<%BANNER%>

Habitat Mediated Community Structure within Spring-fed, Coastal Rivers

Permanent Link: http://ufdc.ufl.edu/UFE0043606/00001

Material Information

Title: Habitat Mediated Community Structure within Spring-fed, Coastal Rivers
Physical Description: 1 online resource (179 p.)
Language: english
Creator: Lauretta, Matthew V
Publisher: University of Florida
Place of Publication: Gainesville, Fla.
Publication Date: 2011

Subjects

Subjects / Keywords: communities -- fishes -- habitat -- invertebrates -- macrophytes -- rivers
Forest Resources and Conservation -- Dissertations, Academic -- UF
Genre: Fisheries and Aquatic Sciences thesis, Ph.D.
bibliography   ( marcgt )
theses   ( marcgt )
government publication (state, provincial, terriorial, dependent)   ( marcgt )
born-digital   ( sobekcm )
Electronic Thesis or Dissertation

Notes

Abstract: Vegetation plays a central role in structuring aquatic ecosystems by altering biogeochemical processes and mediating trophic interactions between fishes and invertebrates. The loss of key vegetative habitat components can alter community structure, and lead to the loss of ecosystem function and services. The goal of this study was to quantitatively assess the effects of macrophyte loss on fish and invertebrate populations within spring-fed, coastal rivers. To accomplish this, I conducted a comparative ecosystem study of two rivers, the Chassahowitzka and Homosassa rivers, where vegetation loss has been disparate over the last 12 years. I sampled aquatic vegetation, invertebrates, and fishes in each river over a three-year period to estimate the community composition and biomass, and examined the diet habits of freshwater and marine fishes. Using empirically derived estimates of community biomass and trophic interactions, I constructed a trophic mass-balance model of the Chassahowitzka River food web and ran time-dynamic simulations to predict the response of fish and invertebrate populations to the extirpation of macrophytes. I compared predicted estimates with the observed community structure of the Homosassa River, where macrophytes have been absent for nearly a decade. Overall, macrophyte extirpation was predicted to result in a 60% reduction in invertebrate biomass and 11% reduction in fish biomass, whereas restoration was predicted to increase invertebrate biomass by 152% and fish biomass by 73%. Observed spatial patterns between rivers validated model predictions for most taxa, including the local extinction of select freshwater groups. This research exemplified the complex trophic interactions that structure aquatic food webs. As vegetative communities shift from highly-structured macrophyte dominated assemblages to boom-and-bust filamentous algae production, an associated shift in primary and secondary food bases is expected to have compound effects on predator populations, including altered prey composition and population dynamics. Predators that forage on a wide range of fish and invertebrate taxa are likely to switch dominant prey types, while specialist species may decline or, in extreme cases, be extirpated from the system. The long-term ecological and socioeconomic consequences of the predicted changes in community structure of coastal river ecosystems remain unknown.
General Note: In the series University of Florida Digital Collections.
General Note: Includes vita.
Bibliography: Includes bibliographical references.
Source of Description: Description based on online resource; title from PDF title page.
Source of Description: This bibliographic record is available under the Creative Commons CC0 public domain dedication. The University of Florida Libraries, as creator of this bibliographic record, has waived all rights to it worldwide under copyright law, including all related and neighboring rights, to the extent allowed by law.
Statement of Responsibility: by Matthew V Lauretta.
Thesis: Thesis (Ph.D.)--University of Florida, 2011.
Local: Adviser: Frazer, Tom K.
Local: Co-adviser: Pine, William.

Record Information

Source Institution: UFRGP
Rights Management: Applicable rights reserved.
Classification: lcc - LD1780 2011
System ID: UFE0043606:00001

Permanent Link: http://ufdc.ufl.edu/UFE0043606/00001

Material Information

Title: Habitat Mediated Community Structure within Spring-fed, Coastal Rivers
Physical Description: 1 online resource (179 p.)
Language: english
Creator: Lauretta, Matthew V
Publisher: University of Florida
Place of Publication: Gainesville, Fla.
Publication Date: 2011

Subjects

Subjects / Keywords: communities -- fishes -- habitat -- invertebrates -- macrophytes -- rivers
Forest Resources and Conservation -- Dissertations, Academic -- UF
Genre: Fisheries and Aquatic Sciences thesis, Ph.D.
bibliography   ( marcgt )
theses   ( marcgt )
government publication (state, provincial, terriorial, dependent)   ( marcgt )
born-digital   ( sobekcm )
Electronic Thesis or Dissertation

Notes

Abstract: Vegetation plays a central role in structuring aquatic ecosystems by altering biogeochemical processes and mediating trophic interactions between fishes and invertebrates. The loss of key vegetative habitat components can alter community structure, and lead to the loss of ecosystem function and services. The goal of this study was to quantitatively assess the effects of macrophyte loss on fish and invertebrate populations within spring-fed, coastal rivers. To accomplish this, I conducted a comparative ecosystem study of two rivers, the Chassahowitzka and Homosassa rivers, where vegetation loss has been disparate over the last 12 years. I sampled aquatic vegetation, invertebrates, and fishes in each river over a three-year period to estimate the community composition and biomass, and examined the diet habits of freshwater and marine fishes. Using empirically derived estimates of community biomass and trophic interactions, I constructed a trophic mass-balance model of the Chassahowitzka River food web and ran time-dynamic simulations to predict the response of fish and invertebrate populations to the extirpation of macrophytes. I compared predicted estimates with the observed community structure of the Homosassa River, where macrophytes have been absent for nearly a decade. Overall, macrophyte extirpation was predicted to result in a 60% reduction in invertebrate biomass and 11% reduction in fish biomass, whereas restoration was predicted to increase invertebrate biomass by 152% and fish biomass by 73%. Observed spatial patterns between rivers validated model predictions for most taxa, including the local extinction of select freshwater groups. This research exemplified the complex trophic interactions that structure aquatic food webs. As vegetative communities shift from highly-structured macrophyte dominated assemblages to boom-and-bust filamentous algae production, an associated shift in primary and secondary food bases is expected to have compound effects on predator populations, including altered prey composition and population dynamics. Predators that forage on a wide range of fish and invertebrate taxa are likely to switch dominant prey types, while specialist species may decline or, in extreme cases, be extirpated from the system. The long-term ecological and socioeconomic consequences of the predicted changes in community structure of coastal river ecosystems remain unknown.
General Note: In the series University of Florida Digital Collections.
General Note: Includes vita.
Bibliography: Includes bibliographical references.
Source of Description: Description based on online resource; title from PDF title page.
Source of Description: This bibliographic record is available under the Creative Commons CC0 public domain dedication. The University of Florida Libraries, as creator of this bibliographic record, has waived all rights to it worldwide under copyright law, including all related and neighboring rights, to the extent allowed by law.
Statement of Responsibility: by Matthew V Lauretta.
Thesis: Thesis (Ph.D.)--University of Florida, 2011.
Local: Adviser: Frazer, Tom K.
Local: Co-adviser: Pine, William.

Record Information

Source Institution: UFRGP
Rights Management: Applicable rights reserved.
Classification: lcc - LD1780 2011
System ID: UFE0043606:00001


This item has the following downloads:


Full Text

PAGE 1

1 HABITAT MEDIATED COMMUNITY STRUCTURE WI THIN SPRING FED, COASTAL RIVERS By MATTHEW VINCENT LAURETTA A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS F OR THE DEGREE OF DOCTOR OF PHILOSOPHY UNIVERSITY OF FLORIDA 2011

PAGE 2

2 201 1 Matthew Vincent Lauretta

PAGE 3

3 ACKNOWLEDGMENTS This research was funded by the Florida Fish and Wildlife Conservation Wildlife Legacy Initiative and the U.S. State Wildlife Grant Program. I thank the Linton E. Grinter Graduate Research Fellow ship Program, which greatly improved my quality of life at UF Travel grants to present research findings at seven professional c onferences w ere provided by the Florida Chapter of the American Fisheries Society ( five grants) the Students United in the Research of Fisheries ( one grant) and the University of Florida Graduate Student Council ( one grant) I thank E Nagid, W Strong and T Tuten for assistance with sampling and logistic al support I thank A. Dutterer and M Edwards for helping with all aspects of data collection I thank C Miller, A Williams, Z Martin, A Cichra E. Buttermore, E Camp, and J Bernatis for helping with sample collection, laboratory analys e s and data processing. I thank B Baker D Gwinn and J Flowers for assistance with sampling I thank J Tetzlaff, M Rogers and M Catalano for creative feedback on the project I th ank my graduate advisors T K. Fraze r, W E. Pine, M S. Allen, C J. Walters and M J Cohen, for challenging me to think critically and providing endless opportunities for learning and professional development. I thank M Boyle and W. Leibfried, for professional mentorship. I thank my un dergraduate advisors at Northern Arizona University, G Koch, N Cobb, and R Foust for introducing me to chemical and ecological research. I thank my family for a lifetime of love and support. I thank my friends for adventure, laughter and lightheartedne ss. Finally, I thank my wife, Sarah, for showing me true love and happiness

PAGE 4

4 TABLE OF CONTENTS P age ACKNOWLEDGMENTS ................................ ................................ ................................ .. 3 LIST OF TABLES ................................ ................................ ................................ ............ 6 LIST OF FIGURES ................................ ................................ ................................ .......... 8 LIST OF ABBREVIATIONS ................................ ................................ ........................... 11 ABSTRACT ................................ ................................ ................................ ................... 12 CHAPTER 1 GENERAL INTRODUCTION ................................ ................................ .................. 14 Introduction ................................ ................................ ................................ ............. 14 Study Objectives ................................ ................................ ................................ ..... 18 2 GEAR CATCHABILITY OF FI SHES AND INVERTEBRATES IN COASTAL RIVERS ................................ ................................ ................................ .................. 22 Introduction ................................ ................................ ................................ ............. 22 Methods ................................ ................................ ................................ .................. 24 Capture Recapture Electrofishing Analysis ................................ ...................... 27 Seine and Throw Trap Removal Sampling Analysis ................................ ......... 28 Distributions of Gear Catc hability Estimates ................................ ..................... 28 Tests for Spatial, Temporal and Interspecific Heterogeneity in Catchability ..... 29 Estimation of Mean Populati on Densities of Fishes and Invertebrates ............. 30 Results ................................ ................................ ................................ .................... 30 Electrofishing Catchability ................................ ................................ ................ 30 Seine Catchability ................................ ................................ ............................. 31 Throw Trap Catchability ................................ ................................ .................... 32 Relative Abundances versus Absolute Densities ................................ ............. 33 Discussion ................................ ................................ ................................ .............. 34 3 THE COMPOSITION AND BIOMASS OF THE AQUATIC COMMUNITIES WITHIN THE CHASSAHOWITZKA AND HOMOSASSA RIVERS ......................... 46 Introduction ................................ ................................ ................................ ............. 46 Methods ................................ ................................ ................................ .................. 49 Submersed Aquatic Vegetation Sampling and Analyses ................................ .. 50 Invertebrate Sampling and Analyses ................................ ................................ 50 Fish Sampling and Analyses ................................ ................................ ............ 53 Results ................................ ................................ ................................ .................... 56 Submersed Aquatic Vegetation ................................ ................................ ........ 56 Invertebrates ................................ ................................ ................................ .... 57 Small bodied Fishes ................................ ................................ ......................... 59 Large bodied Fishes ................................ ................................ ......................... 60 Discussion ................................ ................................ ................................ .............. 61

PAGE 5

5 4 THE DIET HABITS OF FRESHWATER AND MARINE FISHES IN COASTAL RIVERS ................................ ................................ ................................ .................. 90 Introduction ................................ ................................ ................................ ............. 90 Methods ................................ ................................ ................................ .................. 95 Diet Sampling and Laboratory Procedures ................................ ....................... 95 Prey Composition Indices ................................ ................................ ................. 97 Prey Selectivity Indices ................................ ................................ ..................... 99 Relative Foraging Success ................................ ................................ ............... 99 Results ................................ ................................ ................................ .................. 101 Lepomis punctatus ................................ ................................ ......................... 101 Prey composition ................................ ................................ ..................... 101 Prey selectivity ................................ ................................ ......................... 103 Relative foraging su ccess ................................ ................................ ........ 104 Micropterus salmoides ................................ ................................ .................... 104 Prey composition ................................ ................................ ..................... 104 Prey se lectivity ................................ ................................ ......................... 107 Relative foraging success ................................ ................................ ........ 107 Lagodon rhomboides ................................ ................................ ...................... 108 Prey composition ................................ ................................ ..................... 108 Prey selectivity ................................ ................................ ......................... 109 Relative foraging success ................................ ................................ ........ 109 Lutjanus griseus ................................ ................................ ............................. 110 Prey composition ................................ ................................ ..................... 110 Prey selectivity ................................ ................................ ......................... 112 Relative foraging success ................................ ................................ ........ 113 Discussion ................................ ................................ ................................ ............ 113 5 VEGETATIVE HABITAT LOSS EFFECTS ON FISH AND INVERTEBRATE COMMUNITY STRU CTURE IN SPRING FED, COASTAL RIVERS .................... 136 Introduction ................................ ................................ ................................ ........... 136 Methods ................................ ................................ ................................ ................ 139 Trophic Mass balance Model of the Chassahowitzka River Food Web .......... 139 Time dynamic Simulation of Alternative Management Scenarios ................... 141 Results ................................ ................................ ................................ .................. 143 Discussion ................................ ................................ ................................ ............ 145 6 SYNTHESIS AND FUTURE RESEARCH ................................ ............................. 163 LIST OF REFERENCES ................................ ................................ ............................. 167 BIOGRAPHICAL SKETCH ................................ ................................ .......................... 179

PAGE 6

6 LIST OF TABLES Table P age 2 1 List of e quations used to e stimate e lectrofishing c atchability from c losed m ark r ecapture and removal s ampling ................................ ............................... 40 2 2 Akaike information criteria and model weighting of spatial, temporal and interspecific heterogeneity in the catchabi lity of fishes and invertebrates ........... 41 2 3 Mean catchability and 95 th percentile lower and upper limits of catchability for individual taxa sampled by e lectrofish ing, seining and throw trapping ................ 42 3 1 Mean estimated biomass of plants, algae, invertebrates and fishes within the Chassahowitzk a and Homosassa r ivers, Florida ................................ ................ 65 3 2 Freshwater fish species captured within th e Chassahowitzka River, Florida ...... 66 3 3 Saltwater fish species captured within th e Chassahowitzka Riv er, Florida ......... 67 3 4 Freshwater fish species captured within the Homosassa River, Flor ida ............. 68 3 5 Saltwater fish species captured w ith in the Homosassa River, Florida ................ 69 4 1 Mean proportion by dry weight of common prey taxa observed in stomachs of Lepomis punctatus from the Chassahowitzka and Homosassa r ivers, Florida. 119 4 2 Percent frequency of occurrence of common prey taxa observed in stomachs of Lepomis punctatus from the Chassahowitzk a and Homosassa r ivers, Florida ................................ ................................ ................................ .............. 120 4 3 Manly Chesson prey selectivity indices for Lepomis punctatus from the Chassahowitzk a and Homosassa rivers, Florida ................................ .............. 121 4 4 Mean proportion by dry weight of com mon prey taxa observed in stomachs of Micropterus salmoides from the Chassahowitzka and Homosassa r ivers Florida ................................ ................................ ................................ .............. 122 4 5 Percent frequency of occurrence of common prey taxa observed in stomac hs of Micropterus salmoides from the Chassahowitzk a and Homosassa r ivers, Florida ................................ ................................ ................................ .............. 123 4 6 Manly Chesson prey selectivity indices for Micropterus salmoides from the Chassahowitzk a and Homosassa rivers, Florida ................................ .............. 124 4 7 Mean proportion by dry weight of common prey taxa observed in stomachs of Lagodon rhomboides from the Chassahowitzk a and Homosassa r ivers, Florida ................................ ................................ ................................ .............. 125

PAGE 7

7 4 8 Percent frequency of occurrence of common prey taxa observed in stomachs of Lagodon rhomboides from the Chassahowitzk a and Homosassa r ivers, Florida ................................ ................................ ................................ .............. 126 4 9 Manly Chesson prey selectivity indices for Lagodon rhomboides from the Chassahowitzk a and Homosassa rivers, Florida ................................ .............. 127 4 10 Mean proportion by dry weight of common prey taxa observed in stomachs of Lutjanus griseus from the Chassahowitzk a and Homosassa r ivers, Florida ..... 128 4 11 Percent frequency of occurrence of common prey taxa observed in stomachs of Lutjanus gri seus from the Chassahowitzka and Homosass a r ivers, Florida 129 4 12 Manly Chesson prey selectivity indices for Lutjanus griseus from the Chassahowitzka and Homosassa rivers, Florida ................................ .............. 130 5 1 Trophic groups and taxa composition included in the Ecopath trophic mass balance model of th e Chassahowitzka River food web ................................ .... 153 5 2 Data sources for the Ecopath trophic mass balance of th e Chassahowitzka River food web ................................ ................................ ................................ .. 154 5 3 Basic inputs for the Ecopath trophic mass balance model of th e Chassahowitzka River food web ................................ ................................ ....... 155 5 4 Diet composition of consumers within th e Chassahowitzka and Homosassa r iver s Florida ................................ ................................ ................................ .... 156 5 5 Detrital fate matrix for the Ecopath trophic mass balance model of th e Chassahowitzka River food web ................................ ................................ ....... 157

PAGE 8

8 LIST OF FIGURES Figure P age 1 1 Location of the Chassahowitzka and Hom osassa rivers in Hernan do and Citrus counties, Florida ................................ ................................ ....................... 19 1 2 Physical and chemical characteristics of the Chassahowitzka and Homosassa rivers between 1998 and 2010 ................................ ....................... 20 1 3 Long term pattern s in submersed aquatic vegetation biomass within the Chassahowitzk a and Homosassa rivers, Florida ................................ ................ 21 2 1 Study reaches within the H omosassa and Chassahowitzka rivers, Florida ....... 43 2 2 Histograms of observed catchability estima tes across all taxa measured .......... 44 2 3 Relative abundance indices and absolute density estimates of taxa commonly detected during boat electrofishing, seine and throw trap sampling within the Chas sahowitzka and Homosassa rivers ................................ ............. 45 3 1 Average biomass of macrophytes within the Chassahowitzka and Homosassa rivers during August 2007 through August 2010 ............................. 70 3 2 Average biomass of filamentous algae within the Chassahowitzka and Homosassa rivers during August 2007 through August 2010 ............................. 71 3 3 Average density and biomass of amphipods within the Chassahowitzka and Homosassa rivers during August 2007 through Feb ruary 2010 ......................... 72 3 4 Average density and biomass of aquatic insects within the Chassahowitzka and Homosassa rivers during Au gust 2007 through February 2010 ................... 73 3 5 Average density and biomass of gastropods within the Chassahowitzka and Homosassa rivers during Au gust 2007 through February 2010 ......................... 74 3 6 Average density and biomass o f isopods within the Chassahowitzka and Homosassa rivers during Au gust 2007 through February 2010 ......................... 75 3 7 Average density and biomass of tanaids within the Chassahowitzka and Homosassa rivers d uring Au gust 2007 through February 2010 ......................... 76 3 8 Average biomass of freshwater small bodied fishes collected at seine depletion sites within the Chassa howitzka and Homosassa rivers ..................... 77 3 9 Average biomass of saltwater small bodied fishes collected at seine depletion sites within the Chas sahowitzka and Homosassa rivers ..................... 78

PAGE 9

9 3 10 Estimated mean biomass of lake chubsucker captured during mark recapture electrofishing sampling within the Chassah owitzka and Homosassa rivers ....... 79 3 11 Estimated mean biomass of Lepomis s pp. captured during mark recapture electrofishing sampling within the Chassa howitzka and Homosassa rivers ....... 80 3 12 Estimated mean biomass of American eel captured during mark recapture electrof ishing sampling within the Chassa howitzka and Homosassa rivers ....... 81 3 13 Estimated mean biomass of gar captured during mark recapture electrofishing sampling within the Chassa howitzka and Homosa ssa rivers ....... 82 3 14 Estimated mean biomass of largemouth bass captured during mark recapture electrofishing sampling within the Chassahowitzka a nd Homosassa rivers ................................ ................................ ............................... 83 3 15 Estimated mean biomass of striped mullet captured during mark recapture electrofishing sampling within the Chassa howitzka and Homosassa rivers ....... 84 3 1 6 Estimated mean biomass of pinfish captured during mark recapture electrofishing sampling within the Chassa howitzka and Homosassa rivers ....... 85 3 17 Estimated mean biomass of sheepshead captured during mark recapture electrofishing sampling within the Chassa howitzka and Homosassa rivers ....... 86 3 18 Estimated mean biomass of gray snapper captured during mark recapture electrofishing sampli ng within the Chassa howitzka and Homosassa rivers ....... 87 3 19 Estimated mean biomass of red drum captured during mark recapture electrofishing sampling within the Chassa howitzka and Homosassa rive rs ....... 88 3 20 Estimated mean biomass of common snook captured during mark recapture electrofishing sampling within the Chassa howitzka and Homosassa rivers ....... 89 4 1 Mean estimated biomass of filamentous algae and macrophytes within the Chassahowitzka and Homosassa rive rs during the period of study .................. 131 4 2 Seasonal and intera nnual pattern s in mean proportion of empty stomach s of Lepomis punctatus, Micropterus salmoides, Lagodon rhomboides and Lutjanus griseus from the Chas sahowitzka and Homosassa rivers .................. 132 4 3 I ntra annual pattern s in mean proportion of empty stomach s of Lepomis punctatus, Micropterus salmoides, Lagodon rhomboides and Lutjanus griseus within the Chas sahowitzka and Homosassa rivers .............................. 133

PAGE 10

10 4 4 Seasonal and interannual pattern s in mean total prey dry weight per predator body weight of Lepomis punctatus, Micropterus salmoides, Lagodon rhomboides and Lutjanus griseus within the Chas sahowitzka and Homosassa rivers ................................ ................................ ............................. 134 4 5 Intra annual pattern s in mean total prey dry weight per predator body weight of Lepomis punctatus, Micropterus salmoides, Lagodon rhomboides and Lutjanus griseus within the Chassahowitzka and Homosassa rivers. ............... 135 5 1 Ecosim forcing functions used to simulate changes in primary production within the Chassahowitzka River, Florida under alternative management scenarios of macrophyte extirpation versus restoration ................................ .... 158 5 2 Ecopath trophic flow diagra m of the Chassahowitzka River ............................. 159 5 3 Predicted ecotrophic efficiency of trophic groups within the Chas sahowitzka River food web model ................................ ................................ ....................... 160 5 4 Comparison of time dynamic ecosystem model predicted changes in mean annual biomass of trophic groups versus observed spatial dif ferences between the Chas sahowitzka and Homosassa rivers ................................ ...... 161 5 5 Comparison of time dynamic ecosys te m model predicted community responses to the extirpation and restoration of macrophy tes in t he Chassahowitzka River ................................ ................................ ...................... 162

PAGE 11

11 LIST OF ABBREVIATION S Equation Variables and Parameters A Sample a rea B Population b iomass C Number of c aptures per s ample D Population d ensity E Sample e ffort LL Log likelihood LN Natural l og arithm M Number of m arked i ndividuals MW Mean p rey p roportion by d ry w eight MWB Mean p rey d ry w eight per p redator b ody w eight N Population a bundance p Probability of c apture q Catchability c oefficient R Number of r ecaptured i ndividuals w Average w eight per i ndividual Water Quality Variables ALK Alkalinity m easured as c alcium c arbonate NO3 Nitrate SAL Salinity SAV Submersed a quatic v egetation ( m acrophytes and f ilamentous a lgae c ombined ) SRP Soluble r eactive p hosphorus TEMP Temperature

PAGE 12

12 Abstract of Dissertati on Presented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy HABITAT MEDIATED COMMUNITY STRUCTURE WITHIN SPRING FED COASTAL RIVERS By Matthew Vincent Lauretta Decemb er 2011 Chair: Thomas K. Frazer Cochair: William E. Pine, III Major: Fisheries and Aquatic Science s V egetation pla ys a central role in structuring aquatic ecosystems by alter ing biogeochemical processes and mediating trophic int eractions between fishes and invertebrates. The loss of key vegetative habitat components can alter community structure and lead to the loss of ecosystem function and services. The goal of this study was to quantitatively assess t he effects of macrophyt e loss on fish and invertebrate population s within spring fed, coastal rivers To accomplish this, I conducted a comparative ecosystem study of two rivers, the Chassahowitzka and Homosassa rivers, where vegetation loss has been disparate over the last 12 years I sampled aquatic vegetation, invertebrates, and fishe s in each river over a three year period to estimate the community composition and biomass, and examine d the diet habits of freshwater and marine fishes Using empirical ly derived estimates of community biomass and trophic interactions I constructed a trophic mass balance model of the Chassahowitzka River food web and ran time dynamic simulations to predict the response of fish and invertebrate population s to the extirpation of macrophytes I compared predicted estimates with the observed community structure

PAGE 13

13 of the Homosassa River, where macrophyte s have been absent for nearly a decade. Overall, macrophyte extirpation was predicted to result in a 60% reduction in invertebrate biomass and 11% reduction in fish biomass wh ereas restoration was predicted to increase invertebrate biomass by 152% and fish biomass by 73%. Observed spatial pattern s between rivers validat ed model predictions for most taxa, including the local extinction of select fre shwater group s This research exemplified the complex trophic interactions that structure aquatic food webs As vegetati ve communities shift from highly structured macrophyte dominated assemblages to boom and bust filamentous alga e production an associa ted shift in primary and secondary food base s is expected to have compound effects on predator population s including altered prey composition and population dynamics Predator s that forage on a wide range of fish and invertebrate taxa are likely to switc h dominant prey types, while specialist species may decline or, in extreme cases, be extirpated from the system. The long term ecological and socioeconomic co nsequences of the predicted changes in community structure of coastal river ecosystems remain unk nown.

PAGE 14

14 CHAPTER 1 GENERAL INTRODUCTION Introduction Habitat loss is a principal cause of species decline and imperilment ( Wilcox and Murphy 1985 ) and a significant factor contributing to the global reduction in biodiversity ( Fahrig 1997 2003 ). Direct e f fects of habitat loss on animal s include lower e d breeding success, dispersal, foraging success and survival ( Fahrig 2003 ) T hese direct effect s in turn, influence species level population dynamic s that can ultimately affect entire communities Indirect effects of h abitat loss on communities occur primarily as a consequence of alter ed trophic interactions among populations ( Taylor and Merriam 1995 ) For example, a r eduction in habitat complexit y associated with the loss of habitat (e.g., macrophyte exti rpation river channelization deforestation ) can increase predation vulnerability and mortality of prey species (Crowder and Cooper 1982 Power et al. 1996 Becker et al. 2009 ) and result in depletion of prey resources followed by subsequent predator pop ulation crash es (Huffa ker 1958 Hastings 1977 Sutherland and Dolman 1994 ). T he combination of refuge, foraging and reproductive habitat loss reduce s the number of specialist species in a community (Munday 2004) and shorten s food chain length ( Power et a l. 1996, Kom o nen et al. 2000) In combination, the direct and indirect effects of habitat loss can alter community and food web structure ( Harrison and Bruna 1999 Coll et al. 2011 ) with negative consequences on ecosystems and the services that they prov ide ( Dobson et al. 2006 ) Quantifyin g the relationship between key habitat components and population s of animal s is essential to predicting (and potentially mitigating) negative species and communit y level respon ses to large scale changes in habitat.

PAGE 15

15 Mac rophytes are a dominant structural element in many freshwater and coastal ecosystems and play a key role in a number of biogeochemical and ecological processes (Carpenter and Lodge 1986, Jeppesen et al. 1998). P lant s alter the physical and chemical condi tions of the water and sediment; influence nutrient cycling, primary production, and the processing of organic matter; and mediate biotic interactions (Jeppesen et al. 1998). As a result of their high productivity and structural complexity vegetative hab itats support a relatively high abundance and diversity of fishes and invertebrates compared to alternative habitats (e.g., Heck et al. 1995, Randall et al. 1996, West and King 1996 Guidetti 2000 ). Many freshwater and marine fishes and invertebrates util ize vegetative habitats at various phase s of their life cycles. For example, seagrass communities along coastlines provide important juvenile rearing habitat for marine fishes and invertebrates including stocks that support economically important fisheri es such as red drum ( Sciaenops ocellatus ) spotted seatrout ( Mycteroperca microlepis ) and blue crab ( Callinectes sapidus ) ( Stunz et al. 2002, Heck et al. 2003, Neahr et al. 2010 Mizerek et al. 2011) Despite t he large amount of research dedicated to id entifying the role of aquatic vegetation in mediating ecosystem processes, considerably less quantitative research has been applied towards understanding the dynamics between vegetative habitat composition and faunal community structure. Changes in the phy sical and chemical properties of the aquatic environment can lead to large scale shift s i n the composition of primary producers with subsequent e ffects on food webs (Deegan et al. 2002) For example, eutrophication driven shifts in the composition of pla nts and algae may result in the replacement of macrophytes with

PAGE 16

16 benthic algae or phytoplankton (Duarte 1995) and potentially alter species interactions and population dynamics (Be t toli et al. 199 2 Deegan et al. 2002). Changes in aquatic vegetation can a ffect trophic interactions by altering refuge and foraging habitat for associated faunal organisms (Gotceitas and Colgan 1989, Heck and Crowder 1991). A quatic vegetation has been shown to inhibit predator foraging (Crowder and Cooper 1982 Savino and Stei n 1982 ), thereby creating refugia for prey species (Gotceitas and Colgan 1989). Conversely, decreased vegetative cover may increase predation mortality and impact the structure of prey populations (Bettoli et al. 1992). Decline s in vegetation composition and biomass can further impact the food base of fishes via competitive exclusion of species (Peterson et al. 1993) and decrease d juvenile rearing habitat (Sass et al. 2006 ). These factors can in turn influence population dynamics (Peterson et al. 1993, Richardson et al. 1998) and species interactions with potential community level consequences (Crowder and Cooper 1982, He and Kitchell 1990, Bettoli et al 1992). Spring fed rivers in Florida serve as model eco systems to study the role that primary producer s play in structuring faunal communit ies Springs in Florida have long been recognized as optimal systems for ecological study due to their relatively sta ble physical and chemical properties ; abundant aquatic vegetation fish and invertebrate communities ; and high rates of primary productivity (Odum 195 3 Odum 195 7 ) Historically, spring fed rivers, including the Chassahowitzka and Homosassa rivers, supported dense assemblages of macrophytes such as Vallisneria americana Sagittaria kurziana and Potamog eton spp Over the last decade, however, a precipitous decline in vegetation biomass has been documented in these systems

PAGE 17

17 (Frazer et al. 2006). The loss of rooted aquatic vegetation which provides forage and refuge habitat is likely to alter predator p rey dynamic s and other important species level interactions ( Crowder and Cooper 1982 ) Such alterations may lead to undesirable shifts in fish and invertebrate community composition and possibly the loss of key species ( Pillay et al. 2010, Nakamura 2010 ) Due to the relatively sta ble abiotic conditions and long term datasets on vegetation composition and biomass, spring fed rivers in Florida provide a unique opportunit y to study how habitat (submersed aquatic vegetation) mediates the composition and troph ic dynamics of faunal groups and ultimately influences ecosystem structure and function. Th e research carried out and described herein allowed for an evaluat ion of the effects of changes in submersed aquatic vegetation and resulting loss of structural h abitat ( extirpation of macrophytes and replacement with filamentous macroalgae) on invertebrate and fish community composition and trophic interactions with in spring fed, coast a l rivers in Florida A comparative ecosystem study of two spring fed rivers al ong the Gulf of Mexico coast (Figure 1 1) was conducted, and a trophic mass balance model was develop ed based on empirical ly derived estimates of key population level parameters and predator prey association s within the Chassahowitzka River The time dyn amic trophic model was used to predict the effects of macrophyte extirpation on fish and invertebrate communit y composition, biomass and trophic dynamic s T he predictions were compared with the observ ed community structure of the Homosassa River, where ma crophytes have been nearly extirpated for the last decade Both study rivers are located within close proximity of each other and maintain relatively similar hydrological and chemical properties (Figure 1 2). Prior to the last decade, these

PAGE 18

18 systems supp orted dense communities of macrophytes fishes and invertebrates (Herald and Strickland 1949, Odum 1953, Frazer et al. 2006). In their current state, however, only one ( the Chassahowitzka River) retains any historical semblance of its former plant communi ty (Figure 1 3 ). The other ( the Homosassa River) has an altered vegetation community due to the extirpation of aquatic macrophytes (Frazer et al. 2006). Ecosystem responses to the change in vegetative habitat, as measured by changes in fish and invertebr ate communities and the predator prey interactions within each river were quantitatively assessed. Th is study comprised 4 study objectives each of which is addressed subsequently in Chapters 2 5, followed by a summary and synthesis of findings in Chapter 6. Study Objectives Objective 1 E stimate the catchability of fishes and invertebrates to standardized sampling gears target ing large bodied fishes, small bodied fishes and invertebrates and e valuate the spatial, temporal and interspecific heterogeneit y in gear catchability. Objective 2 Q uantif y the composition and biomass of fish and invertebrate assemblages with in the Chassahowitzka and Homosassa rivers seasona lly and across years in conjunction with the long term water quality and submersed aquati c vegetation monitoring programs. Objective 3 Estimate the prey composition, prey selectivity, and relative foraging success of freshwater and marine fishes within the Chassahowitzka and Homosassa river s Objective 4 Q uantitatively assess vegetative habitat loss effects on fish and invertebrate community structure using a time dynamic trophic mass balance model and evalu ate model predictions through spatial and temporal comparisons of the observed aquatic communities within the Chassahowitzka and Ho mosassa rivers.

PAGE 19

19 Figure 1 1. Location of the Chassahowitzka (south) and Homosassa (north) rivers in Hernando and Citrus counties, Florida.

PAGE 20

20 Figure 1 2. Physical and chemical characteristics of the Chassahowitzka and Homosassa rivers between 1998 and 2010 (Frazer, unpublished data) Data represent annual mean values based on quarterly sampling along 10 transects in each system (Frazer et al. 2006).

PAGE 21

21 Figure 1 3 Long term pattern s in submersed aquatic vegetation biomass within the Chassahowitzka and Homosassa rivers, Florida.

PAGE 22

22 CHAPTER 2 GEAR CATCHABILITY OF FISHES AND INVERTEBR ATES IN COASTAL RIVERS Introduction Accurate assessment of fish and invertebrate communities requires information on how population indices generated from disparate sampling me thods and gears reflect the ecosystem in terms of absolute abundance, composition and biomass (Nichols 1992, Anderson 2001). Estimation of the capture probability ( p ) of organisms, defined as the probability of captur ing an individual within a populati on during sampling (equivalent to the proportion of the population in the study area captur ed), is important in characterizing the most basic aspects of community structure including composition and biomass While sampling program design and consideratio ns for estimating p are widespread in the sampling literature (Pollock et al. 2002, Seber 2002, Williams et al. 2002), incorporating estimates of p into the quantitative assessment of aquatic communities has emerg ed only recently as an area of research foc us (e.g., Shea and Peterson 2007, Dauwalter et al. 2008, McCargo and Peterson 2010). Two components of sampling design should be considered when estimating the p of populations to sampling gears sampling intensity (i.e. amount of sampling effort applied t o a s tudy area ) and spatial distribution of effort (i.e. number and location of s ample s relative to the study area). Strong biases in estimates of abundance can result from changes in effort (making ani mals more or less likely to be captur ed) or from diff erences in the distribution of samples (e.g., changes in the proportion of study area sampled). Monitoring programs seeking to characterize community composition should account for variable sampling intensity as well as spatial and temporal heterogeneity in p for each population of interest.

PAGE 23

23 The density catchability equation (Hilborn and Walters 1992, Pollock et al. 2002) is a linear model that directly relates relative abundance ( catch per unit of sampling effort) to population density through a catchab ility coefficient ( q ). This model can be utilized to effectively incorporate sampling effort and study area into the estimation of p and absolute density. By definition, q is equal to the ratio of the relative abundance to the absol ute density of the pop ulation. Frequently in the analysis of fisheries data, q is modeled as the ratio of relative abundance to population abundance (e.g., Richards and Schnute 1986, Wang 1999) or biomass (Hilborn and Walters 1992); however, the density catchability model acco unts for differences in the units of effort between sampling gears and effectively scales relative abundance indices to a common metric, absolute density (i.e. population abundance estimates are scaled by sample areas to compare across gear types). Applic ation of this modeling framework allows for density to be estimated for multiple populations across trophic levels by deploying several gears that target various guilds within the aquatic community. Theoretically, if q is constant and known for each organ ism detected in the study area, then community composition can be accurately estimated from catch and effort data when adequate samples are taken. I utilized the density catchability assessment framework to estimate the community composition of fishes and invertebrates within the Chassahowitzka and Homosassa rivers Florida My objectives w ere to estimate the catchability of fishes and invertebrates commonly detected using multiple standard sampling gears (boat electrofishing, seining and throw trap sampl ing), and to assess the spatial, temporal, and interspecific heterogeneity in catchability estimates To accomplish this, I

PAGE 24

24 implemented capture recapture and removal sampling to couple relative abundance indices with absolute density estimates. This asse ssment exemplified the utility of estimating q for multiple populations to scale sample indices to estimates of absolute density and obtain measures of community composition. I discuss how this information can be useful for informing the design of ecosyst em based assessments and improving the management and restoration of populations and communities. Methods To estimat e the relative abundances (catch per unit effort in number of fish per hour electrofishing ) and absolute densities (number of fish per km 2 ) of large bodied fishes, a team of six crew members implemented multi pass capture recapture electrofish ing in each study reach during July 2007, January 2008, July 2008 and January 2009. S horeline and mid stream transects were sampled once per day for thr ee consecutive days with boat mounted Smith Root 9.0 generat or powered pulsed electrofishers with output settings ranging between 170 and 340 volts and between 20 and 50 amps The electrode arrays consisted of stainless steel cathode cables mounted across the bow of the boat and two insulated booms with removable stainless steel cable anode arrays mounted at each corner of the bow on rotating clutches. Shoreline transects included the entire north and south stream banks within a study reach, and transects were sampled from the upstream end of the reach to the downstream end. Mid stream transects were located at the upper reach boundary, middle of the reach, and lower reach boundary. During electrofishing transects, two people stood on the bow of the boat and dip netted stunned fish. All captured fishes greater than 150 mm in total length (TL) were tagged in the dorsal fin pterygiophores with a t bar anchored external tag containing a unique identification number and

PAGE 25

25 released along the shoreline in the ce nter of the reach All fishes greater than 50 mm TL received a batch mark by clipping the terminal end of the left pelvic fin, which created a secondary mark for fishes greater than 150 mm TL. To examine tag loss and handling mortality, a pilot study was conducted in the Santa Fe River, Florida where 6 0 fish were held in two replicate pens for three nights. Each pen contained 1 0 largemouth bass ( Micropterus salmoides ) 1 0 Lepomis spp., and 1 0 lake chubsucker ( Erimyzon sucetta ) or spotted sucker ( Minytrema melanops ) t o estimate tag loss and examine mortality of tagged fish in a confined environment. The observed mortality was expected to be greater than the mortality of handled fish released into the study reaches due to potentially higher predator encount er rates and increased stress on prey species confined with predators. To estimate the capture probability and absolute density (number of fish per 200 m 2 site area) of small bodied fishes, a crew of four conducted multiple pass removal seine depletions at three locations in each study reach during August 2007, February 2008, August 2008 and February 2009. Site locations in each reach included the north river bank, mid stream and south river bank. S ite locations were assigned randomly without replacement to a corresponding electrofishing mid stream transect, so that both banks and one mid stream site were sampled per reach. Sites measured 10 m in width by 20 m in length. To ensure closure of the sampling sites to migration a 60 m block net was set aroun d the perimeter of each site prior to seining. To set the block net, one person guided the boat around the perimeter of the site, while a second person carefully deployed the net over the side of the boat, and a third person secured the net to the stream bottom by placing concrete anchors on the inside of the net at each of the

PAGE 26

26 corners. The anchors were made of formed concrete ( 30.5 cm diameter and 10.2 cm height, constructed by pouring mixed concrete into a form tube) with 2.4 m ( 5.1 cm diameter) poly vi nyl chloride (pvc) pipes attached to the center of the anchor to support the block net at each corner. Depletion removal sampling was conducted with a minimum of three and maximum of seven seine passes using a 21.3 m wide, 1.8 m deep, 3.17 mm delta mesh b ag seine with a 1.8 x 1.8 m center bag. During each sample pass, one crew member disturbed the shoreline, course woody debris habitat, submersed vegetation and any overhanging tree limbs to displace fish and chase them towards the seine while two other cr ew members swept the entire site with the net. Subsampling occurred when the number of fish captured was too large to count all individuals per species, or the amount of detritus, filamentous algae, and other vegetation was too great to sort fish in a tim ely manner. In these cases, I recorded the total sample weight and weighed a subsample (generally 1/10 th to 1 / 20 th the total sample) to take back to the lab for processing. T he number of fish in the subsample was corrected by the proportion of sample mea sured to estimate the total number of fish captured per pass. To estimate the capture probability and absolute density (number of individuals per 1 m 2 site area) of decapods and select small bodied fishes I utilized data from a companion State Wildlife Gr ant project (Camp et al. 201 0 ) where throw trap removal sampling was carried out monthly in the Chassahowitzka River between June 2008 and May 2009 and quarterly between June 2009 and March 2010 During the study, t hrow trap s ampling also occurred monthl y in the Homosassa River during November 2008 through December 2009, and quarterly during January 2010 through March 2010. A n

PAGE 27

27 aluminum throw trap that measured 1 m in width by 1 m in length by 0.75 m in depth was used to sample f ive habitat type s in each study reach, when available, with up to three replicate patches sampled per habitat type. Habitat types included bare substrate, filamentous algae, Vallisneria americana Potamogeton pectinatus and mixed macrophyte/algae patches. T he reaches directly a bove the salt marsh estuary were not sampled with throw traps Throw trap depletion sampling methods are described in detail by Camp et al. (201 1 ). The catchability of 12 commonly captured taxa per gear type was estimated from catch and effort informatio n and estimates of absolute density. For fish species within the same genus, q was estimated for the genus as a whole. For invertebrates within the same family, q was estimated for the family as a whole. Capture R ecapture E lectrofishing A nalysis The fol lowing model assumptions were inherent in my analysis of population density and catchability from capture recapture electrofishing: Study reaches were closed to migration, births and deaths over the three day sample periods, all tagged animals were recorde d upon recapture, capture probability was homogeneous between marked and unmarked animals, and catch per unit effort was directly proportional to population density. Electrofishing r elative abundance indices were calculated as the number of fish captured on the first pass divided by the sampling effort measured as electrofishing pedal time in hours. Population abundance was estimated from the Lincoln Petersen equation (Table 2 1 Equation 1) when at least one marked individual was recaptured during a samp ling event. Population density was estimated by dividing the estimated abundance by the area of the study reach (Table 2 1 Equation 2). Reach area was

PAGE 28

28 estimated by overlaying polygons onto aerial orthophotographs using ArcGIS software. Density estimat es were scaled to fish per km 2 E lectrofishing q was estimated from empirically derived data as the ratio of the relative abundance to the estimated population density (Table 2 1 Equation 4). Seine and T hrow T rap R emoval S ampling A nalysis The following m odel assumptions were inherent in my analysis of capture probability and population density from seine and throw trap removal sampling: Sites were closed to migration, births and deaths during removal sampling, capture probability was constant across sampl e passes, and subsampling of captured fishes provided an accurate sample of the species composition and total catch. For seines and throw trap removal samples, the abundance and p at each site was estimated by multinomial maximum likelihood estimation (Go uld and Pollock 1997; Table 2 1 Equation 10). For these two gear types, q and p estimates were the same, since the sampling intensity (sampling effort divided by area sampled ) was equal to 1 when the ef fort from one pass of sampling was equivalent to the area of the blocked site (Ellis and Wang 2007; Table 2 1 Equation 8). Population density was estimated by dividing the estimated abundance by the site area (200 m 2 for seine samples, and 1 m 2 for throw trap samples ). Distributions of G ear C atchability E stimates To model the distribution of q estimates across taxa, study reaches and sampling events, I calculated the observed frequencies of positive q estimates for each gear type and fit the beta probability density function (beta distribution) to the obs erved I used a Monte Carlo analysis (10,000

PAGE 29

29 iterations of random draws from the beta distribution with solved parameters) to estimate the mode ( q maximum likelihood estimate), mean ( and 95 th percentile lower (95% LL) and upper (95% UL) limits of for each gear type. A two sided Kolmogorov Smirnov test (KS test, significance level = 0.05) was conducted to examine the goodness of fit of the beta distribution to the observed frequency of q estimates. Tests for S patial, T emporal and I nterspecifi c H eterogeneity in C atchability A kaike information criteria (AIC Akaike 1974) was used to evaluate alte rnative models of spatial, temporal and interspecific heterogeneity in q for each gear type (Table 2 3) following methodologies in Anderson (2008) for model based inference. Mo del AIC values were calculated from the number of model parameters and the to tal negative log likelihood (sum of negative log likelihoods of individual samples). An individual electrofishing sample was defined as a recapture event where at least one individual in the population was released in the reach during previous passes ove r the three day sampling period. A seine or throw trap depletion site was considered an individual sample. The log likelihood of electrofishing samples was equal to the natural log of the binomial probability given the number of trials = marks (M), numbe r of successes = recaptures (R) and probability of success = p = q E/A (Table 2 1 Equation 9). All samples were included in the total negative log likelihood, including samples that failed to recapture marked individuals ( i.e. R=0). The log likelihood of a seine or throw trap sample was equal to the multinomial probability at the maximum likelihood estimate of abundance (N) and p given the observed catches on each pass of the depletion (Table 2 1 Equation 10).

PAGE 30

30 The mean catchability ( ) of individual taxa was estimated by maximum likelihood analysis to compare among populations. I estimated the of individual taxa by solving for the maximum likelihood estimate of the total binomial log likelihood across all electrofishing recapture events, and total multinomial log likelihood of seine and throw trap removal sites (i.e. q was allowed to vary by taxa, but was set equal across study reaches and sampling events for a taxa). I calculated the profile likelihoods (Hilborn and Mangel 1997) to obtain 95% lower (95% LL) and upper (95% UL) credible intervals of for each taxon per gear type. Estimation of M ean P opulation D ensities of F ishes and I nvertebrates R elative abundance indices were calculated for each taxon per gear type, river study reach and sampling event as the number of captures divided by samplin g effort in hours (electrofishing) or area (seine and throw trap). The mean and standard deviation of relative abundance indices were calculated for each river across all sampling periods. The estimates and credible intervals for each taxon were used to estimate absolute densities and 95% upper and lower limits from mean relative abundance estimates. The absolute density estimates were graph ed with relative abundance indices for visual comparison. Results Electrofishing Catchability A total of 354 el ectrofishing recapture events were conducted on individual populations within the six study reaches (three reaches per river Figure 2 1 ), and 126 estimates of electrofishing q were obtained (228 recapture events failed to recapture marked individuals whic h resulted in estimates of q = 0). A histogram of the observed distribution of positive electrofishing q estimates (n = 126) across all taxa, rivers,

PAGE 31

31 reaches, seasons and years is presented in Figure 2 2 Maximum likelihood fitting of the beta distributi on to the observed positive q estimates resulted in beta distribution value = 0.015). The maximum likelihood estimate of electrofishing q = 0.0007, electrofishing 95% LL = 0.0002 and 95% UL = 0.0192. I utili zed all electrofishing capture recapture data (n = 354) to compare alternative models of heterogeneity in electrofishing q (i.e. alternative models allowed for constant q and q to vary by river, reach, season, event and taxa). The number of parameters (K ), total log likelihood values ( LL ) from the binomial likelihood calculated information criteria (AIC c ), and model probabilitie s (w i ) for each model are listed in Table 2 3. The highest weighted model allowed electrofishing q to vary by taxa, and alterna tive models contained negligible probability in comparison (Table 2 2 ). Electrofishing for individual taxa (Table 2 3 ) ranged from 0.0003 (95% LL = 0.0001, 95% UL = 0.0006) for L. rhomboides to 0.0160 (95% LL = 0.0070, 95% UL = 0.0291) for S. ocellatus In general, increased for larger bodied fishes in comparison to smaller bodied species, with smaller taxa (< 200 mm TL ) ranging from 0.0003 ( L. rhomboides ) to 0.0014 ( L. griseus ), and larger taxa (> 200 mm TL ) ranging from 0.0017 ( E. sucetta ) t o 0.0160 ( S. ocellatus ). Seine Catchability I conducted 494 seine removals of individual populations within depletion sites, and obtained 307 estimates of seine q ( I failed to obtain valid depletions in 53 removals, and excluded estimates from 124 removals that captured fewer than 5 total individuals to avoid low sample bias ). A histogram of the observed distribution of q estimates (n = 307) across taxa, rivers, reaches, seasons and years is presented in Figure 2 2.

PAGE 32

32 Maximum likelihood fitting of the beta distribution to the observed seine q estimates value = 0.095). The maximum likelihood estimate of seine q = 0.686, seine 0.192), 95% LL = 0.225 and 95% UL = 0.93 3. S eine capture information from all removal samples (n = 494) was used to compare alternative models of heterogeneity in seine q (i.e. alternative model s allowed for constant q and q to vary by river, reach, season, event and taxa). The number of param eters, total log likelihood values from the multinomial likelihood calculated information criteria, and model probabilitie s for each model are listed in Table 2 2 The highest weighted model allowed seine q to vary by taxa, an d alternative models contain ed negligible probability in comparison (Table 2 2 ). Seine for individual taxa (Table 2 3 ) ranged from 0.214 (95% LL = 0.185, 95% UL = 0.239) for G. bosc/M. gulosus to 0.873 (95% LL = 0.869, 95% UL = 0.878) for M. beryllina In general, I observed t he lowest seine for demersal taxa, ranging from 0.214 ( G. bosc/M. gulosus ) to 0.488 ( T. maculatus ), and I measured the highest seine for taxa that were typically captured near the surface of the water column ranging from 0.739 ( E. harengulus ) to 0.8 73 ( M. beryllina ). Estimated seine ranged from 0.261 ( S. scovelli ) to 0.739 ( Fundulus spp.) for taxa associated with vegetation and other structural habitats, includ ing seawalls and woody debris. Throw Trap Catchability A total of 1 957 throw trap remo vals of individual populations were conducted within depletion sites, and 767 estimates of throw trap q were obtained (152 removals fai led to obtain valid depletions and I excluded estimates from 1,038 removals that captured less than 5 total individuals to avoid low sample bias ). A histogram of the

PAGE 33

33 observed distribution of q estimates (n = 767) across taxa, rivers, reaches, seasons and years is presented in Figure 2 2. Maximum likelihood fitting of the beta distribution to the observed throw trap q est value = 0.251). The maximum likelihood estimate of throw trap q = 0.894, throw trap and 95% UL = 0.992. C apture information fr om all removal samples (n = 1,957) was used to compare alternative models of heterogeneity in throw trap q (i.e. alternative models allowed for constant q and q to vary by river, reach, season, event and taxa). The number of parameters, estimated total l og likeli hood values of the multinomial likelihood calculated information criteria, and model probabilitie s for each model are listed in Table 2 2 Similar to electrofishing and seining, the highest weighted model allowed throw trap q to vary by taxa, an d alternative models contained negligible probability in comparison (Table 2 2 ). Throw trap for individual taxa (Table 2 3 ) ranged from 0.225 (95% LL = 0.186, 95% UL = 0.263) for Grapsidae/Xanthidae (marsh or mud crabs) to 0.950 (95% LL = 0.944, 95% UL = 0.957) for Palaemonetes spp. I observed the lowest throw trap for demersal organisms, ranging from 0.225 for Grapsidae to 0.564 (95% LL = 0.528, 95% UL = 0.599) for G. bosc/M. gulosus I estimated relatively high throw trap for organisms that w ere typically captured near the surface of the water column ( E. harengulus throw trap = 0.621, M. beryllina throw trap = 0.926), as well as for taxa associated with vegetative habitats ( ranged from 0.675 for Lepomis spp. to 0.950 for Palaemonetes spp.). Relative Abundances versus Absolute Densities Means and standard deviations of relative abundance indices we re compared with absolute density estimates for each taxon per gear type and river ( Figure 2 3 ) T he

PAGE 34

34 estimated composition of fishes and in vertebrates varied between relative abundance indices compared to absolute density estimates. For example, I measured the highest electrofishing relative abundance for L. griseus in both rivers; however, absolute density estimates indicated that E. hareng ulus was approximately twice as abundant as L. griseus Similarly, seine relative abundance indices indicated similar catch rates of Lucania spp. and M. beryllina ; however, absolute density estimates demonstrated that Lucania spp. w ere more abundant than all other small bodied fishes in both rivers Thus, my results showed that accounting for catchability substantially influenced fish abundance estimates relative to using relative abundance indices. Discussion The application of sampling effort in relatio n to the behavior and spatial distribution of individuals affects the q of populations (Winters and Wheeler 1985, Angelsen and Olsen 1987, Swain and Sinclair 1994, Ellis and Wang 2007). To account for variable sampling effort and study area in the assessm ent of gear efficiency, I calculated relative abundance indices from catch and effort information, scaled absolute abundance estimates by the study area to estimate absolute density, and estimated q as the ratio of relative abundance to absolute density. This assessment framework allowed me to appropriately evaluate interspecific, spatial and temporal heterogeneity in gear catchability across study systems and sampling events. My results indicated heterogeneity in q estimates between taxa for all gears, a nd I attributed some of this heterogeneity to habitat use patterns of individual taxa. Other researchers may find the individual taxa estimates and the overall beta distributions of q useful as priors for estimating population density and community composition within a broad suite of freshwater and marine ecosystems.

PAGE 35

35 A key finding from this study is that q estimates based on populatio n densities are not widely available in the literature, yet are essential for comparison of gear capture probability across space and time. Estimates of q and p have been published for a multitude of species and study systems (e.g., Mann and Penczak 1984, Bayley and Austen 2002); however, estimates are not directly comparable when the reported values are not scaled by sampling intensity (i.e. the amount of effort applied to a study area). Bayley and Austen (2002) reported experimental estimates of electro fishing p for multiple species based on taxa size and environmental covariates, including water depth and macrophyte cover. Using their logistic model and empirical estimates of average fish size, reach depth, and vegetative cover, I predicted electrofish ing q estimates of M. salmoides to be 0.0013 in the Chassahowitzka River and 0.0035 in the Homosassa River (predicted p was scaled by mean sampling effort and study area to estimate the predicted q ). Surprisingly, the predicted values were similar to my o bserved estimates for M. salmoides of 0.0017 in the Chassahowitzka River and 0.0043 in the Homosassa River. Predicted q model for Lepomis macrochirus were 0.0002 and 0.0011 for the Chassahowitzka and Homosassa riv ers, respectively, and my observed values for Lepomis spp. were 0.0008 and 0.0043. Discrepancies in the predicted and observed values for Lepomis spp. could be due to the fact that the majority of my q estimates were for Lepomis punctatus with relatively few estimates for Lepomis macrochirus Studies that report q and p estimates with undefined sampling intensity are not comparable to other study systems. For example, sampling of a population that is distributed over a relatively small area is not expec ted to produce an equivalent p as the same amount of sampling effort

PAGE 36

36 distributed over a much larger area. To account for differences in spatial and temporally heterogeneity in sampling, I recommend estimating q for populations based on density estimates, and predicting p from the estimated taxa q and a defined amount of effort applied to a study area. Potential biases in my electrofishing q estimates may have resulted from violation of the closed capture recapture model assumptions tagging mortality, tag loss, or other sources I examined closure of the study reaches over the three day sampling periods for M. salmoides from a concurrent acoustic telemetry study (Pine and Tetzlaff 2008), and concluded that estimates for M. salmoides were not likely to be b iased as a result of e migration. Other taxa (e.g., M. cephalus ) are highly mobile and movement out of the study reaches between daily sampling events would result in a negative bias in q estimates. Another potential source of bias in electrofishing q est imates is tagging mortality. I assessed sampling and tagging mortality for three taxa ( M. salmoides Lepomis spp. and E. sucetta ) prior to sampling, and determined that tagging mortality was minimal (0 of 20 M. salmoides 1 of 20 Lepomis punctatus and 0 of 20 E. sucetta held in pens died over a three day observation period), although a negative bias in estimates is possible as a result of sampling mortality or increased susceptibility to predation of tagged animals Tag loss was not a source of bias, sin ce all tagged individuals were double marked by clipping the left pelvic fin; I was able to detect all tagged individuals when recaptured. S eine and throw trap estimates were not considered biased as a result of migration, due to closure of each site prio r to sampling. The density estimates of small bodied fishes however, may be biased low as a result of chasing fish away from the site while setting the block net. B ehavioral responses of

PAGE 37

37 organisms during removal sampling may also result in biased estima tes (Peterson et al. 2004). The crew attempted to alleviate the behavioral response bias by scaring fish from structured habitats during each sample pass; however, I was not able to verify that behavioral responses during removal sampling did not negative ly bias density estimates of smal l bodied fishes and decapods. Information on q of organisms is paramount to researchers whose aim is to accurately assess the composition of aquatic communities. Failure to account for heterogeneity in q and p can lead t o erroneous assessment of trends in population and community metrics when q varies spatially, temporally and between taxa; or when changes in effort and area affect the p of populations. To estimate the q and p of populations, I recommend the sampling app roach detailed by Pollock et al. (2002) of coupling survey indices with absolute density estimates. This can be accomplished by incorporating capture recapture, removal or other abundance estimation methods into monitoring program design. For long term p opulation monitoring, I recommend that researchers estimate q during the initial sampling event to examine interspecific and spatial heterogeneity, and periodically throughout the study to evaluate temporal heterogeneity. Large scale programs may require more advanced tagging methods, such as acoustic or satellite tags, to create a known population of individuals from which q can be estimated. When the relative values of q are known, population indices from multiple gears can be appropriately scaled to est imate absolute densities and community composition. My density estimates from electrofishing, seining and throw trapping demonstrate the usefulness of combining density estimates from multiple gears that target different

PAGE 38

38 guilds within the community to pro vide more accurate estimates of community composition compared to sample indice s from individual gears Increased information about the q of populations could lead to a greater ability to assess population trends and communities as a whole, particularly i n systems that are frequently sampled with more than one gear, such as multispecies fisheries. For these and other multiple use resources, I advocate that researchers assess q for each sampling gear to assess the effectiveness of monitoring in detecting s patial and temporal differences in the structure of communities. Accurate assessment of community structure provides insight into the complex trophic dynamics that structure population and community level processes within aquatic ecosystems (e.g., Kitchel l and Crowder 1986, Polis and Strong 1996). Community assessment is a principal component of ecosystem models which provide a abundance, distribution and dynamics (Hall et al. 1992, Link et al. 2002). These models can elucidate dominant interactions associated with common species, including predator prey cycles and whole community shifts (Walters and Martell 2004). Community assessment is therefore a critical step in link ing population metrics to ecosystem processes. As natural resource management moves towards integrated ecosystems approaches, the information needs for population and community level assessments increase greatly. Efficient and broadly applicable sampling approaches that produce timely and accurate estimates of community composition, density and biomass are greatly needed. Incorporating catchability estimation into large scale monitoring

PAGE 39

39 programs provides the framework needed to efficiently assess communi ty composition from monitoring indices. Such as approach could lead to a greater understanding of community level dynamics that influence populations, and provide more accurate information to natural resource management programs which regulate populations of organisms and their habitats.

PAGE 40

40 Table 2 1 List of e quations used to e stimate c atchability and capture probability from c losed m ark r ecapture and removal s ampling. *Substitution of Equation 1 into Equation 2, and Equation 2 into Equation 4 solves to Equation 5 Equation 5 solves to Equation 9 § k is the number of passes in the depletion sample, q = p when effort is equal to the site area

PAGE 41

41 Table 2 2 Akaike information criteria and model probabilities (w i ) for alternative models of spatial, temporal and interspecific heterogeneity in catchability of fishes and invertebrates sampled by electrofishing, seining and throw trapping. Gear n Model K LL c w i Electrofishing 364 q(constant) 1 502 198 0.0 q(river) 2 427 50 0.0 q(reach) 6 415 35 0.0 q(season) 2 498 192 0.0 q(event) 4 494 188 0.0 q(taxa) 12 391 0 1.0 Seine 494 q(constant) 1 34,501 22,101 0.0 q(river) 2 33,610 20,320 0.0 q(reach) 6 23,573 255 0.0 q(season) 2 34,381 21,862 0.0 q(event) 4 30,740 14,585 0.0 q(taxa) 12 23,439 0 1.0 Throw t rap 1,957 q(constant) 1 9,938 4,533 0.0 q(river) 2 9,609 3,877 0.0 q(reach) 4 9, 254 3,171 0.0 q(season) 4 9,908 4,478 0.0 q(event) 20 9,545 3,786 0.0 q(taxa) 12 7,661 0 1.0

PAGE 42

42 Table 2 3 and upper limits of mean catchability for individual taxa sampled by electrofishing, seining and throw trapping.

PAGE 43

43 Figure 2 1 Study reaches within the Homosassa and C hassahow itzka rivers, Florida.

PAGE 44

44 Figure 2 2 Histograms of observed positive catchability estimates (q) across all taxa measured. Non linear curves represent the fitted beta distributions. Note the difference in scale of the x axis between electrofishing and th e other two gears as a result of the scaling by unit s of effort.

PAGE 45

45 Figure 2 3. Relative abundance indices (CPUE = catch per unit effort) and absolute density estimates of taxa commonly detected during boat electrofishing, seine and throw trap sampli ng within the Chassahowitzka and Homosassa rivers.

PAGE 46

46 CHAPTER 3 THE COMPOSITION AND BIOMASS OF THE AQUAT IC COMMUNITIES WITHIN THE CHASSAHOWITZKA AND H OMOSASSA RIVERS Introduction E cosystem based approaches to natural resource management and restoration are w idely recommended ( Grumbine 1994, Li n denmayer et al. 2000, Pew Oceans Commission 2003, Pikitch et al. 2004, Ocean Studies Board 2006) and increasingly required (e.g., Magnuson Stevens Fisheries Conservation and Management Act Section 406, Malone 1995 ) M a ndates for e cosystem research a re motivated by explicit recognition of the roles of predation (Menge and Sutherland 1976, Bowlby and Roff 1986, Estes et al. 1998), competition (MacArthur 1958, Connell 1961, Schoener 1983), trophic dynamics (Hairston et al. 1960, Paine 1980, Carpenter et al. 1985), and environmental conditions (Sandoey and Nilssen 1987, Ritchie 2000) in structuring populations. This recognition has emphasized the need to link assessments of individual species with broader scale community an d ecosystem studies (Link 2002, Pikitch et al. 2004, Walters et al. 2005). E cosystem research including the quantitative assessment of communit y structure and the function of ecosystem components is central to the development of effective management and restoration programs for renewable natural resources The assessment of community structure is a central component of ecosystem studies (Karr 1987) and is essential for the development of food web models aimed at evaluat ing management policy options and predicting ecosystem effects of environmental changes (Christensen and Pauly 1992). At a minimum, a quantitative characterization of community structure requires concurrent sampling of populations from multiple trophic levels to estimate the composition of producers, primary

PAGE 47

47 consumers, and predators ; and an examin ation of trophic interactions among key taxa. Information on community structure provides insight into the complex trophic dynamics that influence a broad suite of population and community leve l proce sses within aquatic ecosystems (Polis and Strong 1996 Kitchell and Crowder 1986 Levin and Paine 1974) Furthermore, c ommunity assessment is a principal component of ecosystem models which provide a centralized framework for linking biotic and abi population abundance, distribution and dynamics (Hall et al. 1992, Link et al. 2002). These models can elucidate dominant interactions that occur among common species, including predator prey relationship s and whole community shif ts (Walters and Martell 2004). Community assessment is therefore a critical step in linking population metrics to ecosystem processes such as habitat loss The experimental removal of key habita t component s has been shown to alter the community structure of fishes within aquatic ecosystems (Bettoli et al. 199 3 Deegan et al. 2002, Sass et al. 2006). Bettoli et al. ( 1993 ) demonstrated a shift in fish community structure associated with the large scale removal of aquatic vegetation resulting from the stock ing of grass carp in Lake Conroe, Texas including the decline or collapse of small phytophilic populations Sass et al. ( 2006 ) showed that the removal of course woody debris from the littoral zone of lakes in Wisconsin altered growth, predation and recr uitment of fish populations These ecosystem manipulations demonstrate d how removal of critical habitats affect ed fish populations disparately leading to a shift in species assemblages and community dynamics In general, h abitat loss is expected to alte r the community structure within aquatic ecosystems; however, the effects of macrophyte loss on stream faunal communities have not been quantitatively assessed.

PAGE 48

48 Since t he e xperimental removal of vegetative habitat is not always a viable option for assessi ng community effects of habitat loss a comparative analysis of community structure between ecosystems with contrasting vegetative habitat s may elucidate key consequences of habitat loss to fish and invertebrate populations. I quantified the composition an d biomass of the aquatic communities within two spring fed, coastal rivers in Florida to make spatial and temporal comparisons between a highly vegetated river, the Chassahowitzka River, and one where macrophytes have been largely absent for nearly a decad e, the Homosassa River Historically, t hese systems were reported to be some of the most productive ecosystems in the world (Odum 195 7 ) supporting unique, oligohaline communities comprised of marine and freshwater plants, fishes and invertebrates ( Herald and Strickland 1949, Odum 195 3 ). I hypothesized that large scale habitat loss in the Homosassa River resulted in an altered composition of fishes compared to qualitative observations of the fish community recorded 60 years ago (Herald and Strickland 1949 ) I also hypothesized that the current fish and invertebrate community compositions in the Homosassa River differ s from that of the Chassahowitzka River which supports greater abundance of macrophytes including Vallisneria americana Potamogeton spp., Najas guadalupensis Myriophyllum spicatum (non native) and Hydrilla verticillata (non native ) To test this, I sampled fishes invertebrates, macrophyte s and algae using multiple gears (electrofishing, seines, throw traps, benthic cores, invertebrate nets and vegetation quadrats) and utilized information on taxa specific gear catchability (Chapter 2), to estimate the absolute density and biomass of select trophic groups and obtain estimates of community composition I examin e d the seasonal and spat ial composition

PAGE 49

49 of the aquatic communities for the purpose of identifying key differences between rivers with disparate vegetative habitat composition and biomass, and qualitatively evaluate changes in the Homosassa River compared to historical observation s prior to habitat loss While c ommunity level assessment s are relatively rare quantitative estimates are n ecessary for inferr ing how vegetative habitat loss may affect fish and invertebrate communities in stream ecosystems Methods To assess the compos ition and biomass of the aquatic communit ies within the Chassahowitzka and Homosassa rivers, I utilized data from multiple gear types that targeted different guilds of fishes and invertebrates (i.e. benthic invertebrates, plant associated invertebrates, de capods, small bodied fishes, and large bodied fishes) All sampling was conducted in conjunction with the long term submersed aquatic vegetation (SAV) monitoring program (Frazer et al. 200 6). Standardized quadrat sampling methods utilized in the long ter m vegetation monitoring program were implemented to estimate the biomass of macrophytes and filamentous algae in each of the study reaches (Figure 2 1). Sediment cores and 300 mesh net s were us ed to collect grab samples of macroinvertebrates associated with benthic habitats and SAV In addition to sediment cores and vegetation nets, I utilized data collected during a concurrent throw trap sampling program for decapods (Camp et al. 2011) including blue crabs ( Callinectes sapidus ) crayfish (Cambaridae) mud crabs (Grapsidae and Xanthidae) and grass shrimp ( Palaemonetes spp.) To sample the fish community in the rivers, seine sampling was us ed to capture small bodied fishes and electrofishing was us ed to capture large bodied fishes. Sampling occurred a cross multiple spatial

PAGE 50

50 (stratified study reaches in each river) and temporal (monthly, seasonally and yearly) scales to more accurately estimate the average biomass of fishes and invertebrates within the rivers and assess variation in estimates over the p eriod of study Submersed Aquatic Vegetation Sampling and Analyses To estimate the percent cover and biomass of macrophytes and filamentous algae in coastal rivers I utilized standardized quadrat sampling methods outlined by Frazer et al. (2006) for long term SAV monitoring Sampling was conducted biannually in August and February during years one and two of the project and monthly during year three I assumed perfect detection ( p = 1) for plants and algae collected in quadrats. Biomass was estimated a s wet weight of plants or algae per quadrat area. I calculated the mean and standard deviation of macrophyte and filamentous alg ae biomass for each study reach I scaled all biomass e stimates to g per 1 00 m 2 area. Invertebrate Sampling and Analyses I sam pled a quatic invertebrate s associated with sediments and above bottom portions of SAV in the three study reaches of both rivers during August and February of year one and collected invertebrates inhabiting SAV in R eaches 1 and 2 of years two and three S ampling occurred concurrently with SAV monitoring along three fixed transects with in each of the study reaches. I sampled f ive stations ( equally spaced ) along each transect. I collected benthic invertebrate s with a 5 cm inner diameter acrylic push core ( sediment surface area sampled = 20 cm 2 ). To obtain a sample, the core was firmly pushed into the sediments to a depth of 10 cm (volume sampled = 200 cm 3 ) and then carefully withdrawn. I then extruded the sample from the push core into a 1 L container or 1 gallon s ealable, labeled plastic bag and rinsed a ny sample portions remaining inside the push core into the sample container.

PAGE 51

51 To estimate the density and biomass of m acro invertebrates associated with SAV, I sampled 15 uniformly stratified sites per reac h with a 300 mesh, netted ring sampler (inner ring diameter = 252 mm, 0.05 m 2 area). I obtained s amples by placing the open bottom ring of the sampler over a portion of SAV, closing the bottom of the net, and cutting the SAV just above the sediment/wat er interface. I rinsed t he sample into a 1 L sample container or 1 gallon sealable plastic bag and labeled it with the sample location and date I placed a ll samples on ice immediately after collection and transported them to the F lorida Fish and Wildlif e Conservation Commi s sion, Gainesville Fisheries Research Laboratory or University of Florida, Florida Rivers Research Laboratory for processing and taxonomic identification. In the laboratory, individual samples were rinsed from containers into a 300 mesh sieve to remove water, placed in 1 L wide mouth plastic or glass jar s and preserved with 95 % ethanol ( year one ) or were froze n (years two and three ). During year one entire samples were processed by placing small portio ns into a petri dish, cove ring each portion with water, and inspecting the contents under a stereo dissecting microscope with magnification to 63x. I nvertebrates were removed from petri dishes with forceps, identified to major taxonomic group, enumerated and then preserved in labe led vials with 95% ethanol A laboratory sheet was prepared listing taxa and counts for each sample. During years two and three invertebrate samples were white panned to remove and enumerate visible macroinvertebrates. T he SAV sample was then rinsed ov er the white pan, sieved, weighed and subsampled (by wet weight). I nvertebrates were removed from the SAV subsample and individual taxa were enumerated under a stereo dissecting microscope. T he fine particles and remaining periphyton in the white

PAGE 52

52 pan we re then sieved weighed and subsampled (by wet weight). T he invertebrates from the fine material subsample were removed with forceps and enumerated by individual taxa under a stereo dissecting microscope. I corrected i nvertebrate counts for each subsampl e by dividing by the proportion of sample measured and summed the estimate with the counts from the white pan. I assumed perfect detection ( p = 1) of invertebrates sampled with benthic cores and vegetat ion nets. I calculated i nvertebrate density as the mean number of individuals per taxa per sample divided by the sample d surface area. I calculated t he mean density and standard deviation of invertebrates for each study reach and scaled the estimate to 1 00 m 2 area. S eparate a nalyses were conducted for b enthic and SAV substrates. I calculated b iomass estimates of selected taxa by multiplying the estimated density by mean individual mass. I obtained d ry mass estimates for individual taxa from samples that were sorted, enumerated, weighed and dried, or fr om published length mass regressions (Benke et al. 1999) and measurements of mean individual length (total length was measured for amphipods, insects, tanaids, and isopods; and shell length was measured for gastropods ) D ry mass estimates were converted t o wet mass using conversion factors published by Ricciardi and Bourget (1998). Blue crabs, mud crabs, crayfish and grass shrimp were not effectively captured by benthic cores or vegetation nets; therefore, I utilized data from Camp et al. (2011) to obtain biomass estimates of these invertebrates The absolute densities were estimated from throw trap depletion samples taken within the study reaches during the period of study. Absolute densities in each study reach were estimated by the multinomial likeliho od approach for depletion sampling described by Gould and Pollock

PAGE 53

53 (1997). The absolute biomasses were estimated by multiplying the estimated density by average mass per individual. Estimates for each study reach were scaled to g per 100 m 2 area and the a verage biomass in the river was estimated across study reaches (Reaches 1 and 2) The lower study reach es in both rivers w ere not sampled with throw traps. Fish Sampling and Analyses I deployed t wo gear types (electrofishing and seining) to estimate the a bundance and biomass of large and small bodied fish es. Three day mark recapture electrofishing events and three day block net seine depletion sampling occurred during four sample periods (summer 2007, winter 2008, summer 2008, and winter 2009) in each ri ver. Electrofishing occurred biannually during the second and third weeks of July and January of years one and two During biannual sampling, each study reach was electrofished once per day for three consecutive days During year three s ingle pass ele ctrofishing and seine surveys were conducted monthly Standardized sample locations are shown for each gear type in Figure 2 1 Electrofishing reaches included four shoreline transects and three mid stream transects. I defined o ne shoreline transect as the section of littoral stream bank between long term SAV monitoring transects and mid stream transects overlapped the SAV monitoring transects During years one and two of the project I sampled nine multi pass seine depletion sites biannually (August 200 7, February 2007, August 2008, and February 2009) i n each river to assess the small bodied fish community and obtain estimates of gear catchability I sampled t hree sites in each reach at fixed locations that coincide d with electrofishing, long term SAV m onitoring, and invertebrate sampling transects. Sites ranged in size between 200 and 600 m 2 during the first sampling event, but were

PAGE 54

54 standardized at 20 m in length and 10 m in width during subsequent sampling. I chose t he location of each seine depletio n site within a reach randomly without replacement and assign ed one of three possible locations: river right, mid st ream or river left. A ll three locations were sampled at separate transects within a study reach. A 2.4 m deep block net was set around eac h site, and multiple pass sampling ( three to seven passes per site were completed until a d ecline in catches was observed) was executed with a 21.3 m wide, 1.8 m deep, 3.17 mm delta mesh bag seine with a 1.8 x 1.8 m center bag. During year three monthly single pass seine surveys were conducted at block netted site s At most sites, subsampling occurred when either the number of fish captured was too large to count all individuals per species, or the amount of detritus, filamentous algae, and other vegetat ion was too great to sort fish in a timely manner. When subsampling occurred, I recorded the total weight of the sample and weighed a portion to take back to the lab for processing. I then corrected t he number of fish in the subsample by the proportion o f sample measured to estimate the total number of fish captured per pass. A ll fish captured were identified to species when possible otherwise fish were identified to the lowest possible taxonomic resolution. A ll fish were measured for total length (TL) and weight (weights were not taken when windy conditions prevented acc urate measurement) and released, with the exception of fishes kept for diet and growth analyses During electrofishing sampling, every fish greater than 150 mm TL was tagged in the dor sal fin pterygiophores with a t bar external tag containing a unique identification number. T he right pelvic fin of every fish greater than 50 mm TL was

PAGE 55

55 clipped as a secondary mark for externally tagged fish and primary batch mark for fish between 50 and 150 mm in TL Electrofishing c atch per unit effort (CPUE) was calculated for each taxon (i) per sampling event (t) as the number of fish captured (C) within a reach divided by sampling effort (E) in hours : (3 1) E lectrofis hing CPUE was averaged across day s for multiple pass sampling during years one and two Seine and throw trap CPUE were calculated as the number of individuals captured on the first pass of netting divided by area swept (seine unit of effort = 100 m 2 ; thro w trap unit of effort = 1 m 2 ). S eine and throw trap CPUE were averag ed across sites within a reach. A bsolute densit y w as estima ted for each taxon as mean CPUE divided by mean catchability ( ) (Table 2 3): (3 2) T he sampled mean CPUE s w ere assumed equal to the population mean s and confidence intervals of density estimates were estimated from 95% confidence intervals of mean catchability (Table 2 3) The bi omass (B) of each taxon was estimated by multiplying the density times the mean weight of individuals (w) captured within the reach : (3 3) All estimates were scaled to g per 100 m 2 area for comparison between reaches and rivers.

PAGE 56

56 Results Submersed Aquatic Vegetation (SAV) Average SAV cover varied between 11 and 52% during August sampling in Reaches 1 and 2 of the Chassahowitzka River, and varied between 47 and 77% during February sampling. I observed minimal SAV cover in Reach 3 of the Chassahowitzka River during August sampling (0 to 2%), but estimated considerable filamentous algae cover during February of each year (21 to 30%). Average filamentous algae cover in Reach 1 of the Homosassa River peaked in August 2007 (29% ), and ranged between 7 and 25% during the other sample periods. I documented the highest average percent cover in Reaches 2 and 3 of the Homosassa River during February 2008 and 2009 (27 to 56%). The estimated biomass of macrophytes was distinctly differ ent between the Chassahowitzka and Homosassa rivers (Figure 3 1), and both systems demonstrated a strong seasonality in filamentous algae biomass (Figure 3 2). SAV within the Chassahowitzka River during August of each year was comprised primarily of macro phytes (mean plant biomass = 782 1433, and 1728 g/m in 2007, 2008 and 2009, respectively; mean algae biomass = 35, 34, 17 g/m in 2007, 2008 and 2009, respectively). I measured the greatest mean plant biomass in Reach 1 (782 to 1728 g/m), with lower bi omass es observed in Reaches 2 (140 to 1246 g/m) and 3 (0 to 9 g/m) (Figure 3 1 ). The Homosassa River was nearly devoid of macrophytes across all sample periods (0 to 45 g/m) (Figure 3 1) I documented seasonally high biomass of filamentous algae durin g winter (2008 and 2009) and spring (2010) sampling periods in both rivers ( Table 3 1, Figure 3 2 ).

PAGE 57

57 Invertebrates Nematodes, ostracods, oligochaetes, polychaetes, amphipods, copepods, and chironomids were common in sediment samples collected from both riv ers during August 2007 and February 2008. Sediment samples were not processed after the initial two sample periods due to logistical and funding constraints. Benthic invertebrate biomass (measured as the sum of ostracods, oligochaetes, polychaetes, amphi pods, nematodes, copepods, and bivalves) averaged 67 g/100 m 2 during August 2007 and 166 g/100 m 2 during February 2008 within the Chassahowitzka River, and 41 g/100 m 2 during August 2007 and 493 g/100 m 2 during February 20 08 within the Homosassa River. The most numerous taxa associated with SAV samples included amphipods, ostracods, gastropods, copepods, isopods, nematodes chironomids and other insect larvae I estimated higher total biomass of invertebrates associated with SAV (measured as the sum of gast ropods, insects, isopods and tanaids ) during February sampling periods (mean biomass in the Chassahowitzka River = 578 g/100 m 2 Homosassa River = 5023 g/100 m 2 ) compared to August (mean biomass in the Chassahowitzka River = 315 g/100m 2 Homosassa River = 103 g/100 m 2 ) in both rivers Biomass estimates for vegetation associated invertebrates included samples from all biannual sampling periods; however, I was not able to process monthly invertebrate samples from year three due to logistical and funding con straints Amphipods were analyzed separately from other vegetation associated invertebrates due to their relatively high biomass in both rivers compared to other taxa (mean August biomass = 189 g/100 m 2 in the Chassahowitzka River and 191 g/100 m 2 in the Homosassa River) large increases in biomass associated with winter sampling (mean February biomass =

PAGE 58

58 2316 g/100 m 2 in the Chassahowitzka River and 1562 g/100 m 2 in the Homosassa River) (Figure 3 3 ) and importance as prey for fishes (Chapter 4) Vegetati ve net sampling did not effectively capture larger invertebrate taxa, including blue crab ( Callinectes sapidus ), crayfish (Cambaridae), mud crab (Grapsidae and Xanthidae combined) and grass shrimp ( Palaemonetes spp.). Density and biomass estimates o f thes e larger invertebrates were obtained from throw trap sampling data collected during a concurrent study ( Camp et al. 2011 ). Mean biomass estimates of each invertebrate tax on by river and season are listed in Table 3 1 The density and biomass of invertebra tes associated with SAV was greatest during winter sampling periods when fi lamentous algae biomass was high (Figures 3 3 through 3 7). This pattern was apparent for the most abundant taxa of invertebrates with the exception of insects (Figure s 3 3 through 3 7 ). In fact, insect density and biomass was similar across all sampling peri ods in the Chassahowitzka River and I observed a relatively high biomass of insects in the Homosassa River during February 2008 when filamentous algae mats were prevalent (Figu re 3 4 ) Insects, particularly chironomids, were abundant in both filamentous algae and macrophyte samples Of all taxa sampled amphipods and blue crabs demonstrated the greatest biomass in the Chassahowitzka River, and amphipods and mud crabs were most abundant in the Homosassa with peak biomass occurring during winter periods (Table 3 1) One surprising result was the observed higher density and biomass of gastropods associated with filamentous algae during winter in the Homosassa River compared to t he Chassahowitzka River which supports vegetation year round (Figure 3 5)

PAGE 59

59 Small bodied Fishes Overall, freshwater small bodied fishes were less abundant in the Homosassa River compared to the Chassahowitzka River (Figure 3 8 ); however, many saltwater spec ies showed similar biomass between the two rivers with the exception of pinfish ( Lagodon rhomboides ) which occurred in higher biomass within the Chassahowitzka River and gobies ( Gobiosoma bosc Microgobius gulosus ) which occurred in higher biomass in the u pper reaches of the Homosassa River. I estimated greater density and biomass of small bodied fishes in the upper two study reaches of the Chassahowitzka River during August sampling events compared to the Homosassa River (Figures 3 8 and 3 9 ). Small bo died fish density and biomass declined between summer and winter sampling in the Chassahowitzka River during all years, which may be attributed, in part, to decreased biomass of freshwater species (Figure 3 8) In contrast, I did not observe a higher dens ity and biomass of small bodied fishes during summer periods in the Homosassa River. Many small bodied species showed a strong seasonality in their density and biomass, with the greatest biomass observed in late spring through summer, and relatively low biomass during fall and winter (Figures 3 8 and 3 9) Seine sampling within the Chassahowitzka River during August primarily captured rainwater killifish ( Lucania parva ), followed by inland silverside ( Menidia beryllina ), tidewater mojarra ( Eucinostomus h arengulus ), bluefin killifish ( Lucania goodei ), and young of the year spotted sunfish ( Lepomis punctatus ). February sampling within the Chassahowitzka River predominantly captured rainwater killifish, tidewater mojarra, pinfish, needlefish ( Strongylura sp p.) and gray snapper ( Lutjanus griseus ). Seining within the Homosassa River during August produced mostly rainwater killifish, inland silverside, tidewater

PAGE 60

60 mojarra, clown goby ( Microgobius gulosus ) and naked goby ( Gobiosoma bosc ). February sampling in th e Homosassa River captured tidewater mojarra, rainwater killifish, mosquitofish ( Gambusia holbrooki ), bay anchovy ( Anchoa mitchilli ), inland silverside, clown goby and naked goby. Large bodied Fishes The estimated biomass of freshwater fishes was signific antly greater in the Chassahowitzka River compared to the Homosassa River for most sampling periods (Figure s 3 10 through 3 14 ). Mean biomass estimates of large bodied fishes are listed by species o r trophic group in Table 3 1 Total freshwater and saltw ater large bodied fish biomass was greatest in Reach 1 of both rivers with lower biomass observed in downstream reaches. I estimated significantly lower biomass of lake chubsucker (Figure 3 10), Lepomis spp. (Figure 3 11), and adult largemouth bass (Figur e 3 14) in the Homosassa River relative to the Chassahowitzka River during most sampling events; however, Florida gar ( Lepisosteus platyrhincus ) were more abundant in the Homosassa River (Figure 3 13) and comprised a large proportion of the freshwater, lar ge bodied fish biomass. I measured a large increase in the biomass of lake chubsucker (Figure 3 10) and Lepomis spp. ( Figure 3 11 ) between January 2008 and July 2009 within the Chassahowitzka River, corresponding with relatively strong cohorts of young of the year captured during summer 2008 and subsequent sampling events. I documented high densities and biomass of saltwater, large bodied fishes during winter sampling periods of each year in both rivers (Figures 3 15 through 3 20) with t he greatest bioma ss surveyed during January 2008 within Reach 1 of the Homosassa River.

PAGE 61

61 Common fishes captured by electrofishing during August in both rivers included pinfish, spotted sunfish, largemouth bass, striped mullet ( Mugil cephalus ) and American eel ( Anguilla ro strata ). Lake chubsucker were also commonly captured in the Chassahowitzka River, but were rarely encountered in the Homosassa River (5 total young of the year were captured during the period of study). Gray snapper were the most abundant species captur ed during January within both rivers, followed by spotted sunfish, pinfish, largemouth bass and lake chubsucker within the Chassahowitzka River; and striped mullet, spotted sunfish and common snook ( Centropomus undecimalis ) within the Homosassa River. A c omplete list of scientific and common names of freshwater and saltwater fish species captured during electrofishing and seine sampling within each river is provided in Tables 3 2, 3 3, 3 4 and 3 5. The spatial and temporal variability in biomass estimates is illustrated in Figure s 3 10 through 3 20 Discussion The Chassahowitzka River support ed greater vegetative habitat cover and biomass year round in the upper reaches compared to the Homosassa River as a result of the perennial cover and biomass of mac rophytes Filamentous algae were prevalent in both systems during winter sampling periods, creating a seasonally abundant habitat for invertebrates and small bodied fishes, such as amphipods, isopods, gastropods, and killifish. Vegetative habitat wa s low in abundance during the rest of the year in the Homosassa River In areas with higher flows, algae mats we re transported to downstream reaches where they senesce d in areas with lower velocity and higher salinities (Frazer et al. 2006) This may have res ult ed in displacement of organisms utilizing the vegetative habitat in the Homosassa River to alternative habitats such as

PAGE 62

62 littoral areas or benthic substrates, whereas invertebrates and fishes in the Chassahowitzka River may use alternative perennial veg etative habitats I documented relatively large declines in freshwater species density and biomass during winter sampling periods, coincident with immigration of saltwater species that utilize these systems likely as thermal refugia, including gray snapp er, common snook and red drum. Declines in large bodied freshwater species density during winter periods are due, in part, to migration into tributaries, canals and headwater areas, as evidenced by resighting observations of marked fish outside of the stu dy reaches during subsequent months after sampling (Frazer, unpublished data) The sharp decline in small bodied fishes during winter may be a result of increased predation by saltwater piscivores or migration out of the study area, which in turn may r elease predation pressure on small invertebrates and increase the density and biomass of taxa that are exploited as prey by small bodied fishes. Additionally, filamentous algae mats may provide temporary refuge for small invertebrates allowing population s densities to increase under lower predation pressure; however, mortality estimates of invertebrates were not con ducted as part of this study. Overall, I observed similar pattern s in fish density and biomass as those observed for estimated biomass of SAV in the study systems (i.e. reaches and sampling periods with high er biomass of SAV including macrophytes and filamentous algae had a greater estimated density and biomass of invertebrates and fishes). A comparison of invertebrate and fish assemblage s between rivers provided insight into community level changes that may occur if macrophytes are lost from a system. Species that rely on vegetation for foraging, refuge or reproduction will likely be

PAGE 63

63 negatively affected by large scale habitat loss. For ex ample, I estimated greater densities and biomass of multiple freshwater species in the Chassahowitzka River, including crayfish ( Camp et al. 2011 ), grass shrimp ( Camp et al. 2011 ), rainwater killifish, bluefin killifish, Notropis spp., spotted sunfish, lak e chubsucker and largemouth bass, that were less abundant in the Homosassa River. Furthermore, I documented large cohorts of fishes surviving to older age classes in the Chassahowitzka River over the study period. In the Homosassa River, cohorts of age 0 largemouth bass a nd spotted sunfish were observed ; however, few individuals were captured in subsequent sampling events at older age classes, contrary to observations in the Chassahowitzka River (unpublished data) F ew age 0 lake chubsucker were captured in the Homosassa River during the first sampling event and none were captured in the study reaches during the following sampling periods with the exception of June 2010 following high production of filamentous algae in March and April 2010 I observed th e greatest densities of juvenile and small bodied fishes in Reach 1 of the Homosassa River during June 2010, subsequent to the increased production of filamentous algae. These data indicate that macrophytes may be important for recruitment of many species in coastal rivers by providing year round forage and refug e habitat for larvae and juveniles Historic observations of the fish community in the Homosassa River (Herald and Strickland 1949) indicate d that select phytophilic species were once common but m y results showed that these species have been near ly extirpated from reaches where macrophyte loss has been substantial L ake chubsucker and Notropis spp specifically, were observed in the Homosassa River prior to macrophyte loss (Herald and Strickland 1949); however, I rarely encountered either of these species over the period of study.

PAGE 64

64 Although baseline data on the fish communities within coastal rivers are sparse, this qualitative comparison of fish community composition prior to large scale habitat loss corroborates the assertion that macrophytes provide essential habitat for phytophilic species, and that the loss of this key habitat component may result in the decline or extirpation of these taxa S imilar pattern s of fish community effects were ob served by Whitfield (1986) after the loss of aquatic macrophytes in a coastal lake, Bettoli et al. (1993) following the removal of aquatic vegetation from a reservoir community and Deegan et al. (2002) from experimental habitat manipulations in seagrass c ommunities The spatial and temporal comparisons of the aquatic communities within the Chassahowitzka and Homosassa rivers demonstrate d how loss of a key habitat component may affect multiple trophic groups Macrophytes provide predation refuge for small bodied fishes and invertebrates such as Notropis spp. and aquatic insects; create a substrate for the colonization and production of periphyton which serves as a food base for chubsuckers and grass shrimp; and contribute to the detrital base which is uti lized by crayfish and detri ti vorous fishes such as striped mullet. In c ontrast, filamentous algae and its associated periphyton provides habitat for grazing amphipods and other invertebrates, which I observed in greater biomass during periods of increase d filamentous algae production, despite higher densit ies of saltwater fish predators. The extirpation of macrophytes and replacement with filamentous algae production may have cascading food web effects resulting in an altered community structure depende nt on benthic and algal food bases.

PAGE 65

65 Table 3 1 Mean e stimated biomass (g 100 m 2 ) of plants, algae, invertebrates and fishes within the Chassahowitzka and Homosassa r ivers, Florida. 1 Frazer et al. 2006 2 Frazer unpublished data 3 Camp et al. 20 11

PAGE 66

66 Tabl e 3 2 Freshwater fish species captured within the Chassahowitzka River, Florida. Scientific Name Common Name Ameiurus natalis Yellow bullhead Ameiurus nebulosus Brown bullhead Anguilla rostrata American eel Cyprinodon variegatus Sheepshead minnow Er imyzon sucetta Lake chubsucker Fundulus seminolis Seminole killifish Gambusia holbrooki Eastern mosquitofish Heterandria formosa Least killifish Lepisosteus osseus Longnose gar Lepisosteus platyrhincus Florida gar Lepomis gulosus Warmouth Lepomis ma crochirus Bluegill Lepomis microlophus Redear sunfish Lepomis punctatus Spotted sunfish Lucania goodei Bluefin killifish Lucania parva Rainwater killifish Menidia beryllina Inland silverside Micropterus salmoides Largemouth bass Notemigonus crysoleu cas Golden shiner Notropis harperi Redeye chub Notropis petersoni Coastal shiner Poecilia latipinna Sailfin molly

PAGE 67

67 Table 3 3 Saltwater fish species captured within the Chassahowitzka River, Florida. Scientific Name Common Name Anchoa mitchilli Bay a nchovy Archosargus probatocephalus Sheepshead Ariopsis felis Hardhead catfish Bairdiella chrysoura Silver perch Brevoortia sp. Menhaden Caranx hippos Crevalle jack Centropomus undecimalis Common snook Cynoscion nebulosus Spotted seatrout Dasyatis s p. Stingray Elops saurus Ladyfish Eucinostomus harengulus Tidewater mojarra Eucinostomus gula Silver jenny Fundulus confluentus Marsh killifish Fundulus grandis Gulf killifish Gobiosoma bosc Naked goby Lagodon rhomboides Pinfish Leiostomus xanthuru s Spot Lutjanus griseus Gray snapper Microgobius gulosus Clown goby Mugil cephalus Striped mullet Mugil curema White mullet Oligoplites saurus Leatherjacket Opsanus beta Gulf toadfish Sciaenops ocellatus Red drum Strongylura marina Atlantic needlef ish Strongylura notata Redfin needlefish Strongylura timucu Timucu Syngnathus scovelli Gulf pipefish Synodus foetens Lizardfish Trinectes maculatus Hogchoker

PAGE 68

68 Table 3 4 Freshwater fish species captured within the Homosassa River, Florida. Scientif ic Name Common Name Ameiurus natalis Yellow bullhead Ameiurus nebulosus Brown bullhead Anguilla rostrata American eel Cyprinodon variegatus Sheepshead minnow Erimyzon sucetta Lake chubsucker Esox niger Chain pickerel Fundulus seminolis Seminole kill ifish Gambusia holbrooki Eastern mosquitofish Heterandria formosa Least killifish Lepisosteus osseus Longnose gar Lepisosteus platyrhincus Florida gar Lepomis macrochirus Bluegill Lepomis microlophus Redear sunfish Lepomis punctatus Spotted sunfish Lucania goodei Bluefin killifish Lucania parva Rainwater killifish Menidia beryllina Inland silverside Micropterus salmoides Largemouth bass Notemigonus crysoleucas Golden shiner Notropis harperi Redeye chub Notropis petersoni Coastal shiner Poecil ia latipinna Sailfin molly

PAGE 69

69 Table 3 5 Saltwater fish species captured within the Homosassa River, Florida. Scientific Name Common Name Anchoa mitchilli Bay anchovy Archosargus probatocephalus Sheepshead Ariopsis felis Hardhead catfish Bagre marinus Gafftopsail catfish Bairdiella chrysoura Silver perch Brevoortia sp. Menhaden Caranx hippos Crevalle jack Centropomus undecimalis Common snook Cynoscion nebulosus Spotted seatrout Dasyatis sp. Stingray Eugerres plumieri Striped mojarra Echeneis sp. Sharksucker Elops saurus Ladyfish Eucinostomus gula Silver jenny Eucinostomus harengulus Tidewater mojarra Fundulus confluentus Marsh killifish Fundulus grandis Gulf killifish Gobiosoma bosc Naked goby Lagodon rhomboides Pinfish Leiostomus xanthur us Spot Lutjanus griseus Gray snapper Microgobius gulosus Clown goby Mugil cephalus Striped mullet Mugil curema White mullet Oligoplites saurus Leatherjacket Opsanus beta Gulf toadfish Pogonias cromis Black drum Sciaenops ocellatus Red drum Sphyra ena barracuda Barracuda Strongylura marina Atlantic needlefish Strongylura notata Redfin needlefish Strongylura timucu Timucu Syngnathus scovelli Gulf pipefish Synodus foetens Lizardfish Trinectes maculatus Hogchoker

PAGE 70

70 Figure 3 1. Average biomas s ( g m 2 SD ) of macrophytes within the Chassahowitzka and Homosassa rivers during August 2007 through August 2010 (n=15 samples per reach in each river). Biannual time series are shown for the period of study and monthly time series are shown for year th ree Reach 1 Reach 2 Reach 3

PAGE 71

71 Figure 3 2. Average biomass ( g m 2 SD ) of filamentous algae within the Chassahowitzka and Homosassa rivers during August 2007 through August 2010 (n=15 samples per reach in each river). Biannual time series are shown for the period of study an d monthly time series are shown for year three Reach 1 Reach 2 Reach 3

PAGE 72

72 Figure 3 3. Average density (invertebrates m 2 SD ) and biomass ( g m 2 SD ) of amphipods within the Chassahowitzka and Homosassa rivers during August 2007 through February 2010. Reach 1 Reach 2 Reach 3

PAGE 73

73 Figure 3 4. Avera ge density (invertebrates m 2 SD ) and biomass ( g m 2 SD ) of aquatic insects within the Chassahowitzka and Homosassa rivers during August 2007 through February 2010 Reach 1 Reach 2 Reac h 3

PAGE 74

74 Figure 3 5. Average density (invertebrates m 2 SD ) and biomass ( g m 2 SD ) of g astropods within the Chassahowitzka and Homosassa rivers during August 2007 through February 2010 Reach 1 Reach 2 Reach 3

PAGE 75

75 Figure 3 6. Average density (invertebrates m 2 SD ) and biomass ( g m 2 SD ) of isopods within the Chassahowitzka and Homosassa rivers during August 20 07 through February 2010 Reach 1 Reach 2 Reach 3

PAGE 76

76 Figure 3 7. Average density ( invertebrates m 2 SD ) and biomass ( g m 2 SD ) of tanaids within the Chassahowitzka and Homosassa rivers during August 2007 through February 2010 Reach 1 Reach 2 Reach 3

PAGE 77

77 Figure 3 8 Average biomass (g 100 m 2 SD ) of freshwater small bodied fishes collected at seine depletion sites within the Chassahowitzka and Homosassa rivers Biannual time series are shown for the period of study and monthly time series are shown for year three Reach 1 Reach 2 Reach 3

PAGE 78

78 Figure 3 9 Average bi omass (g 100 m 2 SD ) of saltwater small bodied fishes collected at seine depletion sites within the Chassahowitzka and Homosassa rivers. Biannual time series are shown for the period of study and monthly time series are shown for year three Reach 1 Reach 2 Reach 3

PAGE 79

79 Figure 3 10. Estimated mean biomass (g 100 m 2 ) of lake chubsucker captured during mark recapture electrofishing sampling within the Chassahowitzka and Homosassa rivers. Error bars represent 95% confidence intervals of the mean Biannual time series are shown for the period of study and monthly time series are shown for year three Reach 1 Reach 2 Reach 3

PAGE 80

80 Figure 3 11. Estimated mean biomass (g 100 m 2 ) of Lepomis spp. captured during mark recapture electrofishing sampling within the Chassahowitzka and Homosassa rivers. Error b ars represent 95% confidence intervals of the mean Biannual time series are shown for the period of study and monthly time series are shown for year three Reach 1 Reach 2 Reach 3

PAGE 81

81 Figure 3 12. Estimated mean biomass (g 100 m 2 ) of American eel captured during mark recaptu re electrofishing sampling within the Chassahowitzka and Homosassa rivers. Error bars represent 95% confidence intervals of the mean Biannual time series are shown for the period of study and monthly time series are shown for year three Reach 1 Reach 2 Reach 3

PAGE 82

82 Figure 3 1 3. Estimated mean biomass (g 100 m 2 ) of gar c aptured during mark recapture electrofishing sampling within the Chassahowitzka and Homosassa rivers. Error bars represent 95% confidence intervals of the mean Biannual time series are shown for the period of study and monthly time series are shown for year three Reach 1 Reach 2 Reach 3

PAGE 83

83 Figure 3 14. Estimated mean biomass (g 100 m 2 ) of largemouth bass captured during mark recapture electrofishing sampling within the Chassahowitzka and Homosassa rivers. Error bars represent 95% confidence intervals of the mean Biannual time series are shown for the period of study and monthly time series are shown for year three Reach 1 Reach 2 Reach 3

PAGE 84

84 Figure 3 15. Estimated mean biomass (g 100 m 2 ) of striped mullet captured during mark recapture electrof ishing sampling within the Chassahowitzka and Homosassa rivers. Error bars represent 95% confidence intervals of the mean Biannual time series are shown for the period of study and monthly time series are shown for year three Reach 1 Reach 2 Reach 3

PAGE 85

85 Figure 3 16. Estimat ed mean biomass (g 100 m 2 ) of pinfish captured during mark recapture electrofishing sampling within the Chassahowitzka and Homosassa rivers. Error bars represent 95% confidence intervals of the mean Biannual time series are shown for the period of stud y and monthly time series are shown for year three Reach 1 Reach 2 Reach 3

PAGE 86

86 Figure 3 17. Estimated mean biomass (g 100 m 2 ) of sheepshead captured during mark recapture electrofishing sampling within the Chassahowitzka and Homosassa rivers. Error bars represent 95% confide nce intervals of the mean Biannual time series are shown for the period of study and monthly time series are shown for year three Reach 1 Reach 2 Reach 3

PAGE 87

87 Figure 3 18. Estimated mean biomass (g 100 m 2 ) of gray snapper captured during mark recapture electrofishing samplin g within the Chassahowitzka and Homosassa rivers. Error bars represent 95% confidence intervals of the mean Biannual time series are shown for the period of study and monthly time series are shown for year three Reach 1 Reach 2 Reach 3

PAGE 88

88 Figure 3 19. Estimated mean biomas s (g 100 m 2 ) of red drum captured during mark recapture electrofishing sampling within the Chassahowitzka and Homosassa rivers. Error bars represent 95% confidence intervals of the mean Biannual time series are shown for the period of study and monthly time series are shown for year three Reach 1 Reach 2 Reach 3

PAGE 89

89 Figure 3 20. Estimated mean biomass (g 100 m 2 ) of common snook captured during mark recapture electrofishing sampling within the Chassahowitzka and Homosassa rivers. Error bars represent 95% confidence interval s of the mean Biannual time series are shown for the period of study and monthly time series are shown for year three Reach 1 Reach 2 Reach 3

PAGE 90

90 CHAPTER 4 THE DIET HABITS OF FRESHWATER AND MARINE FISHES IN COASTAL RIVERS Introduction Predator prey interactions have been a central focus of ecology for nearly a century (Berryman 1992), and much ecological insight has been gained from examining how predators interact with their prey and vice versa. From the pioneer ing work of Lotka (1925) and Volterra (1931) in constructing differen tial equations mathematically describing predator prey interaction to the development of more complex theories to account for behavioral responses (Holling 1959, Preisser et al. 2005), habitat spatial structure (Huffaker 1958, McArthur and Pianka 1966) and trophic cascades ( Carpenter et al. 1985, Pace et al. 1999 ), the study of predation has provided valuable insight into principle factors that influence the distributions and densities of populations. Predators exert a strong control over prey populations through direct consumption and intimidation (Preisser et al. 2005) particularly in open water food webs (Carpenter and Kitchell 199 6, Mi cheli 1999) Predator abundance and foraging is coupled to prey availability through density dependent rates of reprod uction, survival, consumption, or growth (Soloman 1949, Holling 1959, Murdoch 1971). Controlled experiments on predator prey interactions have demonstrated that predators left unchecked can deplete prey populations, resulting in population crashes (Huffak er 1958) or prey switching (Murdoch 1969). For predator and prey species to coexist, systematic check s and balances must exist within ecosystems to prevent the rapid extinction of populations. Prey species face the serious dilemma of acquiring enough foo d to grow and reproduce without being preyed upon while foraging (Walters and Martell 2004). This dilemma represents a behavioral trade off between feeding in areas with available prey,

PAGE 91

91 and resting in habitats with lower predation risk ( Gilliam and Frase r 1987, Walters and Martell 2004). The presence of predators can cause spatially distinct habitat use patterns of prey populations, including the avoidance of areas with high predator abundance (Preisser et al. 2005). This avoidance behavior can result i n greater food resources for prey species in areas that overlap predators. Similarly, prey populations can quickly deplete food resources within refug e habitats, and must therefore move to areas with greater food availability to forage These coupled be havioral dynamics a re often not effectively captured by mass action mode ls of predator prey interaction (Walters and Martell 2004) ; however, models that incorporate predation vulnerability associated with spatially restricted foraging arenas ( Walters and J uanes 1993, Walter s and Martell 2004) have greatly improved the ability to predict population responses to changes in both food availability and predation This predictive ability is useful for modeling food web interactions and multispecies dynamics in a quatic ecosystems and predicting ecosystem responses to manipulations in predators (i.e. top down, fishing), prey (i.e. bottom up, prey composition and abundance), or ecosystem attributes that ma y impact predator prey dynamics (e.g., habitat structure) ( Ch ristensen and Pauly 1993, Walters and Martell 2004). Quantitative characterization of predator prey interactions is central to trophic dynamic models aimed at understanding population and community level effects of changes in producer and consu mer popula tions If such interactions can be accurately characterized, then predictive models can be developed and used to assess and screen policy options related to managing and restoring ecosystems through manipulation of predators or

PAGE 92

92 prey populations directly ( e.g., harvest regulations) or indirectly through mediating predator prey interactions (e.g habitat modification or restoration). Habitat complexity has been shown to have large decoupling effects on predator prey interactions within aquatic ecosystems. Highly structured habitats can inhibit predator foraging and decrease predation efficiency (Huffaker 195 8 Crowder and Cooper 1982, Savino and Stein 1982), creating distinct refuge patches for prey species. Predators may aggregate and forage in patches c ontaining a high availability of prey, and move to other, more profitable patches when prey populations become less abundant (Murdoch 1969, Valiela 1995). Patch connectivity allows for recolonization of depleted patches from densely populated ones. Withi n highly connected patches, the rate of colonization can offset the rate of predator depletion (local prey extinction). These patch dynamics can stabilize prey populations by lowering the probability of extinction (Fahrig and Merriam 1985, Namba et al. 19 99). P redation on individual prey species has been shown to be lower in ecosystems with highly heterogeneous habitats that support multiple prey populations due to lower predator encounter rate per prey species ( Baalen et al. 2001 ) When prey encounters become rare, predators may exploit alternative populations that occur in greater abundance (Murdoch 1969). This p rey switching behavior can i ncrease the persistence of prey populations (Comins and Hassel 1975, van Baalen et al. 2001) as a result of decre ased predation rates at low prey densities Habitats that provide abundant food resources, low predation risk, and support a high diversity of prey species, are therefore thought to be important for maintaining predat or and prey population viability. How ever, how predator communities respond to large scale changes in habitat, and ultimately

PAGE 93

93 how this relates to predator prey dynamics and ecosystem effects is most often assessed at small scales (aquaria or mesocosms) with few predators or prey ( Jacobson and Berg 1989, Savino and Stein 1989 ), but has recently been expanded experimentally to small lakes (Sass et al. 2006). At larger spatial scales, manipulative experiments of predators, prey populations and habitat s may not be logistically feasible, particul arly in lotic ecosystems. In these ecosystems, c omparison of predator prey interactions across spatially and temporally variably habitat availability and composition can provide insight into how predator populations respond to large scale changes in habit at structure and prey availability. An ecosystem study of coastal rivers in Florida provided an opportunity to assess how spatially and temporally dynamic changes in vegetative habitat affected the prey availability, composition, selectivity and foraging s uccess of freshwater and marine fishes. Spring fed, coastal rivers are unique, highly autochthonous ecosystems comprised of diverse communities of oligohaline and marine plants, algae, invertebrates and fishes ( H erald and Strickland 1949, Odum 1953, Odum 1957). Several rivers along the west coast of Florida including the Chassahowitzka and Homosassa rivers, historically supported dense assemblages of aquatic macrophytes; however, macrophyte fragmentation and loss have been significant during the last dec ade (Frazer et al. 2006). Currently, the Chassahowitzka River support s approximately half of the biomass of macrophytes compared to estimates from a decade ago, and nearly all macrophytes, including Vallisneria americana Potamogeton spp., and Sagittaria kurziana have declined significantly or been extirpated from the Homosassa River (Figure 4 1). Furthermore, large scale seasonal blooms of filamentous algae, including

PAGE 94

94 Chaetomorpha sp., Gracilaria sp., and Lyngbya sp., cover extensive portions of the riv erbeds in late winter through spring (Figure 4 1). The Homosassa River is nearly devoid of aquatic vegetation during summer and fall months, whereas the Chassahowitzka supports macrophytes throughout the year These plant and algae dynamics have resulted in structurally different and seasonally variable vegetative habitats within each river (Figure 4 1). Comparative analysis of predator prey interactions between fishes and invertebrates within these systems over time may help elucidate predominant e ffe cts of macrophyte extirpation and seasonally abundant algae habitat on the food habits of fishes. I examined the diet patterns of four species of fishes, Lepomis punctatus Micropterus salmoides Lagodon rhomboides and Lutjanus griseus within the Chassah owitzka and Homosassa rivers to evaluate the spatia l and temporal heterogeneity in prey composition, selection, and foraging success associated with large scale changes in vegetative habitat and prey availability These species were selected because they represent a spectrum of functional feeding guilds and were common in both rivers during the period of study. I compared diet information between a highly vegetated river, the Chassahowitzka River, and one where macrophytes have been largely absent since 2 006, the Homosassa River. I evaluated a combination of diet indices (Chipps and Garvey 2007) to evaluate (1) differences in prey composition between rivers and seasons, ( 2) prey selectivity of each species and ( 3 ) p redator relative foraging success in ea ch river By coupling quantitative estimates of aquatic vegetation and prey biomass es (Chapter 3) with information on the food habits of fishes, I provide an assessment of community structure necessary to evaluate combined

PAGE 95

95 bottom up and top down effects o f altered vegetation composition and biomass on fish and invertebrate populations in coastal aquatic ecosystems Methods Diet Sampling and Laboratory Procedures The diet contents of L. punctatus, M. salmoides, L. rhomboides and L. griseus were sampled usi ng a nonlethal stomach flushing method (Kamler and Pope 2001) for individuals greater than or equal to 150 mm in total length, and by sacrificing individuals less than 150 mm in total length for diet analysis by dissection Diet sampling was conducted in conjunction with electrofishing sampling for large bodied fishes (Chapter 3); sampling occurred during January and July of years one and two and monthly during year three The gastric lavage apparatus used to flush stomachs comprised 1.2 m of 9.5 mm viny l tubing attached to a 1,9 00 or 2,80 0 liter per minute bilge pump at one end, and a pistol grip, plastic hose nozzle at the opposite end. V inyl tubing (3.1 mm diameter) was attached to the output of the nozzle for insertion through the esophagus of the fish. The bilge pump was anchored to the bottom of a 19 L plastic bucket and connected to a 12 V marine cell battery with a switch installed on the cathode wire. The bucket was filled with freshwater so that the pump generated a steady stream of water w hen the switch was turned on and the trigger was compressed. The 3.1 mm with water, and the stomach contents were flushed into a plastic funnel by holding the fish mouth down and pr essing its stomach inward until the fish extruded the water and gut contents. The plastic collection funnel had a rubber stopper inserted into the bottom to prevent items from washing through, with holes drilled into the lower half of the funnel and cover ed with a 300 filter mesh fabric to allow water drainage but prevent the

PAGE 96

96 loss of prey items larger than 300 The samples were washed from the funnel into a sealable plastic bag by pulling the rubber stopper and rinsing the inside of the funnel. Each sample was labeled with a unique diet identification number Samples were transported on ice to the laboratory and frozen until process ed In the laboratory, diet sample s were thawed, and then rinsed in to a 300 sieve The contents were removed from the sieve a nd placed in a petri dish for examination under a stereo dissecting microscope with magnification to 43x. For sacrificed fish, the stomach was removed and the contents were emptied into a petri dish for examination under the dissecting scope. I ndividual diet items were identified to the lowest possible taxonomic unit, dried in an oven at 70 C for a period of 24 h and then weighed at room temperature When individual diet items could not be separated effectively, I recorded the approximate percent composi tion of each diet item along with the combined weight of all items. I then multiplied the percent composition by total diet weight to approximate the individual weight of each diet item. Filamentous algae, detritus, plant material, and sediment were exclu ded from diet indices of L. punctatus M. salmoides and L. griseus ; however these prey groups were encountered frequently in mouths ( M. salmoides particularly) and stomach samples. Detritus and sediment were excluded from the analysis of L. rhomboides ; however, filamentous algae and plants were included in the diet indices since this species has been shown to be omnivorous in vegetated habitats ( Montgomery and Targett 1992 ) Several invertebrate and fish taxa were combined into prey groups, including in vertebrates associated with vegetation benthic invertebrates, freshwater small bodied fishes, saltwater small bodied fishes, terrestrial invertebrates and

PAGE 97

97 terrestrial vertebrates. Vegetation associated invertebrates comprised copepods, gastropods, insect pupae/larvae, isopods, and tanaids. Benthic invertebrates comprised bivalves, nematodes, oligochaetes, ostracods, and polychaetes. Several invertebrate taxa were collected in both benthic and vegetative habitats (e.g., copepods, nematodes, ostracods, in sect larvae); these taxa were grouped by habitat in which they were observed in greater abundance. Freshwater small bodied fishes comprised Fundulus spp., Gambusia sp., Lucania spp., Menidia sp., Notropis spp., and Poecilia sp. Saltwater small bodied fis hes comprised Anchoa sp., Brevoortia sp., Eucinostomus spp., Gobiosoma sp., Microgobius sp., Strongylura spp., and Trinectes sp. Terrestrial invertebrates comprised arachnids, coleopterans, diplopodans, hymenopterans, lepidopterans, and orthopterans. Ter restrial vertebrates comprised bullfrogs, lizards, snakes, and juvenile waterfowl (observed in Micropterus salmoides within the Homosassa River). Prey Composition Indices The prey composition of each fish species was estimated for the Chassahowitzka and Ho mosassa rivers during summer months (June, July, and August) and winter months (December, January, and February) using mean proportion by dry weight and frequency of occurrence indices (Chipps and Garvey 2007). Diet samples were pooled across years to est imate mean proportion by river and season. The mean proportion by dry weight (MW i ) provided an estimate of the relative importance of each prey group to the predator, and was calculated as:

PAGE 98

98 (4 1) where: P = number of fish with food in their stomachs, i = prey type, j = fish stomach sample, W = dry weight of prey item, Q = total number of prey types The mean frequency of occurrence (O i ) provided an estimate of how often individual prey groups were observed in fish diets, an d was estimated as: (4 2) where: i = prey type, J = number of fish containing prey i P = number of fish with food in their stomachs the test statistic) ( Zar 1999 ) to test for significant differences in prey composition, measured as proportion by dry weight, between rivers and between seasons by river for each fish species. Significance level was 0.05 for all analysis of variance tests. I n these analyses, the proportion by dry weight of each prey group was the response variable, individual fish were treated as replicates (pooled across years), and river or river*season was the treatment variable(s). I then conducted an analysis of varianc e for individual prey groups to assess which groups were significantly different between the rivers and seasons.

PAGE 99

99 Prey Selectivity Indices To examine predator preference for individual prey groups within each river, I calculated prey selectivity indices (Ma nly et al. 197 2 Chesson 198 3 ) using empirical estimates of average prey group biomass (Table 3 1) and mean prey proportion by dry weight in diets Prey selectivity indices ( i ) were calculated separately by river and season (winter and summer) to assess how production of filamentous algae during winter affected the prey selection of predators. Prior to calculating the selectivity indices, estimates of prey biomass and mean p roportion by dry weight were normalized across all prey groups for each season and river. Prey selectivity indices were calculated as: (4 3) where: MW i = mean proportion by dry weight of prey group i B i = estimated biomass of prey group i Q = total number of prey groups i Predator foraging was assumed to be nonselective when i = 1/Q across prey groups. Values of i >1/Q indicate d preference for that prey group (mean proportion in diet > proportion of estimated prey biomass), and values of i <1/Q indicate d avoidance (mean proportion in diet < proportion of estimated prey bioma ss). Relative Foraging Success To examine spatial and temporal pattern s in predator foraging success, I calculated the proportion of empty stomachs (i.e. how many fish sampled contained

PAGE 100

100 prey) and the mean total prey dry weight per predator body weight. Th ese metrics were calculated for individual sampling events to assess seasonal (years one two and three ), interannual (years one two and three ) and intra annual ( year three ) pattern s in stomach emptiness and relative prey consumption within each river, and to compare spatially between the rivers. The proportion of empty stomachs (E j ) was calculated as: (4 4) where: P j = Number of fish of predator j containing prey N j = Total number of fish sampled of predator j The me an prey dry weight per predator body weight (MBW j ) provided a relative measure of the amount of prey consumed per predator, and was calculated as: (4 5) where: P = number of fish with food in their stomachs, j = fish sto mach sample, W = total dry weight of all prey items, B = predator body weight Predator body weight was calculated from measurements of predator total length and length weight regression parameters from empirical length and weight data collected from the C hassahowitzka and Homosassa rivers (Lauretta, unpublished data) The estimates of proportion of empty stomachs and mean prey dry weight per predator body weight were plotted to visually compare between rivers and across sampling events. I use d these comp arisons to test whether predator foraging success is lower in the Homosassa River compared to the Cha ssahowitzka River

PAGE 101

101 Results A total of 1,115 diet samples of L. punctatus 1,155 diet samples of M. salmoides 863 diet samples of L. rhomboides and 863 d iet samples of L. griseus were collected from the Chassahowitzka River between July 2007 and June 2010. A total of 411 diet samples of L. punctatus 705 diet samples of M. salmoides 393 diet samples of L. rhomboides and 699 diet samples of L. griseus we re collected from the Homosassa River during that period. Many diet samples contained unidentifiable prey items, on average about one out of every ten diets across species, with approximately five percent of the prey mass measured being indistinguishable between invertebrate or fish prey, on average. Over half of the diet samples of M. salmoides contained unidentifiable fishes. Invertebrates were more commonly identified by their hard parts and less error was associated with diet determination of L. punc tatus, L. rhomboides and L. griseus As unidentified prey items may cause bias in the prey composition and selectivity indices unidentified crustaceans, total crustaceans, unidentified invertebrates, total invertebrates, unidentified fish, and total fis h prey were included as distinct categories in the diet composition indices. In addition, I calculated selectivity indices for freshwater and saltwater invertebrates and fishes separately. The bias from unidentified items is least for the coarsest resolu tion index (total fishes vs. total invertebrates) since I wa s able to effectively incorporate unidentified groups as either invertebrate or fish prey taxa. Results for individual species follow. Lepomis punctatus Prey composition Diets of L. punctatus fro m the Chassahowitzka River contained a high proportion of amphipods, followed by vegetati on associated invertebrates (Table 4 1). Samples

PAGE 102

102 from the Homosassa River contained a lower proportion of amphipods in comparison, especially during summer months (Ta ble 4 1). Mud crabs (Grapsidae and Xanthidae) and terrestrial invertebrates were consumed in greater proportion during summer in the Homosassa River (Table 4 1) when vegetation and invertebrate biomass was lower. Overall, aquatic invertebrates comprised 80 to 90% of prey taxa for L. punctatus in both rivers, followed by terrestrial invertebrates and small bodied fishes. Multivariate analysis of variance indicated significantly different diet compositions between rivers (p<0.001). Analysis of variance b y individual prey groups indicated that mean proportion by dry weight was significantly different between rivers for Amphipoda (p<0.001), Cambaridae (p value=0.02), Grapsidae/Xanthidae (p<0.001), vegetation associated invertebrates (p<0.001), benthic inver tebrates (p<0.001), and terrestrial invertebrates (p value=0.006); and not significantly different for Callinectes sapidus (p value=0.45) or Palaemonetes spp. (p value=0.92). Multivariate analysis of variance indicated significant differences in the seas onal diet composition of L. punctatus in the Chassahowitzka River (p<0.001) and the Homosassa River (p<0.001). Analysis of variance on individual prey groups indicated that Amphipoda (p<0.001), Palaemonetes spp. (p value=0.04), Cambaridae (p value=0.03), and terrestrial invertebrate (p<0.001) mean proportion by dry weight differed between seasons in the Chassahowitzka River; no significant differences in mean proportion by dry weight were detected between seasons for Grapsidae/Xanthidae (p value=0.24), Cal linectes sapidus (p value=0.54), vegetation associated invertebrates (p value=0.10), or benthic invertebrates (p value=0.17). In the Homosassa River, mean proportion by dry weight differed significantly between

PAGE 103

103 seasons for Amphipoda (p<0.001), Grapsidae/X anthidae (p<0.001), and terrestrial invertebrates (p<0.001). No significant difference between seasons was indicated for Palaemonetes spp. (p = 0.61), Cambaridae (p value=0.48), Callinectes sapidus (p value=0.37), vegetation associated invertebrates (p va lue=0.63), or benthic invertebrates (p value=0.20). Amphipods were present in 80 to 90% of L. punctatus samples within the Chassahowitzka River during summer and winter. Diets of L. punctatus from the Homosassa River during summer had a lower frequency of occurrence of amphipods (34%) compared to winter (71%). Vegetation associated invertebrates were also commonly encountered in both rivers and during both sampl ing seasons (Table 4 2). Other common taxa observed during summer included terrestrial inverte brates in the Chassahowitzka River, and Grapsidae/Xanthidae, benthic invertebrates, and terrestrial invertebrates in the Homosassa River (Table 4 2). Prey selectivity Manly Chesson indices indicated that L. punctatus selectively foraged on freshwater inver tebrates in the Chassahowitzka River, particularly Amphipoda and other invertebrates associated with vegetation during both the summer and winter months (Table 4 3). L epomis punctatus in the Homosassa River also selectively foraged on invertebrates, with distinct differences between seasons (Table 4 3). During summer, fish in the Chassahowitzka River selected for benthic invertebrates and those associated with vegetation and during winter fish selected for Amphipoda and Palaemonetes spp. Fish in the Hom osassa River selectively foraged on saltwater invertebrates in addition to freshwater invertebrates during summer months.

PAGE 104

104 Relative foraging success The majority of L. punctatus diets from both rivers contained prey, with <0.05 proportion of empty stomach s observed for most sampling events. The proportion of empty stomach s from the Homosassa River was equal to or less than the proportion from the Chassahowitzka River for all sample periods with the exception of the December 2009 and January 2010 (Figures 4 2 and 4 3). Mean total prey dry weight per predator body weight indices for L. punctatus in the Homosassa River were equal to or greater than indices from the Chassahowitzka River across all sample events with the exception of February 2010 and March 2010 sample periods (Figures 4 4 and 4 5). A seasonal increase in mean prey dry mass per predator body weight was observed during late spring and early summer in both rivers (Figure 4 5), with a higher mean observed in the Homosassa River during summer period s compared to winter (Figures 4 4 and 4 5). Micropterus salmoides Prey composition Diets of M. salmoides from the Chassahowitzka River during summer contained a high proportion of Cambaridae, freshwater small bodied fishes, and unidentified fish (Table 4 4 ). Winter sampling indicated a higher composition of Amphipoda, saltwater small bodied fishes and Palaemonetes spp., and lower proportion of Cambaridae and freshwater small bodied fishes, compared to summer. Samples from the Homosassa River contained a lo wer proportion of Cambaridae, particularly during summer months; a higher proportion of Grapsidae/Xanthidae, Palaemonetes spp., and saltwater small bodied fishes; and an overall higher proportion of fish during both seasons (Table 4 4).

PAGE 105

105 Multivariate anal ysis of variance indicated significantly different diet compositions between rivers (p<0.001). Analysis of variance of individual prey groups indicated that mean proportion by dry weight was significantly different between rivers for Amphipoda, (p value=0 .002), Palaemonetes spp. (p<0.001), Cambaridae (p<0.001), Grapsidae/Xanthidae (p<0.001), vegetation associated invertebrates (p<0.001), saltwater small bodied fishes (p<0.001), and juvenile Micropterus salmoides (p value=0.03). No significant difference in mean proportion by dry mass was detected for Callinectes sapidus (p value=0.34), benthic invertebrates (p value=0.28), freshwater small bodied fishes (p value=0.76), Lepomis spp. (p value=0.48), Erimyzon sucetta (p value=0.39), Lagodon rhomboides (p va lue=0.25), Lutjanus griseus (p value=0.34), terrestrial invertebrates (p value=0.92), or terrestrial vertebrates (p value=0.42). Multivariate analysis of variance indicated significant differences in the seasonal diet composition of M. salmoides in the C hassahowitzka River (p<0.001) and the Homosassa River (p value=0.002). Analysis of variance on individual prey groups indicated that Amphipoda (p<0.001), Palaemonetes spp. (p value=0.001), Cambaridae (p<0.001), Callinectes sapidus (p value=0.05), freshwat er small bodied fishes (p value=0.006), saltwater small bodied fishes (p<0.001), Lutjanus griseus (p value=0.01) and terrestrial vertebrates (p value=0.04) proportion by dry weight differed between seasons in the Chassahowitzka River. No significant diffe rences in mean proportion by dry weight were detected between seasons for Grapsidae/Xanthidae (p value=0.48), vegetation associated invertebrates (p value=0.22), benthic invertebrates (p value=0.42), Lepomis spp. (p value=0.63), Erimyzon sucetta (p value=0 .21), juvenile Micropterus salmoides (p value=0.43), Lagodon rhomboides (p value=0.26), Mugil

PAGE 106

106 cephalus (p value=0.43), Archosargus probatocephalus (p value=0.21), and terrestrial invertebrates (p value=0.23). In the Homosassa River, mean proportion by dry weight differed significantly between seasons for Amphipoda (p<0.001) and saltwater small bodied fishes (p value=0.04). No significant difference between seasons was indicated for Palaemonetes spp. (p = 0.35), Cambaridae (p value=0.11), Grapsidae/Xanthid ae (p value=0.21), Callinectes sapidus (p value=0.,08), vegetation associated invertebrates (p value=0.37), benthic invertebrates (p value=0.39), freshwater small bodied fishes (p value=0.34), Lepomis spp. (p value=0.91), juvenile Micropterus salmoides (p value=0.71), Lagodon rhomboides (p value=0.55), Archosargus probatocephalus (p value=0.52), Lutjanus griseus (p value=0.12), terrestrial invertebrates (p value=0.33), or terrestrial vertebrates (p value=0.30). Unidentified fish remains w ere the most common diet item encountered in M. salmoides diets in the Chassahowitzka and Homosassa rivers during both seasons. The highest frequency of occurrence of identified prey groups in the Chassahowitzka River was observed for Cambaridae, freshwater small bodied fis hes, vegetation associated invertebrates and Amphipoda during summer months; and Amphipoda, freshwater small bodied fishes, Cambaridae, saltwater small bodied fishes, and Palaemonetes spp. during winter months (Table 4 5). In the Homosassa River during s ummer, freshwater small bodied fishes, Palaemonetes spp., saltwater small bodied fishes and Cambaridae were the most commonly encountered identifiable prey groups (Table 4 5). During winter, Amphipoda had the highest frequency of occurrence, followed by f reshwater small bodied fishes, saltwater small bodied fishes, and

PAGE 107

107 Palaemonetes spp., all of which occurred in at least one out of every ten diet samples (Table 4 5). Prey selectivity Manly Chesson indices indicated that M. salmoides selectively foraged on freshwater small bodied fishes across both seasons and in both rivers (Table 4 6). M icropterus salmoides also selectively cannibalized juveniles during summer months in both rivers. Other prey groups that were selected for during summer included Cambarid ae, L rhomboides vegetation associated invertebrates and Amphipoda in the Chassahowitzka River; and Lepomis spp. and vegetation associated invertebrates in the Homosassa River. Overall, M. salmoides selectively foraged on fishes, especially freshwater fishes, during summer in the Chassahowitzka and winter in both rivers. Selection was greater for invertebrates (freshwater taxa) during summer months in the Homosassa River. Relative foraging success The proportion of empty stomach s increased during winte r months in the Chassahowitzka River, but this pattern was not apparent across years in the Homosassa River (Figures 4 2 and 4 3). A seasonal pattern in mean total prey dry weight per predator body weight was observed in the Homosassa River; means were hi gher during summer months compared to winter (Figures 4 4 and 4 5). This pattern was not apparent in the Chassahowitzka River. Overall, I did not observe a higher proportion of empty stomachs or lower mean total prey dry mass per predator body weight in the Homosassa River, despite lower prey biomass (Cambaridae and freshwater small bodied fishes, in particular).

PAGE 108

108 Lagodon rhomboides Prey composition Diets of L. rhomboides from the Chassahowitzka and Homosassa rivers during summer and winter contained most ly filamentous algae, Amphipoda, and vegetation associated invertebrates (Table 4 7). Diets taken during winter from the Homosassa River contained a higher proportion of Amphipoda compared to summer months (Table 4 7). Multivariate analysis of variance indicated significantly different diet compositions between rivers (p<0.001). Analysis of variance of individual prey groups indicated that mean proportion by dry weight was significantly different between rivers for filamentous algae, (p<0.001), macrophy tes (p<0.001), Cambaridae (p value=0.006), Grapsidae/Xanthidae (p<0.001), and benthic invertebrates (p value=0.05). No significant difference in mean proportion by dry mass was detected for Amphipoda (p value=0.49), Palaemonetes spp. (p value=0.25), Cal linectes sapidus (p value=0.10), vegetation associated invertebrates (p value=0.76), saltwater small bodied fishes (p value=0.49), or terrestrial invertebrates (p value=0.34). Multivariate analysis of variance indicated significant differences in the sea sonal diet composition of L. rhomboides in the Chassahowitzka River (p<0.001) and the Homosassa River (p<0.001). Analysis of variance on individual prey groups indicated that macrophytes (p<0.001), Amphipoda (p value=0.02), Palaemonetes spp. (p value=0.00 7), Cambaridae (p value=0.005), and benthic invertebrates (p value=0.05) proportion by dry weight differed between seasons in the Chassahowitzka River; no significant differences in mean proportion by dry weight were detected between seasons for filamentou s algae (p value=0.08), Callinectes sapidus (p value=0.15),

PAGE 109

109 vegetation associated invertebrates (p value=0.77), saltwater small bodied fishes (p value=0.47), or terrestrial invertebrates (p value=0.23). In the Homosassa River, mean proportion by dry weigh t differed significantly between seasons for filamentous algae (p<0.001), Amphipoda (p<0.001), and Grapsidae/Xanthidae (p value=0.01). No significant difference between seasons was indicated for macrophytes (p value=0.73), Palaemonetes spp. (p = 0.36), Ca mbaridae (p value=0.41), vegetation associated invertebrates (p value=0.23), benthic invertebrates (p value=0.07), or terrestrial invertebrates (p value=0.56). The most frequently identified prey groups of L. rhomboides in the Chassahowitzka River were fil amentous algae, Amphipoda, and vegetation associated invertebrates during both summer and winter (Table 4 8). In the Homosassa River, filamentous algae and amphipods were also commonly encountered in diets; however the frequency of Amphipoda in the diets was more than double during winter than during summer (Table 4 8). Prey selectivity Manly Chesson indices indicated that L. rhomboides selectively foraged on Amphipoda and benthic invertebrates in both rivers during summer and winter (Table 4 9). Fish al so selected for vegetation associated invertebrates during winter in the Chassahowitzka River and summer in the Homosassa River. Relative foraging success L agodon rhomboides diets were rarely empty, as a result of a high proportion of diets containing fila mentous algae and the inclusion of this prey group in the diet indices. Overall, I did not observe a higher proportion of empty stomachs in the Chassahowitzka River during any sampling event with the exception of November 2009

PAGE 110

110 (Figures 4 2 and 4 3). Mean total prey dry weight per predator body weight was less variable in the Chassahowitzka River compared to the Homosassa River, with a seasonally higher mean in the Homosassa River during summer months (Figure s 4 4 and 4 5). In general, mean total prey dry mass per predator body weight was greater in the Homosassa River for most sampling events with the exception of January 2009 and January 2010. Lutjanus griseus Prey composition Diets of L. griseus from the Chassahowitzka and Homosassa rivers during summer contained approximately half fishes and half invertebrates with a large proportion of unidentified fish; samples from winter contained greater than 75% invertebrates, mostly crustaceans (Table 4 10). L. griseus consumed a range of crustaceans, freshwater fishes, and sal twater fishes during summer in both rivers. In winter, mean prey proportion by dry weight was greatest for Amphipoda in the Chassahowitzka River, and Grapsidae/Xanthidae and Amphipoda in the Homosassa River. Multivariate analysis of varian ce indicated significantly different diet compositions between rivers (p<0.001). Analysis of variance of individual prey groups indicated that mean proportion by dry weight was significantly different between rivers for Amphipoda, (p<.001), Cambaridae (p< 0.001), Grapsidae/Xanthidae (p<0.001), and vegetation associated invertebrates (p value=0.01). No significant difference in mean proportion by dry mass was detected for Palaemonetes spp. (p value=0.07), Callinectes sapidus (p value=0.99), benthic inverteb rates (p value=0.92), freshwater small bodied fishes (p

PAGE 111

111 value=0.60), Lepomis spp. (p value=0.86), saltwater small bodied fishes (p value=0.06), Lagodon rhomboides (p value=0.20), or terrestrial invertebrates (p value=0.36). Multivariate analysis of varia nce indicated significant differences in the seasonal diet composition of L. griseus in the Chassahowitzka River (p<0.001) and the Homosassa River (p<0.001). Analysis of variance on individual prey groups indicated that Amphipoda (p<0.001), Palaemonetes s pp. (p value=0.02), vegetation associated invertebrates (p value=0.04), freshwater small bodied fishes (p<0.001), Lepomis spp. (p value=0.007), and L rhomboides (p value=0.03) proportion by dry weight differed between seasons in the Chassahowitzka River. No significant differences in mean proportion by dry weight were detected between seasons for Cambaridae (p value=0.83), Grapsidae/Xanthidae (p value=0.37), Callinectes sapidus (p value=0.27), benthic invertebrates (p value=0.13), saltwater small bodied f ishes (p value=0.24), or terrestrial invertebrates (p value=0.71). In the Homosassa River, mean proportion by dry weight differed significantly between seasons for Amphipoda (p<0.001), Palaemonetes spp. (p value=0.01), and freshwater small bodied fishes ( p value=0.05). No significant difference between seasons was indicated for Cambaridae (p value=0.08), Grapsidae/Xanthidae (p value=0.32), Callinectes sapidus (p value=0.85), vegetation associated invertebrates (p value=0.91), benthic invertebrates (p valu e=0.53), Lepomis spp. (p value=0.10), or saltwater small bodied fishes (p value=0.50). Unidentified fish remains w ere the most common diet item encountered in L. griseus diets in the Chassahowitzka and Homosassa rivers during summer. The highest frequency of occurrence of identified prey groups in the Chassahowitzka River

PAGE 112

112 was observed for Amphipoda, unidentified crustacean, freshwater small bodied fishes, Palaemonetes spp., Grapsidae/Xanthidae, Cambaridae, and vegetation associated invertebrates during sum mer months; and Amphipoda, vegetation associated invertebrates Cambaridae and Grapsidae/Xanthidae during winter months (Table 4 11). In the Homosassa River during summer, freshwater small bodied fishes, Grapsidae/Xanthidae, Palaemonetes spp., and vegetat ion associated invertebrates were the most commonly encountered identifiable prey groups (Table 4 11). During winter, the highest frequency of occurrence was observed for Amphipoda and Grapsidae/Xanthidae, followed by vegetation associated invertebrates a nd Palaemonetes spp. (Table 4 11). Prey selectivit y Manly Chesson indices indicated that L. griseus selectively foraged on Grapsidae/Xanthidae, Amphipoda, L rhomboides freshwater small bodied fishes and Palaemonetes spp. during summer in the Chassahowitz ka River (Table 4 12). During winter, L. griseus selectively foraged on Grapsidae/Xanthidae and freshwater small bodied fishes. L utjanus griseus in the Homosassa River selected for Grapsidae/Xanthidae during summer months, and Palaemonetes spp., freshwat er small bodied fishes, and Amphipoda during winter. Overall, L. griseus selectively foraged on fishes in the Chassahowitzka River during summer and freshwater invertebrates during winter. Selection was greatest for saltwater invertebrates (mainly Grapsi dae/Xanthidae) during summer months in the Homosassa River, and freshwater invertebrates, saltwater invertebrates, and freshwater fishes during winter months.

PAGE 113

113 Relative foraging success The proportion of empty stomach s was not higher in the Homosassa River compared to the Chassahowitzka River across most sampling periods (Figure 4 2), although a seasonal increase in the proportion was observed during November, December, and January in the Homosassa River (Figure 4 3) when the density of L. griseus increased greatly as fish migrated into the rivers from the Gulf of Mexico (Figure 3 18 ). No seasonal pattern in mean total prey dry weight per predator body weight was detected across years in either river (Figure 4 4), but an increase in the mean was observed dur ing late spring in 2010 (Figure 4 5) when filamentous algae production was high in both rivers (Figure 4 1). Overall, I did not observe a higher proportion of empty stomachs or lower mean total prey dry mass per predator body weight in the Homosassa River compared to the Chassahowitzka River. Discussion Vegetative habitat in coastal rivers influences the prey composition, selection, and relative consumption of fishes through bottom up controls on crustaceans and other invertebrate prey, and by mediating to p down controls by freshwater and marine predators on prey populations. I found that aquatic vegetation composition and biomass influenced the prey composition, selectivity and consumption of fishes in coastal rivers by providing a primary food source and refuge habitat for high nutritional quality prey populations, including freshwater small bodied fishes, Amphipoda, Cambaridae, and other invertebrate taxa. Specifically I documented that diet composition was significantly different between rivers with d ifferent vegetation communities and between seasons with different prey communities for all four predator species assessed. I also documented that predatory fish were selectively feeding on

PAGE 114

114 prey resources in both rivers despite differences in habitat and prey availability and that that predator foraging success in a river where macrophytes are largely absent (Homosassa River) is equal to or greater than foraging success in a river with high macrophyte biomass (Chassahowitzka River). Combined, these result s suggest that the loss of macrophytes and associated decline in associated prey populations likely results in prey switching by predators to taxa associated with alternative habitats such as Amphipoda (filamentous algae habitat), Grapsidae/Xanthidae (bent hic substrates and filamentous algae habitats), and saltwater small bodied fishes (pelagic and demersal habitats) depending on the seasonal production of filamentous algae and availability of saltwater prey species. This switching behavior may allow for t he persistence of prey populations in coastal rivers (Comins and Hassel 1975, Baalen et al. 2001), d espite low habitat availability (and likely increased predation risk) and decreased population densities. As native macrophytes have declined in the Homos assa River, the role of filamentous algae in providing predation refuge is likely key in str ucturing predator prey dynamics Production of filamentous algae creates temporary habitat patches that are rapidly colonized by invertebrates, including Amphipoda and vegetation associated invertebrates These dense vegetated habitats may inhibit predator foraging and provide abundant food resources (algae and associated periphyton, including diatoms) for grazing invertebrates allowing prey population densities to increase greatly in filamentous algae patches. This is evidenced by observed seasonal increases in density and biomass of these prey taxa during periods of high algae production, despite increased density of predators in the river, particularly L. griseu s (Chapter 3)

PAGE 115

115 Furthermore, predator foraging success was lower during winter in both rivers despite higher biomass of preferred prey groups, perhaps as a result of interference among predators at high densities, or decreased vulnerability of prey populat ions in a highly structured and productive habitat ( Camp 2010, Camp et al. 201 1 ). Alternatively, predation of small bodied fishes by migratory predators ( L. griseus in particular) and subsequent population declines in fall through winter each year may ha ve resulted in decreased predation pressure on invertebrates in vegetative habitats, since small bodied fish predation on invertebrates is less likely to be inhibited by aquatic vegetation compared to larg er bodied fishes. Fishes in the Chassahowitzka Rive r foraged on a significantly higher proportion of Amphipoda ( L. punctatus and L. griseus ) and Cambaridae ( M. salmoides and L. griseus ), especially during summer months, compared to the Homosassa River where fishes consumed a significantly higher proportion of Grapsidae/Xanthidae ( L. punctatus L. rhomboides and L. griseus ), Palaemonetes spp. ( M. salmoides ) and saltwater small bodied fishes ( M. salmoides ). Interestingly, I did not detect a difference in the proportion of freshwater small bodied fishes in di ets between rivers, despite large differences in the estimated biomass of this prey group in each river ( Chapter 3 ). Similarly, the proportion of vegetation associated invertebrates and Palaemonetes spp. in diets was not lower in fishes from the Homosassa River (proportion of Palaemonetes spp. was higher in M. salmoides diets ), despite the disparity in macrophyte habitat. Based on these results, I infer that macrophyte habitat loss affects the prey composition of freshwater and saltwater fishes in coastal rivers through decreased availability and consumption of select phytophilic crustaceans, primarily Amphipoda and Cambaridae.

PAGE 116

11 6 Frequency of occurrence indices confirmed seasonal differences in prey items for all predator species examined. Fishes in the Cha ssahowitzka River foraged on a significantly higher proportion of Cambaridae, Palaemonetes spp., and freshwater small bodied fishes during summer months. Fishes in the Homosassa River consumed a significantly higher proportion of Grapsidae/Xanthidae, Pala emonetes spp., freshwater small bodied fishes and terrestrial prey during summer months. A significantly higher proportion of Amphipoda and saltwater small bodied fishes was observed in diets from the Chassahowitzka and Homosassa rivers during winter mont hs coincident with high production of filamentous algae and increased density of saltwater fishes. These results suggest that seasonal production of filamentous algae indirectly affects the prey composition of fishes in coastal rivers by providing increas ed habitat and food availability for select prey groups, especially Amphipoda. In addition, larger predators preyed upon a greater proportion of freshwater small bodied fishes in the summer and saltwater small bodied fishes in the winter, providing furthe r evidence of prey switching to seasonally abundant prey resources. I found that predators were selectively foraging on prey groups in each river, but that this selection differed between rivers. L. punctatus selected for Amphipoda and vegetation associat ed invertebrates during both seasons in the Chassahowitzka River but only during winter in the Homosassa River. This is likely because this species selected for vegetative and benthic invertebrates in the Homosassa River during summer when filamentous al gae and Amphipoda biomass were relatively low. I found that M. salmoides selectively foraged on freshwater small bodied fishes and select crustaceans, with distinct differences in the type of crustaceans foraged on between

PAGE 117

117 rivers and seasons. Selection f or some crustacean prey could not be quantified in the Homosassa River because these taxa were not captured during sampling efforts to quantify availability (a gear selectivity issue for Cambaridae and Palaemonetes spp.), although these taxa were observed in the predator diets, indicating high selectivity (prey consumed in high proportion at low biomass). L agodon rhomboides selectivity indices indicated avoidance of filamentous algae and plants. This species selectively consumed Amphipoda, vegetation asso ciated invertebrates and benthic invertebrates in both rivers. Similar to M. salmoides L. griseus selectively foraged on freshwater small bodied fishes and select crustaceans, with differences in prey selection between seasons and rivers. These results clearly show that diet selection of fishes in the coastal rivers was for prey items with high caloric value, mostly small fishes and crustaceans. Tetzlaff (2008) documented larger home ranges and daily movement patterns of M. salmoides in the Homosassa co mpared to Chassahowitzka River likely due to increased searching time when foraging as a result of decreased prey availability Tetzlaff et al. (2009) found higher M. salmoides prey consumption rates in the Homosassa River compared to the Chassahowitzka likely driven by increase d foraging activity i n the Homosassa River, and increased prey vulnerability in an unstructured river I found that the proportion of empty predator stomach s was not lower in the Homosassa River for any of the predator species ev aluated, nor was the average prey dry weight per predator body weight lower. In fact, the proportion of empty stomach s was lower, and the average prey dry weight per predator body weight was generally higher for each species in the Homosassa River. This suggests that predator foraging

PAGE 118

118 success in the Homosassa River is equal to or greater than foraging success in the Chassahowitzka River. This result is somewhat surprising given the large differences in prey biomass between rivers (Chapter 3) and is likel y a result of the utilization of alternative prey groups, including saltwater invertebrates and fishes. However, there may be an increased energy cost associated with foraging in an open water environment compared to a vegetated one with higher prey densi ty when prey encounters are less frequent (Savino and Stein 1989, Tetzlaff 2008 Tetzlaff et al. 2010).

PAGE 119

119 Table 4 1 Mean proportion by dry weight of common prey taxa observed in stomachs of Lepomis punctatus from the Chassahowitzka and Homosassa r ivers, Fl orida.

PAGE 120

120 Table 4 2 Percent frequency of occurrence of common prey taxa observed in stomachs of Lepomis punctatus from the Chassahowitzka and Homosassa r ivers, Florida.

PAGE 121

121 Table 4 3. Manly Chesson prey selectivity indices for Lepomis punctatus from t he Chassahowitzka and Homosassa rivers, Florida.

PAGE 122

122 Table 4 4 Mean proportion by dry weight of common prey taxa observed in stomachs of Micropterus salmoides from the Chassahowitzka and Homosassa r ivers, Florida.

PAGE 123

123 Table 4 5 Percent frequency of occ urrence of common prey taxa observed in stomachs of Micropterus salmoides from the Chassahowitzka and Homosassa r ivers, Florida.

PAGE 124

124 Table 4 6. Manly Chesson prey selectivity indices for Micropterus salmoides from the Chassahowitzka and Homosassa rivers, F lorida.

PAGE 125

125 Table 4 7 Mean proportion by dry weight of common prey taxa observed in stomachs of Lagodon rhomboides from the Chassahowitzka and Homosassa r ivers, Florida.

PAGE 126

126 Table 4 8. Percent frequency of occurrence of common prey taxa observed in stomac hs of Lagodon rhomboides from the Chassahowitzka and Homosassa r ivers, Florida.

PAGE 127

127 Table 4 9. Manly Chesson prey selectivity indices for Lagodon rhomboides from the Chassahowitzka and Homosassa rivers, Florida.

PAGE 128

128 Table 4 10 Mean proportion by dry weigh t of common prey taxa observed in stomachs of Lutjanus griseus from the Chassahowitzka and Homosassa r ivers, Florida.

PAGE 129

129 Table 4 11 Percent frequency of occurrence of common prey taxa observed in stomachs of Lutjanus griseus from the Chassahowitzka and H omosassa r ivers, Florida.

PAGE 130

130 Table 4 12. Manly Chesson prey selectivity indices for Lutjanus griseus from the Chassahowitzka and Homosassa rivers, Florida. Space s in the table separate the different prey grouping methods, which include identifiable taxa, saltwater and freshwater invertebrates and fishes, and invertebrates and fishes.

PAGE 131

131 Figure 4 1. Mean estimated biomass of filamentous algae and macrophytes (measured as wet weight) within sampled reaches of the Chassahowitzka and Homosassa rivers d uring the period of study. Error bars represent one standard deviat ion of the mean total weight of aquatic vegetation.

PAGE 132

132 Figure 4 2. Seasonal and interannual pattern s in mean proportion of empty stomach s of Lepomis punctatus, Micropterus salmoides, Lago don rhomboides and Lutjanus griseus from the Chassahowitzka and Homosassa rivers.

PAGE 133

133 Figure 4 3. Intra annual pattern s in mean proportion of empty stomach s of Lepomis punctatus, Micropterus salmoides, Lagodon rhomboides and Lutjanus griseus within the C hassahowitzka and Homosassa rivers.

PAGE 134

134 Figure 4 4. Seasonal and interannual pattern s in mean total prey dry weight per predator body weight of Lepomis punctatus, Micropterus salmoides, Lagodon rhomboides and Lutjanus griseus within the Chassahowitzka and Homosassa rivers.

PAGE 135

135 Figure 4 5. Intra annual pattern s in mean total prey dry weight per predator body weight of Lepomis punctatus, Micropterus salmoides, Lagodon rhomboides and Lutjanus griseus within the Chassahowitzka and Homosassa rivers.

PAGE 136

136 CHAPTER 5 VEGETATIVE HABITAT LOSS EFFECTS ON FISH AND INVERTEBRATE COMMUNITY STRUCTURE IN SPRING FED, COASTAL RIVERS Introduction Autotroph s play a central role in the ecology of aquatic ecosystems by contributi ng to ecosystem production, modifying biogeochemical processes, and mediati ng biotic interactions (Carpenter and Lodge 1986, Jeppesen et al. 1998, Duarte 2002). A utotrophs including rooted macrophytes, directly support the production of higher trophic levels and provide a fundamental control on the abunda nce and diversity of faunal organisms ( Power 1995). Macrophytes provide a substrate for periphyton, which serve as a primary food base in many aquatic ecosystem s (Jones et al. 1998) The production of p eriphyton is especially important in stream ecosyste ms (Minshall 1978), where downstream current s and low water residence times inhibit the production of phytoplankton and other suspended algae (Wetzel 2001). Furthermore, macrophytes link benthic substrates to the overlying water through the uptake of sedi ment bound nutrients and transport of organic matter, minerals and gases to both the water and benthic environments (Barko and James 1998, Caraco et al. 2006). Vegetative cover can decrease sediment erosion and resuspension by reducing water velocity and turbulence at the water sediment interface (Gregg and Rose 1982, Barko and James 1998, Dodds and Biggs 2002), resulting in increased water clarity and light availability and providing a positive feedback loop for p rimary production Macrophytes mediate predator prey interactions between fishes and invertebrates by providing refuge habitat for prey populations and decreasing prey encounter rates of predators, allowing predator and prey populations to coexist at relatively high densities (Crowder and Coope r 1982). Through a combination of biological and physical controls on faunal

PAGE 137

137 organisms and the aquatic environment, primary producer s have a strong influence on the structure of aquatic communities. Human alterations of the landscape and associated change s in the physical and chemical properties of aquatic environments have resulted in a loss of macrophytes from many shallow aquatic ecosystems around the world (Duarte 2002). For example in eutrophic systems, a utotrophs capable of rapid nutrient uptake an d growth can dominate and potentially ex clude slow growing macrophyte s with low nutrient uptake and assimilation rates (Duarte 1995). This phenomenon is exemplified by fast growing cyanobacteria and phytoplankton in nutrient enriched lakes (Smith et al. 1 999), macroalgae in nutrient enriched estuaries (Valiela et al. 1997), and filamentous algae in nutrient enriched streams (Huntsman 1948, Elwood et al. 1981). The effects of macrophyte loss and replacement by algal speci es on the faunal communities that t hey support are not currently well understood. Population responses to vegetative habitat loss are likely to result from multiple coupled factors, including changes in food base (Chapter 4) loss of fish and invertebrate reproductive and juvenile rearing habitat, and altered trophic interactions. D istinct shifts in the composition and biomass of primary producers have been documented within spring fed, coastal rivers in Florida as a result of changes in watershed land use, including increased agricultural and streamside development, and associated changes in streamflow and water quality (Frazer et al. 2006). Of particular concern is the rapid decline and extirpation of macrophytes, including Vallisneria americana Potamogeton spp., and Sagittaria kurziana from several systems and the widespread proliferation of filamentous algae, including Chaetomorpha sp., Gracilaria

PAGE 138

138 sp., and Lyngbya sp. (Frazer et al. 2006, Stevenson et al. 2004). The loss of macrophytes which provide forage and refuge habitat may alte r invertebrate grazer communities, predator/prey dynamic s of fishes, and other important population level interactions. Such alterations may lead to undesirable shifts in fish and in vertebrate communities and possibly the loss of key species. Spring fed rivers serve as model ecosystems to study the effects of vegetative habitat loss on fish and invertebrate populations due to their relatively stea dy streamflow, stable water temperatures, high rates of primary production (Odum 195 3 ), and diverse communitie s of oligohaline and marine plants, algae, invertebrates and fishes (Herald and Strickland 1949, Odum 1957). The purpose of this study was to develop an ecosystem model of a spring fed, coastal river based on empirical data from the Chassahowitzka River, F lorida for the purpose of predict ing the response s of fish and invertebrate populations to changes in submersed aquatic vegetation (SAV) and resulting loss of habitat ( extirpation of macrophytes and replacement with filamentous macroalgae) The model pred ictions were compared with the observed differences between the community structures of a highly vegetated river (the Chassahowitzka River) and one where rooted macrophytes have declined rapidly over the last decade and have been largely absent since 2006 (the Homosassa River). In addition, the predicted responses of fishes and invertebrate populations to an alternative policy option, macrophy te restoration were compared with the predictions under a macrophyte extirpation scenario. These simulations will prove useful for understanding the ecological changes associated with vegetative habitat loss

PAGE 139

139 and the benefits of ecosystem restoration in terms of fish and invertebrate communities and the goods and services they provide. Methods To assess fish and inver tebrate population responses to vegetative habitat availability in coastal rivers, I utilized a three step assessment approach that included, ( 1) time dynamic simulation of an ecosystem model based on empirical observations of a coastal river food web, (2) model validation by spatial comparative analysis of fish and invertebrate community structure in two coastal rivers with contrasting vegetative habitats, and (3) comparison of alternative policy options for coastal rivers including no action resulting in macrophyte extirpation versus ecosystem restoration resulting in increased macrophyte biomass in the Chassahowitzka River. Trophic M ass balance M odel of a Coastal River Food W eb A trophic mass balance model of the aquatic food web within the Chassahowitzka River was develop ed using the Ecopath with Ecosim software ( Walters et al. 1997 Pauly et al. 2000 Christensen and Walters 2004). The software is available for free download at www.ecopath.org Th e Ecopath modelin g framework balances the annual production in biomass of individual trophic groups with losses to predation, harvest and migration and net change s in biomass (Walters and Martell 2004) Model inputs for each trophic group included the proportion of the s tudy area occupied estimated biomass, production to biomass ratio, consumption to biomass ratio, prey composition harvest information, and proportion of biomass contributed to the detrital pool versus biomass exported from the system Table 5 1 lists th e trophic groups of producers and consumers included in the model and the scientific names of the taxa comprising each trophic group. Empirical biomass estimates of trophic groups were

PAGE 140

140 acquired from vegetation quadrat sampling (macrophytes and filamentous algae), invertebrate sampling (benthic invertebrates, vegetation associated invertebrates and amphipods), throw trap sampling (crayfish, blue crabs, mud crabs, and shrimp; Camp et al. 2011 ), block netted seine sampling (freshwater and saltwater small bod ied fishes), boat electrofishing (lake chubsucker, Lepomis spp., largemouth bass, American eel, gar, striped mullet, pinfish, catfish, sheepshead, gray snapper, red drum and common snook), or data from long term vege tation monitoring (periphyton; Frazer et al. 2006) and other samp ling efforts (sediment diatoms; Frazer unpublished data). Methods for vegetation biomass estimation are described in Frazer et al. (2006), and methods for fish and invertebrate biomass estimation are described in the methods secti on of Chapter 3. All estimates were scaled to biomass in g per 100 m 2 (Table 3 1). The proportion of study area occupied was set equal to one for all trophic groups. Biomass estimates of each trophic group were based on seasonal estimates average d acros s three years of sampling in the Chassahowitzka River, depending on whether the group was more abundant in winter (filamentous algae, select invertebrates and saltwater trophic groups) or summer (freshwater fishes a nd select invertebrate trophic groups ) in the study reaches. Estimates of p roduction to biomass ratio s were determined from published literature (Table 5 2), or estimated from growth at a ge and mortality data (select freshwater fishes). Estimates of c onsumption to biomass ratio s were determined from published literature (Table 5 2) or inferred from estimates of trophic groups within similar trophic guilds. A summary of the basic input parameters for the Ecopath model is provided in T able 5 3.

PAGE 141

141 Empirical diet data for individual fish trophic gr oups w ere pooled across rivers and sampling events, and summarized by percent composition of total dry mass. Diet sampling methods are described in the methods section of Chapter 4. Diet information for invertebrates was synthesized from published litera ture; a list of references is included in Table 5 2 A predator prey matrix was constructed to summarize the proportion of dietary items by prey group for each consumer trophic group (Table 5 4) To account for seasonality of migratory saltwater trophic groups foraging within the rivers, a gulf food base prey group was included in the model and diet composition of saltwater trophic groups was assumed to consist of 50% gulf food base. The contribution of saltwater fishes to the detritus in the rivers was a ssumed to be zero since these fishes utilize the rivers seasonally and all freshwater trophic groups were assumed to contribute fully to the detrital pool (Table 5 5) Since the objective was to assess the effects of vegetative habitat on faunal populati ons, I did not include harvest in the model. The ecotrophic efficiency (proportion of production accounted for by predation, harvest, and net change in biomass within the model ) of each trophic group was solved for using the Ecopath mass balance parameter ization H igh ecotrophic efficiency values may imply competition among predators/fisheries for particular trophic groups, while low values imply low predation/fishing mortality on that trophic group, or inadequate accounting of sources of mortality in the model (Waters and Martell 2004). T ime d ynamic S imulation of A lternative M anagement S cenarios The Ecosim module in the Ecopath with Ecosim program was used to simulate a long term time series of trophic group biomass pattern s under alternative policy optio ns of no management action resulting in the continued decline and eventual extirpation of macrophytes from the Chassahowitzka River and replacement by seasonal production

PAGE 142

142 of filamentous algae versus ecosystem restoration resulting in increased biomass of macrophytes and a reduction in filamentous algae biomass The Ecosim framework simulate s trophic group biomass rates of change over time based on gains from prey consumption times food conversion efficiency (proportion of prey consumed converted to biomas s), and losses to mortality, including predation, fishing (assumed zero for the coastal river model), and unexplained natural mortality ( Waters et al. 1997 ). A 60 year time series was simulated for the Chassahowitzka River to simulate the continued decline and eventual extirpation of macrophytes and replacement by seasonally abundant filamentous algae A forcing function was applied to macrophytes and associated periphyton trophic groups that simulated (1) an initial 20 year period of constant macrophyte b iomass equal to the summer average over the study period in the Chassahowitzka River, (2) a 20 year period of steady linear decline in biomass from the initial biomass to complete extirpation, and (3) a 20 year period with macrophytes and associated periph yton extirpated from the system (Figure 5 1 ) A second forcing function was applied to filamentous algae that simulated (1) a 6 0 year period of cyclical filamentous algae blooms occurring seasonally based on observed monthly biomass pattern s during year t hree of monitoring in the Chassahowitzka and Homosassa rivers (peak biomass was set equal to the mean observed winter biomass in the Chassahowitzka River) (Figure 5 1 ) The relative change in biomass of each trophic group was estimated as the difference b etween the average annual biomass of the initial 10 year period of the simulation and the average annual biomass of the terminal 10 year period of the simulation. The relative biomass change of each trophic group

PAGE 143

143 from the time dynamic simulation was compa red with the observed spatial differences in biomass between the Chassahowitzka and Homosassa rivers. A long term restoration scenario was simulated to examine the community level effects of restoring macrophytes and reducing filamentous algae to observed pre disturbance levels. Two forcing functions were used in this simulation. The first forcing function simulated (1) an initial 20 year period of constant macrophyte biomass equal to the summer average over the study period in the Chassahowitzka River, ( 2) a 20 year period of steady linear increase to twice the initial biomass, and (3) a 20 year terminal period with macrophyte biomass equal to twice the initial biomass (Figure 5 1 ) The first forcing function was also applied to periphyton associated wit h macrophytes. A second forcing function was applied to filamentous algae that simulated (1) an initial 20 year time series of filamentous algae equal to the observed seasonal pattern of filamentous algae biomass in the Chassahowitzka River with the peak production each year equal to the mean biomass observed during winter sampling and (2) a 4 0 year period with constant filamentous algae biomass equal to the observed mean during summer in the Chassahowitzka River (approximately one fifth the initial perio d peak winter biomass) (Figure 5 1 ). The mean annual biomass from the initial 10 year period of the simulation was compared with the terminal 10 year period mean annual biomass for each trophic group. Results The Ecopath trophic mass balance model illustr ated the complexity of trophic interactions within the Chassahowitzka River (Figure 5 2 ). To balance the ecosystem model, several production to biomass estimates of invertebrates and small bodied fishes were adjusted to higher values than the initial va lues from published literature

PAGE 144

144 (Table 5 2). These results are not surprising due to the relatively warm water temperatures year round high nutrient loading, and the high primary production rates documented for spring fed system s in Florida compared to ot her ecosystems (Odum 1957). The balanced trophic model predicted high transfer of invertebrate and small bodied fish production to freshwater and marine fishe s (Figure 5 3 ) Time dynamic simulation of macrophyte extirpation and increased filamentous alg ae production predicted a strong negative response by many trophic groups of fishe s and invertebrates, including gray snapper saltwater catfishes, striped mullet, American eel largemouth bass, Lepomis spp., lake chubsucker, freshwater small bodied fishes blue crabs, crayfish, mud crabs, grass shrimp amphipods, and vegetation associated invertebrates (Figure 5 4 ) Positive responses to changes in submersed aquatic vegetation were predicted for several saltwater fishes ( common snook, red drum, sheepshead pinfish and small bodied species), select freshwater fishes (Florida gar), and select invertebrates ( sediment invertebrates). Comparisons of observed spatial differences in mean trophic group biomass between the Chassahowitzka and Homosassa rivers (Figur e 5 4 ) corroborated model predicted responses for many trophic groups, including common snook, red drum sheepshead, saltwater small bodied fishes, Florida gar, largemouth bass Lepomis spp., lake chubsucker, blue crabs, crayfish, grass shrimp, amphipods, and benthic invertebrates. The predicted and observed changes in biomass w ere similar in direction and magnitude for a few trophic groups, including lake chubsucker, crayfish, and grass shrimp The predicted magnitude of change was greater than the obser ved difference between rivers for largemouth bass, Lepomis spp., freshwater small bodied fishes, and

PAGE 145

145 amphipods. The model predicted change s in biomass that were considerably less than observed spatial differences between rivers for multiple trophic groups including common snook, red drum, sheepshead, saltwater small bodied fishes, gar, and sediment associated invertebrates (Figure 5 4 ) I n fact, the observed increase in biomass was often orders of magnitude greater than the predicted change For several taxa/ trophic groups ( gray snapper, saltwater catfish pinfish striped mullet, American eel mud crab s and vegetation associated invertebrates ), the predicted response was opposite the observed differences in biomass between rivers (Figure 5 4) Time dyn amic simulation of macrophyte restoration indicated a strong positive response by the majority of trophic groups to increased macrophyte and periphyton production (Figure 5 5 ), with the strongest responses predicted for catfish, pinfish, all freshwater fis hes crayfish, grass shrimp, and vegetation associated invertebrates Restoration of aquatic vegetation was predicted to result in a decrease in biomass of mud crabs and amphipods as algae production dec rease d Surprisingly, a couple of trophic groups we re predicted to respond positively under either scenario of macrophyte extirpation or restoration, including common snook, red drum, sheepshead, pinfish, saltwater small bodied fishes, Florida gar, and benthic invertebrates. Discussion T he Ecopath trophic mass balance model of the Chassahowitzka River indicated that invertebrates particularly crustaceans, and small bodied fishes are c entr al to coastal river food web s providing direct energy transfer from primary producers (periphyton, filamentous algae, a nd detrit us from macrophytes, in particular) to large bodied predator s The model estima t ed that the majority of invertebrate and small bodied fish production wa s accounted for by fish predation within the river These

PAGE 146

146 results are consistent wit h empiric al diet patterns of fishes which demonstrat ed that freshwater and marine fishes are select ively foraging on crustaceans vegetation associated invertebrates and small bodied fishes ( Chapter 4 ). P redation by fishes on these trophic groups is also validated by the observed declin es in prey group biomass during winter sampling periods when marine predator density and biomass increased greatly (Chapter 3 ) Similar to these results, o ther researchers have reported high ecotrophic efficiencies of small bodied f ishes, crustaceans, and other invertebrate taxa in coastal aquatic food webs (Christ e nsen 1995 Freire et al. 2008, Coll et al. 2009 ) K ey findings from time dynamic model simulation of macrophyte loss and extirpation included a predicted decline or extir pation of select freshwater and marine trophic group s ( largemouth bass Lepomis spp., lake chubsucker, freshwater small bodied fishes, crayfish, and grass shrimp ) increase d biomass of select fish and invertebrate trophic groups ( common snook red drum, sh eepshead, saltwater small bodied fishes, and sediment invertebrates ) and a resultant shift in faunal community composition. Overall, total invertebrate biomass was predicted to de crease by approximately 60 % if macrophytes we re extirpated from the Chassah owitzka River primarily as a result of de creased biomass of crayfish and grass shrimp T otal fish biomass was predicted to decline by 11% as a n indirect result of macrophyte extirpation These results demonstrate that the loss of macrophytes from the co astal river ecosystem affects the composition of the aquatic community and food web structure, and results in a net decline in the biomass of fishes and invertebrates Similar effects of macrophyte loss have been demonstrated for other freshwater (Bettoli et al. 1993) and marine communitie s (Deegan et al. 2002 Coll et al. 2011 )

PAGE 147

147 M acrophyte extirpation was predicted to result in a shift in the seasonality of primary production from a relatively steady perennial primary producer community with associated in vertebrate s and fishe s to cyclical population dynamics relat ed to boom and bust algae production and corresponding bottom up responses of select invertebrate and fish populations. Empirical observations over the period of study (Chapter 3) indicated that vegetation biomass, freshwater invertebrate and fish trophic group biomass and relative foraging success of fishes (Chapter 4) was less variable seasonally and between years in the Chassahowitzka River, whil e invertebrate and freshwater fish biomass in th e Homosassa River demonstrated seasonal patterns related to the availability of filament ous algae habitat Monthly monitoring of vegetation and fishes in the Homosassa River during year three of the study show ed that large scale algal production during Ma rch and April resulted in a sharp increase in young of the year freshwater and saltwater fishes (Chapter 3); however, the longer term response of invertebrate and fish communities was not measured since the widespread alga e blooms occurred during the end o f the period of study. Based on ecosystem model predictions and empirical observations faunal population dynamics are expected to be more stable in a river dominated by aquatic macrophytes versus one dependent on the production of filamentous algae exc lusively O bserved difference s in the estimated biomass of select species between the Chassahowitzka and Homosassa rivers validated several model predictions under the macrophyte extirpation scenario including losses o f lake chubsucker, crayfish and grass shrimp from the Homosassa River; decreased biomass of largemouth bass Lepomis spp., freshwater small bodied fishes, blue crabs, and amphipods in the

PAGE 148

148 Homosassa River compared to the Chassahowitzka River ; increased biomass of common snook, red drum, sheeps head, saltwater small bodied fishes Florida gar, and benthic invertebrates in the Homosassa River ; and observed boom and bust dynamics of invertebrates associated with filamentous algae and fishes that forage on these invertebrate groups Overall, the E cosim model accurately predicted the direction of faunal group responses for most freshwater and marine taxa; however, the magnitude of change was inaccurate for nearly all taxa measured with the exception of those taxa that were found to be nearly extirpa ted from the Homosassa River (lake chubsucker, crayfish, and grass shrimp) The anomalous predictions in species biomass pattern s provide interesting cases of counterintuitive population responses. The increase in biomass of select saltwater fishes is n ot surp rising since their recruitment may be in dependent of the habitat and trophic dynamics within the rivers and the Homosassa River is deeper, has greater discharge, and provides a larger volume of freshwater that serves as winter habitat for thermally sensitive species (Odum 1953) The contradictory responses of freshwater trophic groups, on the other hand, present areas of potential future ecological research. For example, the observed biomass of Florida gar in the Homosassa River was 72 times greate r than in the Chassahowitzka River; however, the ecosystem model predicted a change in biomass that was considerably less than the observed difference. One possible explanation is increased spawning success of Florida gar I observed gar successfully sp awning on filamentous algae patches during the spring of year three and captured a higher abundance of young of the year in the following sampling events.

PAGE 149

149 effect (Walters and Martell 2004), where the suppressed in one ecosystem state (in this case a highly structured macrophyte dominated system) and whose population bottleneck is released in an altered state (an unstructured a lgae dominated system). Alternative hypotheses that may account for the observed difference in biomass include improved foraging success in an unstructured river or high tolerance to decreased water quality (Kushlan 1974) The ecosystem model incorrectl y predicted the local extinction of several freshwater fish trophic groups in response to the loss of macrophytes Largemouth bass, Lepomis spp. and freshwater small bodied fishes demonstrated a lowered biomass in the Homosassa R iver compared to the Cha ssahowitzka River ; however, the trophic group declin es were less than predicted by the model Since macrophytes have been largely absent from the Homosassa River for approximately the last five years, it is possible that the full extent of population resp onses to vegetative habitat loss were not detected over the last three years of study The response may be much greater following several generations of decreased recruitment and other changes in population dynamics (Lauretta, unpublished data) While I observed relatively large numbers of young of the year largemouth bass and Lepomis spp. during each year in both rivers, there was a clear difference in the number of juveniles surviving to larger age classes. In the Chassahowitzka River, a greater propor tion of juveniles survived compared to the Homosassa River, where older individuals were rare. Alternatively, prey switching by largemouth bass and Lepomis spp. could explain the sustained populations in the Homosassa River as evidenced by spatial differ ence s in diet composition between populations in the river s and the seasonal difference s in prey composition in the

PAGE 150

150 Homosassa River related to availability of filamentous algae and associated invertebrate prey groups (Chapter 4). The loss of grazers and d etritivores from coastal rivers, including lake chubsucker and crayfish, could result in a negative feedback on macrophyte production. For example, lake chubsucker require vegetative habitat to successfully reproduce and forage, and were historically comm on in the Homosassa River when macrophytes communities were prevalent (Herald and Strickland 1949). The decline of this habitat likely had negative effects on chubsucker reproduction foraging success, and survival, potentially leading to the extirpation of this species from portions of the river. The population decline of this key fish herbivore could have decreased grazing of periphyton on plant stems, resulting in increased shading of plants, which may have further accelerated plant loss in the ecosyst em (Roberts et al. 2003) resulting in a vegetative community comprised exclusively of seasonally abundant filamentous algae. Alternative hypotheses for ecosystem change that were not captured in the trophic dynamic model include shifts in system energy dy namics, such as increased allocht h onous input and terrestria l l y based prey items Diet information for several species of freshwater fishes (e.g., largemouth bass and Lepomis spp.) indicated that individuals within the Homosassa River consumed terrestrial organisms, including lizards, waterfowl and terrestrial insects, during summer months when producer and invertebrate biomass was relatively low (Chapter 4) In addition, observed differences in mud crabs and vegetation associated invertebrates particula rly gastropods, between systems and the opposite population trends from the predicted responses demonstrate d that the population ecology (feeding and production) of these trophic groups was not

PAGE 151

151 likely accurately represented in the ecosystem model. Additio n al studies that would increase the understanding of trophic dynamics in the coastal rivers and likely improve the predictability of the ecosystem model include diet composition of small bodied fishes and invertebrates and annual production estimates of pr oducer and invertebrate trophic groups. Macrophyte restoration was predicted to have beneficial effect s on the majority of fish and invertebrate populations especially freshwater trophic groups w ith negative effects predicted for relatively few groups, i ncluding mud crabs and amphipods in response to decreased algae production. These results demonstrated the importance of macrophytes in sustaining the freshwater aquatic food web, as well as the benefit of macrophyte habitat to saltwater fish taxa and th e aquatic communit ies as a whole Since several of these trophic groups support recreational fisheries within the rivers and the Gulf of Mexico, the restoration of macrophytes may have beneficial economic effects as well as ecological benefits These riv ers represent important juvenile rearing and overwintering habitats for economically valuable marine s tock s, including common snook, red drum, and gray snapper. Habitat improvement in the river may benefit these stocks through increased juvenile recruitme nt resulting from greater food availability and lowered predation risk in macrophyte dominated habitats, as well as increased adult survival during winter periods when water temperatures in the Gulf of Mexico can drop below species tolerance thresholds. O verall, macrophyte restoration was predicted to result in an increase in total invertebrate biomass of approximately 152% and an increase in total fish biomass of approximately 73%. In addition to benefits to fish and invertebrate populations the restor ation of macrophytes may help improve and maintain

PAGE 152

152 water quality within the rivers providing a positive feedback for ecosystem primary production, and increasing the water quality and aesthetic value of the rivers to the streamside communit ies and recreat ional boaters

PAGE 153

153 Table 5 1. Trophic groups and taxa composition included in the Ecopath trophic mass balance model of the Chassahowitzka River food web.

PAGE 154

154 Table 5 2. Data sources for the Ecopath trophic mass balance of the Chassahowitzka River food web.

PAGE 155

155 Table 5 3. Basic inputs for the Ecopath trophic mass balance model of the Chassahowitzka River food web

PAGE 156

156 Table 5 4. Diet composition of consumers within the Chassahowitzka and Homosassa r iver s Florida. Prey groups are listed in rows, and predat or groups are listed in columns with reference numbers corresponding to the group name listed in column 1

PAGE 157

157 Table 5 5 Detrital fate matrix for the Ecopath trophic mass balance model of the Chassahowitzka River food web

PAGE 158

158 Figure 5 1. Ecosim forcing functions used to simulate changes in primary production within the Chassahowitzka River, Florida under alternative management scenarios of macrophyte extirpation versus restoration.

PAGE 159

159 Figure 5 2 Ecopath trophic flow diagram of the Chassahowitzka Rive r. The size of the circle is relative to the biomass of the trophic group. Trophic level

PAGE 160

160 Figure 5 3 Predicted ecotrophic efficiency (proportion of production consumed by predators) of trophic groups within the Chassahowitzka River food web model.

PAGE 161

161 Figure 5 4 Comparison of time dynamic ecosystem model predicted changes in mean annual biomass of trophic groups versus observed spatial differences between the Chassahowitzka and Homosassa rivers. 4 48 29 11 72 13 5 Model Predicted Observed

PAGE 162

162 Figure 5 5 Comparison of time dynamic ecosys te m model predicted community responses to the extirpation and restoration of macrophytes in the Chassahowitzka River.

PAGE 163

163 CHAPTER 6 SYNTHESIS AND FUTURE RESEARCH Spring fed systems in Florida have been historically described as homeostatic in their chemical, physical and biol ogical characteristics (Odum 1957). The data collected during this study demonstrate d that spring fed, coastal rivers are spatially and temporally dynamic in vegetative, invertebrate, and fish community composition an d biomass. Based on river wide compar isons of faunal community composition, biomass, diet of fishes, and ecosystem time dynamic simulation, I infer that vege tative habitat loss negatively affec ts species that rely on this habitat type for foraging, refuge or reproduction, including crayfish, grass shrimp, small bodied freshwater fishes lake chubsu cker, pinfish, spotted sunfish and largemouth bass. Species that do not have a strong affinity for structural habitat (SAV in particular) will be less affected by large scale changes in vegetation, such as mud crab s select saltwater small bodied fishes gray snapper Florida gar and longnose gar ( Lepisosteus osseus ) The observed differences and model predicted responses in population biomass and diet of fishes are evidence that changes in vegetati ve habitat affect indivi dual species disproportionately, and continued changes are likely to alter the fish and invertebrate communities in these coastal eco systems Other researchers have documented community level shifts associated with the removal of ke y habitat components in aquatic ecosystems Sass et al. (2006 ) demonstrated that the removal of a structural ly complex habitat (course woody debris) in a Wisconsin lake significantly impacted species interactions and decreased fish abundance, survivorship recruitment and growth resulting in a shift in the community composition of fishes The results of this research demonstrated similar effects of

PAGE 164

164 structural habitat loss on fish populations, although the production loss from the removal of plants may ha ve disproportionately affected grazer communities in coastal rivers, compared to the removal of woody debris observed by Sass et al. (2006). Be t toli et al. (199 3 ) showed that removal of the macrophyte community in a Te xas reservoir by introduced grass car p led to shift in community structure from a predominantly benthic to a pelagic based food web and reported an associated decline in phytophilic fish abu ndance In coastal rivers, the downstream transport of material and decreased water residence time co mpared to lentic ecosystems likely inhibited the production and biomass of plankton in the river s, resulting in a decreased primary food base driven by filamentous algae production Furthermore, the observ ed communit ie s in coastal rivers are influenced by the colonization of marine fishes and invertebrates from the Gulf of Mexico and are not limited to within system recruitment compared to closed ecosystems. Therefore, community changes in the open coastal ecosystem cannot be predicted by in stream proces ses alone, as demonstrated by the considerable differences between predicted and observed biomasses of marine fishes. The experimental remov al of macroalgae in a coastal estuary resulted in a positive response in macrophyte, decapod and fish biomass (Deeg an et al. 2002). Via a nitrogen tracer experiment, Deegan et al. (2002) demonstrated that macroalgae contributed little to secondary consumer production in the estuary. The ir results support my conclusions that vegetative habitat loss results in decrease d invertebrate and fish biomass and indicate that production by filamentous algae may not compensate for the loss of macrophytes and associated periphyton in coastal aquatic ecosystems

PAGE 165

165 Coastal seagrass meadows are some of the most productive ( Odum 195 7 ) and ecologically valuable ecosystems in the world (Duarte 2002). E utrophication and other human related disturbances to aquatic ecosystems have resulted in the global decline of seagrasses and other aquatic macrophytes (Short 1996, McGlathery 2001). The consequences of seagrass loss to the coastal faunal communities are not fully understood; although there is a general agreement that this habitat loss decreases biodiversity and biomass of fishes and invertebrates and alters food web structure ( Nakamura 20 10, Pillay et al. 2010, Coll et al. 2011) The results of this study corroborated those conclusions In general, t he loss of seagrass communities from coastal ecosystems is predicted to result in altered fish and invertebrate community structure and a de cline in consumer biomass Using the coastal river ecosystem as a model of community structure effects, the loss of macrophytes and shift to seasonal filamentous algae production wa s predicted to result in a n overall 6 0% decrease in invertebrate biomass a nd 11 % decrease in fish biomass, while restoration to historical biomass wa s predicted to result in an increase of 152% in invertebrate biomass and 75% increase in fish biomass. Since m any developed and undeveloped nations depend on marine fisheries for f ood and revenue, the restoration of macrophyte communities is expected to have significant socioeconomic benefits through increased biomass of fishes and invertebrates that support coastal fisheries. Future research that could increase the overall understa nding of trophic dynamics in coastal ecosystem s and likely improve the a c c urac y of ecosystem model predictions include diet composition information for small bodied fishes and invertebrates, annual production estimates of producers and invertebrates and

PAGE 166

166 information on fish recruitment under variable vegetative habitat composition and biomass. Diet information of small bodied fishes and invertebrates may provide greater insight into the transfer of energy from producers to predators. Since the scope of t his study was limited to diet sampling of large bodied fishes, there is considerable uncertainty in the contribution of each producer to the aquatic food web. Production estimates of producers and invertebrates that identify the proportion of primary prod uction contributed by macrophytes periphyton and filamentous algae could provide more accurate estimates of the e ffects of vegetation composition on secondary production. Finally, defining the relationship between vegetative habitat and fish recruitment would increase the ability to predict community responses to shifting habitat composition and biomass in coastal aquatic ecosystems.

PAGE 167

167 LIST OF REFERENCES Akaike, H. 1974. A new look at the statistical model identification. IEEE Transactions on Automatic C ontrol 19 :716 723. Anderson, D. R. 2001. The need to get the basics right in wildlife field studies. Wildlife Society Bulletin 29 :1294 1297. Anderson, D. R. 2008. Model Based Inference in the Life Sciences. Springer, Fort Collins Colorado, USA Ange lsen, K. and S. Olsen. 1987. Impact of fish density and effort level on catching efficiency of fishing gear. Fisheries Research 5 :271 278. Barko, J.W., and W. F. James. 1998. Effects of submerged aquatic macrophytes on nutrient dynamics, sedimentatio n, and resuspension. In The Structuring Role of Submerged Macrophytes in Lakes. Ecological Studies 131 :197 216. Bayley, P. B. and D. J. Austen. 2002. Capture efficiency of a boat electrofisher. Transactions of the American Fisheries Society 131 :435 4 51. Becker, C. G., C. F. Ponseca, and C. F. B. Haddad. 2009. Habitat split as a cause of local population declines of amphibians with aquatic larvae. Conservation Biology 24 :287 294. Benke, A. C., A. D. Huryn, L. A. Smock, and J. B. Wallace. 1999. Len gth mass relationships for freshwater macroinvertebrates in North America with particular reference to the Southeastern United States. Journal of the North American Benthological Society 18 :308 343. Berryman, A. A. 1992. The origins and evolution of pre dator prey theory. Ecology 73 :1530 1535. Bettoli, P. W., M. J. Maceina, R. L. Noble, and R. K. Betsill. 1992. Piscivory in largemouth bass as a function of aquatic vegetation abundance. North American Journal of Fisheries Management 12 :509 516. Bettoli P. W., M. J. Maceina, R. L. Noble, and R. K. Betsill. 1993. Response of a reservoir fish community to aquatic vegetation removal. North American Journal of Fisheries Management 13 :110 124. Bowlby, J. N ., and J. C. Roff. 1986. Trophic structure in so uthern Ontario streams. Ecology 67 :1670 1679. Camp, E. V. 2010. Relationships between small bodied fishes and crustaceans and submersed aquatic vegetation: implications of habitat change. Retrieved August 17, 2011, from http://purl.fcla.edu/fcla/etd/UFE0041392

PAGE 168

168 Camp, E. V. D. C. Gwinn, M. V. Lauretta, W. E. Pine, and T. K. Frazer. 201 1 Use of recovery probabilities can improve sampling efficiency for throw traps in vegetated habitats. Transac tions of the American Fisheries Society 140 :164 169. Caraco, N., J. Cole, S. Findlay, and C. Wigand. 2006. Vascular plants as engineers of oxygen in aquatic systems. Bioscience 56 :219 225. Carpenter, S. R., and J. F. Kitchell. 1996. The Trophic Cascad e in Lakes. Cambridge University Press, New York, USA. Carpenter, S. R., J. F. Kitchell and J. R. Hodgson. 1985. Cascading trophic interactions and lake productivity. Bioscience 35 :634 639. Carpenter, S. R. and D. M. Lodge. 1986. Effects of submers ed macrophytes on ecosystem processes. Aquatic Botany 26 :341 370. Chesson, J. 1983. The estimation and analysis of preference and its relationship to foraging models. Ecology 64 :1297 1304. Chipps, S. R. and J. E. Garvey. 2007. Assessment of Diet and Feeding Patterns. Pages 473 514 in C. S. Guy and M. L. Brown, editors. Analysis and Interpretation of Freshwater Fisheries Data. American Fisheries Society, Bethesda, Maryland USA Christensen, V. 1995. A model of trophic interactions in the North S ea in 1981, the year of the stomach. Dana 11 :1 28. Christensen, V. and D. Pauly. 1992. ECOPATH II a software for balancing steady state ecosystem models and calculating network characteristics. Ecological Modelling 61 :169 185. Christensen, V. and D. Pauly (Editors). 1993. Trophic models of aquatic ecosystems. ICLARM Conference Proceedings 26 :1 390. Christensen, V., and C. J. Walters. 2004. Ecopath with Ecosim: methods, capabilities and limitations. Ecological Modelling 172 :109 139. Coll, M., I. Palomera, and S. Tudela. 2009. Decadal changes in a NW Mediterranean Sea food web in relation to fishing exploitation. Ecological Modelling 220 :2088 2102. Coll, M., A. Schmidt, T. Romanuk, and H. K. Lotze, 2011. Food web structure of seagrass communit ies across different spatial scales and human impacts. PLoS ONE 6 :e22591. doi:10.1371/journal.pone.0022591 Comins, H. N. and M. P. Hassell. 1976. Predation in multi prey communities. Journal of Theoretical Biology 62 :93 114.

PAGE 169

169 Connell, J. H. 1961. Th e influence of interspecific competition and other factors on the distribution of the barnacle Chthamalus stellatus Ecology 42 :710 723. Crowder, L. B. and W. E. Cooper. 1982. Habitat structural complexity and the interaction between bluegills and thei r prey. Ecology 63 : 1802 1813. Dauwalter, D. C., D. K. Splinter, W. L. Fisher, and R. A. Marston. 2008. Biogeography, ecoregions, and geomorphology affect fish species composition in streams of eastern Oklahoma, USA. Environmental Biology of Fishes 82 :2 37 249. Deegan, L. A., A. Wright, S. G. Ayvazian, J. T. Finn, H. Golden, R. R. Merson, and J. Harrison. 2002. Nitrogen loading alters seagrass ecosystem structure and support of higher trophic levels. Aquatic Conserv ation of Mar ine and Freshw ater Ecosy st ems 12 :193 212. Dobson, A., H. M. Mooney, J. A. Rusak, O. Sala, V. Wolters, D. Wall, R. Winfree, and M. A. Xenopoulos. 2006. Habitat loss, trophic collapse, and the decline of ecosystem services. Ecology 87 :1915 1924. Dodds, W. K., and J. F. Biggs. 2 002. Water velocity attenuation by stream periphyton and macrophytes in relation to growth form and architecture. Journal of the North American Benthological Society 21 :2 15. Duarte, C. M. 1995. Submerged aquatic vegetation in relation to different nut rient regimes. Ophelia 41 :87 112. Duarte, C.M. 2002. The future of seagrass meadows. Environmental Conservation 29 :192 206. Elwood, J. W., J. D. Newbold, A. F. Trimble, and R. W. Stark. 1981. The limiting role of phosphorus in a woodland stream ecosy stem: effects of P enrichment on leaf decomposition and primary producers. Ecology 62 :146 158. Ellis, N. and Y G. Wang. 2007. Effects of fish density distribution and effort distribution on catchability. ICES Journal of Marine Science 64 :178 191. Este s, J. A., M. T. Tinker, T. M. Williams and D. F. Doak. 1998. Killer whale predation on sea otters linking oceanic and nearshore ecosystems. Science 282 :473 476. Fahrig, L. 1997. Relative effects of habitat loss and fragmentation on population extinct ion. The Journal of Wildlife Management 61 :603 610. Fahrig, L. 2003. Effects of habitat fragmentation on biodiversity. Annual Review of Ecology, Evolution, and Systematics 34 :487 515. Fahrig, L., and G. Merriam. 1985. Habitat patch connectivity and p opulation survival. Ecology 66 :1762 1768.

PAGE 170

170 Florida Geological Survey (FGS). 2004. Springs of Florida. Bulletin No. 66. Tallahassee, Florida USA Frazer, T. K., S. K. Notestein and W. E. Pine, III 2006. Changes in the physical, chemical and vegetat ive characteristics of the Homosassa, Chassahowitzka and Weeki Wachee Rivers. Final Report to the Southwest Florida Water Management District, Surface Water Improvement and Management Program. Tampa, Florida USA Freire, K. M. F., V. Christensen, and D. Pauly. 2008. Description of the East Brazil Large Marine Ecosystem using a trophic model. Scientia Marina 72 :477 491. Gotceitas, V. and P. Colgan. 1989. Predator foraging success and habitat complexity: quantitative test of the threshold. Oecologia 80 :158 166. Gould, W. R. and K. H. Pollock. 1997. Catch effort maximum likelihood estimation of important population parameters. Can adian J ournal of Fish eries and Aquat ic Sci ence s 54 :890 897. Gregg, W. W., and F. L. Rose. 1982. The effects of aquati c macrophytes in the stream micro environment. Aquatic Botany 14 :309 324. Grumbine, R. E. 1994. What is ecosystem management? Conservation Biology 8 :27 38. Guidetti, P. 2000. Differences among fish assemblages associated with nearshore Posidonia ocea nica seagrass beds, rocky algal reefs, and unvegetated sand habitats in the Adriatic Sea. Estuarine, Coastal and Shelf Science 50 :515 529. Hairston, N. G., F. E. Smith and L. B. Slobodkin. 1960. Community structure, population control, and competition. The American Naturalist 94 :421 425. Hall, C. A. S., J. A. Stanford, and F. R. Hauer. 1992. The distribution and abundance of organisms as a consequence of energy balances along multiple environmental gradients. Oikos 65 :377 390. Harrison, S. and E. Br una. 1999. Habitat fragmentation and large scale conservation: what do we know for sure? Ecography 22 : 225 232. Hastings, A. 1977. Spatial heterogeneity and the stability of predator prey systems. Theoretical Population Biology 12 : 37 48. He, X. and J Kitchell. 1990. Direct and indirect effects of predation on a fish community: a whole lake experiment. Transactions of the American Fisheries Society 119 :825 835.

PAGE 171

171 Heck, K. L., Jr., K. W. Able, C. T. Roman, and M. P. Fahay. 1995. Composition, abunda nce, biomass, and production of macrofauna in a New England Estuary: comparisons among eelgrass meadows and other nursery habitats. Estuaries 18 : 379 389. Heck, K. L., Jr. and L. B. Crowder. 1991. Habitat structure and predator prey interactions in veg etated aquatic systems. Pages 281 299 in Habitat Structure: The Physical Arrangements of Objects in Space. Eds. Bell, S.S., E.D. McCoy and H.R. Mushinsky. Chapman and Hall, New York USA Heck, K. L., Jr., G. Hays, and R. J. Orth. 2003. Critical evalu ation of the nursery role hypothesis for seagrass meadows. Marine Ecology Progress Series 253 :123 136. Herald, E. S. and R. R. Strickland. 1949. An annotated list of the fishes of Homosassa Springs, Florida. Journal of Florida Academy of Sciences 11 :9 9 109. Hilborn, R. and M. Mangel. 1997. The Ecological Detective: Confronting Models with Data. Princeton University Press. Princeton, New Jersey USA Hilborn, R. and C. J. Walters. 1992. Quantitative Fisheries Stock Assessment: Choice, Dynamics a nd Uncertainty Kluwer Academic Publishers, Boston USA Holling, C. S. 1959. Th e components of predation as re vealed by a study of s mall mammal predation of the Eu ropean pine sawfly. Canadian Entomologist 91 :293 320. Hoyer, M. V., T. K. Frazer, S. K. Not estein and D. E. Canfield. 2004. Vegetative characteristics of three low lying Florida coastal rivers in relation to flow, light, salinity and nutrients. Hydrobiologia 528 :31 43. Huffaker, C. B. 1958. Experimental studies on predation: dispersal factors and predator prey oscillations. Hilgardia 27 : 343 383. Huntsman, A. G. 1948. Fertility and fertilization of streams. Journal of the Fisheries Research Board of Canada 7 :248 253. Jacobsen, L., and S. Berg. 1998. Diel variation in habitat use by plank tivores in field enclosure experiments: the effect of submerged macrophytes and predation. Journal of Fish Biology 53 :1207 1219. Jeppesen, E., M. Sondergaard, M. Sondergaard, and K. Christoffersen. 1998. The Structuring Role of Submerged Macrophytes in Lakes. Ecological Studies Vol. 131. Springer, New York USA Jones, J. I., B. Moss, and J. O. Young. 1998. Interactions between periphyton, nonmolluscan invertebrates, and fish in standing waters. In The Structuring Role of Submerged Macrophytes in La kes. Ecological Studies 131 :197 216.

PAGE 172

172 Kamler, J. F. and K. L Pope. 2001. Nonlethal methods of examining fish stomach contents. Nebraska Cooperative Fish and Wildlife Research Unit, University of Nebraska, Lincoln USA Karr, J. R. 1987. Biological mon itoring and environmental assessment: a conceptual framework. Environmental Management 11 :249 258. Kitchell, J. F. and L. B. Crowder. 1986. Predator prey interactions in Lake Michigan: model predictions and recent dynamics. Environmental Biology of Fi shes 16 :205 211. Komonen, A., R. Penttila, M. Lindgren, and I. Hanski. 2000. Forest fragmentation truncates a food chain based on an old growth forest bracket fungus. Oikos 90 :119 126. Kushlan, J. A. 1974. Effects of a natural fish kill on the water q uality, plankton, and fish population of a pond in the Big Cypress Swamp, Florida. Transactions of the American Fisheries Society 103 :235 243. Levin, S. A. and R. T. Paine. 1974. Disturbance, patch formation, and community structure. Proc eedings of the Nat ional Acad emy of Sci ence 71 :2744 2747. Lindenmayer, D. B., C. R. Margules, and D. B. Botkin. 2000. Indicators of biodiversity for ecologically sustainable forest management. Conservation Biology 14 :941 950. Link, J. S. 2002. Ecological considerati ons in fisheries management. Fisheries 27 :10 17. Link, J. S., J. K. T. Brodziak, S. F. Edwards, W. J. Overholtz, D. Mountain, J. W. Jossi, T. D. Smith, and M. J. Fogarty. 2002. Marine ecosystem assessment in a fisheries management context. Can adian J ou rnal of Fish eries and Aquat ic Sci ence s 59 :1429 1440. Lotka, A. J. 1925. Elements of physical biology. Williams & Wilkins, Baltimore, Maryland, USA. MacArthur, R. H. 1958. Population ecology of some warblers of northeastern coniferous forests. Ecology 39 :599 619. MacArthur, R. H. and E. R. Pianka. 1966. On optimal use of a patchy environment. The American Naturalist 100 :603 609. Malone, C. R. 1995. Ecosystem management: status of the federal initiative. Bulletin of the Ecological Society of America 76 :158 161. Manly, B. F. J., P. Miller, and L. M. Cook. 1972. Analysis of a selective predation experiment. The American Naturalist 106 :719 736.

PAGE 173

173 Mann, R. H. K. and T. Penczak. 1984. The efficiency of a new electrofishing technique in determining fish number in a large river in Central Poland. J ournal of Fish Biol ogy 24 :173 185. McCargo, J. W., and J. T. Peterson. 2010. An evaluation of the influence of seasonal base flow and geomorphic stream characteristics on Coastal Plain stream fish assemblages Transactions of the American Fisheries Society 139 :29 48. McGlathery, K. 2001. Macroalgal blooms contribute to the decline of seagrass in nutrient enriched coastal waters. Journal of Phycology 37 :453 456. Menge, B. A. and J. P. Sutherland. 1976. Sp ecies diversity gradients: synthesis of the roles of predation, competition, and temporal heterogeneity. The American Naturalist 110 :351 369. Micheli, F. 1999. Eutrophication, fisheries, and consumer resource dynamics in marine pelagic ecosystems. Scie nce 285 :1396 1398. Minshall, G. W. 1978. Autotrophy in stream ecosystems. Bio S cience 28 :767 771. Mizerek, T., H. M. Regan, and K. A. Hovel. 2011. Seagrass habitat loss and fragmentation influence management strategies for a blue crab Callinectes sapid us fishery. Marine Ecology Progress Series 427 : 247 257. Montgomery, J. L. M., and T. E. Targett. 1992. The nutritional role of seagrass in the diet of the omnivorous pinfish Lagodon rhomboides (L.). Journal of Experimental Marine Biology and Ecology 15 8 :37 57. Munday, P.L. 2004. Habitat loss, resource specialization, and extinction on coral reefs. Global Change Biology 10 :1642 1647. Murdoch, W. W. 1969. Switching in general predators: experiments on predator specificity and stability of prey popula tions. Ecological Monographs 39 :335 354. Murdoch, W. W. 1971. The developmental response of predators to changes in prey density. Ecology 52 :132 137. Nakamura, Y. 2010. Patterns in fish response to seagrass bed loss at the southern Ryukyu Islands, Ja pan. Marine Biology 157 :2397 2406. Namba, T., A. Umemoto, and E. Minami. 1999. The effects of habitat fragmentation on persistence of source sink metapopulations in systems with predators and prey or apparent competitors. Theoretical Population Biology 56 :123 137. Neahr, T. A., G. W. Stunz, and T. J. Minello. 2010. Habitat use patterns of newly settled spotted seatrout in estuaries of the north western Gulf of Mexico. Fisheries Management and Ecology 17 :404 413.

PAGE 174

174 Nichols, J. D. 1992. Capture recaptu re models. BioScience 42 :94 102. Oceans Research Board. 2006. A Review of the Draft Ocean Research Priorities Plan: Charting the Course for Ocean Science in the United States. National Academies Press, Washington, D.C USA Odum, H. T. 1953. Factors controlling marine invasion into Florida fresh waters. Bulletin of Marine Science of the Gulf and Caribbean 3 : 134 156. Odum, H. T. 1957. Primary production measurements in eleven Florida springs and a marine turtle grass community. Limnology and Oceano graphy 2 :85 97. Pace, M. L., J. J. Cole, S. R. Carpenter, and J. F. Kitchell. 1999. Trophic cascades revealed in diverse ecosystems. Trends in Ecology and Evolution 14 :483 488. Paine, R. T. 1980. Food webs: linkage, interaction strength and community infrastructure. Journal of Animal Ecology 49 :666 685. Pauly, D., V. Christensen, and C. Walters. 2000. Ecopath, Ecosim, and Ecospace as tools for evaluating ecosystem impacts of fisheries. ICES Journal of Marine Science 57 :697 706. Peterson, B. J., L. Deegan, J. Helfrich, J. E. Hobbie, M. Hullar, B. Moller, T. E. Ford, A. Hershey, A. Hiltner, G. Kipphut, M. A. Lock, D. M. Fiebig, V. McKinley, M. C. Miller, J. R. Vestal, R. Ventullo, and G. Volk. 1993. Biological responses of a tundra river to fertiliz ation. Ecology 74 :653 672. Peterson, J. T., R. F. Thurow, and J. W. Guzevich. 2004. An evaluation of multipass electrofishing for estimating the abundance of stream dwelling salmonids. Transactions of the American Fisheries Society 133 :462 475. Pew Oce for Sea Change. A Report to the Nation. Pew Oceans Commission, Arlington, Virginia, USA Pikitch, E. K., C. Santora, E. A. Babcock, A. Bakun, R. Bonfil, D. O. Conover, P. Dayton, P Doukakis, D. Fluharty, B. Heneman, E. D. Houde, J. Link, P. A. Livingston, M. Mangel, M. K. McAllister, J. Pope and K. J. Sainsbury. 2004. Ecosystem based fishery management. Science 305 :346 347. Pillay, D., G. M. Branch, C. L. Griffiths, C. Williams, and A. Prinsloo. 2010. Ecosystem change in a South African marine reserve (1960 2009): role of seagrass loss and anthropogenic disturbance. Marine Ecology Progress Series 415 :35 48. Pine, W. E. III, K. H. Pollock, J. E. Hightower, T. J. Kwak, and J. A Rice. 2003. A review of tagging methods for estimating fish population size and components of mortality. Fisheries 28 :10 23.

PAGE 175

175 Pine, W. E. III, and J. C. Tetzlaff 2008. Rivers as conduits: An evaluation of riverine fish movement and habitat use betw een FCWCS identified critical habitats in a Florida river. Final Report: Florida Fish and Wildlife Conservation Commission, Tallahassee Florida, USA Polis, G. A. and D. R. Strong. 1996. Food web complexity and community dynamics. The American Natural ist 147 :813 846. Pollock, K. H., J. D. Nichols, T. R. Simons, G. L. Farnsworth, L. L. Bailey and J. R. Sauer. 2002. Large scale wildlife monitoring studies: statistical methods for design and analyses. Environmetrics 13 :105 119. Power M. E. 1995. Top down and bottom up forces in food webs: Do plants have primacy? Ecology 73 :733 746. Power, M. E., W. E. Dietrich, and J. C. Finlay. 1996. Dams and downstream aquatic biodiversity: potential food web consequences of hydrologic and geomorphic change. En vironmental Management 20 :887 895. Preisser, E. L., D. I. Bolnick, and M. F. Benard. 2005. Scared to death? The effects of intimidation and consumption in predator prey interactions. Ecology 86 :510 509. Randall, R. G., C. K. Minns, V. W. Cairns, and J. E. Moore. 1996. The relationship between and index of fish production and submerged macrophytes and other habitat features at three littoral areas in the Great Lakes. Canadian Journal of Fisheries and Aquatic Sciences 53 :35 44. Ricciardi, A. and E. Bo urget. 1998. Weight to weight conversion factors for marine benthic macroinvertebrates. Marine Ecology Progress Series 163 : 245 251. Richards, L. J. and J. T. Schnute. 1986. An experimental and statistical approach to the question: Is CPUE an index of abundance? Can adian J ournal of Fish eries and Aquat ic Sci ences 43 :1214 1227. Ricker, W. E. 1975. Computation and interpretation of biological statistics of fish populations. Bull etin of the Fish eries Res earch Board of Can ada 191 :78. Ritchie, M. E. 200 0. Nitrogen limitation and trophic vs. abiotic influences on insect herbivores in a temperate grassland. Ecology 81 :1601 1612. Roberts, E., J. Kroker, S. Korner, and A. Nicklisch. 2003. The role of periphyton during the re colonization of a shallow lak e with submerged macrophytes. Hydrobiologia 506 509 :525 530. Sandoey, S. and J. P. Nilssen. 1987. Cyclopoid copepods in marginal habitats: Abiotic control of population densities in anthropogenic acidic lakes. Arch ives of Hydrobiology 76 :236 255.

PAGE 176

176 Sass, G. C., J. F. Kitchell, S. R. Carpenter, T. R. Hrabik, A. E. Marburg, and M. G. Turner. 2006. Fish community and food web responses to a whole lake removal of coarse woody habitat. Fisheries 31 :321 330. Savino, J. F. and R. A. Stein. 1982. Preda tor prey interaction between largemouth bass and bluegills as influenced by simulated, submersed vegetation. Transactions of the American Fisheries Society 11 :255 266. Schoener, T. W. 1983. Field experiments on interspecific compet ition. The American N aturalist 122 :220 285. Seber, G. A. F. 2002. The Estimation of Animal Abundance. The Blackburn Press, New Jersey USA Shea, C. P., and J. T. Peterson. 2007. An evaluation of the relative influence of habitat complexity and habitat stability on fish a ssemblage structure in unregulated and regulated reaches of a large southeastern warmwater stream. Transactions of the American Fisheries Society 136 :943 958. Short, K. T 1996. Natural and human induced disturbance of seagrasses. Environmental Conserv ation 23 :17 27. Skalski, G. T., and J. F. Gilliam. 2001. Functional responses with predator interference: viable alternatives to the Holling Type II model. Ecology 82 :3083 3092. Smith, M. D. and A. K. Knapp. 2003. Dominant species maintain ecosystem f unction with non random species loss. Ecology Letters 6 :509 517. Smith, V. H., G. D. Tilman, and J. C. Nekola. 1999. Eutrophication: impacts of excess nutrient inputs on freshwater, marine, and terrestrial ecosystems. Environmental Pollution 100 :179 19 6. Soloman, M. E. 1949. The natural control of animal populations. Journal of Animal Ecology 18 :1 35. Stevenson, R. J., A. Pinowska, and Y. Wang. 2004. Ecological condition of algae and nutrients in Florida springs. Final Report submitted to the Flor ida Department of Environmental Protection, Tallahassee, Florida, USA. Stunz, G. W., T. J. Minello, and P. S. Levin. 2002. A comparison of early juvenile red drum densities among various habitat types in Galveston Bay, Texas. Estuaries 25 :76 85. Sutherl and, W. J. and P. M. Dolman. 1994. Combining behavior and population dynamics with applications for predicting consequences of habitat loss. Proceedings: Biological Sciences 225 :133 138.

PAGE 177

177 Swain, D. P. and A. F. Sinclair. 1994. Fish distribution and cat chability: what is the appropriate measure of distribution? Can adian J ournal of Fish eries and Aquat ic Sci ences 51 :1046 1054. Tetzlaff, J. C. 2008. Energetic consequences of habitat loss: trade offs in energy acquisition and energy expenditure by Micropt erus salmoides Retrieved August 24, 2011, from http://purl.fcla.edu/fcla/etd/UFE0022729 Tetzlaff, J. C., W. E. Pine, III, and T. K. Frazer. 2010. Comparison of bioenergetics parameters from two spring fed riverine largemouth bass populations. The Open Fish Science Journal 3 :87 100. Valiela, I. 1995. Marine Ecological Processes. Second Edition. Springer, New York USA Valiela, I., J. McClelland, J. Hauxwell, P. J. Behr, D. Hersh, and K. Fore man. 1997. Macroalgal blooms in shallow estuaries: controls and ecophysiological and ecosystem consequences. Limnology and Oceanography 42 :1105 1118. van Baalen, M., V. Krivan, P. C. J. van Rijn, and M. W. Sabelis. 2001. Alternative food, switching pr edators, and the persistence of predator prey systems. The American Naturalist 157 :1 13. Volterra, V. 1931. Variations and fluctuations of the number of individuals in animal species living together. Translated from 1928 edition by R. N. Chapman. Animal E cology. Arno, New York, New York, USA. Walters, C. J., V. Christensen, S. J. D. Martell and J. F. Kitchell. 2005. Possible ecosystem impacts of applying MSY policies from single species assessment. ICES Journal of Marine Science 62 :558 568. Walters, C J., V. Christensen, and D. Pauly. 1997. Structuring dynamic models of exploited ecosystems from trophic mass balance assessments. Reviews in Fish Biology and Fisheries 7 :139 172. Walters, C. J., and F. Juanes. 1993. Recruitment limitation as a conse quence of natural selection for use of restricted feeding habitats and predation risk taking by juvenile fishes. Canadian Journal of Fisheries and Aquatic Sciences 50 :2058 2070. Walters, C. J. and S. J. D. Martell. 2004. Fisheries Ecology and Managemen t. Princeton University Press, New Jersey USA Wang, Y. 1999. A maximum likelihood method for estimating natural mortality and catchability coefficient from catch and effort data. Mar ine and Freshwater Res earch 50 :307 311.

PAGE 178

178 West, R. J. and R. J. King. 1996. Marine, brackish, and freshwater fish communities in the vegetated and bare shallows of an Australian coastal river. Estuaries 19 :31 41. Wetzel, R. G. 2001. Limnology: Lake and River Ecosystems. 3 rd Edition. Academic Press, San Die go, CA, USA Whitfield, A. K. 1986. Fish community structure response to major habitat changes within the littoral zone of an estuarine coastal lake. Environmental Biology of Fishes 17 :41 51. Wilcox B. A. and D. D. Murphy 1985. Conservation strategy: the effe cts of fragmentation on extinction. The American Naturalist 125 :879 887. Williams, B. R., J. D. Nichols, and M. J. Conroy. 2002. Analysis and Management of Animal Populations. Academic Press, San Diego USA Winters, G. H. and J. P. Wheeler. 1985. I nteraction between stock area, stock abundance, and catchability coefficient. Can adian J ournal of Fish eries and Aquat ic Sci ences 42 :989 998. Zar, J. H. 1999. Biostatistical Analysis. Fourth Edition. Pearson Education, New Jersey, USA.

PAGE 179

179 BIOGRAPHICAL SKETCH Matthew Lauretta was born and raised in Phoenix, Arizona where he spent his youth exploring the desert, swimming and playing basketball. After graduating high school, he moved to Flagstaff to enjoy the college lifestyle at Northern Arizona Universi ty, located on the cool and mountainous Colorado Plateau. At NAU, Matt hew studied e nvironmental c hemistry and worked as an undergraduate research technician for the Pinyon Ecology Research Group in the Department of Biological Sciences He received two u ndergraduate research awards for his independent study on the use of stable nitrogen isotopes in predicting pin yon tree susceptibility to insect infestation. After graduation, he worked as a field technician monitoring insect, bird, reptile, amphibian and mammal populations along the Colorado River in Grand Canyon. That experience led to a great opportunity to study the native fishes in the Colorado River including the endangered humpback chub ( Gila cypha ). Between the ages of 19 and 26, Matt hew spent over 525 days sleeping under the stars while researching the Grand Canyon ecosystem In addition, he was involved in fish monitoring and environmental assessment studies in several other rivers in Arizona and New Mexico including the Rio Grande, Pecos and Gila rivers. Matt hew received his Ph.D. in fisheries and aquatic sciences from the University of Florida in December 2011, where his research focused coastal river ecosystems.