 1, 2008. 2003 Lake Tohopekaliga habitat management project-a fishery management program. http://www.floridaconservation.org/fishing/pdf/toho-drawdown-flier.pdf Florida Fish and W ildlife Conservation Commission. 2004. Accessed October 1, 2008. KCOL Highlights: 2004 Lake Tohopekaliga extreme dr awdown and habitat management project has concluded. http://www.floridaconservation.or g/fishing /pdf/Kiss-hl-aug-04.pdf Florida Fish and Wildlife Conservation Co mmission. 2008. Accessed September 18, 2008. Fish Identification. http://myfwc.com/Fishing/Fishes/index.html Franz, R. 1995. An introduction to the am phibi ans and reptiles of the Katharine Ordway Preserve-Swisher Memorial Sanctuary. Flor ida Museum of Natural History, Putnam County, FL. 38 Frodge, D. J., G. L. Thomas and G. B. Paul ey. 1990. Effects of canopy formation by floating and submergent aquatic macrophytes on the water quality of two shallow Pacific Northwest lakes. Aquatic Botany 38:231-248. Funderberg, J. B. and D. S. Lee. 1967. Distribution of the Lesser Siren, Siren intermedia in central Florida. Herpetologica 23:65. Gibbons, J. W., J. L. Greene and J. D. C ongdon. 1983. Drought-related responses of aquatic turtle populations. Journa l of Herpetology 17:242-246. Goldstein, R. M., and T. P. Simon. 1999. Toward a united definition of guild structure for feeding ecology of North American freshwater fishes. pp. 123-202 I n T.P. Simon (ed). Assessing the sustainability and biological integrity of water resources using fish communities. CRC Press, Boca Raton, Florida. Halwart, M: 1994. The golden apple snail Pomacea canaliculata in Asian rice farming systems: present impact and future th reat. International Journal of Pest Management 40:199-206. Hanlin, H. G. 1978. Food habits of the greater siren, Siren lacertina in an Alabama coastal plain pond. Copeia 1978:358-360.
122 Hanski, I. 1999. Metapopulation eco logy. Oxford University Press. Harding, C. 2003. Fish and wildlife recreation cr eates huge economic boon for Florida-economic impacts. Accessed October 5, 2008. http://www.floridaconservation.org/econom ic/FWC_Economic_Impact_2003.pdf. Harrison, S. and A. D. Taylor 1997. Empirical evidence for meta population dynamics: a critical review. p. 27-42. In Hanski, I. and A. D. Gilpin (ed.) Metapopulation Dynamics: ecology, genetics and evolution. Academic Press, San Diego, CA, USA. Hassan-Williams, C. and T. H. Bonner. 2007 .Accessed September 15, 2008.Texas Freshwater Fishes. http://www.bio.txstate.edu/~tbonner/txfishes/index.htm Havens, K. E., L. A. Bull, G. L. W arren, T. L. Crisman, E. J. Philips and J. P. Smith. 1996. Food web structure in a subtropical lake ecosystem. Oikos 75:20-32. Havens, K. E., E. D. Fox, S. Gornak and C. Hanlon. 2005. Aquatic vegetation and largemouth bass population responses to water-level variation in Lake Okeechobee, Florida (USA). Hydrobiologia 539. Hillborn, R. 2002. The dark side of reference point s. Bulletin of Marine Science 70(2): 403-408. Hines, J. E. 2006. Software to estimate patch occupancy and related pa rameters. USGS-PWRC. Hoover, J. J., K. J. Killgore and A. F. Cofrancesco. 2004. Suckermouth catfishes: Threats to aquatic ecosystems of the United States? A quatic Nuisance Species Research Program. United States Army Corps of Engineers. Houlahan, J. E, C. S. Findlay, B. R. Schmidt, A. H. Meyer, S. L. Kuzmin. 2000. Quantitative evidence for global amphibian population declines. Nature 404:752-755. Hoyer, M. V., R. W. Bachmann, D. E. Canfield Jr. 2008. Lake management (muck removal) and hurricane impacts to the trophic state of Lake Tohopekaliga, Florida. Lake and Reservoir Management 24:57-68. Hoyer, M. V. and J. R. Canfifeld. 1996. Largem outh bass abundance and aquatic vegetation in Florida lakes: an empirical analysis. Jour nal of Aquatic Plant Management 34:23-32. Jackson, D. A. and H. H. Harvey. 1997. Qualita tive and quantitative sampling of lake fish communities. Canadian Journal of Fish eries and Aquatic Sciences 54:2807-2813. Johnson, K. G., M. S. Allen, K. E. Havens. 2007. A review of littoral vegetation, fisheries and wildlife responses to hydro logic variation at Lake Okeechobee. Wetlands 27:110-126.
123 Johnson, S. 2005. Amphibians and Reptiles. p. 51-58 In Havens, K., M. Allen, M. Clark, D. Gawlik, J. Gore, S. Johnson and W. Kitche ns. Peer review for Kissimmee Chain-ofLakes long-term management plan conceptu al ecosystem model development. South Florida Water Management District. Karunaratne, L. B., P. C. Darby and R. E. Bennetts. 2006. The effects of wetland habitat structure on Florida apple sn ail density. Wetlands 26:1143-1150. Killgore, K. J., R. P. Morgan, N. B. Rybi cki. 1989. Distribution and abundance of fishes associated with submersed aquatic plants in the Potomac River. North American Journal of Fisheries Management 9:101-111. Klemens, M. W. 2000. Turtle Conservation. Smith sonian Institution Press, Washington, DC. Kohler, C. C., R. J. Sheehan and J. J. Sweatman. 1993. Largemouth bass hatching success and first-winter survival in two Illinois reser voirs. North American Journal of Fisheries Management 13:125-133. Kolar, C. S., and D. M. Lodge. 2001. Progress in invasion biology: predic ting invaders. Trends in Ecology and Evolution 16:199. Larimore, Kenneth D. 1957. Ecological life histor y of the warmouth (Centrarchidae). Illinois Natural History Survey, Bulletin 27(1):1-83. Lee, D. S. 1980. Lepomis microlophus (Gnther), Redear sunfish. pp. 601 In D. S. Lee, et al. Atlas of North American Freshwater Fishes N. C. State Museum of Natural History, Raleigh, NC. Ligas, F.J. 1960. The Everglades bullfrog lif e history and management. Florida Game and Freshwater Fish Commission, Tallahassee, Florida. MacKenzie, D. I., Nichols, J. D., Lachman, G. B., Droege, S., Royle, J. A., Langtimm, C. A. 2002. Estimating Site Occupancy Rates When Detection Probabilities are Less Than One. Ecology 83:2248-2255. MacKenzie, D. I. 2005. What are the issues with presence-absence data for wildlife managers? Journal of Wildlife Management 69:849-860. Mackenzie, D. I., Nichols, J. D., Royle, J. A ., Pollock, K. H., Bailey, L. L., Hines, J. E. 2006. Occupancy estimation and modeling: inferr ing patterns and dynamics of species occurrence. Academic Press, Burlington, MA. MacRae, P.S.D. and D.A. Jackson. 2006. Character izing north temperate lake littoral fish assemblages: a comparison between distan ce sampling and minnow traps. Canadian Journal of Fisheries and Aquatic Sciences 63:558-568.
124 Marchand, M. N. and J. A. Litvaitis. 2004. Effect s of habitat features and landscape composition on the population structure of a common aqua tic turtle in a re gion undergoing rapid development. Conservation Biology 18:758-767. Martof, B. S. 1969. Prolonged inanition in Siren lacertina Copeia 1969:285-289. Martof, B. S., W. M. Palmer, J. R. Bailey a nd J. R. Harrison III. 1980. Amphibians and reptiles of the Carolinas and Virginia. University of North Carolina Press, Chapel Hill. Mazerolle, M. J., L. L. Bailey, W. L. Kendall, J. A. Royle, S. J. Converse, J. D. Nichols. 2007. Making great leaps forward: accounting for dete ctability in herpetological field studies. Journal of Herpetology 41:672-689. Miranda, L. E., M. P. Driscoll and M. S. Allen. 2000. Transient physiochemical microhabitats facilitate fish survival in inhospitable a quatic plant stands. Freshwater Biology 44:617628. Miranda, L. and K. B. Hodges. 2000. Role of aquatic vegetation coverage on hypoxia and sunfish abundance in bays of a eutrop hic reservoir. Hydrobiologia 427:51-57. Miranda, L. E. and L. L. Pugh. 1997. Relationshi p between vegetation coverage and abundance, size and diet of juvenile largemouth bas during winter. North American Journal of Fisheries Management 17:601-610. Mitchell, D. F. 1982. Effects of water level fluctuation on reproduction of largemouth bass, Micropterus salmoides at Millerton Lake, California, in 1973. California Fish and Game 68:68-77. Moyer, E. J., M. W. Hulon, R. S. Butler, and R. W. Hujik. 1989. Kissimmee chain of lakes studies: study 1. Lake Tohopeka liga investigations. Florida Game and Freshwater Fish Commission, Tallahassee, FL. Moyer, E. J., M. W. Hulon, J. J. Sweatman, R. S. Butler and V. P. Williams. 1995. Fishery responses to habitat restorat ion in Lake Tohopekaliga, Florida. North American Journal of Fisheries Management 15:591-595. Muench, A. M. 2004. Aquatic vertebrate usage of littoral habitat prior to extreme habitat modification in Lake Tohopekaliga, FL. Ma sters Thesis. University of Florida, Gainesville, Florida, USA. Mushinsky, H. R., J. J. Hebrard, D. S. Vodopich. 1982. Ontogeny of water snake foraging ecology. Ecology 63:1624-1629. Nico, L. G., S. J. Walsh and R. H. Robi ns. 1996. An introduced population of the South American Callichthyid catfish Hoplosternum littorale in the Indian River Lagoon system, Florida. Florida Scientist 59:189-200.
125 Office of Program Policy Analysis and Government Accountability. 2001. Justification review of the Fish and Wildlife Conservation Commission. Tallahassee, FL. 01-48. Office of Program Policy Analysis and Governme nt Accountability. 2003. Progress Report. Tallahassee, FL. 03-38 Olson, M. H., S. R. Carpenter, P. Cunningham, S. Gafny, B. R. Herwig, N. P. Nibbelink, T. Pellett, C. Storlie, A. S. Trebitz and K. A. Wilson. 1998. Managing macrophytes to improve fish growth: a multi-lake expe riment. Fisheries Management 23:6-12. Pelicice, F. M., A. A. Agostinho, S. M. Thomaz. 2005. Fish assemblages associated with Egeria in a tropical reservoir: investigating the effects of plant biomass and diel period. Acta Oecologica 27:9-16. Power, A. 2007. Invasive species: island apple sn ail. University of Georgia, Georgia, USA. Accessed October 13, 2008. http://www.shellfish.uga.edu/invasive%20w ebitem s/meet%20guests/apple%20snail.pdf. Rawlings, T. A., K. A. Hayes, R. H. Cowie and T. M. Collins. 2007. The identity, distribution, and impacts of non-native apple snails in the continental United States. BMC Evolutionary Biology 7:97. Rey, J. B. 1996. Electrofishing s econd edition. American Fisherie s Society, Bethesda, Maryland. Rhymer, J. M and D. Simberloff. 1996. Ex tinction by hybridizati on and introgression. Annual Review of Ecology and Systematics 27:83-109. Roe, J.H., B.A. Kingsbury and N.R. Herbert. 2003. Wetland and upland us e patterns in semiaquatic snakes: implications for we tland conservation. Wetlands 23:1003-1014. Roe, J. H., B. A. Kingsbury and N. R. He rbert. 2004. Comparative water snake ecology: conservation of mobile animals that use temporally dynamic resources. Biological Conservation 118:79-89. Rozas, L. P. and W. E. Odum. 1988. Occupation of submerged aquatic vegetation by fishes: testing the roles of food a nd refuge. Oecologia 77:101-106. Savino, J. F. and R. A. Stein. 1982. Bluegill growth as modified by plant density: an exploration of underlying mechanisms Oecologia 89:153-160. Schiffer, D. 1998. Hydrology of central Florida lake s: a primer. United States Geological Survey. Library of Congress Cataloging. Denver, CO, USA. Accessed September 15, 2008. http://fl.water.usgs.gov/PDF_files/c1137_schiffer.pdf Schramm H. L., K. J. Jirka, M. V. Hoyer. 1987. Epiphytic macroinvertebrates on dominant macrophytes in two central Florida lakes. Journal of Freshwater Ecology 4:151-161.
126 Schramm, H. L. and K. J. Jirka. 1989. Epiphy tic macroinvertebrates as a food resource for bluegill in Florida lakes. Transactions of the American Fisheries Soceity 118:416-426. Schriver, P., J. Bogestrand, E. Jeppesen, M. Sondergaard. 1995. Impact of submerged macrophytes on fish-phytoplankton-zooplankton interactions: large-scale enclosure experiments in a shallow eutrophic lake. Freshwater Biology 33:255-270. Sculthorpe, C. D. 1985. The biology of aquatic vascular plants. Palgrave Macmillan, London. Semlitsch, R. D. and J. R. Bodie. 1998. Are small isolated wetlands expendable? Conservation Biology 12:1129-1133. Semlitsch, R. D. and J. R. Bodie. 2003. Biologi cal criteria for buffer zones around wetlands and riparian habitats for amphibians and reptiles. Conservation Biology 17:1219-1228. Shireman, J. V., D. E. Colle and D. F. Durant 1981. Efficiency of rotenone sampling with large and small block nets in vegetated and open-wa ter habitats. Transactions of the American Fisheries Society 110:77-80. Sinsch, U. 1990. Migration and orientation in anuran amphibians. Ethology Ecology and Evolution 2:65-79. Smith, M.A. and D.M. Green. 2005. Dispersal an d the metapopulation paradigm in amphibian ecology and conservation: are all ampibi an populations metapopulations? Ecography 28:110-128. Snodgrass, J. W., J. W. Ackerman, A. L. Bryan, Jr. and J. Burger. 1999. Influence of hydroperiod, isolation and hete rospecifics on the distributi on of aquatic salamanders (Siren and Amphiuma) among depre ssion wetlands. Copeia 1999:107-113. Sorensen, K. 2003. Trapping succe ss and population analysis of Siren lacertina and Amphiuma means Masters Thesis, University of Florida, Gainesville, FL. Soupir, C. A., M. L.Brown, C. C. Stone, J. P. Lott. 2006. Comparison of creel survey methods on Missouri River reservoirs. North American Journal of Fisheries Management 26:338-350. South Florida Water Management District. 2003. Kissimmee Basin Water Supply Plan Support Document. Chapter 4: Natural Resources, p. 39-50. Stuart, S. N., J. S. Chanson, N. A. Cox, B. E. Yo ung, A. S. L. Rodrigues, D. L. Fischman and R. L. Waller. 2004. Status and trends of amphibian declines and extinctions worldwide. Science 306:1783-1786.
127 Trebitz, A. S. Carpenter, P. Cunningham, B. John son, R. Lillie, D. Marshall, T. Martin, R. Narf, T. Pellett, S. Stewart, C. Storlie, J. Unmuth. 1997. A model of bluegill-largemouth bass interactions in relation to aquatic vegetati on and its management. Ecological Modelling 94:139-156. Tugend, K. I. 2001. Changes in the plant and fish communities in enhanced littoral areas of Lake Kissimmee, Florida, following a major habita t management. Masters Thesis, University of Florida, Gainesville, FL. Turner, R. 1996. Use of stems of emergent plan ts for oviposition by the Florida apple snail Pomacea paludosa and implications for marsh manage ment. Florida Scientist 59:35-49. Ugarte, C. A. 2004. Human impacts on pig frog ( Rana grylio ) populations in South Florida wetlands: harvest, water management and me rcury contamination. Ph.D. Dissertation. Florida International University, Miami, FL. Vermaat, J. E., L. Santamaria, P. J. Roos. 2000. Water flow across and sediment trapping in submerged macrophyte beds of contrasting growth form. Archiv fur Hydrobiologie 148:549-562. Virginia Department of Game and Inland Fisheries (VGDIF) webpage. Accessed September 15, 2008. http://www.dgif.virginia.gov/wildlife/ Waters, D. S. and R. L. Noble. 2004. Spawning season and nest fidelity of largem outh bass in a tropical reservoir. North American Jo urnal of Fisheries Management 24:1240-1251. Wegener, W. and V. P. Williams. 1974. Water le vel manipulation project completion report for Lake Tohopekaliga drawdown study. Florida Ga me and Freshwater Fish Commission, Kissimmee, FL. Wegener, W. and V. P. Williams. 1975. Fish population responses to improved lake habitat utilizing an extreme drawdown. p. 144-161. Proceedings of the Annual Conference Southeastern Association of Game and Fish Commissioners. Welch, Z. 2004. Littoral vegetation of Lake Tohope kaliga community descriptions prior to a large-scale fisheries habitat-management project. Masters Thesis. University of Florida, Gainesville, Florida USA. Whitter, T. R. and R. M. Hughes. 1998. Evaluation of fish species tolerances to environmental stressors in lakes in the northeastern United States. North American Journal of Fisheries Management 18:236-252. Wigand, C., J. C. Stevenson, J. C. Cornwell. 1997. Effects of different submersed macrophytes on sediment biogeochemistry. Aquatic Botany 56:233-244.
128 Wilbur, R. L. 1969. The redear sunfish in Fl orida. Fishery Bulletin. Florida Game and Freshwater Fish Commission. 5:1-64 Wiley, M. J., R. W. Gorden, S. W. Waite and T. Powless. 1984. The relationship between aquatic macrophytes and sport fish production in Il linois ponds: a simple model. North American Journal of Fisheries Management 4:111-119. Williams, V. P. 2001. Effects of point-source rem oval on lake water quality: a case history of Lake Tohopekaliga, Florida. Lake and Reservoir Management 17:315-329. Wood, K. V., J. D. Nichols, H. F. Percival and J. E. Hines. 1998. Size-sex variation in survival rates and abundance of pig frogs, Rana grylio in northern Florida wetlands. Journal of Herpetology 32:527-535. Wright, A. H. and A. A. .Wright. 1949. Handbook of Frogs and Toads of the United States and Canada Third edition. Comstock Pub lishing Associates, Ithaca, New York.
129 BIOGRAPHICAL SKETCH Melissa DeSa received two bachelors degree s: one in biological sciences from Brock University in St. Catherines, Ontario; and the other in wildlife biology from the University of Guelph in Guelph, Ontario. After receiving these de grees, she pursued field work in this area for a couple of years before beginning graduate sc hool. This marks over 4 years of a greatly privileged opportunity and expe rience at the Florida Fish and Wildlife Cooperative Research Unit in Gainesville, Florida. Upon completi on Melissa hopes to pursue a career involving the human dimensions of envi ronmental decision-making.