This item is only available as the following downloads:
1 EVALUATION OF ELECTROFISHING FOR INDEXING FISH ABUND AN CE IN FLORIDA LAKES By MATT A. HANGSLEBEN A THESIS PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DE G REE OF MASTER OF SCIENCE UNIVERSITY OF FLORIDA 2011
2 2011 Matt A. Hangsleben
3 To my brother Staff Sergeant Mark Hangsleben
4 ACKNOWLEDGMENTS I thank my graduate advisor, Dr. Mike Allen, without whom none of this would have been possible. I al so thank my committee members, Dr. Bill Pine and Jim Estes for their guidance and advice. I would especially like to thank Dan Gwinn for all his help and advice. I would also like to thank the following individuals for all their help on this project Jani ce Kerns, Bryan Matthias, Robert Harris, Matt Lauretta, and Stephanie Shaw. Last but not least I would like to thank my family and Nicole Kirchner for their unconditional love and support.
5 TABLE OF CONTENTS page ACKNOW LEDGMENTS ................................ ................................ ................................ .. 4 LIST OF TABLES ................................ ................................ ................................ ............ 6 LIST OF FIGURES ................................ ................................ ................................ .......... 7 ABSTRACT ................................ ................................ ................................ ..................... 8 CHAPTER 1 INTRODUCTION ................................ ................................ ................................ .... 11 2 METHODS ................................ ................................ ................................ .............. 15 Objective 1: Variation in q Acr oss Lakes and Seasons for Different Fish Species .. 15 Objective 2: Effects of Submersed Vegetation on q ................................ ................ 20 Objective 3: Evaluating the Effects of Variable q on the Ability to Monitor Abundance in Fish Stocks with CPUE Data from Boat Electrofishing ................. 21 3 RESULTS ................................ ................................ ................................ ............... 25 4 DISCUSSION ................................ ................................ ................................ ......... 39 5 F URTHER STUDY ................................ ................................ ................................ .. 44 APPENDIX : PLANT FREQUENCY OCCURANCE IN LAKES ................................ ...... 45 LIST OF REFERENCES ................................ ................................ ............................... 52 BIOGRAPHICAL SKETCH ................................ ................................ ............................ 56
6 LIST OF TABLES Table page 2 1 Size, mean depth, average secchi depth, and vegetation characteristics in each of the five lakes sampled. ................................ ................................ .......... 24 2 2 Instantaneous natural mortality rates per year for each species an d length group ................................ ................................ ................................ ................. 24 3 1 Number of marking and recapture events with each gear type in each lake. ..... 30 3 2 Numbers of fish marked an d recaptured with each gear type in each lake before adjusting for tagging and natural mortality. ................................ .............. 31 3 3 Models used in all species AIC comparison and AIC values. ........................... 32 3 4 Models used in largemouth bass AIC comparison and AIC values. ................. 32 3 5 Models used in lake chubsucker AIC com parison and AIC values. .................. 32 3 6 Models used in bluegill AIC comparison and AIC values. ................................ 32
7 LIST OF FIGURES Figure page 3 1 Observed catchability for largemouth bass in Lakes Speckled Perch, Devils ................................ ........................ 33 3 2 Observed catchability for lake chubsucker during the fall, spring and summer recapture events. ................................ ................................ ................................ 34 3 3 Observed catchability for bluegill in each lake during fall, spring, and summer recapture events. ................................ ................................ ................................ 35 3 4 Catchability for largemouth bass and bluegill in ponds with abundant vegetation and ponds with little to no vegetation. ................................ ............... 36 3 5 Simulated ability to detect a change in largemouth bass abundance with variable and constant catchability for electrofishing. ................................ ........... 37 3 6 Simulated ability to detect a change in lake chubsucker abundance with varia ble and constant catchability for electrofishing. ................................ ........... 38
8 Abstract of Thesis Presented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Master of Sc ience EVALUATION OF ELECTROFISHING FOR INDEXING FISH ABUNDANCE IN FLORIDA LAKES By Matt A. Hangsleben August 2011 Chair: Mike Allen Major: Fisheries and Aquatic Sciences Electrofishing catch per unit effort (CPUE) data are commonly used to index tempor al trends in abundance in fish monitoring programs, but the reliability of this index requires the assumption that the fraction of fish stock caught per unit effort (catchability, q ) is relatively precise and constant through time. A wide range of biologi cal, environmental, and technical factors can affect catchability, potentially violating these assumptions. To understand if CPUE data can be used to index abundance through time for Florida lakes I evaluated how electrofishing catchability varies tempor ally with different biotic and abiotic factors in five small lakes in north central Florida. I also evaluated the influence of variable electrofishing catchability on the ability of a monitoring program to detect a true change in abundance in response to a perturbation, such as could occur following changes in water levels or disruption to vegetation via hurricane using a simul ation. Lastly, I evaluated the e ffect of submersed aquatic vegetation on mean electrofishing catchability using a series of hatcher y ponds. Electrofishing catchability in the lakes study varied season, lake, and species. Catchability was higher but substantially more variable for largemouth bass Micropterus salmoides and lake chubsucker Erimyzon sucetta than for bluegill Lepomis
9 mac rochirus Catchability was highly variable between years in the same season for both largemouth bass and lake chubsucker in some of the lakes, which could preclude the use of CPUE as a reliable index of abundance. Catchability for bluegill was low but pr ecise; indicating that electrofishing CPUE could monitor abundance for this species. Simulation results revealed that statistical power decreased and the Type I error rate (i.e., the probability of detecting a difference when in fact no difference occurre d) increased substantially if q varies through time as I observed for largemouth bass and lake chubsucker Type I error rates were well above the expected value of 0.05, reaching as high as 0.7 for largemouth bass and 0.6 for lake chubsucker at high sampl e sizes. This resulted because increasing sample size improves the ability to detect real changes, but also increases the probability of detecting spurious changes due to variable q (i.e., Type I error). Thus, variable catchability hinders our ability to use CPUE data to index trends in fish abundance. Mean electrofishing catchability in the hatchery pond study showed no difference between the pond treatments, indicating that relatively high coverages of submersed aquatic plants had no substantial influe nce on average electrofishing catchability. Mean catchability was higher for largemouth bass than bluegill, similar to the lake results. Mean catchability was more variable for both largemouth bass and bluegill in ponds with abundant vegetation than in t hose with low aquatic plant abundance. This suggests that abundant vegetation does not influence average q values, but it does increase the variability in electrofishing catchability and thus could increase the uncertainty in CPUE data used to index fish abundance. These results indicated that variable electrofishing catchability hinders our ability to detect trends in abundance using CPUE data for two of the three species I evaluated Further
10 research should evaluate the temporal variability in electrof ishing catchability and explore alternate sampling methods and data sources for their reliability for monitoring fishery trends.
11 CHAPTER 1 INTRODUCTION Fish monitoring programs often seek to assess fish abundance and community composition. Most monitorin g programs use catch data to evaluate temporal trends in fish community composition, and catch per unit effort (CPUE) data are often used as an index of abundance for individual fish species. The validity of using CPUE to index abundance relies on assumin g a constant and linear relationship between CPUE and abundance as per (Harley et al. 2001): (1) where C = catch, E = effort, q = catchability, and N = abundance. For this relationship to hold true, q has to be relatively precise and constant through time. By rearranging equation 1, catchability is defined as the fraction of a fish stock collected (capture probability, C/N) per unit effort: (2) Electrofishing is widely used in f isheries management to characterize fish communities, including estimates of relative abundance, community composition, and size/age structure (Reynolds 1996). Electrofishing CPUE is easily obtained and is probably the most widely used relative abundance index in freshwater systems. However, the assumption that electrofishing CPUE is an index of abundance relies on q being relative precise and constant through time. Zalewski and Cowx (1990) argue that factors affecting q can be placed into three categori es: biological (e.g. fish size,
12 abundance, species, etc.), environmental (e.g. water clarity and temperature, substrate, aquatic plants, season, etc.), and technical (e.g. personnel and equipment), and each of these could influence the ability of CPUE data to index fish abundance. Factors affecting electrofishing catchability are interrelated and their combined effects can be difficult to isolate. Nonetheless, studies have showed that electrofishing catchability is influenced by fish abundance (McInerny an d Cross 2000; Schoenebeck and Hansen 2005), habitat (Simpson 1978; Price and Peterson 2010), water clarity (Kirkland 1962; Simpson 1978; Gilliland 1987), water temperature (Danzmann et al. 1991), power output ( Miranda and Dolan 2003 ), and personnel (Hardin and Conner 1992). Fisheries agencies often standardize factors they can control such as power output, the number of p ersonnel, and sampling season. However, many of these factors are difficult to standardize ( e.g. habitat and water clarity), influence ca tchability and thus could influence effectiveness of CPUE data to monitor fish abundance. For example, e lectrofishing catchability has been shown to vary by species. For example, Price and Peterson (2010) found electrofishing capture efficiencies to var y by species in streams. Electrofishing catchability can also vary by the habitat preference of a species. For example fish species that are located in the littoral zone (e.g. bluegill Lepomis macrochirus ) are more susceptible to electrofishing than spec ies located in the limnetic zone (e.g. gizzard shad Dorosoma cepedianum ). Fish size is another factor that has been shown to effect electrofishing (Dolan and Miranda 2003). Dolan and Miranda (2003) also suggested that many of the inconsistencies in elect rofishing immobilization thresholds for species may be the result
13 of differences in body sizes. Fish species also exhibit season specific habitat use which can affect electrofishing catchability. Many monitoring programs standardize by season to make the ir catches less variable because fish species change habitats across seasons. For example, Mesing and Wicker (1986) showed that largemouth bass Micropterus salmoides in two Florida lakes moved to inshore areas to spawn, making them more susceptible to ele ctrofishing. Coutant (1975) also showed that largemouth bass typically move to shallow water to spawn, but will move to deeper cooler water when temperatures exceed a certain threshold making them less susceptible to electrofishing. Water temperatures a lso change by season and Danzmann et al. (1991) showed a positive relationship between catchability of largemouth bass and bluegill with water temperatures. O ther factors shown to affect electrofishing catchability are habitat characteristics (e.g., abunda nce of submersed vegetation, woody debris etc. ), which var y across water bodies. For example, water clarity can affect electrofishing CPUE for largemouth bass (Kirkland 1962; Simpson 1978; Gilliland 1987) and bluegill (Simpson 1978). Other studies have also linked water clarity with a change in fish habitat selection (Miner and Stein 1996). Simpson (1978) found that electrofishing q increased in ponds with cover relative to those devoid of cover. Price and Peterson (2010) also found that electrofishing capture efficiency for 50 stream dwelling species varied by habitat characteristics and habitat complexity. Abundance of aquatic macr ophytes has also been shown to a ffect electrofishing catchability (Chick et al. 1999 ; Bayley and Austen 2002 ).
14 Understand ing how q varies for a given set of conditions is important to understanding whether CPUE can reliably index fish abundance. An ability to reliably index fish abundance is a key tool to assess fish population responses to management actions such as change s in harvest regulations or unplanned actions such as nonnative species introductions. The purpose of this study was to evaluate how electrofishing catchability may vary for fishes in Florida lakes. I hypothesize d electrofishing catchability would vary b y species, season, presence of aquatic vegetation, and water body (i.e., differences in depth, water clarity, etc.) Thus, my objectives were to 1) evaluate how q varied across lakes that differed in habitat characteristics, within and between seasons for three different fish species, 2) evaluate how q varied with the presence or absence of abundant submersed vegetation using a hatchery pond experiment, 3) evaluate the effects of the observed variation in q on the ability to monitor abundance in fish stock s with CPUE data from boat electrofishing.
15 CHAPTER 2 METHODS Objective 1: Variation in q Across L akes and Seasons for D ifferent F ish S pecies I used five Florida lakes for objective 1, to evaluate how electrofishing catchability varied by season and specie s. None of these lakes had high coverages of submersed aquatic macrophytes, so objective 2 was addressed with a separate pond study. The lakes used for objective 1 ranged in size from 3.6 to 21 ha, mean dep th from 1.86 to 4.64 m and average secchi depth from 0.98 to 5.23 m (Table 2 1). Average width of floating and emergent vegetation ranged from 5.8 to 17.1 m, percent area coverage (PAC) ranged from 21 to 66 percent, and percent volume inhabited (PVI) ranged from 4.8 to 18 percent (Table 2 1). Average emergent plant biomass ranged from 1.2 to 6.35 kg wet weight/m 2 average floating leaved plant biomass ranged from 0.01 to 4.64 kg wet weight/m 2 and average submersed plant biomass ranged from 0.1 to 3.51 kg wet weight/m 2 (Table 2 1). All plant species o bserved while sampling in each transect were listed according to the frequency that they occurred (Appendix A). All aquatic vegetation sampling took place July 27 29, 2010 and followed Florida LAKEWATCH aquatic plant sampling procedures (Florida LAKEWATCH 2009). My approach to estimate q was to establish marked fish populations in all lakes, and then conduct standardized electrofishing using F lorida Fish and Wildlife Commission (FWC) protocols ( Bonvechio 2009) to obtain estimates of catchability for each s pecies. Marked populations of fish were established in each lake using electrofishing and angling. Marking events were done with a 4.88 m aluminum boat equipped with a Smith Root VIA generator powered pulsator (GPP), one boom with eight droppers made of inch stainles s steel cable, and 1 2 netters. All species were
16 identified and measured for total length (TL) to the nearest millimeter (mm). Largemouth bass greater than 249 mm TL received a passive integrated transponder (PIT) tag and a right pelvic fi n clip (RP2). Largemouth bass were PIT tagged in the abdominal cavity following the procedure developed by Harvey and Campbell (1989). Largemouth bass between 100 mm and 249 mm TL received a left pelvic fin clip (LP2). Bluegill and lake chubsucker Erimy zon sucetta greater than 49 mm TL received a LP2 clip. I estimated short term tagging mortality for each size group to adjust the marked population available for recapture. Subsamples of fish from each 50 mm length group were placed in holding pens. Hold ing pens consisted of a PVC rectangular frame measuring 3.0 m by 1.25 m. The mesh measured 19.3 mm stretched and extended a depth of approximately 1.5 m. Fish were held for 48 h and then evaluated for mortality. Tagging mortality was estimated for each cage replicate as the number of fish dead at the end of the experiment divided by the total number of fish found alive at the end of the experiment. Mean tagging mortality and associated confidence intervals were obtained using 1,000 nonparametric bootstr ap simulations, resampling the mortality estimates of the cage replicates with replacement using Poptools software (Hood 2009). Confidence intervals were approximated as the 2.5 and 97.5 percentile of the bootstrap samples. Electrofishing catchability was measured during recapture events that were conducted following FWC sampling program protocols. Recapture trips were conducted two times in fall (early December ) of 2009 spring (February March), and summer (June) and three time s in fall of 2010 and then in the spring of 2011. Catchability was
17 not evaluated in three of the lakes in the fall of 2009 due to lower numbers of marked fish, as the project was just underway. Electrofishing was done using a 5.5 m boat equipped with a Smith Root 9.0 GPP, two boom s with eight droppers made from inch stainless steel cable, and two netters. A recapture event was defined as one circle around the entire perimeter of each lake, broken down into 600 second transects to mimic FWC protocol. Catchability was measured usi ng only marked fish, but unmarked fish captured during recapture events were also marked for future recapture events. manual for lentic systems (Bonvechio 2009). Catch ability for each recapture event was calculated for the entire lake by taking the number of recaptures ( R ) divided by the total number of marks available ( M a ), multiplying by the area ( A ) of the lake per 10 ha, then dividing by the effort ( E ) in hours (M. Lauretta, University of Florida, personal communication): (3) Equation 3 corrects values of q for lake size and sampling effort, such that the y were comparable across lakes. Because this study spanned about two years, it was i mportant to correct the number of tagged fish available for natural mortality between mark and recapture events Expected rates of natural mortality were used to adjust marks available through time following Lorenzen (2000), which predicts natural mortali ty t o decrease with fish length as: (4)
18 where M l is the instantaneous natural mortality rate at length l M r is the instantaneous natural mortality rate at reference length l r and c is the allometric exponent of the mortali ty length relationship. I set the allometric exponent of the mortality length relationship ( c ) to 0.4 (Gwinn and Allen 2010), which causes a gradual decline in natural mortality with fish length. Reference mortalities for largemouth bass and bluegill we re obtained from the literature for Florida systems (Renfro et al 1997; Crawford and Allen 2006). Reference lengths were also obtained from the literature and were based on the age of fish used in the mortality estimates and the length at age of fish in the study lakes (Canfield and Hoyer 1992). Each species was separated into two length groups based on natural breaks in length frequency data. Natural mortality (M l ) for each length group was calculated using median lengths ( l ) from each length group and reference values (summarized in Table 2 2). Only one previous study has estimated instantaneous total natural mortality for lake chubsucker (Winter 1984). I suspected that their mortality estimates measured in Nebraska would not be appropriate for lake c hubsucker populations in Florida, and thus I obtained an empirical estimate from fish at my study lakes. Lake chubsucker instantaneous natural mortality was obtained using a linear regression model developed by Hoenig (1983 cited by Hewitt and Hoenig 2005 ): (5) where M is the instantaneous natural mortality rate and t max is the maximum age observed. Because Eberts et al. (1998) showed no significant difference in length at age between male and female lake chubsucker, maximum age was obtained by taking s agittal otoliths from a sample of the largest lake chubsucker captured from each lake.
19 Sectioned otoliths were aged by two readers and disagreements were evaluated by a third reader. The maximum age was used in equatio n 5 to obtain an estimate of natural mortality which was used with the mean length of lake chubsucker from all lakes as the reference valu es in equation 4 (Table 2 2). I evaluated the effects of lake, season, and species on catchability with a logistic mod el formulated as: (6) where q is the catchability, 0 is the intercept, 1i is the slope of the variable of interest and x 1i is the variable of interest (e.g lake type). My catch data conformed to a negative bin omial distribution, which explains c atch data with a dispersion parameter k that is estimated when fitted to the data. The parameters 1i were estimated iteratively by maximizing the negative binomial log likelihood function: (7) where k is the dispersion parameter, x is the observed catch (i.e., recaptures) and is the predicted catch Predicted catch was given by: (8) where is the predicted catchability from equati on 6, E is the effort (hrs per 10 ha) and M a is the adjusted number of marks available for each recapture event (adjusted for natural mortality) To evaluate my hypothesis that electrofishing q varied by species, season, and lake I confronted my data with nine models: 1) null model where q is constant across all variables, models 2 5) q varies by single factors season, lake, year, or species, models 6 8) season by species interaction, lake by season interaction, and species by lake
20 interaction, and model 9) q varies by a season, lake and species interaction. information criterion (AIC) was used to evaluate which model explained the observed variation in the best (Akaike 1973, cited by Anderson 2008) criterion is given by : (9) where p is the number of estimated parameters in the model criterion was used because it selects the most parsimonious model, considering the tradeoff between the variance explained b y the model ve rsus the number of parameters. Anderson (2008) suggested that models with AIC values less then four have a lot of empirical support and Bolker (2008) suggested that models with AIC values less then two a part are essentially equivalent. T herefore, I considered models that had AIC values close to zero and had AIC values less than two apart and chose th e model with fewest parameters. Model probabilities were also calculated for eac h model following Anderson (2008) and are conditional of t he model set. If interaction models were chosen, I considered sub models t o evaluate the interactions. Objective 2: Effects of Submersed V egetation on q Evaluating catchability in hatchery ponds allowed us to compare how q varies between ponds with abunda nt submersed vegetation and those with little to no vegetation. The lake study could not be used to evaluate the effects of vegetation on electrofishing catchability because submersed aquatic plant abundance was low and very similar among lakes (Table 2 1 ). Electrofishing catchability was evaluated for largemouth bass (100 470 mm TL) and bluegill (50 230 mm TL) in ten hatchery ponds approximately 0.4 hectares in size with a maximum depth of two meters. Ponds were
21 grouped into two categories, ponds with a bundant vegetation (N=6) and ponds with little to no vegetation (N=4). The abundant vegetation ponds had percent area coverage ranging from 50 95% of primarily hydrilla Hydrilla verticillata where the low vegetation ponds had little to no aquatic vegetat ion due to grass carp stocking. The abundant vegetation ponds were electrofished three times in September of 2009 following the same recapture protocol as the lake sampling. The low vegetation ponds were electrofished twice at the end of July 2010 follow ing the same protocols. All species were identified and measured for total length to the nearest millimeter. Each pond was then drained to obtain true abu ndances for each fish species (see Allen et al. In P ress for de tails of the pond draining ) Observe d catchability was estimated for each pond using equation 2. I compared the mean q for each species between the two groups of ponds to explore effects of submersed vegetation on the average catchability, as well as the variation in catchability. Objective 3: Evaluating the E ffects of V ariable q on the Ability to Monitor Abundance in Fish Stocks with CPUE Data from Boat E lectrofishing I evaluated how the observed variation in q would influence the use of electrofishing CPUE data to detect changes in relativ e abundance with a simulation. This simulation is designed to inform resource managers how variable q affects statistical power and the Type I error rate. I assumed that variability in q among the lakes and seasons from the small lakes study (objective 1 ) would approximate the variability expect ed in one lake over time Because my lakes varied moderately in vegetation abundance, depth, and water clarity (Table 2 1, Appendix A), I felt this assumption was reasonable as it could approximate the variation i n depth, water clarity,
22 and littoral habitat complexity that would occur with changes in water level and water chemistry though time. The simulation estimated the expected statistical power and Type I error rates when comparing CPUE between two blocks of y ears. Estimating the Type I error rate, observe a change when a change has not occurred, and the statistical power, observe a change when one has occurred, is important to natural resource managers because it show s them with what probability they can dete ct a change of a certain size and with what probability that change is correct. To predict the Type I error and statistical power I simulated multi year datasets of electrofishing catch under the hypotheses of both constant and variable catchability. Cat ch was simulated as a random draw from a negative binomial distribution such that multiple draws for a given year would represent samples from multiple electrofishing transects. The negative binomial draws were expressed as: (10) where q y is the catchability in year y is the expected average catch, and k is the negative binomial dispersion parameter that influences the variation in C n across replicate samples ( ). Higher values of k decrease the variation among mu ltiple draws and vice versa, allowing me to mimic common among transect variation in electrofishing catches of fish for Florida lakes. Variation in q among years was simulated by drawing a separate value of q y from a beta distribution parameterized to mim ic the predicted variability in q from the best AIC model from the s mall lakes study (objective 1).
23 Parameter inputs for the simulation were the expected average catch the shape parameters of the beta distribution ( a and b ), and the negative binomial dispersion parameter k I set and k to 500 and 1, respectively. I chose these values because they produced catches similar to what you would expect from electrofishin g catch data in Florida. I set the shape parameters of the beta distribution to values that would produce similar coefficients of variation to the q values predicted by equation 6 for each species. To evaluate the ability of a monitoring program to detect a change in abundance in response to a perturbation, such as could occur following water level changes and/or disruption to vegetation, I induced a 100% (doubling) increase in the population for the second half of the simulated blocks of years of the data set. I compared the average catch pre and post change for each dataset with a two tailed t test with equal variance to determine the probability that the 100% increase in the population would be detected (i.e., statistical power). I repeated this analy sis with zero change in the population between blocks of years to determine with what probability a spurious change in relative abundance would be detected (i.e., Type I error rate). A test was considered significa nt I evaluated the effects of evaluating pre and post evaluation periods of 3, 5, and 10 years with sample sizes (i.e., electrofishing transects) ranging from six to 90. The analysis was repeated on 1,000 simulated da tasets to evaluate the influence of variable catchability on statistical power and Type I error rates.
24 Table 2 1. Size, mean depth, average secchi depth, average width of emergent and floating leaved zone, percent area coverage and volume infested, and av erage emergent, floating leaved, and submersed biomass in each of the five lakes sampled. Lake Size (ha) Average depth (m) Average secchi depth (m) Average width of emergent and floating leaved zone (m) Percent area covered Percent volume infested Average emergent plant biomass (kg wet wt/m 2 ) Average floating leaved plant biomass (kg wet wt/m 2 ) Average submersed plant biomass (kg wet wt/m 2 ) Devils Hole 11.5 4.64 4.48 10.60 66.4 4.8 6.35 4.24 3.51 Speckled Perch 12.6 1.86 1.57 9.20 21.0 7.4 2.27 1.64 0.77 Big Fish 3.0 3.25 3.02 4.10 28.2 1.7 0.76 0.03 3.49 Keys Pond 3.6 2.92 2.90 5.80 39.6 5.2 1.20 0.01 0.60 Johnson's Pond 20.8 2.45 0.98 17.10 36.9 18.0 5.31 4.64 0.10 Table 2 2. Instantaneous natural mortality rates for each species and length group wi th reference values and lengths used to correct for the number of tags present at each sampling event. Species Length group (mm) Median length (mm) M Reference M Reference length (mm) Bass 100 249 145 0.71 0.52 for ages 3 6 (Renfro et al. 1997) 32 5 for age 4 5 (Canfield and Hoyer 1992) 250+ 325 0.53 Bluegill 50 149 80 0.71 0.5 for ages 2 6 (Crawford and Allen 2006) 195 for age 4 (Canfield and Hoyer 1992) 150+ 170 0.52 Lake chubsucker 50 209 125 0.86 0.62 275 210+ 290 0.61 *Estimated in this study
25 CHAPTER 3 RESULTS Objective 1: Variation in q Across Lakes and Seasons for Different Fish S pecies Extensive effort (i.e., over 25 electrofishing boat trips per lake) was exerted to establish marked populations of each spec ies at each lake (Table 3 1 ). Total unadjusted number of marked and recaptured fish from both marking and recapture events varied by lake, with more fish marked and recaptured in the large lakes (Table 3 2). Lake chubsucker was not collected at Big Fish Lake, and very few were marked at Johnson Pond (Table 3 2). Tagging mortality was negligible for largemouth bass, lake chubsucker, and large bluegill. No lake chubsucker or large bluegill died in cage experiments and very few bass died resulting in taggin g mortalities of 1% or less. Tagging mortality for bluegill 50 100 mm and 101 150 mm in length averaged 19 % and 2 % respectively. Thus, only bluegill less than 150 mm were corrected for short term tagging mortality. My analysis of lake chubsucker otolit hs indicated that the maximum age of lake chubsucker was seven years. Using equation 4 and this maximum age obtained from the sectioned otoliths, I estimated total instantaneous natural mortality ( M ) at 0.62 for lake chubsucker. This estimate was used as the reference mortality in equation 3. The instantaneous natural mortality rate for each length group was 0.86 for lake chubsucker between 50 209 mm and 0.61 for lake chubsucker greater than 210 mm. Natural mortality rates used for largemouth bass and bl uegill are shown in Table 2. indicated that the variation in electrofishing catchability in the lake study was best explained by the species, season, lake interaction model (Table 3 3) w ith a model probability of 0.99 indica ting that the three way
26 interaction model had substantially more support than any other model. This model indicated that q varied significantly among species, seasons, and lakes but that the differences were not consistent across the levels of each t reatm ent. The species and lake interaction model had marginal support with a AIC value of 8.5; however, all the other models had very little support (i.e., AIC >10, Table 3 3). To dissect this three way interaction of species, season, and lake, I evaluated h ow q varied among season and lake for each species. Although the season and lake interaction model had the lowest AIC value for larg emouth bass; the lake model had substantial support with a AIC value of 1.89 (Table 3 4, Figure 3 1). Because a AI C < 2 indicates that the models are essentially equivalent, I selected the model that included only the lake variable (fewer parameters) as the best model to explain how q varied for largemouth bass. All other models for largemouth bass had very little s uppor t (i.e., AIC > 10). E lectrofishing catchability varied substantially across lakes for largemouth bass, with Devils Hole and Big Fish Lakes having the lowest q values and Johnson Pond the highest (Figure 3 1). T he season model and the season x lake interaction mod el both had substantial support for explaining how q varied for lake chubsucker ( AIC = 0.00 and 0.03, respec tively, Table 3 5, Figure 3 2). I selected the model including only season as a variable because it has the lowest number of parameters. Thus, el ectrofishing catchability for lake chubsucker varied by season. Lake chubsucker q values were marginally higher in spring than in the other seasons (Figure 3 2) but also exhibited high variability similar to the largemouth bass data. All other models fo r lake chubsucker had marginal support.
27 The null model had the most support f or blueg ill (Table 3 6), which indicated that q did not vary among lakes and seasons for bluegill (Figure 3 3). Values of q were consistently low across all lakes for bluegill (F igure 3 3) and did not vary with lake or season. Thus, my results showed differences in catchability among the species with largemouth bass varying by lake, lake chubsucker varying by season and bluegill catchability as constant across seasons and lakes. Objective 2: Effects of Submersed V egetation on q My evaluation of the effects of submersed vegetation on catchability corroborated the results of my AIC model selection because mean electrofishing catchability was greater and more variable for largemouth bass than for bluegill (Figure 3 4). Mean electrofishing catchability for largemouth bass and bluegill was slightly higher in ponds with abundant vegetation than in ponds with little to no vegetation (Figure 3 4); however, 95% confidence intervals overlap ped Additionally, the 95% confidence interval for catchability for both species in ponds with abundant vegetation was much larger than in pon ds with little to no vegetation. This indicated that catchability for largemouth bass and bluegill was more vari able in ponds with abundant vegetation than in those with scarce plants. Thus, submersed aquatic vegetation tended to affect the variability of electrofishing catchabi lity but did not influence the mean q values. Objective 3: Evaluating the E ffects of V ar iable q on the Ability to Monitor Abundance in Fish S tocks wi th CPUE Data from Boat E lectrofishing Simulations were only run for largemouth bass and lake chubsucker because the best model for bluegill was the null model where q was constant across seasons and lakes. Because catchability is constant for bluegill Type I error rates will not increase and statistical power will increase with sample size. Furthermore, constant q infers that
28 electrofishing CPUE could be used to index abundance for bluegill; alt hough given the low q values obtaining adequate sample size for size/age information c ould be more difficult for bluegill than for largemouth bass and lake chubsucker. Sample size (i.e., number of electrofishing transects) and the number of sample years i nfluenced the reliability of CPUE data to index largemouth bass and lake chubsucker abundance. Variable catchability affected the ability to detect changes in abundance by reducing the probability of detecting a real change (i.e., statistical power) and b y increasing the probability of detecting a spurious change (i.e., Type I error rate; Figures 3 5 and 3 6). This pattern was true for both largemouth bass and lake chubsucker, however, there was a higher increase in Type I error rate for largemouth bass. For example, the highest realized for largemouth bass under variable q was approximately 70% (Figure 3 5) where the highest realized for lake chubsucker under variable q was approximately 55% (Figure 3 6). My model inputs resulted in a coefficient of variation of q of 58% for largemouth bass and a coefficient of variation of q of 28% for lake chubsucker. Thus, the higher the levels of variation in catchability results in a higher Type I error rate. Increased sample size increased both statistical pow er and Type I error rate; whereas the number of years pre and post perturbation influenced statistical power (Figures 3 5 and 3 6). For example, Type I error rates for largemouth bass increased from approximately 15% at small sample sizes to approximatel y 70% for very large sample sizes when comparing between two years (Figure 3 5). Type I error rates remained unchanged as the number of years compared increased (Figure 3 5). Conversely, statistical power for largemouth bass increased when sample size
29 in creased but also increased as the number of years compared increased (Figure 3 5). Thus, with the variability in catchability I observed, increasing sample size for a given year will improve statistical power but the probability of finding spurious differ ences in abundance (Type I error) also increases substantially. Increasing the number of sample years increased the statistical power but had no influence on the Type I error rate, meaning that more sample years improved the reliability of CP UE data for i ndexing abundance. Further simulations with a tenfold increase in the population increased statistical power but had no effect on the Type I error rate. Thus, by increasing the size of the change to be detected the probability of seeing a change when one has not occurred does not decrease.
30 Table 3 1. Number of marking and recapture events with each gear type in each lake. Electrofishing Angling Lake Marking events Recapture events Marking events Devils Hole 15 14 7 Speckled Perch 16 14 2 Big Fish 17 12 5 Keys 17 12 5 Johnson's Pond 18 12 0
31 Table 3 2. Numbers of fish marked and recaptured with each gear type in each lake before adjusting for tagging and natural mortality. Electrofishing Angling Total Lake Species Marked Recapt ured Marked Recaptured Marked Recaptured Devils Hole Largemouth bass 838 233 59 24 897 257 Bluegill 1 629 67 0 0 1 629 67 Lake chubsucker 866 502 0 0 866 502 Speckled Perch Largemouth bass 786 327 18 4 804 331 Bluegill 1 574 73 0 0 1 574 73 Lake chubsucker 535 276 0 0 535 276 Big Fish Largemouth bass 419 137 18 3 437 140 Bluegill 1 351 56 0 0 1 351 56 Lake chubsucker 0 0 0 0 0 0 Keys Largemouth bass 118 55 25 10 143 65 Bluegill 1 236 42 3 0 1 239 42 Lake chubsuc ker 549 316 0 0 549 316 Johnson's Pond Largemouth bass 816 264 0 0 816 264 Bluegill 1 701 29 0 0 1701 29 Lake chubsucker 88 6 0 0 88 6
32 Table 3 3. Models used in all species AIC comparison AIC values, and model probabilities. Model Negat ive Loglikelihood Parameters AIC Wi Species*season*lake 382.75 10 785.50 0.00 0.99 Species*lake 389.05 8 794.10 8.60 0.01 Species*season 395.18 6 802.36 16.86 0.00 Species 401.11 4 810.23 24.72 0.00 Lake*season 445.37 8 906.74 121.24 0.0 0 Lake 448.28 6 908.56 123.05 0.00 Season 453.47 4 914.94 129.43 0.00 Year 455.42 3 916.85 131.35 0.00 Null 457.31 2 918.62 133.12 0.00 Table 3 4. Models used in largemouth bass AIC comparison, AIC values and model probabilities Model Negative Loglikelihood Parameters AIC Wi Season*lake 157.24 8 330.48 0.00 0.72 Lake 160.18 6 332.36 1.89 0.28 Season 168.73 4 345.46 14.99 0.00 Null 171.17 2 346.34 15.87 0.00 Table 3 5. Models used in lake chubsucker AIC comparison, AIC values, a nd model probabilities Model Negative Loglikelihood Parameters AIC Wi Season 116.62 4 241.24 0.00 0.50 Season*lake 113.64 7 241.27 0.04 0.49 Null 123.36 2 250.73 9.49 0.00 Lake 120.65 5 251.30 10.06 0.00 Table 3 6. Models used in bluegi ll AIC comparison, AIC values, and model probabilities Model Negative Loglikelihood Parameters AIC Wi Null 103.49 2 210.98 0.00 0.64 Season 103.04 4 214.09 3.11 0.13 Lake 100.75 6 213.50 2.52 0.18 Season*lake 100.09 8 216.18 5.20 0.05
33 Figure 3 1. Observed catchability (fraction of fish caught per unit effort) for largemouth Keys.
34 Figure 3 2. Observed catchability (fraction of fish caught per unit effort) for lake chubsucker during the fall, spring and summer recapture events.
35 Figure 3 3. Observed catchability (fraction of fish caught per unit effort) fo r bluegill in each lake during fall, spring, and summer recapture events.
36 Figure 3 4. Catchability (fraction of fish caught per unit effort) for largemouth bass and bluegill in ponds with abundant vegetation and ponds with little to no vegetation.
37 Figure 3 5 Simulated ability to detect a change in largemouth bass abundance using e lectrofishing CPUE data with variable and constant catchability: (A) comparing one year to one year; (B) comparing three years to three years; (C) comparing five years to five years; (D) comparing 10 years to 10 years.
38 Figure 3 6 Simulated ability to detect a change in lake chubsucker abundance using e lectrofishing CPUE data with variable and constant catchability : (A) comparing one year to one year; (B) comparing three years to three yea rs; (C) comparing five years to five years; (D) comparing 10 years to 10 years.
39 CHAPTER 4 DISCUSSION My results provided evidence that electrofishing catchability can be quite variable for Florida lakes and demonstrate how variable catchability can reduce the usefulness of CPUE indices for evaluating changes in fish abundance. Electrofishing catchability varied by species, season, and lake. Constant catchability for bluegill suggests that electrofishing CPUE data may be used to index abundance of bluegill and thus electrofishing would be a viable method for measuring bluegill abundance in Florida lakes. Alternately, my simulation showed that the variability in catchability observed for largemouth bass and lake chubsucker substantially increased the Type I error rates an d decreased the ability to dete ct a real change in abundance. These results highlight the inherent problems in a pplying existing conventions, such as constant catchability, without f irst confirming their validity. Many researchers have argued that electrofishing CPUE data can be used to index abundance under certain conditions (Hall 1986; Coble 1992; Hill and Willis 1994; Bayley and Austen 2002; Schoenebeck and Hansen 2005) Schoenebeck and Hansen (2005) showed that population density can be estimated from electrofishing CPUE data for walleye Sander vitreus largemouth bass, smallmouth bass Micropterus dolomieu nor thern pike Esox lucius and muskellunge E. masquinongy during specific seasons in Wisconsin lakes, assuming that catchability is density independent and that the effects of relevant environmental variables are known. However, they only evaluated the possi bilities of hyperstability, hyperdepletion, or proportional relationships between CPUE and abundance (Hilborn and Walters 1992) and did not consider the possibility that q could vary substantially across sequential sampling events. My results showed
40 high variation in q across lakes for largemo uth bass, and thus indicated that electrofishing CPUE could misrepresent changes in abundance in many instances. Bayley and Austen (2002) argued that CPUE data can be corrected to produce unbiased estimates of density but only with a standardized protocol and constant environmental and target fish conditions. They proposed a linear regression model that predicts mean q from mean fish length, mean lake depth and macrophyte coverage. Using their model to predict mean q in my five study lakes produced close estimates to the measure mean q for four out of the five lakes. However, like Schoenebeck and Hansen (2005) Bayley and Austen (2002) did not account for the possibility that q can vary substantially across sequenti al sampling events Evaluating how q varies temporally is critical in understanding whether CPUE data can be used to index trends in abundance through time in monitoring programs. Some studies have found that macrophytes affect electrofishing CPUE (Mirand a and Pugh 1997; Chick et al. 1999), while others have found no relationship (Bain and Boltz 1992). However, no studies have evaluated how q varies with different macrophyte coverages. My results showed that q is more variable in systems with higher macr ophyte coverages for largemouth bass and bluegill, but that there is no difference in the mean q between systems with abundant plants and those devoid of q will decrease for largemouth bass as macrophyte cover increases from 0% to 50%. Further research needs to focus on how electrofishing q varies with different macrophytes coverages, but my results indicated little difference in mean q between my s ubmersed vegetation tr eatments.
41 A key assumption of my simulation was that the variability in q observed among lakes and seasons would approximate within lake trends through time for a monitoring program I assert that this approximation was valid because the variation in vege tation abundance, depth, and water clarity in my lakes could represent the changes in depth, water clarity, and littoral habitat complexity that would occur in a single lake through time due to changes in water level and water chemistry. There are substan tial temporal changes in vegetation abundance, water clarity, and water level in Florida lakes (Bowes et al. 1979; Nagid et al. 2001; Hoyer et al 2005). Bowes et al. (1979) found that hydrilla abundance and water chemistry changed substantially through t ime in three Florida lakes. Nagid et al. (2001) and Hoyer et al (2005) documented that water level in some Florida lakes varies temporally and can affect water chemistry and macrophyte abundance. My assumption may not be valid for systems that have very little fluctuations in depth, water clarity, and vegetation abundance. However, I believe that the variability I observed among lakes would represent temporal habitat changes within some lakes, and thus changes in q that could be observed. An example of w here this assumption may not be valid is Lake Carlton, Florida. Brandon Thompson (Florida Fish and Wildlife Commission, personnel communication) found estimates of annual mean q to be more precise in the spring for largemouth bass over three years of an e lectrofishing mark recapture study, when compared to the variability I observed across lakes. However, the mean q decreased by about two thirds between the first two years at Lake Carlton. I saw a similar decrease in the spring mean q for largemouth ba ss at Lake Speckled Perch in my study, when comparing between two years. q values were very
42 similar between years in the spring. Thus, I saw evidence in some cases of large variation in q between sampling years, even when standardizing all other fish collection methods. Even using the more precise Lake Carlton q estimates for largemouth bass in my simulation, the Type I error rate was only reduced from 0.5 to 0.4 when comparing between two years with 25 samples. Thes e examples demonstrate that CPUE data could be misleading in monitoring trends in abundance through time. F urther research needs to focus on how electrofishing catchability varies temporally. Long term monitoring programs frequently use electrofishing CPU E data to index trends in abundance. However, many monitoring programs often neglect two sources of variability, spatial variation and detectability, which influence the use of CPUE data as an index of abundance (Yoccoz et al. 2001). Yoccoz et al. (2001) recommend that monitoring programs should incorporate estimating detection probabilities into their sampling, instead of relying on indices to draw temporal inferences. However, incorporating estimates of detection probabilities would require substantial ly more effort and cost. In many cases this is not feasible due to the limited resources and time many monitoring agencies are faced with. This study evaluated one source of variation (i.e. detectability) in monitoring programs; however, spatial variation also needs to be accounted for Mesing and Wicker (1986) showed in two Florida lakes that largemouth bass not only move to inshore areas to spawn but have considerable i nshore offshore movement through out the year. L argemouth bass also have considerabl e inshore offshore movement in Lake Santa Fe, Florida (B. Matthias, personal communication, University of Florida) I nshore offshore movement of fish could be a major cause of the variabili ty I observed in
43 electrofishing catchability. If monitoring progr ams seek to index abundance of fish species using electrofishing CPUE data spatial variation and detectability need to be accounted for.
44 CHAPTER 5 FURTHER STUDY My results indicate that if electrofishing catchability is highly variable then CPUE data shoul d not be used to index abundance. However, a major assumption of my analysis was that the variability I observed in electrofishing catchability among lakes and seasons would approximate the temporal variability in one lake over a large number of years. T herefore, I recommend multiple years of sampling on a few systems similar to this study are needed to further evaluate the temporal variability i n electrofishing catchability. I showed how variable catchability would influence the ability of CPUE data to i other fisheries metrics. Other fisheries metrics have assumptions like constant vulnerability that that could be violated if catchability is highly variable. Therefore f urther analysis could explore how variation in q could influence other fisheries metrics, such as estimation of size/age structure from electrofishing. If electrofishing catchability varies substantially through time causing unacceptable levels of stati stical power and Type I error rate, then CPUE data should not be used to index fish abundance. However, monitoring trends in abundance is not the only way to understand how fish populations change through time. I recommend that alternate sampling methods and data sources (e.g., creel surveys, age sampling, etc.) be explored for their reliability for monitoring trends in abundance or the rates that affect abundance (e.g. recruitment, mortality, and growth) A simulation model would be helpful to explore the utility of alternate data sources to monitor fish stocks.
45 APPENDIX PLANT FREQUENCY OCCU RANCE IN LAKES Devils Hole Lake Aquatic plant data collected on July 27, 2010 Frequency that plant species occur in 8 evenly spaced transect around the lake. Com mon Name Scientific Name Frequency (%) spatterdock Nuphar lutea 100 maidencane Panicum hemitomon 100 pickerelweed Pontederia cordata 100 lemon bacopa Bacopa caroliniana 100 willow Salix spp. 87.5 leafy bladderwort Utriculari a foliosa 87.5 road grass Eleocharis baldwinii 75 Triadenum virginicum 75 buttonbush Cephalanthus occidentalis 62.5 green algae Chlorphyta 62.5 yellow eyed grass Xyris spp. 50 Hypericum spp. 37 .5 unidentified # 8 Family: Lamiaceae 37.5 banana lily Hymphoides aquatica 25 unidentified # 7 25 water moss Fontinalis spp 25 water pennywort Hydrocotyle umbellata 12.5
46 redroot Lachnanthes caroliniana 12.5 unidentified # 9 12.5 unidentified # 10 Family: Poaceae 12.5
47 Speckled Perch Lake Aquatic plant data collected on July 28, 2010 Frequency that plant species occur in 8 evenly spaced transect around the lake. Common Name Scientific Name Frequency(%) road grass Eleocharis baldwinii 100 Hypericum spp. 100 spatterdock Nuphar lutea 100 banana lily Nymphoides aquatic 100 maidencane Panicum hemitomon 100 leafy bladderwort Utricularia foliosa 87.5 unidentified # 27 50 butt onbush Cephalanthus occidentalis 37.5 fascicled beaksedge Rhynchospora fascicularis 37.5 rush fuirena Fuirena scirpoidea 25 torpedograss Panicum repens 25 knotweed Polygonum spp. 25 pickerelweed Pontederia cordata 25 ort Triadenum virginicum 25 giant spikerush Eleocharis interstincta 12.5 willow Salix spp. 12.5 yellow eyed grass Xyris spp 12.5 green algae Chlorphyta 12.5
48 unidentified 8 Family: Lamiaceae 12.5 tape grass Vallisneria amer icana 12.5 little bluestem Schizachyrium scoparium 12.5
49 Aquatic plant data collected on July 29, 2010 Frequency that plant species occur in 8 evenly spaced transect around the lake. Common Name Scientific Name Frequency (%) spatter dock Nuphar lutea 100 maidencane Panicum hemitomon 75 water pennywort Hydrocotyle umbellate 62.5 sawgrass Cladium jamaicense 50 small duckweed Lemma valdiviana 50 giant cut grass Zizaniopsis miliacea 50 big floating bladderwort U tricularia inflata 37.5 southern naiad Najas guadalupensis 25 duck potato Sagittaria lancifolia 25 cattail Typha spp. 25 buttonbush Cephalanthus occidentalis 12.5 flat sedge Cyperus odoratus 12.5 common waterweed Egeria densa 12.5 knotweed Polygonum spp. 12.5 unidentified # 27 12.5 tangled bladderwort Utricularia biflora 12.5
50 Big Fish Lake Aquatic plant data collected on July 28, 2010 Frequency that plant species occur in 8 evenly spaced transect around the la ke. Common Name Scientific Name Frequency (%) musk grass Chara spp 100 rush fuirena Fuirena scirpoidea 100 green algae Chlorphyta 100 spadeleaf Centella asiatica 87.5 water pennywort Hydrocotyle umbellata 87.5 piedmont primrose Ludwigia arcuata 25 frogs fruit Phyla nodiflora 25 maidencane Panicum hemitomon 25 sweetscent Pluchea odorata 12.5 unidentified # 33 12.5
51 Keys Lake Aquatic plant data collected on July 28, 2010 Frequency that plant species occur i n 8 evenly spaced transect around the lake. Common Name Scientific Name Frequency(%) road grass Eleocharis baldwinii 100 Hypericum spp. 100 yellow eyed grass Xyris spp. 100 green algae Chlorphyta 100 rush fuirena Fuirena scirpoidea 100 florida bladderwort Utricularia floridana 87.5 maidencane Panicum hemitomon 62.5 spadeleaf Centella asiatica 25 hatpin Eriocaulon spp. 25 fox tail club moss Lycopodium alopecu roides 12.5
52 LIST OF REFERENCES A kaike, H. 1973. Information theory as an exte nsion of the maximum likelihood principle. Pages 267 281 in B. N. Petrov and F. Csaki editors. Sec ond international symposium on information theory. Akademiai Kiado, Budapest. Allen, M. S., M. W. Rogers, M. J. Catalano, D. G. Gwinn, and S. J. Walsh. In press Evaluating the potential for stock size to limit recruitment in largemouth bass. Transaction of the American Fisheries Society. Anderson, D. R. 2008. Model based inferences in the life sciences: a primer o n evidence. Springer, New York. Bain, M. B. and S. E. Boltz. 1992. Effect of aquatic plant control on the microdistribution and population characteristics of largemouth bass. Transactions of the American Fisheries Society 121:94 103. Bayley, P. B. and D. J Austen. 2002. Capture efficiency of a boat electrofisher. Transactions of the American Fisheries Society 131:435 451. Bolker, B. 2008. Ecological models and data in R. Princeton University Press, Princeton, New Jersey. Bonvechio, K. Standardized sampling manual for lentic systems. February 27, 2009 version 5. Florida Fish and Wildlife Conservation Commission: Fish and Wildlife Research Institute Freshwater Fisheries Research Section. Bowes, G. A., S. Holaday, and W. T. Haller. 1979. Seasonal variation i n the biomass, tuber density, and photosynthetic metabolism of hydrilla in three Florida lakes. Journal of Aquatic Plant Management 17:61 65. Canfield, D. E., Jr., and Hoyer, M. V. 1992. Aquatic macrophytes and their relat ion to the limnology of Florida l akes. Final Report submitted to the Bureau of Aquatic Plant Management, Florida Department of Natural Resources, Tallahassee FL Chick, J. H., S. Coyne, and J. C. Trexler. 1999. Effectiveness of airboat electrofishing for sampling fishes in shallow, vege tated habitats. North American Journal of Fisheries Management 19:957 967. Coble, D. W. 1992. Predicting population density of largemouth bass from electrofishing catch per effort. North American Journal of Fisheries Management 12:650 652. Coutant, C. C. 1 975. Responses of bass to natural and artificial temperature regimes. Pages 272 285 in H. Clepper, editor. Black bass biology and management. Sport Fishing Institute, Washington, D.C.
53 Crawford S. and M. S. Allen. 2006. Fishing and natural mortality of blue gills and redear sunfish at Lake Panasoffkee, Florida: implications for size limits. North American Journal of Fisheries Management 26:42 51. Danzmann, R. G., D. S. MacLennan, D. G. Hector, P. D. N. Hebert, and J. Kolasa. 1991. Acute and final temperature preferenda as predictors of Lake St. Clair fish catchability. Canadian Journal of Fisheries and Aquatic Sciences 48:1408 1418. Dolan, C. R. and L. E. Miranda. 2003. Immobilization thresholds of electrofishing relative to fish size. Transactions of the Amer ican Fisheries Society 132:969 976. Eberts, R. C., Jr., V. J. Santucci, Jr., and D. H. Wahl. 1998. Suitability of the lake chubsucker as prey for largemouth bass in small impoundments. North American Journal of Fisheries Management 18:295 307. Florida LA KEWATCH. 2009. Long Term Fish, Plants, and Water Quality Monitoring Program: 2009. Final Report. Department of Fisheries and Aquatic Sciences. Institute of F ood and Agricultural Sciences, University of Florida, Gainesville, Florida. Gilliland, E. 1987. population structure indices. Proceeding of the Annual Conference Southeastern Association of Fish and Wildlife Agencies 39(1985):277 287. Gwinn, C. D. and M. S. Allen. 2010. Exploring pop ulation level effects of fishery closures during spawning: an example using largemouth bass. Transactions of the American Fisheries Society 139:626 634. Hall, T. J. 1986. Electrofishing catch per hour as an indicator of largemouth bass density in Ohio impo undments. North American Journal of Fisheries Management 6:397 400. Hardin, S. and L. L. Connor. 1992. Variability of electrofishing crew efficiency, and sampling requirements for estimating reliable catch rates. North American Journal of Fisheries Managem ent 12:612 617. Harley, S. J., R. A. Myers, and A. Dunn. 2001. Is catch per unit effort proportional to abundance? Canadian Journal of Fisheries and Aquatic Sciences 58:1760 1772. Harvey, W. D., and D. L. Campbell. 1989. Technical notes: retention of passi ve integrated transponder tags in largemouth bass brood fish. The Progressive Fish Culturist 51:164 166. Hewitt, D. A. and J. M. Hoenig. 2005. Comparison of two approaches for estimating natural mortality based on longevity. Fishery Bulletin 103:433 437. Hilborn, R., and C. J. Walters. 1992. Quantitative fisheries stock assessment: choice, dynamics, and uncertainty. Chapman and Hall, New York.
54 Hill, T. D. and D. W. Willis. 1994. Influence of water conductivity on pulsed AC and pulsed DC electrofishing catc h rates for largemouth bass. North American Journal of Fisheries Management 14:202 207. Hoenig, J. M. 1983. Empirical use of longevity data to estimate morality rates. Fishery Bulletin 82:898 903. Hood, G. M. 2009. PopTools version 3.1.1. http://www.poptools.org Hoyer, M. V., C. A. Horsburgh, D. E. Canfield, Jr., and R. W. Bachmann. Lake level and trophic state variables among a population of shallow Florida lakes and within individual lakes. Canadian Journal of Fisher ies and Aquatic Sciences 62:2760 2769. Kirkland, L. 1965. A tagging experiment on spotted and largemouth bass using an electric shocker and the Petersen disc tag. Proceedings of the Annual Conference Southeastern Association of Game and Fish Commissioners 16(1962):424 432. Lorenzen, K. 2000. Allometry of natural mortality as a basis for assessing optimal release size in fish stocking programmes. Canadian Journal of Fisheries and Aquatic Sciences 57:2374 2381. McInerny, M. C., and T. K. Cross. 2000. Effects of sampling time, intraspecific density, and environmental variables on electrofishing catch per effort or largemouth bass in Minnesota lakes. North American Journal of Fisheries Management 20:328 336. Mesing C. L., and A. M. Wicker. 1986. Home range, spaw ning migrations and home of radio tagged Florida largemouth bass in two central Florida lakes. Transactions of the American Fisheries Society 115:286 295. Miner, F. G., and R. A. Stein. 1996. Detection of predators and habitat choice by small bluegills: ef fects of turbidity and alternative prey. Transactions of the American Fisheries Society 125:97 103. Miranda, L. E. and C. R. Dolan. 2003. Test of a power transfer model for standardized electrofishing. Transactions of the American Fisheries Society 132:117 9 1185 Miranda, L. E. and L. L. Pugh. 1997. Relationship between vegetation coverage and abundance, size, and diet of juvenile largemouth bass during winter. North American Journal of Fisheries Management 17:601 610. Nagid, E. J., D. E. Canfield, Jr., and M. V. Hoyer. Wind induced increases in trophic state characteristics of a large (27 km 2 ), shallow (1.5m mean depth) Florida lake. Price, A. L. and J. T. Peterson. 2010. Estimation and modeling of electrofishing capture efficiency for fishes in wadeable wa rmwater streams. North American Journal of Fisheries Management 30:481 498.
55 Renfro, D. J., W. F. Porak, and S. Crawford. 1999. Angler exploitation of largemouth bass determined using variable reward tags in two central Florida lakes. Proceedings of the Ann ual Conference Southeastern Association of Fish and Wildlife Agencies 51 :1997, pp. 175 183. Reynolds, J. B. 1996. Electrofishing. Pages 221 253 in B. R. Murphy and D. W. Willis, editors. Fisheries techniques, 2 nd edition. American Fisheries Society. Bethe sda, Maryland. Schoenebeck, C. W. and M. J. Hansen. 2005. Electrofishing catchability of walleyes, largemouth bass, smallmouth bass, northern pike, and muskellunge in Wisconsin lakes. North American Journal of Fisheries Management 25:1341 1352. Simpson, D E. 1978. Evaluation of electrofishing efficiency for largemouth bass and Winter, R. L. 1984. An assessment of lake chubsuckers as a forage for largemouth bass in a small Nebraska po nd. Nebraska Game and Parks Commission, Technical Series 16, Lincoln. Yoccoz, N. G., J. D. Nichols, and T. Boulinier. 2001. Monitoring of biological diversity in space and time. Trends in Ecology and Evolution 16:446 453. Zalewski, M., and I.G. Cowx. 1990. Factors affecting the efficiency of electric fishing. Pages 89 111 in I. G. Cowx and P. Lamarque, editors. Fishing with electricity, applications in freshwater fisheries management. Fishing News Books. Oxford, UK
56 BIOGRAPHICAL SKETCH Matt Allen Hangsleben was born in Shawnee Mission, Kansas in 1983. His passion for the outdoors led him to seek a career in natural resources. He earned his B achelor of S cience in animal ecology in 2006 with an emphasis in fisheries and aquatic sciences at Iowa State Univer sity He spent his summers in college working in various fisheries jobs in Iowa, Wyoming, and Arkansas. While working in western Wyoming and Arkansas his love of the outdoors grew. After working a kids fishing clinic in Wyoming and seeing the smiles o n all the kids faces whenever they caught a fish he was certain this is what he wanted to do with his life. After graduation he took a job working for the Idaho Game and Fish in a remote part of north central Idaho. Then he took a job working in the Gr and Canyon for the Arizona Game and Fish. Finally he moved to Florida to join the Allen lab for a few months before starting working on electrofishing catch ability issues. Matt complete 2011.