University of Florida Thesis Or Dissertation
Total Page:16
File Type:pdf, Size:1020Kb
USE OF STRONG HABITAT-ABUNDANCE RELATIONSHIPS TO ASSESS POPULATION STATUS OF CRYPTIC FISHES: AN EXAMPLE USING HARLEQUIN DARTER By KATHRYN M. HARRIGER A THESIS PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF MASTER OF SCIENCE UNIVERSITY OF FLORIDA 2018 © 2018 Kathryn M. Harriger To the imperiled fishes of the Southeast ACKNOWLEDGMENTS I am especially appreciative of the Florida Fish and Wildlife Conservation Commission (FWC) for providing me with the opportunity to pursue a Master’s degree. My committee members, Dr. Micheal S. Allen, Howard Jelks (U.S. Geological Survey), and Dr. Jeffery Hill, provided support and valuable insights to improve this thesis. I greatly appreciate the time Paul Schueller (FWC) spent helping me with all aspects of the analysis of this study. This project would not have been completed without the hard work from fellow FWC employees John Knight, Amanda Mattair, Matt Wegener, and Neil Branson who assisted in the mark-recapture study. I also greatly appreciate the guidance in use of side scan sonar from Adam Kaeser (U.S. Fish and Wildlife Service) and geoprocessing of side scan sonar images by Cameron Bodine (FWC). Robert Dorazio (U.S. Geological Survey) provided valuable direction with statistical analyses. 4 TABLE OF CONTENTS page ACKNOWLEDGMENTS ...............................................................................................................4 LIST OF TABLES ...........................................................................................................................6 LIST OF FIGURES .........................................................................................................................7 ABSTRACT .....................................................................................................................................8 CHAPTER 1 INTRODUCTION ..................................................................................................................10 2 METHODS .............................................................................................................................13 Study Site ................................................................................................................................13 Mark-Recapture ......................................................................................................................14 Quantifying In-Stream Wood ..................................................................................................15 Analysis ..................................................................................................................................17 3 RESULTS ...............................................................................................................................22 Mark-Recapture ......................................................................................................................22 Quantifying In-Stream Wood ..................................................................................................22 Analysis ..................................................................................................................................23 4 DISCUSSION .........................................................................................................................28 APPENDIX A SUPPLEMENTARY TABLES ..............................................................................................32 B R CODE FOR THE HIERARCHICAL BAYESIAN MODEL .............................................36 C JAGS CODE FOR THE HIERARCHICAL BAYESIAN MODEL ......................................40 D INPUT FILES FOR THE HIERARCHICAL BAYESIAN MODEL ....................................42 LIST OF REFERENCES ...............................................................................................................43 BIOGRAPHICAL SKETCH .........................................................................................................47 5 LIST OF TABLES Figure page 2-1 Wood index used to map wood in sonar images in Big Escambia and Pine Barren creeks .................................................................................................................................21 3-1 Potential models for the hierarchical Bayesian analysis with AIC scores .........................26 3-2 Parameter mean, standard deviation and credible intervals for the posterior distribution of the multinomial mixture model ..................................................................27 A-1 Capture histories for Big Escambia Creek. ........................................................................32 A-2 Capture histories for Pine Barren Creek. ...........................................................................33 A-3 Site abundance estimates for Big Escambia Creek. ...........................................................34 A-4 Site abundance estimates for Pine Barren Creek ...............................................................35 6 LIST OF FIGURES Figure page 2-1 Survey reaches in Pine Barren and Big Escambia creeks. .................................................20 2-2 Three GIS layers used to organize data prior to analysis ...................................................21 3-1 Visual depiction of the number of wood pieces and darter abundance at each unsampled site in Big Escambia Creek. .............................................................................25 3-2 Visual depiction of the number of wood pieces and darter abundance at each unsampled site in Pine Barren Creek. ................................................................................26 3-3 Expected Harlequin Darter abundance as a function of woody debris. .............................27 7 Abstract of Thesis Presented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Master of Science USE OF STRONG HABITAT-ABUNDANCE RELATIONSHIPS TO ASSESS POPULATION STATUS OF CRYPTIC FISHES: AN EXAMPLE USING HARLEQUIN DARTER By Kathryn M. Harriger May 2018 Chair: Micheal S. Allen Major: Fisheries and Aquatic Sciences Understanding trends in abundance is important to fisheries conservation, but techniques for estimating stream-wide abundance of cryptic fish with strong habitat-abundance relationships are not well established. I developed techniques to address this need using Harlequin Darter Etheostoma histrio, which is a small, cryptic freshwater fish associated with woody structure in streams where it occurs. Specifically, my objectives were to (1) determine how darter abundance and in-stream wood were related at sampled sites, and (2) to use this relationship to estimate darter abundance at unsampled sites and extrapolate for stream-wide darter abundance estimates and associated uncertainty. I conducted mark-recapture studies using visual surveys to sample Harlequin Darters from Big Escambia and Pine Barren creeks (Escambia River tributaries in northwest Florida). The amount of in-stream wood in both creeks was quantified and mapped using side scan sonar and geographic information system tools. These darter and wood data were used in a hierarchical Bayesian model (multinomial mixture model) to determine site abundance of Harlequin Darters, the effect of in-stream wood on darter abundance, and to extrapolate darter abundance stream-wide. I found a positive relationship between wood and darter abundance at both creeks, and there were more wood pieces in Pine Barren Creek than Big 8 Escambia Creek. Therefore, the total site abundance and extrapolated stream-wide abundance of Harlequin Darters were both greatest in Pine Barren Creek. The extrapolated stream-wide abundance estimates were 7,238 darters (95% credible interval = 5,746–9,220) in Big Escambia Creek and 8,804 darters (95% credible interval = 7,684–10,116) in Pine Barren Creek. My methods were effective for estimating stream-wide abundance of a small, cryptic fish that uses complex woody habitat, and my findings may assist in the conservation of Harlequin Darters. 9 CHAPTER 1 INTRODUCTION Understanding trends in abundance is important for conservation of fishes, especially for imperiled species (a species in decline regardless of their official listing status; Rahel et al. 1999). A common goal of imperiled species management is to determine population status and maintain or improve a species’ status (Rahel et al. 1999; FWC 2013). Since ecological processes vary depending on scale (Lewis et al. 1996; Rabeni and Sowa 1996), choosing the correct scale is important to addressing specific conservation issues. Appropriate sampling and analysis of fish and habitat is important to extrapolate results to extensive reaches of streams. Estimating abundance of a fish species at a large scale can be logistically difficult and costly. To address these difficulties, many studies follow a general two-step process to first stratify sampling by habitat type within a stream or watershed, estimate abundance of the fish at these sites, and then extrapolate results to the rest of the stream or watershed (similar to the Basinwide Visual Estimation Technique: Hankin and Reeves 1988; Dolloff et al. 1993). However, effective and efficient methods for estimating abundance of imperiled fishes, which are often cryptic, elusive, and patchily distributed, may require modifications of this general technique. Toepfer et al. (2000) adjusted methods to account for nonlinear longitudinal distribution