<%BANNER%>

Effects of Introduced Groundwater on Water Chemistry and Fish Assemblages in Central Florida Lakes


PAGE 1

EFFECTS OF INTRODUCED GROUNDW ATER ON WATER CHEMISTRY AND FISH ASSEMBLAGES IN CENTRAL FLORIDA LAKES By PATRICK COONEY 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 2004

PAGE 2

Copyright 2004 by Patrick Cooney

PAGE 3

To my dad, Michael Leo Cooney, you are mi ssed, and I will never stop loving and thinking about you.

PAGE 4

ACKNOWLEDGMENTS I first thank Dr. Mike Allen for serving as my advisor and committee chair. I am very grateful for his guidance throughout this project. I also thank Doug Leeper at the Southwest Florida Water Management District (SWFWMD) for his time and providing necessary access and information, and Dr. Daniel Canfield Jr. and Dr. Tom Frazer for their advice while serving as members of my graduate committee. Similarly, I thank Mark Hoyer of the Department of Fisheries and Aquatic Sciences for his advice. I especially express my gratitude to the following people for their tremendous support in the field and lab: M. Bennett, T. Bonvechio, S. Cooney, K. Dockendorf, D. Dutterer, J. Harris, K. Henry, G. Kaufman, S. Larson, C. Mwatela, E. Nagid, M. Rogers, N. Trippel and G. Warren. I also thank all of the people in the LAKEWATCH laboratory for their time and use of laboratory equipment. Most importantly, I thank and dedicate this to the people closest to me. Seans help on my project meant a lot to me. As his younger brother, I have greatly appreciated his guidance and encouragement throughout my life. The foundation that my mom and dad built and the constant support they both provided along the way made me who I am today. I never could have attained this goal without all of them and the rest of my family. Finally, I thank Julie. She has been with me every step of the way for the duration of this venture. Aside from the tremendous help she provided on this project, I most appreciated when she was there at the end of the day to make me smile. iv

PAGE 5

TABLE OF CONTENTS page ACKNOWLEDGMENTS.................................................................................................iv LIST OF TABLES.............................................................................................................vi LIST OF FIGURES..........................................................................................................vii ABSTRACT.....................................................................................................................viii INTRODUCTION...............................................................................................................1 METHODS..........................................................................................................................4 Study Sites....................................................................................................................4 Electrofishing................................................................................................................7 Fish Population Measures.............................................................................................8 Fish Population Analysis............................................................................................10 RESULTS..........................................................................................................................14 Comparison of Limnological Variables......................................................................14 Groundwater Pumping History...................................................................................15 Fish Population Comparisons.....................................................................................16 Multiple Regression Analysis.....................................................................................17 Canonical Correspondence Analysis..........................................................................18 Cluster Analysis..........................................................................................................20 DISCUSSION....................................................................................................................35 MANAGEMENT IMPLICATIONS.................................................................................46 APPENDIX COMMONLY HARVESTED FISH SPECIES...........................................48 LIST OF REFERENCES...................................................................................................49 BIOGRAPHICAL SKETCH.............................................................................................55 v

PAGE 6

LIST OF TABLES Table page 1 The county, wellfield in closest proximity, location, surface area, average depth determined with fathometer and year of first groundwater augmentation for the seven study lakes......................................................................................................27 2 The number of wells, the average volume of water pumped each day from all wells combined, and the year of initial service for the wellfields in close proximity to the augmented lakes............................................................................28 3 Mean limnological characteristics in 2003...............................................................29 4 Groundwater pumping history in augmented lakes..................................................30 5 Water chemistry of groundwater from well samples and historical lake water samples prior to initial augmentation.......................................................................31 6 Fish population measures of augmented and nonaugmented lakes..........................32 7 Significant linear regression models predicting dependent fish variables at nonaugmented and augmented lakes combined.......................................................33 8 Results of canonical correspondence analysis (CCA) for 34 fish species abundances, as measured by catch per unit effort, from 43 lakes in Florida...........34 9 Intraset correlation between the limnological variables examined and the three axes in the canonical correspondence analysis (CCA) using 34 fish species abundances, as measured by catch per unit effort, from 43 lakes in Florida...........34 A-1 Commonly harvested fish species, and the total length (mm) at which they are generally first harvested...........................................................................................48 vi

PAGE 7

LIST OF FIGURES Figure page 1 Augmented lakes sampled in three Florida counties................................................22 2 Simple linear regressions of (a) Log 10 transformation of species evenness by number of fish versus Secchi depth, and (b) diversity by number of fish versus Secchi depth.............................................................................................................23 3 Joint Plot of axis 1 versus axis 2 of the canonical correspondence analysis with lakes and species plotted along environmental gradients.........................................24 4 Joint Plot of axis 2 versus axis 3 of the canonical correspondence analysis with lakes and species plotted along environmental gradients.........................................25 5 Cluster analysis of lakes using total alkalinity, chloride, total phosphorus, and Secchi depth.............................................................................................................26 vii

PAGE 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 Science EFFECTS OF INTRODUCED GROUNDWATER ON WATER CHEMISTRY AND FISH ASSEMBLAGES IN CENTRAL FLORIDA LAKES By Patrick Cooney August 2004 Chair: Micheal S. Allen Major Department: Fisheries and Aquatic Sciences Water levels in central Florida lakes have declined since the 1960s as a result of numerous factors. To maintain water levels in these lakes, the Southwest Florida Water Management District (SWFWMD) issued permits to pump water from limestone aquifers into lakes. I assessed effects of groundwater augmentation on limnological variables and fish assemblages in seven Central Florida lakes. Pumping history information indicated that lake level fluctuations were reduced, and pumping volumes could replace the volume of water in a lake multiple times in a single year. Well water samples, when compared with current lake water samples, indicated that well water had higher mean total alkalinity and total phosphorus concentrations, and lower concentrations of total nitrogen and chlorides. The replacement of original lake water with aquifer water indicated similar patterns when comparing current lake water samples to historical samples prior to initial introduction of groundwater. Current lake water samples had higher mean pH, Secchi depth, total viii

PAGE 9

alkalinity, total phosphorus, and chloride concentrations, and lower mean color, nitrogen and chlorophyll concentrations than historical means. Historical fish population studies did not exist on these lakes therefore data from the augmented lakes were compared to 36 nonaugmented lakes in Florida. The mean values for catch per unit effort (CPUE), species richness and biomass of harvestable fishes were lower in augmented lakes than those in nonaugmented lakes. However, significant multiple linear regressions indicated that fish population responses of augmented lakes to environmental variables were similar to nonaugmented lakes with similar limnological characteristics. Canonical correspondence analysis (CCA) was used to examine the relationship between the abundance of individual fish species and measured limnological characteristics. Most fish species and nonaugmented lakes were correlated with axis one of the CCA, whereas augmented lakes were more related to axis two, indicating that augmented lakes were characteristic of high total alkalinity and Secchi depth, and low chloride and phosphorus concentrations. Cluster analysis with these four variables further demonstrated the similarities in limnological characteristics among augmented lakes. Joint plots of the CCA indicated a high probability of a low abundance of individual species in augmented lakes compared to a majority of nonaugmented lakes. One of the augmented lakes had much lower pumping rates than the others, and exhibited less of a shift in limnological variables from historical values, as well as had fish population characteristics more closely resembling those of nonaugmented lakes in the joint plot of the CCA. Therefore, reduced volumes of groundwater introduction could reduce the alteration of limnological and fish population characteristics. ix

PAGE 10

INTRODUCTION Lake water levels in central Florida have drastically decreased since the 1960s due to multiple influences. As a consequence of low precipitation (Stewart and Hughes 1974), groundwater levels were depressed and discharges of inlet streams were significantly reduced, causing lakes to receive little water input (Dooris and Martin 1979). Further, urban development changed Floridas drainage systems and diverted storm runoff away from lakes (Stewart and Hughes 1974), and agriculture endeavors withdrew water from lakes for citrus irrigation and freeze protection (Dooris and Moresi 1975). Population expansion also increased water demand, resulting in increased pumping of aquifer water at wellfields, subsequently lowering groundwater and lake levels (Stewart 1968; Stewart and Hughes 1974; Allen 1999). Groundwater is utilized for public, industrial and agricultural purposes (Southwest Florida Water Management District [SWFWMD] 1998; Brenner et al. 2000). In the northern Tampa area, groundwater pumping at wellfields began in 1963 to meet the increased demands for water (Stewart and Hughes 1974). Wells ranged from 120 to 180 meters in depth and produced thousands of cubic meters of water a day from the Tampa and Suwannee Limestone Formations, the two uppermost layers of the Floridan Aquifer (Stewart and Hughes 1974; Sinclair 1977). The Floridan Aquifer is comprised of sand and clay in the upper regions, with dolomite comprising the lower regions (Stewart and Hughes 1974; Belanger and Kirkner 1994). 1

PAGE 11

2 Lakes in the vicinity of the wellfields were hydraulically connected to the water table aquifer, meaning that water moved naturally between the lakes and the water table aquifer, a surficial aquifer located two to five meters below land surface (Stewart and Hughes 1974). As groundwater was pumped at wellfields, a localized cone of depression formed in the Floridan Aquifer, inducing an increase in leakage from the water table aquifer to the Floridan Aquifer. As a result, flow increased from the lakes to the water table aquifer, causing a decline in water levels in lakes near wellfields considerably greater than those that would naturally occur in lakes away from wellfields (Stewart 1968; Stewart and Hughes 1974; BRA 1996). Landowners expressed concern with the declining lake water levels, and in order to address the issue, the SWFWMD permitted landowners to construct wells of similar depths to those in the wellfields for pumping water from the Floridan Aquifer into lakes (BRA 1982; Dooris et al. 1982; Belanger and Kirkner 1994; BRA 1996; Allen 1999). Water levels in these lakes are now constantly maintained slightly above original mean lake levels and are not pumped to a degree that will allow spill over (Stewart and Hughes 1974). However, maintaining lake levels at higher than normal levels accelerates evaporation rates, and also increases leakage of lake water, further increasing the permeability of lake-bottom sediments (Stewart and Hughes 1974; Belanger and Kirkner 1994). As lake water leakage increases, even more groundwater is necessary to maintain water levels year round (Stewart and Hughes 1974). Previous investigations evaluated the altered water chemistry of augmented lakes and the consequent change in macrophyte growth and phytoplankton diversity. For example, Martin et al. (1976b) found that the elevated hardness of pumped groundwater

PAGE 12

3 increased the ability of augmented lakes to support hydrilla, Hydrilla verticillata growth, and Dooris et al. (1982) found that phytoplankton diversity was enhanced in augmented lakes due to increased concentrations of inorganic carbon via groundwater input. Little work has assessed effects of groundwater augmentation on fish communities (Bartos 1998; Allen 1999). I evaluated the influence of lake augmentation on limnological charateristics and fish populations in seven augmented lakes in central Florida in the summer of 2002. My objectives were to 1) determine limnological characteristics and pumping history of seven augmented lakes and their corresponding groundwater wells and compare the limnological characteristics between lakes, wells and historical data, 2) estimate fish population parameters in the lakes, and 3) compare augmented lake limnological characteristics and fish populations with those from a data base of 36 nonaugmented Florida Lakes.

PAGE 13

METHODS Study Sites The seven augmented lakes are located in Pasco, Hillsborough, and Polk counties in central Florida (Figure 1). The lake surface areas (SA) were obtained from the Gazetteer of Florida Lakes (Shafer et al. 1986) and unpublished SWFWMD reports, and the locations of the lakes were obtained using global positioning system (GPS) coordinates from a Garmin GPSMAP 76 (Table 1). These lakes are not spring-fed from the Floridan Aquifer but do exchange water with surficial aquifers. Each lake exhibited significant declines in water levels due to reduced rainfall and wellfield pumping, where hundreds of thousands of cubic meters of water were removed from the Floridan Aquifer each day (Table 2), creating the necessity to pump Floridan Aquifer water into each lake to maintain lake levels. Goose Lake was the first of the study lakes to be augmented, (1954), and Loyce Lake the most recent (1996) (Table 1). Limnological Characteristics I assessed some important limnological characteristics of my study lakes in August, 2003. The percent lake area covered by aquatic macrophytes (PAC) was recorded using a boat-mounted Raytheon DE-719 Precision Survey Fathometer (Maceina and Shireman 1980). Seven transects were made across each lake at a constant speed while the fathometer recorded the presence or absence of plants on a paper roll. The total length of paper recorded on for each lake was divided into 100 equally spaced instantaneous samples, and the presence or absence of plants at these locations was recorded. The 4

PAGE 14

5 number of locations with aquatic vegetation present was expressed as a percentage (Canfield and Hoyer 1992). For water chemistry, three 1-liter samples of water were collected at arm depth (~0.5m) in acid-cleaned Nalgene bottles at three mid-lake sampling stations established in each lake in August of 2003. The samples were immediately placed on ice and returned to the laboratory for analysis. Secchi depth (m) was measured at each station and averaged to determine mean Secchi depth. Dissolved oxygen concentration (mg/L) and temperature ( o C) were also measured with a Model 85 Yellow Springs Instrument (YSI) meter at about 40% of the depth at each mid-lake station and at the location of pumped water discharge. Upon arriving at the laboratory (University of Florida, Gainesville, Florida), pH was measured immediately using an Orion Model 601A pH meter calibrated against buffers at pH 4.0, 7.0, and 10.0. Total alkalinity (mg/L as CaCO 3 ) was determined by titration with 0.02 molar H 2 SO 4 (APHA 1989). Chlorophyll concentrations (g/L) were determined spectrophotometrically (method 10200 H (2c), APHA 1989) following pigment extraction with ethanol (Sartory and Grobbelaar 1984). Total phosphorus concentrations (g/L) were determined using procedures of Murphy and Riley (1962) with a persulfate digestion (Menzel and Corwin 1965). Total nitrogen concentrations (g/L) were determined by oxidizing water samples with alkaline persulfate and determining nitrate-nitrogen with second derivative spectroscopy (DElia et al. 1977; Simal et al. 1985; Wollin 1987; Crumpton 1992). Chloride concentrations (mg/L) were determined by titration of the water samples with 0.0141 mole mercuric nitrate and using diphenylcarbazone for determining endpoints (Hach Chemical Company 1975). To

PAGE 15

6 analyze for color (platinum-cobalt units), water samples were first filtered through a Gelman type A-E glass fiber filter. Color was then determined by using the platinum-cobalt method and a spectrophotometer (APHA 1989). Three 1-liter water samples were also collected from each of the wells supplying water to the lakes in August of 2003. The volume of the well delivery pipe was measured, and the pump was run to flush the pipe with at least twice the calculated volume before water samples were taken. These samples were placed on ice and returned to the laboratory and analyzed at the same time as the lake samples for pH, total alkalinity, total phosphorus, total nitrogen, and chloride concentrations. Chlorophyll concentrations and color were not determined for the well samples because natural filtration and lack of sunlight exposure that is characteristic of the Floridan Aquifer makes the levels of these variables negligible. Secchi depth was not measured within the well pipe. The pumping history of each study lake was determined from unpublished SWFWMD reports. The daily, monthly and yearly averages from these reports were used to determine the average amount of groundwater pumped on a yearly basis. The volume of each lake was determined by multiplying the surface area of the lake by the average depth determined from the fathometer transects. The average volume of water pumped per year was then divided by the volume of the lake to determine the amount of times per year the water pumped would replace the current volume of water in each lake. Mountain Lake and Sunset Lake each share pumps with other lakes, and the amount of pumped water for individual lakes was not separately recorded. Therefore, I calculated

PAGE 16

7 the volume of all lakes receiving water from a shared pump, and determined the percentage of the total volume attributed to Mountain and Sunset lakes. Finally, I compared the ranges and means of limnological characteristics of the augmented lakes to the well water, and to the limited amount of historical water chemistry data for the augmented lakes that existed prior to initial pumping of aquifer water. Electrofishing Fish populations in the seven augmented lakes were sampled during the warm season in July or August of 2002 using electrofishing. Electrofishing transects of continuous DC current were conducted for ten minutes to collect fish in the littoral area of each lake with a 4.3 m aluminum jon boat powered by a 15 horsepower outboard motor. Six transects were conducted at Clear, Dan and Sunset lakes, seven transects at Goose, Loyce and Saddleback lakes, and eight at Mountain Lake. The number of 10-minute transects indicate how many transects were necessary to circumnavigate the entire lake. Electrofishing equipment consisted of a generator (5000 Watt AC), pulsator (Coffelt model VVP 15) and a bow-mounted cathode probe supplying an electrical output of approximately seven amps. All collected fish from each transect were counted, measured to the nearest millimeter total length (TL), weighed to the nearest gram, and identified to species. Fish with total lengths less than 20 mm TL were not included in analyses due to selectivity of the gear (Reynolds 1996). Due to the lack of fish population studies on these augmented lakes prior to the initial pumping of water, I compared the data collected from augmented lakes to a data set of 60 nonaugmented Florida lakes (Canfield and Hoyer 1992; Bachmann et al. 1996). Two lakes, Mountain and Gate, were removed from the 60 lake data set because they

PAGE 17

8 were augmented lakes. Three lakes, Apopka, Lochloosa and Harris, were also removed due to their surface areas being orders of magnitude larger than all other study lakes, because lake size influences species richness (Bachmann et al. 1996). To assess the likelihood that more electrofishing transects would add additional species, I constructed curves, using Bachmann et al.s method, for augmented and nonaugmented lakes demonstrating the cumulative number of fish species captured as more transects were conducted (Bachmann et al. 1996). To ensure that the right-hand portion of the curves flattened, the number of species captured in the next to last electrofishing transect was divided by the number of species captured in the last transect, and expressed as a percentage, which was termed the exhaustion index (Bachmann et al. 1996). All seven augmented lakes had exhaustion indexes of 100%, meaning that all captured species from each lake had already been captured by the next to last transect. In 19 of the 55 nonaugmented lakes, the exhaustion indexes were less than 90%, possibly indicating that these lakes were inadequately sampled. Therefore, these 19 lakes were not used in the analyses. The remaining 36 nonaugmented lakes had exhaustion indexes that equaled or exceeded 90%, indicating that the right-hand portion of the curves had flattened. These 36 nonaugmented lakes were used for comparison with the seven augmented lakes. Fish Population Measures For the seven augmented lakes and the remaining 36 nonaugmented lakes, I estimated catch per unit effort (CPUE) and species richness. Catch per unit effort (CPUE) was calculated by dividing the total number of individual fish captured in each transect by 10 minutes (duration of one transect), and then averaging across the total

PAGE 18

9 number of transects conducted in the particular lake (number of fish/minute). Species richness was calculated as the total number of fish species collected in each lake. Evenness was calculated for both the number of individuals and the total weight of each species using Simpsons measure of evenness (Krebs 1999). Evenness attempts to measure how evenly the number of individuals or weight is distributed among all species in a community. The Simpsons measure of evenness (E) is defined as: EpPsi12/(/) (eq. 1) where p i is the number of individuals or total weight of the ith species, P is the total number of individuals or total weight of all species, and s is the total number of species in each lake. This index is relatively unaffected by the rare species in the sample and ranges from 0 to 1 (Krebs 1999). In addition, I also calculated a Shannon-Wiener index of diversity for both the number of individuals and the total weight of each species collected in each lake (Krebs 1999). Diversity attempts to account for evenness and richness by looking at both the number of species and how evenly distributed the number of individuals or weight is amongst the total number of species in each lake. The Shannon-Weiner index of species diversity (H) is defined as: HpPpPiisi'(log12 ) (eq. 2) where p i is the number of individuals or total weight of the ith species, P is the total number of individuals or total weight of all species, and s is the total number of species in each lake. For biological communities, H ranges from zero to five (Krebs 1999), and is expressed in bits per individual (bits/individual).

PAGE 19

10 I also calculated the total biomass of harvestable fish caught per minute in each of the lakes (Canfield and Hoyer 1990). These fish exceeded lengths at which anglers generally harvest the given species (Appendix A). Fish Population Analysis The relationships between fish population variables (CPUE, evenness, diversity, richness and harvestable biomass) and limnological variables (total alkalinity, chlorides, total phosphorus, total nitrogen, lake surface area, Secchi depth, chlorophyll, color and percent composition of submersed aquatic vegetation) were examined for the augmented and nonaugmented lakes using multiple linear regression. Prior to model selection, a Wilk-Shapiro test was performed on all dependent fish variables, except diversity, to test for normality (Procedure UNIVARIATE NORMAL, SAS 1996). The evenness index for both fish weight and number of fish, as well as CPUE and harvestable biomass, were log 10 (x+1) transformed to increase normality. Stepwise model selection procedure was used to create multiple regression models (STEPWISE option, SAS 1996) with a significance level of 0.05 for independent variables to remain in the model. Multiple regression models with only one significant independent variable predicting a fish population variable were graphed. Confidence limits of 95% were placed above and below the predicted regression line for each model, and the points corresponding to augmented lakes were examined for influence or diverging patterns. Models with multiple significant habitat variables predicting a fish population variable were examined for possible influence by augmented lakes. This was done using influence diagnostics, including DFFITS, COVRATIO and studentized-residuals (SAS 1996). The DFFITS value represents the number of estimated standard errors that the fitted value changes if the point is removed from the data set (Myers 1990). A value

PAGE 20

11 close to zero indicates a low influence of the given point. The COVRATIO values display the reduction in the estimated generalized variance of the coefficient over what would be produced without the data point. A value close to one indicates little influence on the estimated generalized variance. Finally, studentized-residuals were used to detect outliers. A value close to zero indicates a minimal residual for the given point, indicating a non-outlier (Myers 1990). I concluded that if augmented lakes did not have extreme values for DFFITS, COVRATIO, and studentized residuals, then the observation would be within the overall pattern for nonaugmented lakes. Canonical correspondence analyses (CCA) (PC-ORD 1999) is a multivariate analysis technique that utilizes data from two matrices to relate community composition to known variation in the environment (Ter Braak 1986). The CCA was used to arrange lakes and species of fish along environmental gradients. Catch per unit effort (fish/min.) was calculated as a measure of relative abundance for each species in each lake and placed in the primary matrix for comparison of community patterns across lake samples (Hinch and Collins 1993). Species observed in less than three of the 43 lakes (seven augmented and 36 nonaugmented) were removed from the analysis to reduce the effects of rare taxa. No rare taxa were found in the augmented lakes. The same limnological variables as in the multiple regressions were placed in the secondary matrix across lakes. Percent data (percent lake area covered by aquatic macrophytes) were arcsine(x/100) transformed (Zarr 1999) and all other directly measured environmental variables were log 10 (x+1) transformed to reduce kurtosis (Palmer 1993). Canonical correspondence analysis is not hampered by high multicollinearity between species, or between environmental variables (Palmer 1993). Therefore,

PAGE 21

12 preprocessing or elimination of multicollinear data is unnecessary. Similarly, CCA estimates the modal locations of highly skewed species distributions quite well, and is robust to violations of assumptions (Palmer 1993). In CCA, lakes were assigned scores determined from weighted-averages of species abundances (Palmer 1993). A multiple linear least-squares regression was then performed with environmental variables as independent variables, and lake scores as the dependent variables. New lake scores were then assigned as the value predicted from the resulting regression equation. The algorithm continued re-standardizing lakes and species scores until they remained constant with progressing iterations. The product was the first ordination axis, which was a linear combination of environmental variables that maximized the correlation between lake and species scores. Second and higher ordination axes also maximized correlations with scores, uncorrelated with the previous axes (Ter Braak 1986). Intraset correlations are the correlation coefficients between the environmental variables and the ordination axes (Ter Braak 1986). The signs and magnitudes of the intraset correlations were examined to assess the relative importance of each environmental variable in structuring the fish community. Intraset values less than -0.350 and greater than 0.350 were considered more highly correlated than those between -0.350 and 0.350. This criterion was arbitrary and was not intended to reflect statistical significance, and all intraset values are presented. The canonical correspondence analysis was graphed on a joint plot, with the first three ordination axes (PC-ORD 1999). The lakes and species were examined for the dominant patterns in community composition as explained by the environmental

PAGE 22

13 variables, and the augmented lakes were further examined for diverging patterns. The intraset values of the CCA were also examined to determine which environmental variables were most strongly correlated with augmented lake scores. Those intraset values less than -0.350 and greater than 0.350 were considered more highly correlated, and were used to construct a cluster diagram (PC-ORD 1999).

PAGE 23

RESULTS Comparison of Limnological Variables Augmented lakes had a smaller average surface area than nonaugmented lakes. The seven augmented lakes in this study had surface areas ranging from 13 ha to 39 ha (Table 1) with an average of 21 ha, whereas the nonaugmented lakes ranged from 1.8 ha to 271 ha, with an average of 83 ha. The range of percent lake area covered by aquatic macrophytes (PAC) displayed large variation for both augmented (10% to 58%) and nonaugmented lakes (0% to 100%) (Table 3). Four of the augmented lakes had Secchi depths greater than the average of the nonaugmented lakes (2.04 m), whereas all of the augmented lakes had pH levels higher than the average for the nonaugmented lakes (7.56) (Table 3). Three of the seven augmented lakes exceeded the range of the total alkalinity for the nonaugmented lakes (0.28 to 106 mg/L as CaCO 3 ), and all augmented lakes had values greater than the nonaugmented average (31.1 mg/L as CaCO 3 ). All augmented lakes but Saddleback had lower chlorophyll concentrations than the average for nonaugmented lakes (26.5 g/L), and all had lower phosphorus and nitrogen concentrations than the average nonaugmented values (30.5 g/L, and 939 g/L respectively). Sunset Lake was the only augmented lake to exceed the average chloride concentration for the nonaugmented lakes (18.5 mg/L). Finally, the color values of the augmented lakes were within the range (1.25 to 57.5 Pt-Co units) and near the average (20.6 Pt-Co units) of the nonaugmented lakes (Table 3). 14

PAGE 24

15 Groundwater Pumping History Clear Lake had the smallest volume of the augmented lakes (1.20 x 10 5 m 3 ), and Mountain Lake had the largest volume (1.13 x 10 6 m 3 ) (Table 4). Similarly, Mountain Lake had the largest average volume of groundwater pumped into the lake (2.67 x 10 6 m 3 /year), whereas, Sunset Lake had a dramatically smaller average than the rest of the lakes (9.97 x 10 4 m 3 /year). The number of times that the volume of pumped groundwater replaced the volume of water in the lakes ranged from 0.238 to 3.28 times/year, with Sunset Lake having the lowest rate and Clear Lake having the highest (Table 4). After investigating the water chemistry of the groundwater pumped from the wells at each lake and the historical data from several of the lakes prior to initial pumping (Table 5), numerous patterns were found. Loyce Lake was the only lake with historical Secchi depths prior to augmentation, and upon the addition of groundwater, the Secchi depth increased from 0.8 to 3.05 meters. In all of the lakes, the total alkalinity and total phosphorus concentrations in the well samples were higher than the current lake water samples, and in every case, the lakes increased in pH, alkalinity and total phosphorus since their historical measurements. Conversely, all but one well water sample, Saddleback lake, had lower total nitrogen concentrations than the current lake samples, coinciding with a decrease in nitrogen when compared to historical water samples (Table 5). The well water samples also had lower chloride concentrations than the current samples from the lakes; however, the two lakes with historical chloride data increased in chloride concentrations since the initiation of augmentation. Loyce, Saddleback, and Sunset lakes experienced a decrease in color, and Loyce Lake exhibited a decrease in chlorophyll when compared with from historical data, coinciding with groundwater pumping.

PAGE 25

16 In each of the augmented lakes, groundwater was pumped at a location about 100 meters from the lake, and either formed a small stream or was run down a pipe where the water was released in an upward fashion, like a fountain. The mean oxygen and temperature levels measured at mid-lake stations were nearly identical to those measured at the end of these streams and pipes, where groundwater is introduced into the lakes. Fish Population Comparisons Catch per unit effort (CPUE) of all fish varied among the augmented lakes, with Goose Lake having the lowest (1.16 fish/minute) and Sunset Lake having the highest (10.7 fish/minute) (Table 6). Goose Lake also had the lowest species richness (5 species), and Clear Lake had the highest (11 species). However, Goose Lake had the highest index of evenness (E) by number of fish per species (0.70), and Sunset Lake had the lowest (0.21). Similarly, Goose Lake had the highest index of evenness by weight of fish per species (0.61), and Mountain Lake had the lowest (0.32). Species diversity (H) by number ranged from 1.37 to 2.41 at Sunset Lake and Clear Lake, respectively, and species diversity by weight ranged from 1.67 to 2.38 at lakes Dan and Sunset, respectively. Biomass of harvestable length fish ranged from 15.1 grams per minute at Clear Lake to 173 grams per minute at Mountain Lake. The averages of mean CPUE and species richness of the nonaugmented lakes exceeded values at six of the seven augmented lakes (Table 6). The ranges and averages of the evenness and diversity variables of both the augmented and nonaugmented lakes were similar. However, the average of the mean harvestable fish biomass for the nonaugmented lakes exceeded the values of all seven augmented lakes (Table 6).

PAGE 26

17 Multiple Regression Analysis The multiple linear regressions with stepwise model selection for the augmented and nonaugmented lakes combined were all significant (P < 0.05) (Table 7). Limnological variables explained 13% to 63% of the variability in fish population variables in the nonaugmented lakes. Secchi depth was negatively related to diversity by number, and positively related to logarithm evenness (E) by number. Species diversity (H) by weight was positively related to color and surface area. Log CPUE was positively related to chlorides and negatively related to Secchi depth and PAC. The log (E) by weight was positively related to total nitrogen as well as Secchi depth. Species richness was negatively related to total nitrogen and Secchi depth and positively related to chlorides and lake surface area. Log (harvestable biomass) was positively related to lake surface area and negatively related to Secchi and PAC. Diversity by weight was the only fish variable that was not related to Secchi depth. The simple linear regressions of Secchi depth versus the logarithm transformation of evenness by number (Figure 2a) and Secchi depth versus diversity by number (Figure 2b) show that the data points for the augmented lakes are not outside of the 95% confidence intervals. However, the Secchi depth values of a majority of the augmented lakes are higher than the majority of the nonaugmented lakes. The remaining four fish population parameters were all determined by multiple habitat variables. Therefore, I examined the influence diagnostics to determine the influence of augmented lakes on the multiple regressions. In each case, the absolute values of the studentized residuals were minimal, the COVRATIO values were close to one, and the absolute values of the DFFITS values were minimal, indicating that the

PAGE 27

18 augmented lakes had little influence on the multiple regressions, similar to the results of the simple linear regressions. Canonical Correspondence Analysis Canonical correspondence analysis (CCA) was used to explore the relationship between the abundance, as determined by catch per unit effort, of fish species, the lakes where they were found, and the tested limnological variables. The first three axes in the CCA, which are linear combinations of limnological variables that maximize the correlation between site and species scores, explained 26% of the variation (Table 8). The first canonical axis, which explained 12% of the variation, was positively correlated (>0.350) with chlorophyll, total phosphorus, total nitrogen, total alkalinity, color, surface area and chlorides, and negatively correlated (<-0.350) with Secchi depth and percent area covered by aquatic macrophytes (Table 9). The second canonical axis, which explained 7.7% of the variation, was positively correlated (>0.350) with chlorides and total phosphorus, and negatively correlated (<-0.350) with total alkalinity and Secchi depth. The third canonical axis, which explained 5.6% of the variation, was negatively correlated (<-0.350) with chlorides and color. By definition, the majority of the variation was explained by the first axis. However, all of the augmented lakes, except for Sunset, were most highly correlated with axis two, in a negative fashion. There were only four other nonaugmented lakes that were most highly correlated with axis two in a negative fashion, demonstrating the similarities of the augmented lakes to each other in relation to environmental gradients. The six augmented lakes correlated with axis two were more characteristic of lower chlorides and total phosphorus, and higher Secchi depths and total alkalinity. In contrast,

PAGE 28

19 Sunset Lake, more negatively correlated to axis three, was characterized by higher chlorides and color (Table 9). The joint plot of axis one versus axis two displayed the general pattern of the nonaugmented lakes along axis one (Figure 3). The six augmented lakes Clear, Dan, Goose, Loyce, Mountain and Saddle, were more negatively correlated to axis two than they were correlated to axis one, and were all located in the same general location of the joint plot (Figure 3). Similarly, in the joint plot of axis two versus axis three (Figure 4), these same augmented lakes were more highly correlated with axis two in a negative fashion than they were with axis three, and again were all located in the same general area of the joint plot, whereas Sunset Lake was more related to axis three, with only a slight negative correlation with axis two. In joint plots, the lake points are found at the centroid of the species points that occur at that lake, allowing inferences to which species are likely to be present at a particular lake (Ter Braak 1986). Also, the species points are approximately the optima of where they are found in highest abundance, hence the abundance or probability of occurrence of a species decreases with distance from its location in the diagram. For example, the lined topminnow, Fundules lineolatus was highly negatively related to axis one, as were two lakes, Turkey Pond and Keys Pond, that the topminnow was found in close proximity to in the joint plot (Figure 3). This species was found in only six of the 43 lakes, and in highest abundance in these two lakes, explaining its closeness to these two lakes (Figure 3). The lined topminnows low abundance or absence in the other lakes demonstrates its distance from those lakes. Bluegill, Lepomis macrochirus was found in all but two lakes, demonstrating this species ability to survive across varying

PAGE 29

20 environmental gradients. Accordingly, bluegill were located near the intersection of all three axes, where there is little correlation to high or low values of environmental variables (Figure 3 and 4). Consequently, those lakes within close proximity to the intersection of the axes, with little correlation to any of the three axes, have a higher probability of having higher abundances of bluegill and other fish species that are found near the intersection than those on the perimeter. Six of the seven augmented lakes were found on the perimeter of the joint plots. Accordingly, all but one species of fish was at relatively large distances in the diagram from the six augmented lakes that were more correlated to axis two (Figure 3 and 4). The one fish species, taillight shiner, Notropis maculatus which was found in close proximity to these six lakes, was only found in one of the augmented lakes, Clear Lake. This fish represented the second highest abundant fish species in Clear Lake, and was only found in four other lakes, at low abundances, explaining the close proximity to Clear Lake. Therefore, the lack of other species in close proximity to these six augmented lakes demonstrates that there is a higher probability that abundance of individual fish species in these six augmented lakes is lower than other lakes more close in proximity to species points. Sunset Lake is closer in proximity to several fish species points indicating a higher probability of a higher abundance of individual fish species in this lake as compared to the other augmented lakes. Cluster Analysis The intraset correlations of axis two of the CCA demonstrate that six of the augmented lakes were correlated with higher total alkalinity and Secchi depth, and lower chlorides and total phosphorus. These four environmental variables were used to create a cluster diagram (Figure 5). The cluster diagram displayed five of the augmented lakes

PAGE 30

21 Clear, Goose, Loyce, Mountain and Dan, in a small cluster. Saddleback Lake, the other augmented lake correlated with axis two of the CCA, is grouped in the next closest small cluster. This further demonstrates the similar limnological characteristics of the augmented lakes with large volumes of groundwater introduction. Sunset Lake is not closely clustered with any of the other augmented lakes, but rather with two other nonaugmented lakes that, like Sunset Lake, were also most correlated with axis three in a negative fashion.

PAGE 31

22 # S# S# S# S# S# S# S PolkHillsboroughPasco Study Sites Loyce Lake Clear Lake Goose Lake N Dan Lake Sunset Lake Saddleback Lake Mountain Lake 40 Kilometers Figure 1. Augmented lakes sampled in three Florida counties.

PAGE 32

23 -1-0.8-0.6-0.4-0.200.20123456Secchi Depth (meters)log10 Evenness by Number 0.4 (a) 00.511.522.533.50123456Secchi Depth (meters)Diversity by Number 4(b) Figure 2. Simple linear regressions of (a) Log 10 transformation of species evenness by number of fish versus Secchi depth (r 2 =0.298, df=42, p-value<0.001), and (b) diversity by number of fish versus Secchi depth (r 2 =0.134, df=42, p-value=0.016), with regression lines () and 95% confidence intervals for an observation around the lines (-) for nonaugmented () and augmented lakes ().

PAGE 33

24 Keys-pon Okahumpk Wauberg Miona Wales Baldwin Susannah Cue Crooked Round-po Hollings Hunter Hartridg Killarny Holden Bell Bonny Live-oak Koon Moore Watertow Patrick Tomahawk Carlton Rowell Turkey-p Fish Bull-Pon Mill-Dam West-Moo 1Mountai Douglas Pasadena Sanitary Little-F Picnic Clear Dan Goose Loyce Mountain Saddle Sunset talk cl tp tn SA Secchi CHLA color PAC Axis 1Axis 2 Bluegill Lined Topminnow Taillight Shiner Figure 3. Joint Plot of axis 1 versus axis 2 of the canonical correspondence analysis with lakes () and species (+) plotted along environmental gradients. Limnological characteristics include: percent area covered by aquatic macrophytes (PAC), Secchi depth (Secchi), total alkalinity (talk), color, total nitrogen (tn), chlorophyll (CHL), total phosphorus (tp), Surface Area (SA) and chloride (cl). Augmented lakes include: Clear, Dan, Goose, Loyce, Mountain, Saddleback (Saddle) and Sunset. The oval contains six augmented lakes that are most related to axis 2. The box contains Sunset Lake.

PAGE 34

25 Keys-pon Okahumpk Wauberg Miona Wales Baldwin Susannah Cue Crooked Round-po Hollings Hunter Hartridg Killarny Holden Bell Bonny Moore Live-oak Koon Watertow Patrick Tomahawk Carlton Rowell Turkey-p Fish Bull-Pon Mill-Dam West-Moo 1Mountai Douglas Pasadena Sanitary Little-F Picnic Clear Dan Goose Loyce Mountain Saddle Sunset talk cl tp tn SA Secchi CHLA color PAC Axis 2Axis 3ill Taillight Shiner Blueg Lined Topminnow Figure 4. Joint Plot of axis 2 versus axis 3 of the canonical correspondence analysis with lakes () and species (+) plotted along environmental gradients. Limnological variables include: percent area covered by aquatic macrophytes (PAC), Secchi depth (Secchi), total alkalinity (talk), color, total nitrogen (tn), chlorophyll (CHL), total phosphorus (tp), surface area (SA) and chloride (cl). Augmented lakes include: Clear, Dan, Goose, Loyce, Mountain, Saddleback (Saddle) and Sunset. The oval contains six augmented lakes that are most related to axis 2. The box contains Sunset Lake.

PAGE 35

Distance (Objective Function)Information Remaining (%) 6.7E-03100 7.6E+0075 1.5E+0150 2.3E+0125 3E+010 Keys-ponRound-poTurkey-pCueCrookedTomahawkMooreKoonMill-DamBull-PonPicnicOkahumpkSanitarySunsetHartridgWest-MooPatrickDouglasRowellFish1MountaiMionaBellPasadenaLive-oakWalesSusannahLittle-FBaldwinKillarnyWatertowSaddleClearGooseLoyceMountainDanWaubergHollingsHunterHoldenBonnyCarlton 26 Figure 5. Cluster analysis of lakes using total alkalinity, chloride, total phosphorus, and Secchi depth. Augmented lakes (boxed) include Clear, Dan, Goose, Loyce, Mountain, Saddleback (Saddle) and Sunset.

PAGE 36

27 Table 1. The county, wellfield in closest proximity, location, surface area, average depth determined with fathometer and year of first groundwater augmentation for the seven study lakes. Mountain Lake is not within the vicinity of a wellfield, but rather, requires augmentation due to its proximity to the highest elevation in peninsular Florida, on the Lake Wales Ridge, increasing its elevation above the surficial aquifer. Lake County Wellfield Latitude Longitude Surface Area Average Depth Year of First Augmentation (ha) (m) Clear Pasco Eldridge-Wilde 28.3625N 82.4789W 16 1.50 1978 Dan Hillsborough Cross Bar Ranch 28.1667N 82.6464W 14 3.18 Early 1970's Goose Pasco Eldridge-Wilde 28.3559N 82.4702W 15 1.49 1954 Loyce Pasco Eldridge-Wilde 28.3758N 82.4958W 18 1.73 1996 Mountain Polk Elevation 27.9348N 81.5898W 39 2.92 1975 Saddleback Hillsborough Section 21 28.1194N 82.4942W 13 2.56 1968 Sunset Hillsborough Cross Bar Ranch 28.1345N 82.6267W 15 2.83 1976

PAGE 37

28 Table 2. The number of wells, the average volume of water pumped each day from all wells combined, and the year of initial service for the wellfields in close proximity to the augmented lakes. Wellfield Number of Wells Average Volume Pumped Initial Year (m 3 /day) Cross Bar Ranch 17 1.00 x 10 5 1981 Eldridge-Wilde 58 8.52 x 10 4 1956 Section 21 8 3.79 x 10 4 1963

PAGE 38

29 Table 3. Mean limnological characteristics in 2003. These include percent lake area covered by aquatic macrophytes (PAC), Secchi depth, ph, total alkalinity, total chlorophyll, total phosphorus, total nitrogen, chloride, and color. Water chemistry means were based on three samples per lake, and PAC was measured using transects with a recording fathometer. Nonaugmented lake means are from 36 Florida lakes. Lake PAC Secchi pH Total Alkalinity Total Chlorophyll Total Phosphorus Total Nitrogen Chloride Color (%) (m) (mg/L as CaCO 3 ) (g/L) (g/L) (g/L) (mg/L) (Pt-Co units) Augmented Clear 58 2.44 7.80 115 5.60 13.3 427 6.00 23.3 Dan 14 1.52 8.00 120 15.6 16.7 890 9.92 53.0 Goose 52 3.35 7.90 107 3.20 12.3 670 6.92 31.7 Loyce 28 3.05 7.80 103 1.60 8.00 463 7.17 16.0 Mountain 44 3.35 9.03 64.7 8.40 14.3 487 9.67 15.0 Saddleback 12 1.52 8.27 76.3 34.4 19.3 737 7.67 42.0 Sunset 10 1.83 7.77 56.0 20.2 18.0 810 27.1 44.0 Mean 31.1 2.44 8.08 91.7 12.7 14.6 641 10.6 32.1 Nonaugmented Mean 44.2 2.04 7.56 31.1 26.5 30.5 939 18.5 20.6 Range 0-100 0.30-5.80 4.30-9.18 0.28-106 0.82-159 0.83-159 99.0-1750 2.50-51.7 1.25-57.5

PAGE 39

30 Table 4. Groundwater pumping history in augmented lakes. Variables include lake volume, the average volume of groundwater pumped per year, the years of historical pumping data averaged, and the number of times the volume of pumped water would replace the volume of water in the lake in one year. Lake Name Lake Volume Average Volume Pumped Years Fill Rate (m 3 ) (m 3 /year) (times/year) Clear 1.20 x 10 5 3.94 x 10 5 1990-98 3.28 Dan 4.77 x 10 5 7.61 x 10 5 1994-98 1.60 Goose 2.23 x 10 5 3.19 x 10 5 1990-98 1.43 Loyce 3.13 x 10 5 3.50 x 10 5 1995-98 1.12 Mountain 1.13 x 10 6 2.67 x 10 6 1989-94 2.36 Saddleback 3.33 x 10 5 4.43 x 10 5 1968-71 1.33 Sunset 4.19 x 10 5 9.97 x 10 4 1977-01 0.24

PAGE 40

31 Table 5. Water chemistry of groundwater from well samples and historical lake water samples prior to initial augmentation. Variables include date of collection (month/year), Secchi depth, pH, total alkalinity, total chlorophyll, total phosphorus, total nitrogen, chloride, and color. Secchi depth, total chlorophyll and color were not determined for groundwater. Blanks for historical lake samples indicate missing data. Lake Name Date Secchi pH Total Alkalinity Total Chlorophyll Total Phosphorus Total Nitrogen Chloride Color (month/year) (m) (mg/L asCaCO 3 ) (g/L) (g/L) (g/L) (mg/L) (Pt-Co units) Well Samples Clear 8/2003 7.7 177 25.0 310 5.95 Dan 8/2003 7.8 210 77.0 440 7.50 Goose 8/2003 7.6 184 40.0 330 5.00 Loyce 8/2003 7.6 178 28.0 420 5.50 Mountain 8/2003 7.7 152 60.0 450 3.00 Saddleback 8/2003 7.2 232 40.0 910 4.50 Sunset 8/2003 7.8 189 68.0 660 8.60 Mean 7.6 189 48.3 503 5.73 Historical Lake Samples Loyce 7/1995 0.80 6.9 36.0 28.0 5.00 2670 5.00 50.0 Loyce 7/1984 6.1 6.00 3.00 78.0 Sunset 5/1976 5.6 20.0 16.5 90.0 Saddleback 3/1968 5.5 14.0 40.0 Saddleback 2/1968 6.1 14.0 60.0

PAGE 41

32 Table 6. Fish population measures of augmented and nonaugmented lakes. These include number of electrofishing transects (N), mean catch per unit effort (CPUE), standard deviation of the catch per unit effort (CPUE SD), species richness, species evenness by number, species evenness by weight, species diversity by number, species diversity by weight and mean harvestable fish biomass. Nonaugmented lake averages are from 36 Florida lakes. Lake N Mean CPUE Richness Evenness Evenness Diversity Diversity Mean CPUE SD by Number by Weight by Number by Weight Harvestable Fish Biomass (transects) (fish/ minute) (species) (grams/ minute) Augmented Clear 6 7.05 5.96 11 0.37 0.40 2.41 2.30 74.7 Dan 6 1.65 0.68 7 0.54 0.33 2.26 1.67 15.1 Goose 7 1.16 0.89 5 0.70 0.66 1.95 1.86 51.9 Loyce 7 1.31 0.51 8 0.57 0.44 2.35 2.13 32.6 Mountain 8 3.38 0.92 9 0.38 0.32 2.18 1.77 173 Saddleback 7 1.71 0.45 10 0.36 0.35 2.38 2.12 38.4 Sunset 6 10.7 3.76 10 0.21 0.47 1.37 2.38 26.5 Mean 6.71 3.85 8.57 0.45 0.42 2.13 2.03 58.9 Nonaugmented Mean 5.26 7.45 10.4 0.44 0.37 2.22 1.98 208 Range 3-6 0.58-36.1 2-18 0.28-0.96 0.17-0.66 0.28-3.15 0.31-2.82 3.87-1150

PAGE 42

33 Table 7. Significant linear regression models predicting dependent fish variables at nonaugmented and augmented lakes combined. Dependent fish variables include log 10 of catch per unit effort (lcpue), log 10 species evenness by number (levennum) and weight (levenwt), species diversity by number (divnum) and weight (divwt), species richness (rich), and log 10 harvestable fish biomass. Independent variables include surface area of lake (ha, sa), percent area coverage of aquatic macrophytes (%, pac), Secchi depth (m, Secchi), total alkalinity (mg/L, talk), total chlorophyll (g/L, chl), total phosphorus (g/L, tp), total nitrogen (g/L, tn), chloride (mg/L, cl), and color (color). Model R-square df P-value lcpue = 0.724 + 0.016(cl) 0.134(Secchi) 0.004(PAC) 0.509 42 <0.001 levennum = -0.561 + 0.072(Secchi) 0.298 42 <0.001 levenwt = -0.669 + 0.00008(tn) + 0.069(Secchi) 0.246 42 0.004 divnum = 2.448 0.017(Secchi) 0.134 42 0.016 divwt = 1.215 + 0.004(sa) +0.017(color) 0.495 42 <0.001 rich = 10.803 + 0.113(cl) 0.002(tn) + 0.019(sa) 1.407(Secchi) 0.626 42 <0.001 lharvest = 2.302 + 0.002(SA) 0.133(Secchi) 0.005(PAC) 0.402 42 <0.001

PAGE 43

34 Table 8. Results of canonical correspondence analysis (CCA) for 34 fish species abundances, as measured by catch per unit effort, from 43 lakes in Florida. Proportional limnological variables were arcsine(x/100) transformed; all other limnological variables were log 10 (x+1) transformed. Species environmental correlations were conducted using Pearson tests. Statistic Axis 1 Axis 2 Axis 3 Eigenvalue 0.335 0.214 0.153 0.875 0.866 0.822 Species-Environmental Correlations 12.2 7.7 5.6 % Variance in species data explained by the axis 12.2 19.9 25.5 Cumulative % of variance in species explained Table 9. Intraset correlation between the limnological variables examined and the three axes in the canonical correspondence analysis (CCA) using 34 fish species abundances, as measured by catch per unit effort, from 43 lakes in Florida. Percent area covered by aquatic macrophytes (PAC) was arcsine(x/100) transformed. All other environmental parameters were log 10 (x+1) transformed. Intraset correlation may help indicate which environmental variables structure the community, as well as help determine which environmental variables are most influential in a site. The higher the absolute value of the intraset correlation, the more the parameter explains the variation. Values below -0.350 and above 0.350 were considered more highly correlated than those between -0.350 and above 0.350, and are followed by a for emphasis. Variable Axis 1 Axis 2 Axis 3 Total Alkalinity 0.643* -0.450* 0.185 Chloride 0.400* 0.526* -0.574* Total Phosphorus 0.793* 0.420* 0.317 Total Nitrogen 0.649* 0.064 0.211 Surface Area of Lake 0.479* 0.262 -0.124 Secchi Depth -0.753* -0.399* -0.296 Chlorophyll 0.888* 0.299 0.239 Color 0.629* -0.027 -0.506* PAC -0.553* 0.117 -0.279

PAGE 44

DISCUSSION Historically, the water chemistry of lakes in central Florida differs significantly from the chemistry of aquifer water in the same vicinity (Dooris and Martin 1979; BRA 1996; Hassell et al. 1997). The historical data from the augmented lakes prior to initial groundwater introduction demonstrated lower levels of pH, alkalinity, and water clarity prior to augmentation, similar to typical small lakes in central Florida that contain acidic, soft, tannin-colored water, with low levels of bicarbonate, inorganic carbon, alkalinity, conductivity, and calcium (Canfield 1981; Canfield and Hoyer 1988; BRA 1996; Hassell et al. 1997). The water chemistry of the augmented lakes in this study shifted to levels resembling the water chemistry of aquifer water upon the introduction of large volumes of groundwater, by which an entire lakes volume can be replaced several times a year. As groundwater is introduced, the replacement of original lake water is generally characterized by increases in clarity, pH, hardness, bicarbonate, inorganic carbon, alkalinity, conductivity, calcium, magnesium, dissolved solids, nitrogen and sodium concentrations, all of which are chemical characteristics of water contained in the Floridan Aquifer (Stewart and Hughes 1974; Martin et al. 1976a; BRA 1982; Dooris et al. 1982; BRA 1996; Hassell et al. 1997). Dooris and Martin (1979) noted that increased pumping of aquifer water caused a shift in the water chemistry of augmented lakes to closely resemble the chemical characteristics of aquifer water. It is apparent that the 35

PAGE 45

36 patterns of increasing pH, alkalinity and clarity in my study lakes are analogous to those of previously studied augmented lakes. Another noted effect of groundwater pumping, as reported by Canfield and Hoyer (1990) in Gate Lake and Mountain Lake, Florida, is the addition of nutrients. The well water samples in the seven augmented lakes had higher mean concentrations of phosphorus than both historical and current lake water sample levels. Conversely, in all sampled augmented lakes but Saddleback Lake, nitrogen concentrations were lower than sampled well water and historical values. However, the nitrogen to phosphorus ratios are much greater than 17, suggesting that phosphorus is the limiting nutrient in each of the seven augmented lakes (Florida Lakewatch 2000). The addition of groundwater demonstrates an increase in the limiting nutrient in the augmented lakes. Despite the overall increase in phosphorus by groundwater introduction, these augmented lakes are still characterized by low total phosphorus (< 20 g/L). Trophic states, according to concentrations of phosphorus, indicate that all of the studied augmented lakes were either oligotrophic or mesotrophic (Forsberg and Ryding 1980). Therefore, there is little expected change in fish population parameters due to increased nutrient introduction in the studied augmented lakes. The cluster analysis (Figure 6) of total alkalinity, chlorides, total phosphorus and Secchi depth versus all 43 lakes demonstrated the similarity in water chemistry of six of the augmented lakes to each other, and of one of the augmented lakes to a group of nonaugmented lakes. Despite being in three different counties, six of the augmented lakes were similarly characterized by higher alkalinity, lower chlorides, lower phosphorus and higher Secchi depths, whereas Sunset Lake was more characterized by

PAGE 46

37 lower total alkalinity and higher chloride concentrations. This was likely a result of the lower average volume of groundwater pumped into Sunset Lake each year (9.97 x 10 4 m 3 /year) compared to the range of yearly averages of the other augmented lakes (3.19 x 10 5 to 2.67 x 10 6 m 3 /year). Therefore, reduced groundwater introduction could decrease the effects of shifted water chemistry, resulting in a lake more characteristic of natural limnological characteristics. Another change that augmented lakes endure is reduced water level fluctuation. For example, Mountain Lake experienced lake level fluctuations of approximately 3 m during the 1940s and 1950s, whereas the recent stage fluctuation indicates a much narrower overall range of variation of about 1 m during the past decade (BRA 2001). Similarly, Saddleback Lake experienced less than 0.5 m of fluctuation in the ten years following augmentation, and showed very little response to heavy rainfall, as caused by the artificial head placed on the lake above the already lowered potentiometric head (Jones 1978). The other augmented lakes were also characterized by comparable lake level fluctuation reduction. The combination of increased water clarity, increased nutrients, increased hardness, and reduced water level fluctuation could change the characteristics of aquatic plant communities in augmented lakes. Increased water clarity increases light penetration, often allowing plants to grow faster and at greater depths (Canfield et al. 1985). Likewise, increased nutrients also increase plant growth. Also, Martin et al. (1976b) found that the elevated hardness of pumped groundwater increased the ability of augmented lakes to support hydrilla, Hydrilla verticillata growth. Consequently, many of the augmented lakes have had a history of aquatic plant problems. For example,

PAGE 47

38 Mountain and Saddleback lakes have been stocked with grass carp, Ctenopharyngodon idella treated with aquatic herbicides and had harvest programs to decrease the amount of plants (Canfield and Hoyer 1990; personal communication with lake residents). Conversely, Goose Lake and Clear Lake are in the middle of a wellfield, and do not have public access or people living on them. Therefore, there is no concern for aquatic vegetation control. These two lakes yielded the highest PAC of the augmented lakes. Little work has assessed effects of groundwater augmentation, and the ensuing alterations to lake characteristics, on fish communities (Allen 1999). Cowx (2000) found that the discharge of groundwater lowered dissolved oxygen concentrations and reduced water temperatures in the River Ouse, Yorkshire, UK, and suggested that the low dissolved oxygen concentrations could cause asphyxiation of fish with possible loss of sensitive species if chronic, and the low water temperatures could reduce fish growth, leading to a decline in stock (Cowx 2000). In Florida, groundwater generally has lower temperatures and dissolved oxygen concentrations than surface water (McKinset and Chapman 1998). However, well introduced groundwater is not pumped directly into the lakes in Florida as is the case in the River Ouse. The streams and the pipes that deliver the groundwater to the lakes in Florida allow the water to warm up and aerate before entering the water body, as demonstrated by the similarity of mid-lake and discharge measurements. Further, the difference in temperature between groundwater and lake water in Florida is less than that of the study in Yorkshire, decreasing this concern in Florida lakes. The effects of the nutrient introduction by means of groundwater pumping in the previously mentioned Canfield and Hoyer (1990) study were not detrimental to the fish

PAGE 48

39 populations, and instead, were possibly beneficial with respect to species diversity, total fish biomass, and sport-fish-abundance. Allen (1999) found similar results with respect to increased fish species diversity in one augmented lake in Florida (Round Lake). However, in contrast to Canfield and Hoyer (1990), Allen (1999) found that Round Lake had significantly lower total fish biomass and density as compared to two nonaugmented lakes. The augmented lakes in my study had little influence on the regressions of fish population parameters versus limnological variables, suggesting that the fish populations in the augmented lakes were similar to fish populations in nonaugmented lakes with similar limnological characteristics. However, there is evidence that the limnological variables shifted from their original levels to those more indicative of aquifer water. Therefore, one must consider that as the limnological variables shifted with groundwater introduction, the fish populations in the augmented lakes responded by shifting correspondingly along the gradient of the regression. The data from the nonaugmented lakes in this study were collected from 1986 to 1990. Therefore, the limnological and fish population parameters could have changed in these lakes over time in a similar fashion to the augmented lakes, making it difficult to compare the two samples. However, upon inspection of numerous limnological variables, including pH, total phosphorus, total nitrogen and total alkalinity, from several of the nonaugmented lakes, the changes were either small or obsolete. For example, both Lake Susannah had the same pH (7.8) in 1988 and 1999, and similar alkalinities over the same time period (30.8 and 30.7 respectively). Also, the magnitudes of change from the few limnological parameters measured historically from the augmented lakes were much

PAGE 49

40 greater than the magnitudes of change for the nonaugmented lakes. Although extensive limnological parameter studies were not performed on the augmented lakes prior to groundwater introduction, several patterns emerged from the limited data available. As groundwater was pumped, the augmented lakes were characterized by increased Secchi depths, total alkalinity, phosphorus and chloride. They were also characterized by decreases in total nitrogen and color, and would most likely all increase in percent area covered by aquatic macrophytes if people did not control their levels. Augmentation also increases surface area. Upon placing these characteristics in the multiple regressions for all lakes, several fish population responses can be predicted for augmented lakes. However, it is difficult to predict exact changes on a temporal scale due to the lack of previous limnological studies on these lakes. For example, there were no previous estimates of PAC on the augmented lakes, and only one lake had a historical secchi depth measurement. Therefore, the following are projected patterns for fish population parameters based on general observed patterns for the limnological parameters. The pattern of increase or decrease for catch per unit effort over time with groundwater introduction is difficult to determine since chloride concentrations, Secchi levels and PAC increase with groundwater pumping, acting as opposing terms in the regression. However, the average catch per unit effort for the augmented lakes (3.85 fish/minute) was much lower than average catch per unit effort of the nonaugmented lakes (7.45 fish/minute), suggesting that catch per unit effort decreased. Evenness by number would increase because Secchi depth increases with groundwater pumping. Evenness by weight would have opposing terms since total nitrogen decreases and Secchi increases, making it difficult to determine

PAGE 50

41 the pattern. Diversity by number would decrease with groundwater pumping due to increasing Secchi depth, opposing the results reported by Canfield and Hoyer (1990) and Allen (1999). Conversely, diversity by weight would increase due to increasing surface area and color. Both richness and harvestable biomass have opposing values as groundwater increases, making their change with groundwater augmentation difficult to determine. However, similar to catch per unit effort, mean species richness and mean harvestable fish biomass were lower for the seven augmented lakes (8.57 species and 58.9 g/minute respectively) than for the 36 nonaugmented lakes (10.4 species and 208 g/minute respectively), indicating a lower average number of species, individuals and weight of fish of harvestable size than those lakes without groundwater being pumped. The multiple regression analyses were useful for indicating that fish populations of augmented lakes did not deviate from the patterns of nonaugmented lakes. However, multiple regressions are unable to indicate the patterns for numerous individual fish species across multiple limnological variables in multiple lakes. The joint plots of the CCA suggested that the abundance, expressed as catch per unit effort, of individual species in six of the augmented lakes had a high probability of being low compared to a majority of nonaugmented lakes, agreeing with the Round Lake study by Allen (1999). This also corresponded with the finding that catch per unit effort of all species and species richness were low in these lakes compared to nonaugmented lakes. Further, a majority of all fish species abundances were more correlated to the first axis of the CCA, whereas the majority of augmented lakes were highly correlated to the second axis, explaining that the gradient of environmental patterns determining the fish community in

PAGE 51

42 these augmented lakes was different than the gradient determining fish communities in a majority of nonaugmented lakes. A study in Max Lake, Wisconsin, examined the effects of groundwater pumping on fish population dynamics for largemouth bass and yellow perch populations (Engel et al. 2000). The groundwater pumping failed to alter growth, abundance, biomass, or mortality of yellow perch and largemouth bass 3 to 7 years old. However, very little water was pumped into the lake as compared to the amount of water pumped into the studied augmented lakes in Florida. Only 5% of the water in Max Lake was replaced each year by augmentation, as compared to the range of about 24% to 328% in the seven augmented lakes. However, similar to the comparison of Sunset Lake to the other studied augmented lakes, low volumes of introduced groundwater could have reduced effects on water chemistry and fish population responses compared to lakes with high volumes of groundwater introduction. Numerous studies discuss fish growth and biomass responses to alterations in water chemistry. Many have focused on the relationship between trophic states (as described by Forsberg and Ryding 1980) and fish communities in lakes. Both fish growth (Larkin and Northcote 1969; Bayne et al. 1994) and fish production (Downing et al. 1990; Ney 1996) are closely correlated with total phosphorous concentrations in lakes. Similarly, total phosphorous concentrations (Kautz 1980; Hanson and Leggett 1982; Yurk and Ney 1989; Hoyer and Canfield 1991; Lee et al. 1991; Bayne et al. 1994; Bachmann et al. 1996), total nitrogen concentrations (Bachmann et al. 1996), and chlorophyll concentrations (McConnell et al. 1977; Ogelsby 1977; Jones and Hoyer 1982; Bachmann et al. 1996), are all positively related to total fish biomass in lakes. Conversely,

PAGE 52

43 oligotrophication, the reversal of the eutrophication process, is accompanied by declines in growth, standing stock, and harvest in fish (Ney 1996). Each of these studies displays a positive correlation between fish productivity, fish biomass, and fish abundance as eutrophication proceeds. The Forsberg and Ryding (1980) guidelines for trophic state indicate that the augmented lakes range from oligotrophic to mesotrophic, possibly explaining for the low CPUE and low harvestable biomass found in these lakes. However, the slight increase in nutrients caused by groundwater introduction may have slightly increased these fish population variables from their original levels. Several investigations have also reported decreases in the relative abundance of piscivorous fish, with possible losses of sensitive species, as a consequence of eutrophication (Larkin and Northcote 1969; Persson et al. 1988; Bachmann et al. 1996; Ney 1996). Opportunistic, eurytolerant, non-piscivorous species are likely to replace stenotolerant, piscivorous fishes with increasing fertility of lakes (Ney 1996), causing higher standing crops of such fish species as gizzard shad, Dorsoma cepadianum threadfin shad, Dorosoma petenense and common carp, Cyprinus carpio in eutrophic and hypereutrophic lakes (Hasler 1947; Larkin and Northcote 1969; Bachmann et al. 1996). No evidence was found in this group of augmented lakes to suggest that piscivorous fish were being replaced by eurytolerant, non-piscivorous fish. For example, bluegill was the most abundant fish species present in six of the seven augmented lakes. Also, threadfin shad were found in only one of the augmented lakes, Sunset Lake, where limnological variables are more closely related to nonaugmented lakes. Further, in Sunset Lake there were healthy populations of reproducing largemouth bass, bluegill,

PAGE 53

44 black crappie, warmouth and redear sunfish of multiple size ranges, while only one school of shad was encountered. Therefore, these lakes are far from being in the hypereutrophic range where these problems occur in fish populations, and there is no evidence that the increase in nutrients from groundwater introduction into the studied augmented lakes caused losses of species. Numerous studies have found that lake size is the dominant factor determining fish species richness in Florida lakes, in which the number of fish species increases with increasing lake surface area (Barbour and Brown 1974; Matuszek and Beggs 1988; Keller and Crisman 1990; Bachmann et al. 1996). In this study, species richness was positively correlated to the surface area of augmented and nonaugmented lakes combined in the multiple regression. Florida lakes are generally shallow, with flat slopes, and a small decrease or increase in lake level stage creates a dramatic alteration in surface area (Dooris 1982), possibly having a major effect on fish communities. For example, Mountain Lake is listed as having a surface area of 55 hectares prior to lake level declines (BRA 2001). Corresponding with the drop in water level was a decrease in the lake surface area, upon which augmentation returned the surface area to 39 hectares. This was a decline from the original surface area, but larger than the surface area without augmentation. Therefore, the lake could display a decrease in species richness from historical fish population data prior to water level declines, but an increase from periods immediately prior to augmentation. Allen (1999) suggested that fish population responses to augmentation may be variable and depend on the original water chemistry of the natural lake relative to the water chemistry of the pumped groundwater. Therefore, prior to initial groundwater

PAGE 54

45 introduction for augmentation of lake levels, a comparison of water chemistry characteristics from lake water and well water from within close proximity, and estimating the volume of pumped water that will be necessary to maintain lake levels, may be useful in determining expected shifts in water chemistry and fish populations. If surface area is also a dominant factor in determining fish population parameters, it may be important to not only monitor water chemistry, but also surface area prior to initial declines and initial pumping.

PAGE 55

MANAGEMENT IMPLICATIONS Hassel (1994) suggested that augmentation is not a good long-term solution to restore lake levels to reasonable levels because of altered environmental factors. He further states that lake augmentation is a short-term remedy for a long-term problem. It is apparent that Hassel is correct in stating that environmental factors have been altered in augmented lakes. However, without augmentation, many of the lakes would go dry, as was the case of Loyce Lake prior to groundwater introduction. Augmentation allows for lakes to be utilized for boating, swimming and other recreational activities. It also allows for lake and wetland hydrology to be maintained and fish and wildlife habitat to be provided. Further, fish, bird, reptilian, mammalian, insect and aquatic plant populations were all seen in the augmented lakes in this study. Without augmentation, it is likely fish would die, and the use of the lakes for recreational purposes would be compromised. The human population in the Tampa area is constantly increasing, along with the demand for water. As the population further expands from Tampa into the suburbs, more wellfields will be created, and more lakes will be affected. Similarly, as the existing wellfields increase the amount of water they withdraw the cones of depression will increase, affecting more lakes in the future. Granting more permits for lake level augmentation with groundwater pumping will further alter limnological characteristics until the demand for groundwater is decreased. However, lakes will still be able to be utilized and fish populations will still be able to exist and reproduce, despite possible 46

PAGE 56

47 shifts with altered environmental patterns. It has also been suggested that these shifts were minimized with reduced levels of groundwater pumping. Another change that could improve fish population parameters, and reduce changes in plant community characteristics, is an increase in water level fluctuation (Bonvechio and Allen in press). A more natural water level regime could be created by augmenting during rainy seasons, and allowing the lakes to decrease in level during the dry season.

PAGE 57

APPENDIX COMMONLY HARVESTED FISH SPECIES Table A-1. Commonly harvested fish species, and the total length (mm) at which they are generally first harvested. Species Name Scientific Name Size (mm) Chain pickerel Esox niger 400 Yellow bullhead Amerius natalis 280 Brown bullhead Ameirus nebulosus 280 White catfish Ameirus catus 280 Channel catfish Ictalurus punctatus 280 Largemouth bass Micropterus salmoides 280 Sunshine bass Morone chrysops x M. saxatilis 280 Black crappie Pomoxis nigromaculatus 240 Redbreast sunfish Lepomis auritus 200 Warmouth Lepomis gulosus 200 Bluegill Lepomis macrochirus 200 Redear sunfish Lepomis microlophus 200 Flier Centrarchus macropterus 200 48

PAGE 58

LIST OF REFERENCES Allen, M. S. 1999. Assessment of fish assemblages in Lakes Dosson, Halfmoon and Round in Hillsborough County, Florida. Southwest Florida Water Management District, Brooksville. Allen, M. S., and K. I. Tugend. 2000. Effects of a large-scale habitat enhancement project on habitat quality for age-0 largemouth bass at Lake Kissimmee, Florida. Pages 265-276 in D. Phillipp and M. Ridgeway, editors. Black Bass: Ecology, Conservation and Management. American Fisheries Society, Bethesda, Maryland. American Public Health Association. 1989. Standard methods for the examination of water and waste water. 17 th edition. New York. Bachmann, R. W., B. L. Jones, D. D. Fox, M. Hoyer, L. A. Bull, and D. E. Canfield, Jr. 1996. Relations between trophic state indicators and fish in Florida (U.S.A.) Lakes. Canadian Journal of Fisheries and Aquatic Sciences 53:842-855. Barbour, C. D., and J. H. Brown. 1974. Fish species diversity in lakes. American Naturalist 108(962):473-489. Bartos, L. F. 1998. Environmental augmentation. The resource regulation newsletter. Southwest Florida Water Management District 10(3). Bayne, D. R., M. J. Maceina, and W. C. Reeves. 1994. Zooplankton, fish and sport fishing quality among four Alabama and Georgia reservoirs of varying trophic status. Lake and Reservoir Management 8:153-163. Belanger, T. V., and R. A. Kirkner. 1994. Groundwater/surface water interaction in a Florida augmentation lake. Lake and Reservoir Management 8:165-174. Biological Research Associates (BRA). 1982. Ecological impact of augmentation in three lakes in Hillsborough County, Florida. Pinellas County Water System, Tampa. Biological Research Associates (BRA). 1996. General limnological assessment of three augmented lakes in Northwest Hillsborough County, Florida. Pinellas County Water System, Clearwater. Biological Research Associates (BRA). 2001. Mountain Lake limnological evaluation and lake management alternatives. Mountain Lake Corporation, Lake Wales, Florida. 49

PAGE 59

50 Brenner, M., J. M. Smoak, M. S. Allen, C. L. Schelske, and D. A. Leeper. 2000. Biological accumulation of 226 Ra in a groundwater-augmented Florida lake. Limnology and Oceanography 45:710-715. Canfield, D. E., Jr. 1981. Guide to the physiographic divisions of Florida. Report to Florida Cooperative Extension Service. Institute of Food and Agricultural Sciences. University of Florida, Gainesville. Canfield, D. E., Jr., K. A. Langeland, S. B. Linda, and W. T. Haller. 1999. Relations between water transparency and maximum depth of macrophyte colonization in lakes. Journal of Aquatic Plant Management 23:25-28. Canfield, D. E., Jr., and M. V. Hoyer. 1988. Regional geology and the chemical andtrophic state characteristics of Florida Lakes. Lake and Reservoir Management 4(1):21-31. Canfield, D. E., Jr., and M. V. Hoyer. 1990. A characterization of fish populations in two central Florida Lakes. Final Report. Florida Turfgrass Association, Gainesville. Canfield, D. E., Jr., and M. V. Hoyer. 1992. Aquatic macrophytes and their relation to the limnology of Florida lakes. Final Report submitted to the bureau of Aquatic Plant Management, Florida Department of Natural Resources, Tallahassee. Cowx, I. G. 2000. Potential impact of groundwater augmentation of river flows on fisheries: a case study from the River Ouse, Yorkshire, UK. Fisheries Management and Ecology 7:85-96. Crumpton, W. G., T. M. Isenhart, and P. D. Mitchell. 1992. Nitrate and organic N analysis with second-derivative spectroscopy. Limnology and Oceanography 37:907-913. DElia, C. F., P. A. Steudler, and N. Corwin. 1977. Determination of total nitrogen in aqueous samples using persulfate digestion. Limnology and Oceanography 22:760-764. Dooris, P. M. 1982. Lake augmentation in northwest Hillsborough County. Technical Report of the Southwest Florida Water Management District, Brookesville. Dooris, P. M., and D. F. Martin. 1979. Groundwater induced changes in lake chemistry. Ground Water 17:324-327. Dooris, P. M., G. M. Dooris, and D. F. Martin. 1982. Phytoplankton responses to ground water addition in central Florida Lakes. Water Resources Bulletin 18:335-337.

PAGE 60

51 Dooris, P. M. and R. J. Moresi. 1975. Evaluation of lake augmentation practices in northwest Hillsborough County, Florida. Technical Report of the Southwest Florida Water Management District, Brookesville. Downing, J. A., C. Plante, and S. Lalonde. 1990. Fish reproduction correlated with primary productivity, not the morphodoedaphic index. Canadian Journal of Fisheries and Aquatic Sciences 47:1929-1936. Engel, S., M. H. Hoff, M. T. Vogelsang Jr., K. E. Bass, and J. S. Anderson. 2000. Fish population dynamics in Max Lake, a softwater Wisconsin lake subject to ground-water pumping. Wisconsin Department of Natural Resources Research Report, Woodruff. Forsberg, C., and S. O. Ryding. 1980. Eutrophication parameters and trophic state indices in 30 Swedish waste-receiving lakes. Archive fur Hydrobiologie 88:189-207. Frey, D. G. 1955. Distribution ecology of the cisco in Indiana. Investigation of Indiana Lakes-Streams 4:177. Gottgens, J. F. 1994. Redistribution of organic sediments in a shallow lake following a short-term drawdown. Hydrobiologia 130:179-194. Hach Chemical Company. 1975. Water and wastewater analysis procedures, third edition. Ames, Iowa. Hanson, J. M., and W. C. Leggett. 1982. Empirical prediction of fish biomass and yield. Canadian Journal of Fisheries and Aquatic Sciences 39: 257-263. Hasler, A. D. 1947. Eutrophication of lakes by domestic drainage. Ecology 28:383-395. Hassell, A. L. 1994. A chemical and biochemical characterization of lakes Cooper, Strawberry, Crystal, Hobbs, Starvation, and Saddleback in Hillsborough County (Florida). Masters thesis submitted to the Department of Chemistry at the University of South Florida, Tampa. Hassell, A. L., P. M. Dooris, and D. F. Martin. 1997. Maucha diagrams and chemical analyses to diagnose changes in lake chemistry. Environmental Chemistry 60:75-80. Hinch, S. G. and N. C. Collins. 1993. Relationships of littoral fish abundance to water chemistry and macrophyte variables in central Ontario lakes. Canadian Journal of Fisheries and Aquatic Sciences 50:1870-1878. Hoyer, M. V., and D. E. Canfield Jr. 1991. A phosphorus-fish standing crop relationship for streams? Lake and Reservoir Management 7(1):25-32.

PAGE 61

52 Jones, J. R., and M. V. Hoyer. 1982. Sportfish harvest predicted by summer chlorophyll a concentration in Midwestern lakes and reservoirs. Transactions of the American Fisheries Society 111:176-179. Jones, K. C. 1978. Lake augmentation alternatives in Northwest Hillsborough Basin. Memorandum. File No. 14-000-REG-31-00. Kautz, E. S. 1980. Effects of eutrophication on the fish communities of Florida lakes. Proceedings of Annual Conference of Southeastern Association of Fish and Wildlife Agencies 34:67-80. Kellar, A. E., and T. L. Crisman. 1990. Factors influencing fish assemblages and species richness in subtropical Florida lakes and a comparison with temperate lakes. Canadian Journal of Fisheries and Aquatic Sciences 47:2137-2146. Krebs, C. J. 1999. Ecological methodology, second edition. Benjamin Cummings, Menlo Park, California. Florida Lakewatch. 2000. A beginners guide to water management nutrients. Information circular #102. Department of Fisheries and Aquatic Sciences, University of Florida, Gainesville. Larkin, P. A., and T. G. Northcote. 1969. Fish as indices of eutrophication. Pages 256-273 in Eutrophication, causes, consequences, correctives. National Academy of Sciences, Washington D. C. Lee, G. F., P. E. Jones, and R. A. Jones. 1991. Effects of eutrophication on fisheries. Review of Aquatic Sciences 5:287-305. Maceina, M. J., and J. V. Shireman. 1980. The use of a recording fathometer for determination of distribution and biomass of Hydrilla. Journal of Aquatic Plant Management. 18:34-39. Martin, D. F., D. M. Victor, and P. M. Dooris. 1976a. Effects of artificially introduced ground water on the chemical and biochemical characteristics of six Hillsborough County (Florida) Lakes. Water Research 10:65-69. Martin, D. F., D. M. Victor, and P. M. Dooris. 1976b. Implications of lake augmentation on Hydrilla growth. Environmental Science Engineering A11:245-253. Matuszek, J. E., and G. L. Beggs. 1988. Fish species richness in relation to lake area, pH, and other abiotic factors on Ontario lakes. Canadian journal of Fisheries and Aquatic Sciences 45:1931-1941. McConnell, W. J., S. Lewis, and J. E. Olson. 1977. Gross photosynthesis as an estimator of potential fish production. Transactions of the American Fisheries Society 106:417-423.

PAGE 62

53 McKinsey, D. M., and L. J. Chapman. 1998. Dissolved oxygen and fish distribution in a Florida spring. Environmental Biology of Fishes 53:211-223. Menzel, D. W., and N. Corwin. 1965. The measurement of total phosphorus in seawater based on the liberation of organically bound fractions of persulfate oxidation. Limnology and Oceanography 10:280-282. Murphy, J., and J. P. Riley. 1962. A modified single solution method for the determination of phosphate in natural waters. Analytica Chimica Acta 27:31-36. Myers, R. H. 1990. Classical and modern regression with applications, second edition. PWS-Kent Publishing Company, Boston, Massachusetts. Ney, J. J. 1996. Oligotrophication and its discontents: effects of reduced nutrient loading on reservoir fisheries. Pages 285-295 in L. E. Miranda and D. R. Devries, editors. Multidimensional approaches to reservoir fisheries management. American Fisheries Society, Bathesda, Maryland. Ogelsby, R. F. 1977. Relationships of fish yield to lake phytoplankton standing crop, production, and morphoedaphic factors. Journal of Fisheries Research Board of Canada 34:2271-2279. Palmer, M. W. 1993. Putting things in even better order: the advantages of canonical correspondence analysis. Ecology 74(8):2215-2230. PC-ORD. 1999. Multivariate analysis of ecological data, version 4. MjM Software Design, Gleneden Beach, Oregon. Persson, L., Andersson, G., Hamrin, S. F., and Johansson, L. 1988. Predator regulation and primary production along the productivity gradient of temperate lake ecosystems. Pages 45-65 in S. R. Carpenter., editor. Complex interactions in freshwater ecosystems. Springer-Verlag, New York. Reynolds, J. B. 1996. Electrofishing. Pages 221-253 in B. R. Murphy and D. W. Willis, editors. Fisheries Techniques, second edition. American Fisheries Society. Bethesda, Maryland. Sartory, D. P., and J. U. Grobbelaar. 1984. Extraction of chlorophyll a from freshwater phytoplankton for spectrophotometric analysis. Hydrobiologia 114:177-187. Statistical Analysis Systems (SAS). 1996. SAS statistics users guide. SAS Institute, Inc., Cary, North Carolina. Shafer, M. D., R. E. Dickinson, J. P. Heaney and W. C. Huber. 1986. Gazetteer of Florida lakes. Florida Water Resources Research Center, Publication 96, Gainesville, Florida.

PAGE 63

54 Simal, J., M. A. Lage, and I. Iglesias. 1985. Second derivative ultraviolet spectroscopy and sulfamic acid method for determination of nitrates in water. Journal Association of Official Analytical Chemists 68:962-964. Sinclair, W. C. 1977. Experimental study of artificial recharge alternatives in Northwest Hillsborough County, Florida. United States Geological Survey: Water-Resources Investigations 77-13. Stewart, J. W. 1968. Hydrologic effects of pumping from the Floridan Aquifer in Northwest Hillsborough, Northeast Pinellas, and Southwest Pasco Counties, Florida. United States Geological Survey, Tallahassee. Stewart, J. W., and G. H. Hughes. 1974. Hydrologic consequences of using groundwater to maintain lake levels affected by water wells near Tampa, Florida. Florida Department of Natural Resources, Tallahassee. Southwest Florida Water Management District (SWFWMD). 1998. Water supply assessment 1995-2020. Resource Projects Department, Southwest Florida Water Management District, Brookesville. Ter Braak, C. J. F. 1986. Canonical correspondence analysis: a new eigenvector technique for multivariate direct gradient analysis. Ecology 67(5):1167-1179. Wollin, K. M. 1987. Nitrate determination in surface waters as an example of the application of UV derivative spectrometry to environmental analysis. Acta Hydrochemica Hydrobiologia 15:459-469. Yurk, J. J., and J. J. Ney. 1989. Phosphorus-fish community biomass relationships in southern Appalachian reservoirs: can lakes be too clean for fish? Lake and Reservoir Management 5(2):83-90. Zar, J. H. 1999. Biostatistical analysis, fourth edition. Prentice-Hall, Upper Saddle River, NJ.

PAGE 64

BIOGRAPHICAL SKETCH Patrick Cooney was born and raised in El Dorado Hills, a small and fast growing town in the Sierra-Nevada Mountain Range in northeastern California. Upon graduating from high school, he moved to Miami, Florida, to pursue a degree in both biology and marine science from the University of Miami. During this time, he enjoyed studying abroad and conducting research on marine life and marine ecosystems in Townsville, Australia, while attending James Cook University, near the Great Barrier Reef. After graduating from the University of Miami, he promptly moved to Bahia de Kino, Mexico (located on the Sea of Cortez), to study an obligate mutualism between the Senita moth and the Senita cactus, and fish for Dorado, Coryphaena hippurus Once this research was completed, he returned to Florida to spend time with friends and conduct research on freshwater fish and water chemistry at the University of Florida in Gainesville. He will continue to feed his hunger for knowledge, and teach others the knowledge he has attained. 55


Permanent Link: http://ufdc.ufl.edu/UFE0005260/00001

Material Information

Title: Effects of Introduced Groundwater on Water Chemistry and Fish Assemblages in Central Florida Lakes
Physical Description: Mixed Material
Copyright Date: 2008

Record Information

Source Institution: University of Florida
Holding Location: University of Florida
Rights Management: All rights reserved by the source institution and holding location.
System ID: UFE0005260:00001

Permanent Link: http://ufdc.ufl.edu/UFE0005260/00001

Material Information

Title: Effects of Introduced Groundwater on Water Chemistry and Fish Assemblages in Central Florida Lakes
Physical Description: Mixed Material
Copyright Date: 2008

Record Information

Source Institution: University of Florida
Holding Location: University of Florida
Rights Management: All rights reserved by the source institution and holding location.
System ID: UFE0005260:00001


This item has the following downloads:


Full Text












EFFECTS OF INTRODUCED GROUNDWATER ON WATER CHEMISTRY
AND FISH ASSEMBLAGES IN CENTRAL FLORIDA LAKES

















By

PATRICK COONEY


A THESIS PRESENTED TO THE GRADUATE SCHOOL
OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT
OF THE REQUIREMENTS FOR THE DEGREE OF
MASTER OF SCIENCE

UNIVERSITY OF FLORIDA


2004

































Copyright 2004

by

Patrick Cooney


































To my dad, Michael Leo Cooney, you are missed, and I will never stop loving and
thinking about you.















ACKNOWLEDGMENTS

I first thank Dr. Mike Allen for serving as my advisor and committee chair. I am

very grateful for his guidance throughout this project.

I also thank Doug Leeper at the Southwest Florida Water Management District

(SWFWMD) for his time and providing necessary access and information, and Dr. Daniel

Canfield Jr. and Dr. Tom Frazer for their advice while serving as members of my

graduate committee. Similarly, I thank Mark Hoyer of the Department of Fisheries and

Aquatic Sciences for his advice.

I especially express my gratitude to the following people for their tremendous

support in the field and lab: M. Bennett, T. Bonvechio, S. Cooney, K. Dockendorf, D.

Dutterer, J. Harris, K. Henry, G. Kaufman, S. Larson, C. Mwatela, E. Nagid, M. Rogers,

N. Trippel and G. Warren. I also thank all of the people in the LAKEWATCH laboratory

for their time and use of laboratory equipment.

Most importantly, I thank and dedicate this to the people closest to me. Sean's help

on my project meant a lot to me. As his younger brother, I have greatly appreciated his

guidance and encouragement throughout my life. The foundation that my mom and dad

built and the constant support they both provided along the way made me who I am

today. I never could have attained this goal without all of them and the rest of my family.

Finally, I thank Julie. She has been with me every step of the way for the duration of this

venture. Aside from the tremendous help she provided on this project, I most appreciated

when she was there at the end of the day to make me smile.

















TABLE OF CONTENTS

page

A C K N O W L E D G M E N T S ................................................................................................. iv

L IST O F TA B LE S ......... ............. ............ ......... ..... ........ ....... vi

L IST O F F IG U R E S .... ......................................................... .. .......... .............. vii

A B S T R A C T .......................................... .................................................. v iii

INTRODUCTION ............................... .................... ...............

M E T H O D S ............................................................................ .4

Stu dy Sites ......................................................................... 4
E lectrofishing ................................................................................... 7
F ish P population M measures ................................................................ ............... 8
Fish Population Analysis .................................. .......................... ... ......10

R E S U L T S ................................................................................14

Comparison of Limnological Variables............................... .... .........14
Groundwater Pumping History ............... .............. ..........15
Fish Population Comparisons ...... ...................................16
M multiple Regression Analysis..............................................................17
Canonical Correspondence Analysis ............. ................................................. 18
Cluster A analysis .................................................. 20

D IS C U S S IO N ......... ........................................................................................ 3 5

MANAGEMENT IMPLICATIONS ....................................................... 46

APPENDIX COMMONLY HARVESTED FISH SPECIES ........................................48

L IST O F R E F E R E N C E S ................................ .......................................................... 49

B IO G R A PH IC A L SK E T C H ....................................................................................... 55





v















LIST OF TABLES


Table pge

1 The county, wellfield in closest proximity, location, surface area, average depth
determined with fathometer and year of first groundwater augmentation for the
sev en stu dy lak es ................................................... ................ 2 7

2 The number of wells, the average volume of water pumped each day from all
wells combined, and the year of initial service for the wellfields in close
proximity to the augm ented lakes. ........................................ ....................... 28

3 M ean limnological characteristics in 2003 ............ ..............................................29

4 Groundwater pumping history in augmented lakes............................................30

5 Water chemistry of groundwater from well samples and historical lake water
samples prior to initial augmentation. ............................... .. ....................... 31

6 Fish population measures of augmented and nonaugmented lakes........................32

7 Significant linear regression models predicting dependent fish variables at
nonaugmented and augmented lakes combined. ............................................... 33

8 Results of canonical correspondence analysis (CCA) for 34 fish species
abundances, as measured by catch per unit effort, from 43 lakes in Florida ..........34

9 Intraset correlation between the limnological variables examined and the three
axes in the canonical correspondence analysis (CCA) using 34 fish species
abundances, as measured by catch per unit effort, from 43 lakes in Florida ..........34

A-i Commonly harvested fish species, and the total length (mm) at which they are
generally first harvested. ............................................... ............................... 48
















LIST OF FIGURES


Figure pge

1 Augmented lakes sampled in three Florida counties ........................................... 22

2 Simple linear regressions of (a) Loglo transformation of species evenness by
number of fish versus Secchi depth, and (b) diversity by number of fish versus
S e c ch i d ep th ....................................................... ................ 2 3

3 Joint Plot of axis 1 versus axis 2 of the canonical correspondence analysis with
lakes and species plotted along environmental gradients.......................................24

4 Joint Plot of axis 2 versus axis 3 of the canonical correspondence analysis with
lakes and species plotted along environmental gradients.......................................25

5 Cluster analysis of lakes using total alkalinity, chloride, total phosphorus, and
Secchi depth. .........................................................................26















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 Science

EFFECTS OF INTRODUCED GROUNDWATER ON WATER CHEMISTRY
AND FISH ASSEMBLAGES IN CENTRAL FLORIDA LAKES

By

Patrick Cooney

August 2004

Chair: Micheal S. Allen
Major Department: Fisheries and Aquatic Sciences

Water levels in central Florida lakes have declined since the 1960s as a result of

numerous factors. To maintain water levels in these lakes, the Southwest Florida Water

Management District (SWFWMD) issued permits to pump water from limestone aquifers

into lakes. I assessed effects of groundwater augmentation on limnological variables and

fish assemblages in seven Central Florida lakes.

Pumping history information indicated that lake level fluctuations were reduced,

and pumping volumes could replace the volume of water in a lake multiple times in a

single year. Well water samples, when compared with current lake water samples,

indicated that well water had higher mean total alkalinity and total phosphorus

concentrations, and lower concentrations of total nitrogen and chlorides. The

replacement of original lake water with aquifer water indicated similar patterns when

comparing current lake water samples to historical samples prior to initial introduction of

groundwater. Current lake water samples had higher mean pH, Secchi depth, total









alkalinity, total phosphorus, and chloride concentrations, and lower mean color, nitrogen

and chlorophyll concentrations than historical means.

Historical fish population studies did not exist on these lakes therefore data from

the augmented lakes were compared to 36 nonaugmented lakes in Florida. The mean

values for catch per unit effort (CPUE), species richness and biomass of harvestable

fishes were lower in augmented lakes than those in nonaugmented lakes. However,

significant multiple linear regressions indicated that fish population responses of

augmented lakes to environmental variables were similar to nonaugmented lakes with

similar limnological characteristics.

Canonical correspondence analysis (CCA) was used to examine the relationship

between the abundance of individual fish species and measured limnological

characteristics. Most fish species and nonaugmented lakes were correlated with axis one

of the CCA, whereas augmented lakes were more related to axis two, indicating that

augmented lakes were characteristic of high total alkalinity and Secchi depth, and low

chloride and phosphorus concentrations. Cluster analysis with these four variables

further demonstrated the similarities in limnological characteristics among augmented

lakes. Joint plots of the CCA indicated a high probability of a low abundance of

individual species in augmented lakes compared to a majority of nonaugmented lakes.

One of the augmented lakes had much lower pumping rates than the others, and

exhibited less of a shift in limnological variables from historical values, as well as had

fish population characteristics more closely resembling those of nonaugmented lakes in

the joint plot of the CCA. Therefore, reduced volumes of groundwater introduction could

reduce the alteration of limnological and fish population characteristics.















INTRODUCTION

Lake water levels in central Florida have drastically decreased since the 1960s due

to multiple influences. As a consequence of low precipitation (Stewart and Hughes

1974), groundwater levels were depressed and discharges of inlet streams were

significantly reduced, causing lakes to receive little water input (Dooris and Martin

1979). Further, urban development changed Florida's drainage systems and diverted

storm runoff away from lakes (Stewart and Hughes 1974), and agriculture endeavors

withdrew water from lakes for citrus irrigation and freeze protection (Dooris and Moresi

1975). Population expansion also increased water demand, resulting in increased

pumping of aquifer water at wellfields, subsequently lowering groundwater and lake

levels (Stewart 1968; Stewart and Hughes 1974; Allen 1999).

Groundwater is utilized for public, industrial and agricultural purposes (Southwest

Florida Water Management District [SWFWMD] 1998; Brenner et al. 2000). In the

northern Tampa area, groundwater pumping at wellfields began in 1963 to meet the

increased demands for water (Stewart and Hughes 1974). Wells ranged from 120 to 180

meters in depth and produced thousands of cubic meters of water a day from the Tampa

and Suwannee Limestone Formations, the two uppermost layers of the Floridan Aquifer

(Stewart and Hughes 1974; Sinclair 1977). The Floridan Aquifer is comprised of sand

and clay in the upper regions, with dolomite comprising the lower regions (Stewart and

Hughes 1974; Belanger and Kirkner 1994).









Lakes in the vicinity of the wellfields were hydraulically connected to the water

table aquifer, meaning that water moved naturally between the lakes and the water table

aquifer, a surficial aquifer located two to five meters below land surface (Stewart and

Hughes 1974). As groundwater was pumped at wellfields, a localized cone of depression

formed in the Floridan Aquifer, inducing an increase in leakage from the water table

aquifer to the Floridan Aquifer. As a result, flow increased from the lakes to the water

table aquifer, causing a decline in water levels in lakes near wellfields considerably

greater than those that would naturally occur in lakes away from wellfields (Stewart

1968; Stewart and Hughes 1974; BRA 1996).

Landowners expressed concern with the declining lake water levels, and in order to

address the issue, the SWFWMD permitted landowners to construct wells of similar

depths to those in the wellfields for pumping water from the Floridan Aquifer into lakes

(BRA 1982; Dooris et al. 1982; Belanger and Kirkner 1994; BRA 1996; Allen 1999).

Water levels in these lakes are now constantly maintained slightly above original mean

lake levels and are not pumped to a degree that will allow spill over (Stewart and Hughes

1974). However, maintaining lake levels at higher than normal levels accelerates

evaporation rates, and also increases leakage of lake water, further increasing the

permeability of lake-bottom sediments (Stewart and Hughes 1974; Belanger and Kirkner

1994). As lake water leakage increases, even more groundwater is necessary to maintain

water levels year round (Stewart and Hughes 1974).

Previous investigations evaluated the altered water chemistry of augmented lakes

and the consequent change in macrophyte growth and phytoplankton diversity. For

example, Martin et al. (1976b) found that the elevated hardness of pumped groundwater










increased the ability of augmented lakes to support hydrilla, Hydrilla verticillata, growth,

and Dooris et al. (1982) found that phytoplankton diversity was enhanced in augmented

lakes due to increased concentrations of inorganic carbon via groundwater input.

Little work has assessed effects of groundwater augmentation on fish communities

(Bartos 1998; Allen 1999). I evaluated the influence of lake augmentation on

limnological characteristics and fish populations in seven augmented lakes in central

Florida in the summer of 2002. My objectives were to 1) determine limnological

characteristics and pumping history of seven augmented lakes and their corresponding

groundwater wells and compare the limnological characteristics between lakes, wells and

historical data, 2) estimate fish population parameters in the lakes, and 3) compare

augmented lake limnological characteristics and fish populations with those from a data

base of 36 nonaugmented Florida Lakes.















METHODS

Study Sites

The seven augmented lakes are located in Pasco, Hillsborough, and Polk counties

in central Florida (Figure 1). The lake surface areas (SA) were obtained from the

Gazetteer ofFlorida Lakes (Shafer et al. 1986) and unpublished SWFWMD reports, and

the locations of the lakes were obtained using global positioning system (GPS)

coordinates from a Garmin GPSMAP 76 (Table 1). These lakes are not spring-fed from

the Floridan Aquifer but do exchange water with surficial aquifers. Each lake exhibited

significant declines in water levels due to reduced rainfall and wellfield pumping, where

hundreds of thousands of cubic meters of water were removed from the Floridan Aquifer

each day (Table 2), creating the necessity to pump Floridan Aquifer water into each lake

to maintain lake levels. Goose Lake was the first of the study lakes to be augmented,

(1954), and Loyce Lake the most recent (1996) (Table 1).

Limnological Characteristics

I assessed some important limnological characteristics of my study lakes in August,

2003. The percent lake area covered by aquatic macrophytes (PAC) was recorded using

a boat-mounted Raytheon DE-719 Precision Survey Fathometer (Maceina and Shireman

1980). Seven transects were made across each lake at a constant speed while the

fathometer recorded the presence or absence of plants on a paper roll. The total length of

paper recorded on for each lake was divided into 100 equally spaced instantaneous

samples, and the presence or absence of plants at these locations was recorded. The









number of locations with aquatic vegetation present was expressed as a percentage

(Canfield and Hoyer 1992).

For water chemistry, three 1-liter samples of water were collected at arm depth

(-0.5m) in acid-cleaned Nalgene bottles at three mid-lake sampling stations established

in each lake in August of 2003. The samples were immediately placed on ice and

returned to the laboratory for analysis. Secchi depth (m) was measured at each station

and averaged to determine mean Secchi depth. Dissolved oxygen concentration (mg/L)

and temperature (C) were also measured with a Model 85 Yellow Springs Instrument

(YSI) meter at about 40% of the depth at each mid-lake station and at the location of

pumped water discharge.

Upon arriving at the laboratory (University of Florida, Gainesville, Florida), pH

was measured immediately using an Orion Model 601A pH meter calibrated against

buffers at pH 4.0, 7.0, and 10.0. Total alkalinity (mg/L as CaCO3) was determined by

titration with 0.02 molar H2SO4 (APHA 1989). Chlorophyll concentrations (tg/L) were

determined spectrophotometrically (method 10200 H (2c), APHA 1989) following

pigment extraction with ethanol (Sartory and Grobbelaar 1984). Total phosphorus

concentrations (tg/L) were determined using procedures of Murphy and Riley (1962)

with a persulfate digestion (Menzel and Corwin 1965). Total nitrogen concentrations

([tg/L) were determined by oxidizing water samples with alkaline persulfate and

determining nitrate-nitrogen with second derivative spectroscopy (D'Elia et al. 1977;

Simal et al. 1985; Wollin 1987; Crumpton 1992). Chloride concentrations (mg/L) were

determined by titration of the water samples with 0.0141 mole mercuric nitrate and using

diphenylcarbazone for determining endpoints (Hach Chemical Company 1975). To









analyze for color (platinum-cobalt units), water samples were first filtered through a

Gelman type A-E glass fiber filter. Color was then determined by using the platinum-

cobalt method and a spectrophotometer (APHA 1989).

Three 1-liter water samples were also collected from each of the wells supplying

water to the lakes in August of 2003. The volume of the well delivery pipe was

measured, and the pump was run to flush the pipe with at least twice the calculated

volume before water samples were taken. These samples were placed on ice and returned

to the laboratory and analyzed at the same time as the lake samples for pH, total

alkalinity, total phosphorus, total nitrogen, and chloride concentrations. Chlorophyll

concentrations and color were not determined for the well samples because natural

filtration and lack of sunlight exposure that is characteristic of the Floridan Aquifer

makes the levels of these variables negligible. Secchi depth was not measured within the

well pipe.

The pumping history of each study lake was determined from unpublished

SWFWMD reports. The daily, monthly and yearly averages from these reports were

used to determine the average amount of groundwater pumped on a yearly basis. The

volume of each lake was determined by multiplying the surface area of the lake by the

average depth determined from the fathometer transects. The average volume of water

pumped per year was then divided by the volume of the lake to determine the amount of

times per year the water pumped would replace the current volume of water in each lake.

Mountain Lake and Sunset Lake each share pumps with other lakes, and the amount of

pumped water for individual lakes was not separately recorded. Therefore, I calculated









the volume of all lakes receiving water from a shared pump, and determined the

percentage of the total volume attributed to Mountain and Sunset lakes.

Finally, I compared the ranges and means of limnological characteristics of the

augmented lakes to the well water, and to the limited amount of historical water

chemistry data for the augmented lakes that existed prior to initial pumping of aquifer

water.

Electrofishing

Fish populations in the seven augmented lakes were sampled during the warm

season in July or August of 2002 using electrofishing. Electrofishing transects of

continuous DC current were conducted for ten minutes to collect fish in the littoral area

of each lake with a 4.3 m aluminum j on boat powered by a 15 horsepower outboard

motor. Six transects were conducted at Clear, Dan and Sunset lakes, seven transects at

Goose, Loyce and Saddleback lakes, and eight at Mountain Lake. The number of 10-

minute transects indicate how many transects were necessary to circumnavigate the entire

lake. Electrofishing equipment consisted of a generator (5000 Watt AC), pulsator

(Coffelt model VVP 15) and a bow-mounted cathode probe supplying an electrical output

of approximately seven amps. All collected fish from each transect were counted,

measured to the nearest millimeter total length (TL), weighed to the nearest gram, and

identified to species. Fish with total lengths less than 20 mm TL were not included in

analyses due to selectivity of the gear (Reynolds 1996).

Due to the lack of fish population studies on these augmented lakes prior to the

initial pumping of water, I compared the data collected from augmented lakes to a data

set of 60 nonaugmented Florida lakes (Canfield and Hoyer 1992; Bachmann et al. 1996).

Two lakes, Mountain and Gate, were removed from the 60 lake data set because they









were augmented lakes. Three lakes, Apopka, Lochloosa and Harris, were also removed

due to their surface areas being orders of magnitude larger than all other study lakes,

because lake size influences species richness (Bachmann et al. 1996).

To assess the likelihood that more electrofishing transects would add additional

species, I constructed curves, using Bachmann et al.'s method, for augmented and

nonaugmented lakes demonstrating the cumulative number of fish species captured as

more transects were conducted (Bachmann et al. 1996). To ensure that the right-hand

portion of the curves flattened, the number of species captured in the next to last

electrofishing transect was divided by the number of species captured in the last transect,

and expressed as a percentage, which was termed the exhaustion index (Bachmann et al.

1996). All seven augmented lakes had exhaustion indexes of 100%, meaning that all

captured species from each lake had already been captured by the next to last transect. In

19 of the 55 nonaugmented lakes, the exhaustion indexes were less than 90%, possibly

indicating that these lakes were inadequately sampled. Therefore, these 19 lakes were

not used in the analyses. The remaining 36 nonaugmented lakes had exhaustion indexes

that equaled or exceeded 90%, indicating that the right-hand portion of the curves had

flattened. These 36 nonaugmented lakes were used for comparison with the seven

augmented lakes.

Fish Population Measures

For the seven augmented lakes and the remaining 36 nonaugmented lakes, I

estimated catch per unit effort (CPUE) and species richness. Catch per unit effort

(CPUE) was calculated by dividing the total number of individual fish captured in each

transect by 10 minutes (duration of one transect), and then averaging across the total









number of transects conducted in the particular lake (number of fish/minute). Species

richness was calculated as the total number of fish species collected in each lake.

Evenness was calculated for both the number of individuals and the total weight of

each species using Simpson's measure of evenness (Krebs 1999). Evenness attempts to

measure how evenly the number of individuals or weight is distributed among all species

in a community. The Simpson's measure of evenness (E) is defined as:


E 1/(pP) (eq. 1)


where p, is the number of individuals or total weight of the ith species, P is the total

number of individuals or total weight of all species, and s is the total number of species in

each lake. This index is relatively unaffected by the rare species in the sample and ranges

from 0 to 1 (Krebs 1999).

In addition, I also calculated a Shannon-Wiener index of diversity for both the

number of individuals and the total weight of each species collected in each lake (Krebs

1999). Diversity attempts to account for evenness and richness by looking at both the

number of species and how evenly distributed the number of individuals or weight is

amongst the total number of species in each lake. The Shannon-Weiner index of species

diversity (H') is defined as:


H'= 1 (log 2 ) (eq. 2)
1=1P P

where p, is the number of individuals or total weight of the ith species, P is the total

number of individuals or total weight of all species, and s is the total number of species in

each lake. For biological communities, H' ranges from zero to five (Krebs 1999), and is

expressed in bits per individual (bits/individual).


Krebs 1999), and is

expressed in bits per individual (bits/individual).









I also calculated the total biomass of harvestable fish caught per minute in each of

the lakes (Canfield and Hoyer 1990). These fish exceeded lengths at which anglers

generally harvest the given species (Appendix A).

Fish Population Analysis

The relationships between fish population variables (CPUE, evenness, diversity,

richness and harvestable biomass) and limnological variables (total alkalinity, chlorides,

total phosphorus, total nitrogen, lake surface area, Secchi depth, chlorophyll, color and

percent composition of submersed aquatic vegetation) were examined for the augmented

and nonaugmented lakes using multiple linear regression. Prior to model selection, a

Wilk-Shapiro test was performed on all dependent fish variables, except diversity, to test

for normality (Procedure UNIVARIATE NORMAL, SAS 1996). The evenness index for

both fish weight and number of fish, as well as CPUE and harvestable biomass, were

logio(x+l) transformed to increase normality. Stepwise model selection procedure was

used to create multiple regression models (STEPWISE option, SAS 1996) with a

significance level of 0.05 for independent variables to remain in the model.

Multiple regression models with only one significant independent variable

predicting a fish population variable were graphed. Confidence limits of 95% were

placed above and below the predicted regression line for each model, and the points

corresponding to augmented lakes were examined for influence or diverging patterns.

Models with multiple significant habitat variables predicting a fish population

variable were examined for possible influence by augmented lakes. This was done using

influence diagnostics, including DFFITS, COVRATIO and studentized-residuals (SAS

1996). The DFFITS value represents the number of estimated standard errors that the

fitted value changes if the point is removed from the data set (Myers 1990). A value









close to zero indicates a low influence of the given point. The COVRATIO values

display the reduction in the estimated generalized variance of the coefficient over what

would be produced without the data point. A value close to one indicates little influence

on the estimated generalized variance. Finally, studentized-residuals were used to detect

outliers. A value close to zero indicates a minimal residual for the given point, indicating

a non-outlier (Myers 1990). I concluded that if augmented lakes did not have extreme

values for DFFITS, COVRATIO, and studentized residuals, then the observation would

be within the overall pattern for nonaugmented lakes.

Canonical correspondence analyses (CCA) (PC-ORD 1999) is a multivariate

analysis technique that utilizes data from two matrices to relate community composition

to known variation in the environment (Ter Braak 1986). The CCA was used to arrange

lakes and species of fish along environmental gradients. Catch per unit effort (fish/min.)

was calculated as a measure of relative abundance for each species in each lake and

placed in the primary matrix for comparison of community patterns across lake samples

(Hinch and Collins 1993). Species observed in less than three of the 43 lakes (seven

augmented and 36 nonaugmented) were removed from the analysis to reduce the effects

of rare taxa. No rare taxa were found in the augmented lakes. The same limnological

variables as in the multiple regressions were placed in the secondary matrix across lakes.

Percent data (percent lake area covered by aquatic macrophytes) were arcsine(x/100)

transformed (Zarr 1999) and all other directly measured environmental variables were

logio(x+l) transformed to reduce kurtosis (Palmer 1993).

Canonical correspondence analysis is not hampered by high multicollinearity

between species, or between environmental variables (Palmer 1993). Therefore,









preprocessing or elimination of multicollinear data is unnecessary. Similarly, CCA

estimates the modal locations of highly skewed species distributions quite well, and is

robust to violations of assumptions (Palmer 1993).

In CCA, lakes were assigned scores determined from weighted-averages of species

abundances (Palmer 1993). A multiple linear least-squares regression was then

performed with environmental variables as independent variables, and lake scores as the

dependent variables. New lake scores were then assigned as the value predicted from the

resulting regression equation. The algorithm continued re-standardizing lakes and

species scores until they remained constant with progressing iterations. The product was

the first ordination axis, which was a linear combination of environmental variables that

maximized the correlation between lake and species scores. Second and higher

ordination axes also maximized correlations with scores, uncorrelated with the previous

axes (Ter Braak 1986).

Intraset correlations are the correlation coefficients between the environmental

variables and the ordination axes (Ter Braak 1986). The signs and magnitudes of the

intraset correlations were examined to assess the relative importance of each

environmental variable in structuring the fish community. Intraset values less than -0.350

and greater than 0.350 were considered more highly correlated than those between -0.350

and 0.350. This criterion was arbitrary and was not intended to reflect statistical

significance, and all intraset values are presented.

The canonical correspondence analysis was graphed on ajoint plot, with the first

three ordination axes (PC-ORD 1999). The lakes and species were examined for the

dominant patterns in community composition as explained by the environmental






13


variables, and the augmented lakes were further examined for diverging patterns. The

intraset values of the CCA were also examined to determine which environmental

variables were most strongly correlated with augmented lake scores. Those intraset

values less than -0.350 and greater than 0.350 were considered more highly correlated,

and were used to construct a cluster diagram (PC-ORD 1999).















RESULTS

Comparison of Limnological Variables

Augmented lakes had a smaller average surface area than nonaugmented lakes.

The seven augmented lakes in this study had surface areas ranging from 13 ha to 39 ha

(Table 1) with an average of 21 ha, whereas the nonaugmented lakes ranged from 1.8 ha

to 271 ha, with an average of 83 ha. The range of percent lake area covered by aquatic

macrophytes (PAC) displayed large variation for both augmented (10% to 58%) and

nonaugmented lakes (0% to 100%) (Table 3). Four of the augmented lakes had Secchi

depths greater than the average of the nonaugmented lakes (2.04 m), whereas all of the

augmented lakes had pH levels higher than the average for the nonaugmented lakes

(7.56) (Table 3). Three of the seven augmented lakes exceeded the range of the total

alkalinity for the nonaugmented lakes (0.28 to 106 mg/L as CaCO3), and all augmented

lakes had values greater than the nonaugmented average (31.1 mg/L as CaCO3). All

augmented lakes but Saddleback had lower chlorophyll concentrations than the average

for nonaugmented lakes (26.5 tg/L), and all had lower phosphorus and nitrogen

concentrations than the average nonaugmented values (30.5 [tg/L, and 939 [tg/L

respectively). Sunset Lake was the only augmented lake to exceed the average chloride

concentration for the nonaugmented lakes (18.5 mg/L). Finally, the color values of the

augmented lakes were within the range (1.25 to 57.5 Pt-Co units) and near the average

(20.6 Pt-Co units) of the nonaugmented lakes (Table 3).









Groundwater Pumping History

Clear Lake had the smallest volume of the augmented lakes (1.20 x 105 m3), and

Mountain Lake had the largest volume (1.13 x 106 m3) (Table 4). Similarly, Mountain

Lake had the largest average volume of groundwater pumped into the lake (2.67 x 106

m3/year), whereas, Sunset Lake had a dramatically smaller average than the rest of the

lakes (9.97 x 104 m3/year). The number of times that the volume of pumped groundwater

replaced the volume of water in the lakes ranged from 0.238 to 3.28 times/year, with

Sunset Lake having the lowest rate and Clear Lake having the highest (Table 4).

After investigating the water chemistry of the groundwater pumped from the wells

at each lake and the historical data from several of the lakes prior to initial pumping

(Table 5), numerous patterns were found. Loyce Lake was the only lake with historical

Secchi depths prior to augmentation, and upon the addition of groundwater, the Secchi

depth increased from 0.8 to 3.05 meters. In all of the lakes, the total alkalinity and total

phosphorus concentrations in the well samples were higher than the current lake water

samples, and in every case, the lakes increased in pH, alkalinity and total phosphorus

since their historical measurements. Conversely, all but one well water sample,

Saddleback lake, had lower total nitrogen concentrations than the current lake samples,

coinciding with a decrease in nitrogen when compared to historical water samples (Table

5). The well water samples also had lower chloride concentrations than the current

samples from the lakes; however, the two lakes with historical chloride data increased in

chloride concentrations since the initiation of augmentation. Loyce, Saddleback, and

Sunset lakes experienced a decrease in color, and Loyce Lake exhibited a decrease in

chlorophyll when compared with from historical data, coinciding with groundwater

pumping.









In each of the augmented lakes, groundwater was pumped at a location about 100

meters from the lake, and either formed a small stream or was run down a pipe where the

water was released in an upward fashion, like a fountain. The mean oxygen and

temperature levels measured at mid-lake stations were nearly identical to those measured

at the end of these streams and pipes, where groundwater is introduced into the lakes.

Fish Population Comparisons

Catch per unit effort (CPUE) of all fish varied among the augmented lakes, with

Goose Lake having the lowest (1.16 fish/minute) and Sunset Lake having the highest

(10.7 fish/minute) (Table 6). Goose Lake also had the lowest species richness (5

species), and Clear Lake had the highest (11 species). However, Goose Lake had the

highest index of evenness (E) by number of fish per species (0.70), and Sunset Lake had

the lowest (0.21). Similarly, Goose Lake had the highest index of evenness by weight of

fish per species (0.61), and Mountain Lake had the lowest (0.32). Species diversity (H')

by number ranged from 1.37 to 2.41 at Sunset Lake and Clear Lake, respectively, and

species diversity by weight ranged from 1.67 to 2.38 at lakes Dan and Sunset,

respectively. Biomass of harvestable length fish ranged from 15.1 grams per minute at

Clear Lake to 173 grams per minute at Mountain Lake.

The averages of mean CPUE and species richness of the nonaugmented lakes

exceeded values at six of the seven augmented lakes (Table 6). The ranges and averages

of the evenness and diversity variables of both the augmented and nonaugmented lakes

were similar. However, the average of the mean harvestable fish biomass for the

nonaugmented lakes exceeded the values of all seven augmented lakes (Table 6).









Multiple Regression Analysis

The multiple linear regressions with stepwise model selection for the augmented

and nonaugmented lakes combined were all significant (P < 0.05) (Table 7).

Limnological variables explained 13% to 63% of the variability in fish population

variables in the nonaugmented lakes. Secchi depth was negatively related to diversity by

number, and positively related to logarithm evenness (E) by number. Species diversity

(H') by weight was positively related to color and surface area. Log CPUE was

positively related to chlorides and negatively related to Secchi depth and PAC. The log

(E) by weight was positively related to total nitrogen as well as Secchi depth. Species

richness was negatively related to total nitrogen and Secchi depth and positively related

to chlorides and lake surface area. Log harvestablee biomass) was positively related to

lake surface area and negatively related to Secchi and PAC. Diversity by weight was the

only fish variable that was not related to Secchi depth.

The simple linear regressions of Secchi depth versus the logarithm transformation

of evenness by number (Figure 2a) and Secchi depth versus diversity by number (Figure

2b) show that the data points for the augmented lakes are not outside of the 95%

confidence intervals. However, the Secchi depth values of a majority of the augmented

lakes are higher than the majority of the nonaugmented lakes.

The remaining four fish population parameters were all determined by multiple

habitat variables. Therefore, I examined the influence diagnostics to determine the

influence of augmented lakes on the multiple regressions. In each case, the absolute

values of the studentized residuals were minimal, the COVRATIO values were close to

one, and the absolute values of the DFFITS values were minimal, indicating that the









augmented lakes had little influence on the multiple regressions, similar to the results of

the simple linear regressions.

Canonical Correspondence Analysis

Canonical correspondence analysis (CCA) was used to explore the relationship

between the abundance, as determined by catch per unit effort, of fish species, the lakes

where they were found, and the tested limnological variables. The first three axes in the

CCA, which are linear combinations of limnological variables that maximize the

correlation between site and species scores, explained 26% of the variation (Table 8).

The first canonical axis, which explained 12% of the variation, was positively correlated

(>0.350) with chlorophyll, total phosphorus, total nitrogen, total alkalinity, color, surface

area and chlorides, and negatively correlated (<-0.350) with Secchi depth and percent

area covered by aquatic macrophytes (Table 9). The second canonical axis, which

explained 7.7% of the variation, was positively correlated (>0.350) with chlorides and

total phosphorus, and negatively correlated (<-0.350) with total alkalinity and Secchi

depth. The third canonical axis, which explained 5.6% of the variation, was negatively

correlated (<-0.350) with chlorides and color.

By definition, the majority of the variation was explained by the first axis.

However, all of the augmented lakes, except for Sunset, were most highly correlated with

axis two, in a negative fashion. There were only four other nonaugmented lakes that

were most highly correlated with axis two in a negative fashion, demonstrating the

similarities of the augmented lakes to each other in relation to environmental gradients.

The six augmented lakes correlated with axis two were more characteristic of lower

chlorides and total phosphorus, and higher Secchi depths and total alkalinity. In contrast,









Sunset Lake, more negatively correlated to axis three, was characterized by higher

chlorides and color (Table 9).

The joint plot of axis one versus axis two displayed the general pattern of the

nonaugmented lakes along axis one (Figure 3). The six augmented lakes Clear, Dan,

Goose, Loyce, Mountain and Saddle, were more negatively correlated to axis two than

they were correlated to axis one, and were all located in the same general location of the

joint plot (Figure 3). Similarly, in the joint plot of axis two versus axis three (Figure 4),

these same augmented lakes were more highly correlated with axis two in a negative

fashion than they were with axis three, and again were all located in the same general

area of the joint plot, whereas Sunset Lake was more related to axis three, with only a

slight negative correlation with axis two.

In joint plots, the lake points are found at the centroid of the species points that

occur at that lake, allowing inferences to which species are likely to be present at a

particular lake (Ter Braak 1986). Also, the species points are approximately the optima

of where they are found in highest abundance, hence the abundance or probability of

occurrence of a species decreases with distance from its location in the diagram. For

example, the lined topminnow, Fundules lineolatus, was highly negatively related to axis

one, as were two lakes, Turkey Pond and Keys Pond, that the topminnow was found in

close proximity to in the joint plot (Figure 3). This species was found in only six of the

43 lakes, and in highest abundance in these two lakes, explaining its closeness to these

two lakes (Figure 3). The lined topminnow's low abundance or absence in the other

lakes demonstrates its distance from those lakes. Bluegill, Lepomis macrochirus, was

found in all but two lakes, demonstrating this species ability to survive across varying









environmental gradients. Accordingly, bluegill were located near the intersection of all

three axes, where there is little correlation to high or low values of environmental

variables (Figure 3 and 4). Consequently, those lakes within close proximity to the

intersection of the axes, with little correlation to any of the three axes, have a higher

probability of having higher abundances of bluegill and other fish species that are found

near the intersection than those on the perimeter.

Six of the seven augmented lakes were found on the perimeter of the joint plots.

Accordingly, all but one species of fish was at relatively large distances in the diagram

from the six augmented lakes that were more correlated to axis two (Figure 3 and 4). The

one fish species, taillight shiner, Notropis maculatus, which was found in close proximity

to these six lakes, was only found in one of the augmented lakes, Clear Lake. This fish

represented the second highest abundant fish species in Clear Lake, and was only found

in four other lakes, at low abundances, explaining the close proximity to Clear Lake.

Therefore, the lack of other species in close proximity to these six augmented lakes

demonstrates that there is a higher probability that abundance of individual fish species in

these six augmented lakes is lower than other lakes more close in proximity to species

points. Sunset Lake is closer in proximity to several fish species points indicating a

higher probability of a higher abundance of individual fish species in this lake as

compared to the other augmented lakes.

Cluster Analysis

The intraset correlations of axis two of the CCA demonstrate that six of the

augmented lakes were correlated with higher total alkalinity and Secchi depth, and lower

chlorides and total phosphorus. These four environmental variables were used to create a

cluster diagram (Figure 5). The cluster diagram displayed five of the augmented lakes






21


Clear, Goose, Loyce, Mountain and Dan, in a small cluster. Saddleback Lake, the other

augmented lake correlated with axis two of the CCA, is grouped in the next closest small

cluster. This further demonstrates the similar limnological characteristics of the

augmented lakes with large volumes of groundwater introduction. Sunset Lake is not

closely clustered with any of the other augmented lakes, but rather with two other

nonaugmented lakes that, like Sunset Lake, were also most correlated with axis three in a

negative fashion.











Loyce Lake Clear Lake

Goose Lake

S Pasco
i-C- -. -


Polk


nset Lake Saddleback Lake
40 Kilometers


Figure 1. Augmented lakes sampled in three Florida counties.


Lake














(a)0.4

0.2



S 0 1 2 3 4 5 6
z
z -0.2 4



S-0.4 "A







-1
Secchi Depth (meters)




(b)4
3.5 ----.


3 0

E 2.5 -
z *
2 2

1.5 5



0 ---o


0
0 1 2 3 4 5 6
Secchi Depth (meters)

Figure 2. Simple linear regressions of (a) Loglo transformation of species evenness by
number of fish versus Secchi depth (r2=0.298, df=42, p-value<0.001), and (b)
diversity by number of fish versus Secchi depth (r2=0.134, df=42, p-
value=0.016), with regression lines (-) and 95% confidence intervals for an
observation around the lines (- -) for nonaugmented (*) and augmented lakes
(A).


n around the lines (- -) for nonaugmented (*) and augmented lakes
(A).












Picnic













Bull-Pon



+
Tomahawk
Round-po Douglas

Crooked
A


Live-oak +
+ 4- PAC,


Lined Topminnow
-+urkey-p
A
Keys-pon
'n


Patrick
+


Figure 3. Joint Plot of axis 1 versus axis 2 of the canonical correspondence analysis with
lakes (A) and species (+) plotted along environmental gradients.
Limnological characteristics include: percent area covered by aquatic
macrophytes (PAC), Secchi depth (Secchi), total alkalinity (talk), color, total
nitrogen (tn), chlorophyll (CHL), total phosphorus (tp), Surface Area (SA)
and chloride (cl). Augmented lakes include: Clear, Dan, Goose, Loyce,
Mountain, Saddleback (Saddle) and Sunset. The oval contains six augmented
lakes that are most related to axis 2. The box contains Sunset Lake.


Rowell
A


++ +


Wauberg
A



++ +


Wauberg
A























Little-F
A

Watertow
A


Hunter
.0 A
x
Hollings

+ Picnic
Wauberg A
A



Holden
A Tomahawk

--


ries
Bull-Pon
3onny
tp
:nA +

Axis 2


SOkahumpk Douglas
Keys-pon A Fi h
A Bell A
ys- A
+ Koon Rowell
A A
Sunset Round-po
Moore A





Live-oak
A




Figure 4. Joint Plot of axis 2 versus axis 3 of the canonical correspondence analysis with
lakes (A) and species (+) plotted along environmental gradients.
Limnological variables include: percent area covered by aquatic macrophytes

(PAC), Secchi depth (Secchi), total alkalinity (talk), color, total nitrogen (tn),
chlorophyll (CHL), total phosphorus (tp), surface area (SA) and chloride (cl).
Augmented lakes include: Clear, Dan, Goose, Loyce, Mountain, Saddleback
(Saddle) and Sunset. The oval contains six augmented lakes that are most
related to axis 2. The box contains Sunset Lake.


nted lakes that are most
related to axis 2. The box contains Sunset Lake.

















Distance (Objective Function)
6.7E-03 7.6E+00 1.5E+01 2.3E+01 3E+01


Information Remaining (%)
100 75 50 25 0

Keys-pon
Round-po
Turkey-p
Crooked
Tomahaw k

Moore
Koon
Mll-Dam
Bull-Pon
Picnic
Okahumpk
Sanitarv
Punset I
West-Moo
Patrick
Douglas
Row ell
Fish
1Mountai
Mona
Bell
Pasadena
Live-oak
Susannah
Little-F
Baldw in
Killarny
Watertow
Haddle
Car







Figure 5. Cluster analysis of lakes using total alkalinity, chloride, total phosphorus, and Secchi depth. Augmented lakes (boxed)
3oose
-oyce i
/buntain
Dan
Hollings
Hunter
Holden
Bonny
Carbon

Figure 5. Cluster analysis of lakes using total alkalinity, chloride, total phosphorus, and Secchi depth. Augmented lakes (boxed)
include Clear, Dan, Goose, Loyce, Mountain, Saddleback (Saddle) and Sunset.














Table 1. The county, wellfield in closest proximity, location, surface area, average depth determined with fathometer and
year of first groundwater augmentation for the seven study lakes. Mountain Lake is not within the vicinity of a
wellfield, but rather, requires augmentation due to its proximity to the highest elevation in peninsular Florida, on
the Lake Wales Ridge, increasing its elevation above the surficial aquifer.


Latitude Longitude


Surface Area Averege
Depth


(ha)


Clear
Dan
Goose
Loyce
Mountain
Saddleback
Sunset


Pasco
Hillsborough
Pasco
Pasco
Polk
Hillsborough
Hillsborough


Eldridge-Wilde
Cross Bar Ranch
Eldridge-Wilde
Eldridge-Wilde
Elevation
Section 21
Cross Bar Ranch


28.3625oN
28.1667oN
28.3559oN
28.3758oN


82.47890W
82.6464oW
82.4702oW
82.4958oW


27.9348oN 81.5898oW
28.1194oN 82.4942oW
28.1345oN 82.6267oW


(m)

1.50
3.18
1.49
1.73
2.92
2.56
2.83


Year of First
Augmentation



1978
Early 1970's
1954
1996
1975
1968
1976


Lake


County


Wellfield







28
Table 2. The number of wells, the average volume of water pumped each day from
all wells combined, and the year of initial service for the wellfields in
close proximity to the augmented lakes.

Average
Wellfield Number of Wells Volume Pumped Initial Year
(m3/day)

Cross Bar Ranch 17 1.00 x 105 1981
Eldridge-Wilde 58 8.52 x 104 1956
Section 21 8 3.79 x 104 1963











Table 3. Mean limnological characteristics in 2003. These include percent lake area covered by aquatic macrophytes (PAC), Secchi
depth, ph, total alkalinity, total chlorophyll, total phosphorus, total nitrogen, chloride, and color. Water chemistry means
were based on three samples per lake, and PAC was measured using transects with a recording fathometer. Nonaugmented
lake means are from 36 Florida lakes.

PAC Secchi pH Total Total Total Total Chloride Color
Lake Alkalinity Chlorophyll Phosphorus Nitrogen
(mg/L as
(%) (m) CaC03) (tg/L) (tg/L) (tg/L) (mg/L) (Pt-Co units)


Augmented
Clear 58 2.44 7.80 115 5.60 13.3 427 6.00 23.3
Dan 14 1.52 8.00 120 15.6 16.7 890 9.92 53.0
Goose 52 3.35 7.90 107 3.20 12.3 670 6.92 31.7
Loyce 28 3.05 7.80 103 1.60 8.00 463 7.17 16.0
Mountain 44 3.35 9.03 64.7 8.40 14.3 487 9.67 15.0
Saddleback 12 1.52 8.27 76.3 34.4 19.3 737 7.67 42.0
Sunset 10 1.83 7.77 56.0 20.2 18.0 810 27.1 44.0
Mean 31.1 2.44 8.08 91.7 12.7 14.6 641 10.6 32.1

Nonaugmented
Mean 44.2 2.04 7.56 31.1 26.5 30.5 939 18.5 20.6
Range 0-100 0.30-5.80 4.30-9.18 0.28-106 0.82-159 0.83-159 99.0-1750 2.50-51.7 1.25-57.5










Table 4. Groundwater pumping history in augmented lakes. Variables include lake
volume, the average volume of groundwater pumped per year, the years of
historical pumping data averaged, and the number of times the volume of
pumped water would replace the volume of water in the lake in one year.


Average
Lake Name Lake Volume Volume Pumped Years Fill Rate
(m3) (m3/year) (times/year)

Clear 1.20 x 105 3.94 x 105 1990-98 3.28
Dan 4.77 x 105 7.61 x 105 1994-98 1.60
Goose 2.23 x 105 3.19x 105 1990-98 1.43
Loyce 3.13 x105 3.50x 105 1995-98 1.12
Mountain 1.13 x 106 2.67 x 106 1989-94 2.36
Saddleback 3.33 x 105 4.43 x 105 1968-71 1.33
Sunset 4.19x 105 9.97 x 104 1977-01 0.24












Table 5. Water chemistry of groundwater from well samples and historical lake water samples prior to initial augmentation.
Variables include date of collection (month/year), Secchi depth, pH, total alkalinity, total chlorophyll, total
phosphorus, total nitrogen, chloride, and color. Secchi depth, total chlorophyll and color were not determined for
groundwater. Blanks for historical lake samples indicate missing data.

Lake Total Total Total Total
Name Date Secchi pH Alkalinity Chlorophyll Phosphorus Nitrogen Chloride Color
(mg/L as (Pt-Co
(month/year) (m) CaCO3) ([tg/L) ([tg/L) ([tg/L) (mg/L) units)

Well Samples
Clear 8/2003 7.7 177 25.0 310 5.95
Dan 8/2003 7.8 210 77.0 440 7.50
Goose 8/2003 7.6 184 40.0 330 5.00
Loyce 8/2003 7.6 178 28.0 420 5.50
Mountain 8/2003 7.7 152 60.0 450 3.00
Saddleback 8/2003 7.2 232 40.0 910 4.50
Sunset 8/2003 7.8 189 68.0 660 8.60
Mean 7.6 189 48.3 503 5.73

Historical Lake Samples


Loyce 7/1995
Loyce 7/1984
Sunset 5/1976
Saddleback 3/1968
Saddleback 2/1968


0.80 6.9
6.1
5.6
5.5
6.1


36.0
6.00
20.0
14.0
14.0


28.0


5.00


2670


5.00
3.00
16.5


50.0
78.0
90.0
40.0
60.0











Table 6. Fish population measures of augmented and nonaugmented lakes. These include number of electrofishing transects (N),
mean catch per unit effort (CPUE), standard deviation of the catch per unit effort (CPUE SD), species richness, species
evenness by number, species evenness by weight, species diversity by number, species diversity by weight and mean
harvestable fish biomass. Nonaugmented lake averages are from 36 Florida lakes.

Lake N Mean CPUE Richness Evenness Evenness Diversity Diversity Mean
CPUE SD by by by by Harvestable
Number Weight Number Weight Fish Biomass
(fish/ (grams/
transectss) minute) (species) minute)

Augmented
Clear 6 7.05 5.96 11 0.37 0.40 2.41 2.30 74.7
Dan 6 1.65 0.68 7 0.54 0.33 2.26 1.67 15.1
Goose 7 1.16 0.89 5 0.70 0.66 1.95 1.86 51.9
Loyce 7 1.31 0.51 8 0.57 0.44 2.35 2.13 32.6
Mountain 8 3.38 0.92 9 0.38 0.32 2.18 1.77 173
Saddleback 7 1.71 0.45 10 0.36 0.35 2.38 2.12 38.4
Sunset 6 10.7 3.76 10 0.21 0.47 1.37 2.38 26.5
Mean 6.71 3.85 8.57 0.45 0.42 2.13 2.03 58.9

Nonaugmented
Mean 5.26 7.45 10.4 0.44 0.37 2.22 1.98 208
Range 3-6 0.58-36.1 2-18 0.28-0.96 0.17-0.66 0.28-3.15 0.31-2.82 3.87-1150











Table 7. Significant linear regression models predicting dependent fish variables at nonaugmented and
augmented lakes combined. Dependent fish variables include logo of catch per unit effort
(lcpue), logo species evenness by number (levennum) and weight (levenwt), species diversity by
number (divnum) and weight (divwt), species richness (rich), and logo harvestable fish biomass.
Independent variables include surface area of lake (ha, sa), percent area coverage of aquatic
macrophytes (%, pac), Secchi depth (m, Secchi), total alkalinity (mg/L, talk), total chlorophyll
([tg/L, chl), total phosphorus ([tg/L, tp), total nitrogen ([tg/L, tn), chloride (mg/L, cl), and color
(color).

Model R-square df P-value

Icpue = 0.724 + 0.016(cl) 0.134(Secchi) 0.004(PAC) 0.509 42 <0.001
levennum = -0.561 + 0.072(Secchi) 0.298 42 <0.001
levenwt = -0.669 + 0.00008(tn) + 0.069(Secchi) 0.246 42 0.004
divnum = 2.448 0.017(Secchi) 0.134 42 0.016
divwt = 1.215 + 0.004(sa) +0.017(color) 0.495 42 <0.001
rich = 10.803 + 0.113(cl) 0.002(tn) + 0.019(sa) 1.407(Secchi) 0.626 42 <0.001
harvest = 2.302 + 0.002(SA) 0.133(Secchi) 0.005(PAC) 0.402 42 <0.001









Table 8. Results of canonical correspondence analysis (CCA) for 34 fish species abundances,
as measured by catch per unit effort, from 43 lakes in Florida. Proportional
limnological variables were arcsine(x/100) transformed; all other limnological
variables were loglo(x+l) transformed. Species environmental correlations were
conducted using Pearson tests.

Statistic Axis 1 Axis 2 Axis 3
Eigenvalue 0.335 0.214 0.153



Species-Environmental 0.875 0.866 0.822
Correlations

% Variance in species data 12.2 7.7 5.6
explained by the axis


Cumulative % of variance 12.2 19.9 25.5
in species explained

Table 9. Intraset correlation between the limnological variables examined and the three axes in
the canonical correspondence analysis (CCA) using 34 fish species abundances, as
measured by catch per unit effort, from 43 lakes in Florida. Percent area covered by
aquatic macrophytes (PAC) was arcsine(x/100) transformed. All other
environmental parameters were loglo(x+l) transformed. Intraset correlation may
help indicate which environmental variables structure the community, as well as
help determine which environmental variables are most influential in a site. The
higher the absolute value of the intraset correlation, the more the parameter explains
the variation. Values below -0.350 and above 0.350 were considered more highly
correlated than those between -0.350 and above 0.350, and are followed by a for
emphasis.

Variable Axis 1 Axis 2 Axis 3
Total Alkalinity 0.643* -0.450* 0.185
Chloride 0.400* 0.526* -0.574*
Total Phosphorus 0.793* 0.420* 0.317
Total Nitrogen 0.649* 0.064 0.211
Surface Area of Lake 0.479* 0.262 -0.124
Secchi Depth -0.753* -0.399* -0.296
Chlorophyll 0.888* 0.299 0.239
Color 0.629* -0.027 -0.506*
PAC -0.553* 0.117 -0.279















DISCUSSION

Historically, the water chemistry of lakes in central Florida differs significantly

from the chemistry of aquifer water in the same vicinity (Dooris and Martin 1979; BRA

1996; Hassell et al. 1997). The historical data from the augmented lakes prior to initial

groundwater introduction demonstrated lower levels of pH, alkalinity, and water clarity

prior to augmentation, similar to typical small lakes in central Florida that contain acidic,

soft, tannin-colored water, with low levels of bicarbonate, inorganic carbon, alkalinity,

conductivity, and calcium (Canfield 1981; Canfield and Hoyer 1988; BRA 1996; Hassell

et al. 1997).

The water chemistry of the augmented lakes in this study shifted to levels

resembling the water chemistry of aquifer water upon the introduction of large volumes

of groundwater, by which an entire lakes volume can be replaced several times a year.

As groundwater is introduced, the replacement of original lake water is generally

characterized by increases in clarity, pH, hardness, bicarbonate, inorganic carbon,

alkalinity, conductivity, calcium, magnesium, dissolved solids, nitrogen and sodium

concentrations, all of which are chemical characteristics of water contained in the

Floridan Aquifer (Stewart and Hughes 1974; Martin et al. 1976a; BRA 1982; Dooris et

al. 1982; BRA 1996; Hassell et al. 1997). Dooris and Martin (1979) noted that increased

pumping of aquifer water caused a shift in the water chemistry of augmented lakes to

closely resemble the chemical characteristics of aquifer water. It is apparent that the









patterns of increasing pH, alkalinity and clarity in my study lakes are analogous to those

of previously studied augmented lakes.

Another noted effect of groundwater pumping, as reported by Canfield and Hoyer

(1990) in Gate Lake and Mountain Lake, Florida, is the addition of nutrients. The well

water samples in the seven augmented lakes had higher mean concentrations of

phosphorus than both historical and current lake water sample levels. Conversely, in all

sampled augmented lakes but Saddleback Lake, nitrogen concentrations were lower than

sampled well water and historical values. However, the nitrogen to phosphorus ratios are

much greater than 17, suggesting that phosphorus is the limiting nutrient in each of the

seven augmented lakes (Florida Lakewatch 2000). The addition of groundwater

demonstrates an increase in the limiting nutrient in the augmented lakes.

Despite the overall increase in phosphorus by groundwater introduction, these

augmented lakes are still characterized by low total phosphorus (< 20 tg/L). Trophic

states, according to concentrations of phosphorus, indicate that all of the studied

augmented lakes were either oligotrophic or mesotrophic (Forsberg and Ryding 1980).

Therefore, there is little expected change in fish population parameters due to increased

nutrient introduction in the studied augmented lakes.

The cluster analysis (Figure 6) of total alkalinity, chlorides, total phosphorus and

Secchi depth versus all 43 lakes demonstrated the similarity in water chemistry of six of

the augmented lakes to each other, and of one of the augmented lakes to a group of

nonaugmented lakes. Despite being in three different counties, six of the augmented

lakes were similarly characterized by higher alkalinity, lower chlorides, lower

phosphorus and higher Secchi depths, whereas Sunset Lake was more characterized by









lower total alkalinity and higher chloride concentrations. This was likely a result of the

lower average volume of groundwater pumped into Sunset Lake each year (9.97 x 104

m3/year) compared to the range of yearly averages of the other augmented lakes (3.19 x

105 to 2.67 x 106 m3/year). Therefore, reduced groundwater introduction could decrease

the effects of shifted water chemistry, resulting in a lake more characteristic of natural

limnological characteristics.

Another change that augmented lakes endure is reduced water level fluctuation.

For example, Mountain Lake experienced lake level fluctuations of approximately 3 m

during the 1940s and 1950s, whereas the recent stage fluctuation indicates a much

narrower overall range of variation of about 1 m during the past decade (BRA 2001).

Similarly, Saddleback Lake experienced less than 0.5 m of fluctuation in the ten years

following augmentation, and showed very little response to heavy rainfall, as caused by

the artificial head placed on the lake above the already lowered potentiometric head

(Jones 1978). The other augmented lakes were also characterized by comparable lake

level fluctuation reduction.

The combination of increased water clarity, increased nutrients, increased hardness,

and reduced water level fluctuation could change the characteristics of aquatic plant

communities in augmented lakes. Increased water clarity increases light penetration,

often allowing plants to grow faster and at greater depths (Canfield et al. 1985).

Likewise, increased nutrients also increase plant growth. Also, Martin et al. (1976b)

found that the elevated hardness of pumped groundwater increased the ability of

augmented lakes to support hydrilla, Hydrilla verticillata, growth. Consequently, many

of the augmented lakes have had a history of aquatic plant problems. For example,









Mountain and Saddleback lakes have been stocked with grass carp, Ctenopharyngodon

idella, treated with aquatic herbicides and had harvest programs to decrease the amount

of plants (Canfield and Hoyer 1990; personal communication with lake residents).

Conversely, Goose Lake and Clear Lake are in the middle of a wellfield, and do not have

public access or people living on them. Therefore, there is no concern for aquatic

vegetation control. These two lakes yielded the highest PAC of the augmented lakes.

Little work has assessed effects of groundwater augmentation, and the ensuing

alterations to lake characteristics, on fish communities (Allen 1999). Cowx (2000) found

that the discharge of groundwater lowered dissolved oxygen concentrations and reduced

water temperatures in the River Ouse, Yorkshire, UK, and suggested that the low

dissolved oxygen concentrations could cause asphyxiation of fish with possible loss of

sensitive species if chronic, and the low water temperatures could reduce fish growth,

leading to a decline in stock (Cowx 2000). In Florida, groundwater generally has lower

temperatures and dissolved oxygen concentrations than surface water (McKinset and

Chapman 1998). However, well introduced groundwater is not pumped directly into the

lakes in Florida as is the case in the River Ouse. The streams and the pipes that deliver

the groundwater to the lakes in Florida allow the water to warm up and aerate before

entering the water body, as demonstrated by the similarity of mid-lake and discharge

measurements. Further, the difference in temperature between groundwater and lake

water in Florida is less than that of the study in Yorkshire, decreasing this concern in

Florida lakes.

The effects of the nutrient introduction by means of groundwater pumping in the

previously mentioned Canfield and Hoyer (1990) study were not detrimental to the fish









populations, and instead, were possibly beneficial with respect to species diversity, total

fish biomass, and sport-fish-abundance. Allen (1999) found similar results with respect

to increased fish species diversity in one augmented lake in Florida (Round Lake).

However, in contrast to Canfield and Hoyer (1990), Allen (1999) found that Round Lake

had significantly lower total fish biomass and density as compared to two nonaugmented

lakes.

The augmented lakes in my study had little influence on the regressions of fish

population parameters versus limnological variables, suggesting that the fish populations

in the augmented lakes were similar to fish populations in nonaugmented lakes with

similar limnological characteristics. However, there is evidence that the limnological

variables shifted from their original levels to those more indicative of aquifer water.

Therefore, one must consider that as the limnological variables shifted with groundwater

introduction, the fish populations in the augmented lakes responded by shifting

correspondingly along the gradient of the regression.

The data from the nonaugmented lakes in this study were collected from 1986 to

1990. Therefore, the limnological and fish population parameters could have changed in

these lakes over time in a similar fashion to the augmented lakes, making it difficult to

compare the two samples. However, upon inspection of numerous limnological

variables, including pH, total phosphorus, total nitrogen and total alkalinity, from several

of the nonaugmented lakes, the changes were either small or obsolete. For example, both

Lake Susannah had the same pH (7.8) in 1988 and 1999, and similar alkalinities over the

same time period (30.8 and 30.7 respectively). Also, the magnitudes of change from the

few limnological parameters measured historically from the augmented lakes were much









greater than the magnitudes of change for the nonaugmented lakes. Although extensive

limnological parameter studies were not performed on the augmented lakes prior to

groundwater introduction, several patterns emerged from the limited data available.

As groundwater was pumped, the augmented lakes were characterized by increased

Secchi depths, total alkalinity, phosphorus and chloride. They were also characterized by

decreases in total nitrogen and color, and would most likely all increase in percent area

covered by aquatic macrophytes if people did not control their levels. Augmentation also

increases surface area.

Upon placing these characteristics in the multiple regressions for all lakes, several

fish population responses can be predicted for augmented lakes. However, it is difficult

to predict exact changes on a temporal scale due to the lack of previous limnological

studies on these lakes. For example, there were no previous estimates of PAC on the

augmented lakes, and only one lake had a historical secchi depth measurement.

Therefore, the following are projected patterns for fish population parameters based on

general observed patterns for the limnological parameters. The pattern of increase or

decrease for catch per unit effort over time with groundwater introduction is difficult to

determine since chloride concentrations, Secchi levels and PAC increase with

groundwater pumping, acting as opposing terms in the regression. However, the average

catch per unit effort for the augmented lakes (3.85 fish/minute) was much lower than

average catch per unit effort of the nonaugmented lakes (7.45 fish/minute), suggesting

that catch per unit effort decreased. Evenness by number would increase because Secchi

depth increases with groundwater pumping. Evenness by weight would have opposing

terms since total nitrogen decreases and Secchi increases, making it difficult to determine









the pattern. Diversity by number would decrease with groundwater pumping due to

increasing Secchi depth, opposing the results reported by Canfield and Hoyer (1990) and

Allen (1999). Conversely, diversity by weight would increase due to increasing surface

area and color. Both richness and harvestable biomass have opposing values as

groundwater increases, making their change with groundwater augmentation difficult to

determine. However, similar to catch per unit effort, mean species richness and mean

harvestable fish biomass were lower for the seven augmented lakes (8.57 species and

58.9 g/minute respectively) than for the 36 nonaugmented lakes (10.4 species and 208

g/minute respectively), indicating a lower average number of species, individuals and

weight of fish of harvestable size than those lakes without groundwater being pumped.

The multiple regression analyses were useful for indicating that fish populations of

augmented lakes did not deviate from the patterns of nonaugmented lakes. However,

multiple regressions are unable to indicate the patterns for numerous individual fish

species across multiple limnological variables in multiple lakes. The joint plots of the

CCA suggested that the abundance, expressed as catch per unit effort, of individual

species in six of the augmented lakes had a high probability of being low compared to a

majority of nonaugmented lakes, agreeing with the Round Lake study by Allen (1999).

This also corresponded with the finding that catch per unit effort of all species and

species richness were low in these lakes compared to nonaugmented lakes. Further, a

majority of all fish species abundances were more correlated to the first axis of the CCA,

whereas the majority of augmented lakes were highly correlated to the second axis,

explaining that the gradient of environmental patterns determining the fish community in









these augmented lakes was different than the gradient determining fish communities in a

majority of nonaugmented lakes.

A study in Max Lake, Wisconsin, examined the effects of groundwater pumping on

fish population dynamics for largemouth bass and yellow perch populations (Engel et al.

2000). The groundwater pumping failed to alter growth, abundance, biomass, or

mortality of yellow perch and largemouth bass 3 to 7 years old. However, very little

water was pumped into the lake as compared to the amount of water pumped into the

studied augmented lakes in Florida. Only 5% of the water in Max Lake was replaced

each year by augmentation, as compared to the range of about 24% to 328% in the seven

augmented lakes. However, similar to the comparison of Sunset Lake to the other studied

augmented lakes, low volumes of introduced groundwater could have reduced effects on

water chemistry and fish population responses compared to lakes with high volumes of

groundwater introduction.

Numerous studies discuss fish growth and biomass responses to alterations in water

chemistry. Many have focused on the relationship between trophic states (as described

by Forsberg and Ryding 1980) and fish communities in lakes. Both fish growth (Larkin

and Northcote 1969; Bayne et al. 1994) and fish production (Downing et al. 1990; Ney

1996) are closely correlated with total phosphorous concentrations in lakes. Similarly,

total phosphorous concentrations (Kautz 1980; Hanson and Leggett 1982; Yurk and Ney

1989; Hoyer and Canfield 1991; Lee et al. 1991; Bayne et al. 1994; Bachmann et al.

1996), total nitrogen concentrations (Bachmann et al. 1996), and chlorophyll

concentrations (McConnell et al. 1977; Ogelsby 1977; Jones and Hoyer 1982; Bachmann

et al. 1996), are all positively related to total fish biomass in lakes. Conversely,









oligotrophication, the reversal of the eutrophication process, is accompanied by declines

in growth, standing stock, and harvest in fish (Ney 1996). Each of these studies displays

a positive correlation between fish productivity, fish biomass, and fish abundance as

eutrophication proceeds. The Forsberg and Ryding (1980) guidelines for trophic state

indicate that the augmented lakes range from oligotrophic to mesotrophic, possibly

explaining for the low CPUE and low harvestable biomass found in these lakes.

However, the slight increase in nutrients caused by groundwater introduction may have

slightly increased these fish population variables from their original levels.

Several investigations have also reported decreases in the relative abundance of

piscivorous fish, with possible losses of sensitive species, as a consequence of

eutrophication (Larkin and Northcote 1969; Persson et al. 1988; Bachmann et al. 1996;

Ney 1996). Opportunistic, eurytolerant, non-piscivorous species are likely to replace

stenotolerant, piscivorous fishes with increasing fertility of lakes (Ney 1996), causing

higher standing crops of such fish species as gizzard shad, Dorsoma cepadianum,

threadfin shad, Dorosoma petenense, and common carp, Cyprinus carpio, in eutrophic

and hypereutrophic lakes (Hasler 1947; Larkin and Northcote 1969; Bachmann et al.

1996).

No evidence was found in this group of augmented lakes to suggest that

piscivorous fish were being replaced by eurytolerant, non-piscivorous fish. For example,

bluegill was the most abundant fish species present in six of the seven augmented lakes.

Also, threadfin shad were found in only one of the augmented lakes, Sunset Lake, where

limnological variables are more closely related to nonaugmented lakes. Further, in

Sunset Lake there were healthy populations of reproducing largemouth bass, bluegill,









black crappie, warmouth and redear sunfish of multiple size ranges, while only one

school of shad was encountered. Therefore, these lakes are far from being in the

hypereutrophic range where these problems occur in fish populations, and there is no

evidence that the increase in nutrients from groundwater introduction into the studied

augmented lakes caused losses of species.

Numerous studies have found that lake size is the dominant factor determining fish

species richness in Florida lakes, in which the number of fish species increases with

increasing lake surface area (Barbour and Brown 1974; Matuszek and Beggs 1988; Keller

and Crisman 1990; Bachmann et al. 1996). In this study, species richness was positively

correlated to the surface area of augmented and nonaugmented lakes combined in the

multiple regression. Florida lakes are generally shallow, with flat slopes, and a small

decrease or increase in lake level stage creates a dramatic alteration in surface area

(Dooris 1982), possibly having a major effect on fish communities. For example,

Mountain Lake is listed as having a surface area of 55 hectares prior to lake level declines

(BRA 2001). Corresponding with the drop in water level was a decrease in the lake

surface area, upon which augmentation returned the surface area to 39 hectares. This was

a decline from the original surface area, but larger than the surface area without

augmentation. Therefore, the lake could display a decrease in species richness from

historical fish population data prior to water level declines, but an increase from periods

immediately prior to augmentation.

Allen (1999) suggested that fish population responses to augmentation may be

variable and depend on the original water chemistry of the natural lake relative to the

water chemistry of the pumped groundwater. Therefore, prior to initial groundwater






45


introduction for augmentation of lake levels, a comparison of water chemistry

characteristics from lake water and well water from within close proximity, and

estimating the volume of pumped water that will be necessary to maintain lake levels,

may be useful in determining expected shifts in water chemistry and fish populations. If

surface area is also a dominant factor in determining fish population parameters, it may

be important to not only monitor water chemistry, but also surface area prior to initial

declines and initial pumping.















MANAGEMENT IMPLICATIONS

Hassel (1994) suggested that augmentation is not a good long-term solution to

restore lake levels to reasonable levels because of altered environmental factors. He

further states that lake augmentation is a short-term remedy for a long-term problem. It is

apparent that Hassel is correct in stating that environmental factors have been altered in

augmented lakes. However, without augmentation, many of the lakes would go dry, as

was the case of Loyce Lake prior to groundwater introduction.

Augmentation allows for lakes to be utilized for boating, swimming and other

recreational activities. It also allows for lake and wetland hydrology to be maintained

and fish and wildlife habitat to be provided. Further, fish, bird, reptilian, mammalian,

insect and aquatic plant populations were all seen in the augmented lakes in this study.

Without augmentation, it is likely fish would die, and the use of the lakes for recreational

purposes would be compromised.

The human population in the Tampa area is constantly increasing, along with the

demand for water. As the population further expands from Tampa into the suburbs, more

wellfields will be created, and more lakes will be affected. Similarly, as the existing

wellfields increase the amount of water they withdraw the cones of depression will

increase, affecting more lakes in the future. Granting more permits for lake level

augmentation with groundwater pumping will further alter limnological characteristics

until the demand for groundwater is decreased. However, lakes will still be able to be

utilized and fish populations will still be able to exist and reproduce, despite possible






47


shifts with altered environmental patterns. It has also been suggested that these shifts

were minimized with reduced levels of groundwater pumping.

Another change that could improve fish population parameters, and reduce changes

in plant community characteristics, is an increase in water level fluctuation (Bonvechio

and Allen in press). A more natural water level regime could be created by augmenting

during rainy seasons, and allowing the lakes to decrease in level during the dry season.














APPENDIX
COMMONLY HARVESTED FISH SPECIES

Table A-1. Commonly harvested fish species, and the total length (mm) at which they
are generally first harvested.

Species Name Scientific Name Size
(mm)

Chain pickerel Esox niger 400
Yellow bullhead Amerius natalis 280
Brown bullhead Ameirus nebulosus 280
White catfish Ameirus catus 280
Channel catfish Ictalurus punctatus 280
Largemouth bass Micropterus salmoides 280
Sunshine bass Morone chrysops x M. saxatilis 280
Black crappie Pomoxis nigromaculatus 240
Redbreast sunfish Lepomis auritus 200
Warmouth Lepomis gulosus 200
Bluegill Lepomis macrochirus 200
Redear sunfish Lepomis microlophus 200
Flier Centrarchus macropterus 200















LIST OF REFERENCES


Allen, M. S. 1999. Assessment of fish assemblages in Lakes Dosson, Halfmoon and
Round in Hillsborough County, Florida. Southwest Florida Water Management
District, Brooksville.

Allen, M. S., and K. I. Tugend. 2000. Effects of a large-scale habitat enhancement
project on habitat quality for age-0 largemouth bass at Lake Kissimmee, Florida.
Pages 265-276 in D. Phillipp and M. Ridgeway, editors. Black Bass: Ecology,
Conservation and Management. American Fisheries Society, Bethesda, Maryland.

American Public Health Association. 1989. Standard methods for the examination of
water and waste water. 17th edition. New York.

Bachmann, R. W., B. L. Jones, D. D. Fox, M. Hoyer, L. A. Bull, and D. E. Canfield, Jr.
1996. Relations between trophic state indicators and fish in Florida (U.S.A.)
Lakes. Canadian Journal of Fisheries and Aquatic Sciences 53:842-855.

Barbour, C. D., and J. H. Brown. 1974. Fish species diversity in lakes. American
Naturalist 108(962):473-489.

Bartos, L. F. 1998. Environmental augmentation. The resource regulation newsletter.
Southwest Florida Water Management District 10(3).

Bayne, D. R., M. J. Maceina, and W. C. Reeves. 1994. Zooplankton, fish and sport
fishing quality among four Alabama and Georgia reservoirs of varying trophic
status. Lake and Reservoir Management 8:153-163.

Belanger, T. V., and R. A. Kirkner. 1994. Groundwater/surface water interaction in a
Florida augmentation lake. Lake and Reservoir Management 8:165-174.

Biological Research Associates (BRA). 1982. Ecological impact of augmentation in
three lakes in Hillsborough County, Florida. Pinellas County Water System,
Tampa.

Biological Research Associates (BRA). 1996. General limnological assessment of three
augmented lakes in Northwest Hillsborough County, Florida. Pinellas County
Water System, Clearwater.

Biological Research Associates (BRA). 2001. Mountain Lake limnological evaluation
and lake management alternatives. Mountain Lake Corporation, Lake Wales,
Florida.









Brenner, M., J. M. Smoak, M. S. Allen, C. L. Schelske, and D. A. Leeper. 2000.
Biological accumulation of 226Ra in a groundwater-augmented Florida lake.
Limnology and Oceanography 45:710-715.

Canfield, D. E., Jr. 1981. Guide to the physiographic divisions of Florida. Report to
Florida Cooperative Extension Service. Institute of Food and Agricultural
Sciences. University of Florida, Gainesville.

Canfield, D. E., Jr., K. A. Langeland, S. B. Linda, and W. T. Haller. 1999. Relations
between water transparency and maximum depth of macrophyte colonization in
lakes. Journal of Aquatic Plant Management 23:25-28.

Canfield, D. E., Jr., and M. V. Hoyer. 1988. Regional geology and the chemical
andtrophic state characteristics of Florida Lakes. Lake and Reservoir Management
4(1):21-31.

Canfield, D. E., Jr., and M. V. Hoyer. 1990. A characterization of fish populations in
two central Florida Lakes. Final Report. Florida Turfgrass Association,
Gainesville.

Canfield, D. E., Jr., and M. V. Hoyer. 1992. Aquatic macrophytes and their relation to
the limnology of Florida lakes. Final Report submitted to the bureau of Aquatic
Plant Management, Florida Department of Natural Resources, Tallahassee.

Cowx, I. G. 2000. Potential impact of groundwater augmentation of river flows on
fisheries: a case study from the River Ouse, Yorkshire, UK. Fisheries Management
and Ecology 7:85-96.

Crumpton, W. G., T. M. Isenhart, and P. D. Mitchell. 1992. Nitrate and organic N
analysis with second-derivative spectroscopy. Limnology and Oceanography
37:907-913.

D'Elia, C. F., P. A. Steudler, and N. Corwin. 1977. Determination of total nitrogen in
aqueous samples using persulfate digestion. Limnology and Oceanography
22:760-764.

Dooris, P. M. 1982. Lake augmentation in northwest Hillsborough County. Technical
Report of the Southwest Florida Water Management District, Brookesville.

Dooris, P. M., and D. F. Martin. 1979. Groundwater induced changes in lake chemistry.
Ground Water 17:324-327.

Dooris, P. M., G. M. Dooris, and D. F. Martin. 1982. Phytoplankton responses to
ground water addition in central Florida Lakes. Water Resources Bulletin 18:335-
337.









Dooris, P. M. and R. J. Moresi. 1975. Evaluation of lake augmentation practices in
northwest Hillsborough County, Florida. Technical Report of the Southwest
Florida Water Management District, Brookesville.

Downing, J. A., C. Plante, and S. Lalonde. 1990. Fish reproduction correlated with
primary productivity, not the morphodoedaphic index. Canadian Journal of
Fisheries and Aquatic Sciences 47:1929-1936.

Engel, S., M. H. Hoff, M. T. Vogelsang Jr., K. E. Bass, and J. S. Anderson. 2000. Fish
population dynamics in Max Lake, a softwater Wisconsin lake subject to ground-
water pumping. Wisconsin Department of Natural Resources Research Report,
Woodruff.

Forsberg, C., and S. O. Ryding. 1980. Eutrophication parameters and trophic state
indices in 30 Swedish waste-receiving lakes. Archive fur Hydrobiologie 88:189-
207.

Frey, D. G. 1955. Distribution ecology of the cisco in Indiana. Investigation of Indiana
Lakes-Streams 4:177.

Gottgens, J. F. 1994. Redistribution of organic sediments in a shallow lake following a
short-term drawdown. Hydrobiologia 130:179-194.

Hach Chemical Company. 1975. Water and wastewater analysis procedures, third
edition. Ames, Iowa.

Hanson, J. M., and W. C. Leggett. 1982. Empirical prediction offish biomass and yield.
Canadian Journal of Fisheries and Aquatic Sciences 39: 257-263.

Hasler, A. D. 1947. Eutrophication of lakes by domestic drainage. Ecology 28:383-395.

Hassell, A. L. 1994. A chemical and biochemical characterization of lakes Cooper,
Strawberry, Crystal, Hobbs, Starvation, and Saddleback in Hillsborough County
(Florida). Master's thesis submitted to the Department of Chemistry at the
University of South Florida, Tampa.

Hassell, A. L., P. M. Dooris, and D. F. Martin. 1997. Maucha diagrams and chemical
analyses to diagnose changes in lake chemistry. Environmental Chemistry 60:75-
80.

Hinch, S. G. and N. C. Collins. 1993. Relationships of littoral fish abundance to water
chemistry and macrophyte variables in central Ontario lakes. Canadian Journal of
Fisheries and Aquatic Sciences 50:1870-1878.

Hoyer, M. V., and D. E. Canfield Jr. 1991. A phosphorus-fish standing crop relationship
for streams? Lake and Reservoir Management 7(1):25-32.









Jones, J. R., and M. V. Hoyer. 1982. Sportfish harvest predicted by summer chlorophyll
a concentration in Midwestern lakes and reservoirs. Transactions of the American
Fisheries Society 111:176-179.

Jones, K. C. 1978. Lake augmentation alternatives in Northwest Hillsborough Basin.
Memorandum. File No. 14-000-REG-31-00.

Kautz, E. S. 1980. Effects of eutrophication on the fish communities of Florida lakes.
Proceedings of Annual Conference of Southeastern Association of Fish and
Wildlife Agencies 34:67-80.

Kellar, A. E., and T. L. Crisman. 1990. Factors influencing fish assemblages and species
richness in subtropical Florida lakes and a comparison with temperate lakes.
Canadian Journal of Fisheries and Aquatic Sciences 47:2137-2146.

Krebs, C. J. 1999. Ecological methodology, second edition. Benjamin Cummings,
Menlo Park, California.

Florida Lakewatch. 2000. A beginner's guide to water management nutrients.
Information circular #102. Department of Fisheries and Aquatic Sciences,
University of Florida, Gainesville.

Larkin, P. A., and T. G. Northcote. 1969. Fish as indices of eutrophication. Pages 256-
273 in Eutrophication, causes, consequences, correctives. National Academy of
Sciences, Washington D. C.

Lee, G. F., P. E. Jones, and R. A. Jones. 1991. Effects of eutrophication on fisheries.
Review of Aquatic Sciences 5:287-305.

Maceina, M. J., and J. V. Shireman. 1980. The use of a recording fathometer for
determination of distribution and biomass of Hydrilla. Journal of Aquatic Plant
Management. 18:34-39.

Martin, D. F., D. M. Victor, and P. M. Dooris. 1976a. Effects of artificially introduced
ground water on the chemical and biochemical characteristics of six Hillsborough
County (Florida) Lakes. Water Research 10:65-69.

Martin, D. F., D. M. Victor, and P. M. Dooris. 1976b. Implications of lake augmentation
on Hydrilla growth. Environmental Science Engineering Al 1:245-253.

Matuszek, J. E., and G. L. Beggs. 1988. Fish species richness in relation to lake area,
pH, and other abiotic factors on Ontario lakes. Canadian journal of Fisheries and
Aquatic Sciences 45:1931-1941.

McConnell, W. J., S. Lewis, and J. E. Olson. 1977. Gross photosynthesis as an estimator
of potential fish production. Transactions of the American Fisheries Society
106:417-423.









McKinsey, D. M., and L. J. Chapman. 1998. Dissolved oxygen and fish distribution in a
Florida spring. Environmental Biology of Fishes 53:211-223.

Menzel, D. W., and N. Corwin. 1965. The measurement of total phosphorus in seawater
based on the liberation of organically bound fractions of persulfate oxidation.
Limnology and Oceanography 10:280-282.

Murphy, J., and J. P. Riley. 1962. A modified single solution method for the
determination of phosphate in natural waters. Analytica Chimica Acta 27:31-36.

Myers, R. H. 1990. Classical and modern regression with applications, second edition.
PWS-Kent Publishing Company, Boston, Massachusetts.

Ney, J. J. 1996. Oligotrophication and its discontents: effects of reduced nutrient loading
on reservoir fisheries. Pages 285-295 in L. E. Miranda and D. R. Devries, editors.
Multidimensional approaches to reservoir fisheries management. American
Fisheries Society, Bathesda, Maryland.

Ogelsby, R. F. 1977. Relationships of fish yield to lake phytoplankton standing crop,
production, and morphoedaphic factors. Journal of Fisheries Research Board of
Canada 34:2271-2279.

Palmer, M. W. 1993. Putting things in even better order: the advantages of canonical
correspondence analysis. Ecology 74(8):2215-2230.

PC-ORD. 1999. Multivariate analysis of ecological data, version 4. MjM Software
Design, Gleneden Beach, Oregon.

Persson, L., Andersson, G., Hamrin, S. F., and Johansson, L. 1988. Predator regulation
and primary production along the productivity gradient of temperate lake
ecosystems. Pages 45-65 in S. R. Carpenter., editor. Complex interactions in
freshwater ecosystems. Springer-Verlag, New York.

Reynolds, J. B. 1996. Electrofishing. Pages 221-253 in B. R. Murphy and D. W. Willis,
editors. Fisheries Techniques, second edition. American Fisheries Society.
Bethesda, Maryland.

Sartory, D. P., and J. U. Grobbelaar. 1984. Extraction of chlorophyll a from freshwater
phytoplankton for spectrophotometric analysis. Hydrobiologia 114:177-187.

Statistical Analysis Systems (SAS). 1996. SAS statistics user's guide. SAS Institute,
Inc., Cary, North Carolina.

Shafer, M. D., R. E. Dickinson, J. P. Heaney and W. C. Huber. 1986. Gazetteer of
Florida lakes. Florida Water Resources Research Center, Publication 96,
Gainesville, Florida.









Simal, J., M. A. Lage, and I. Iglesias. 1985. Second derivative ultraviolet spectroscopy
and sulfamic acid method for determination of nitrates in water. Journal -
Association of Official Analytical Chemists 68:962-964.

Sinclair, W. C. 1977. Experimental study of artificial recharge alternatives in Northwest
Hillsborough County, Florida. United States Geological Survey: Water-Resources
Investigations 77-13.

Stewart, J. W. 1968. Hydrologic effects of pumping from the Floridan Aquifer in
Northwest Hillsborough, Northeast Pinellas, and Southwest Pasco Counties,
Florida. United States Geological Survey, Tallahassee.

Stewart, J. W., and G. H. Hughes. 1974. Hydrologic consequences of using groundwater
to maintain lake levels affected by water wells near Tampa, Florida. Florida
Department of Natural Resources, Tallahassee.

Southwest Florida Water Management District (SWFWMD). 1998. Water supply
assessment 1995-2020. Resource Projects Department, Southwest Florida Water
Management District, Brookesville.

Ter Braak, C. J. F. 1986. Canonical correspondence analysis: a new eigenvector
technique for multivariate direct gradient analysis. Ecology 67(5): 1167-1179.

Wollin, K. M. 1987. Nitrate determination in surface waters as an example of the
application of UV derivative spectrometry to environmental analysis. Acta
Hydrochemica Hydrobiologia 15:459-469.

Yurk, J. J., and J. J. Ney. 1989. Phosphorus-fish community biomass relationships in
southern Appalachian reservoirs: can lakes be too clean for fish? Lake and
Reservoir Management 5(2):83-90.

Zar, J. H. 1999. Biostatistical analysis, fourth edition. Prentice-Hall, Upper Saddle
River, NJ.















BIOGRAPHICAL SKETCH

Patrick Cooney was born and raised in El Dorado Hills, a small and fast growing

town in the Sierra-Nevada Mountain Range in northeastern California. Upon graduating

from high school, he moved to Miami, Florida, to pursue a degree in both biology and

marine science from the University of Miami. During this time, he enjoyed studying

abroad and conducting research on marine life and marine ecosystems in Townsville,

Australia, while attending James Cook University, near the Great Barrier Reef. After

graduating from the University of Miami, he promptly moved to Bahia de Kino, Mexico

(located on the Sea of Cortez), to study an obligate mutualism between the Senita moth

and the Senita cactus, and fish for Dorado, Coryphaena hippurus. Once this research was

completed, he returned to Florida to spend time with friends and conduct research on

freshwater fish and water chemistry at the University of Florida in Gainesville. He will

continue to feed his hunger for knowledge, and teach others the knowledge he has

attained.