Bottlenose Dolphin (Tursiops Truncatus) Stock Structure Within the Estuaries of Southern Georgia

BOTTLENOSE DOLPHIN (TURSIOPS TRUNCATUS) STOCK STRUCTURE WITHIN THE ESTUARIES OF SOUTHERN GEORGIA

Brian C. Balmer

A Dissertation Submitted to the University of North Carolina Wilmington in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy

Department of Biology and Marine Biology

University of North Carolina Wilmington

2011

Approved by

Advisory Committee

Randall Wells Lori Schwacke a

Douglas Nowacek Frederick Scharf a

Amanda Southwood Williard Andrew Westgate a

D. Ann Pabst a Chair

Accepted by

______Dean, Graduate School

TABLE OF CONTENTS

INTRODUCTION...... …………………………………………………………….. iv

ACKNOWLEDGEMENTS …………………………………………………………… xiv

DEDICATION …………………………………………………………………………xvii

LIST OF TABLES …………………………………………………………………… xviii

LIST OF FIGURES …………………………………………………………………... xix

CHAPTER ONE: RELATIONSHIP BETWEEN PERSISTENT ORGANIC POLLUTANTS (POPS) AND RANGING PATTERNS IN COMMON BOTTLENOSE DOLPHINS (TURSIOPS TRUNCATUS) FROM COASTAL GEORGIA, USA……………………………………………………………………….... 1

ABSTRACT……..…………………………………………………………. 2

INTRODUCTION …………………………………………...... 4

METHODS ………………………………………………………………... 8

RESULTS ……………………………………………………………….. 16

DISCUSSION …………………………………………………………… 19

CHAPTER TWO: SEASONAL ABUNDANCE, SITE-FIDELITY, HABITAT USE, AND RANGING PATTERNS OF BOTTLENOSE DOLPHINS (TURSIOPS TRUNCATUS) WITHIN THE ESTUARIES OF SOUTHERN GEORGIA, USA…………………………………...... 29

ABSTRACT……..……………………………………………………….. 30

INTRODUCTION …………………………………………...... 32

METHODS ………………………………………………………………. 37

RESULTS ……………………………………………………………….. 48

DISCUSSION …………………………………………………………… 68

CHAPTER THREE: DEFINING RESIDENCY PATTERNS FOR COMMON BOTTLENOSE DOLPHINS (TURSIOPS TRUNCATUS) WITHIN THE ESTUARIES OF SOUTHERN GEORGIA, USA…………………………………… 80

ABSTRACT……..……………………………………………………….. 81

INTRODUCTION …………………………………………...... 83

METHODS ………………………………………………………………. 89

RESULTS ………………………………………………………………. 92

DISCUSSION …………………………………………………………… 99

CHAPTER FOUR: EVALUATION OF A SINGLE-PIN, SATELLITE-LINKED TRANSMITTER DEPLOYED ON BOTTLENOSE DOLPHINS (TURSIOPS TRUNCATUS) ALONG THE COAST OF GEORGIA, USA…………………….. 105

ABSTRACT……..……………………………………………………… 106

INTRODUCTION …………………………………………...... 108

METHODS …………………………………………………………….. 111

RESULTS ……………………………………………………………… 117

DISCUSSION …………………………………………………………. 121

LITERATURE CITED...... …………………………………………………………. 123

iii

INTRODUCTION

Rationale

Bottlenose dolphins (Tursiops truncatus) are top-level predators and long- lived residents within the estuaries of the southeastern United States (reviewed in Shane et al., 1986; Wells and Scott, 1999). Lipophilic persistent organic pollutants (POPs), which are biomagnified in organisms at higher trophic levels, are stored in dolphins‟ lipid-rich blubber, making the bottlenose dolphin a sensitive indicator for POPs in coastal ecosystems. Extremely high contaminant concentrations of POPs have been measured in bottlenose dolphins sampled in the Turtle/Brunswick River Estuary (TBRE), a region located along the southern coast of Georgia (Pulster et al., 2009). However, little baseline data exist on bottlenose dolphins within this region.

This dissertation focuses on bottlenose dolphins that utilize the estuaries of southern Georgia, a region that has been exposed to extremely high levels of anthropogenic contaminants. The goals of this dissertation were to determine dolphin abundance and distribution, habitat utilization patterns, and anthropogenic contaminant burdens to better understand bottlenose dolphin stock structure within a complex salt marsh ecosystem in southern Georgia. I specifically compared these features across two adjacent research sites - the heavily polluted waters within the TBRE surrounding the city of Brunswick,

Georgia, defined as the Brunswick field site, and the waters surrounding the

Sapelo Island National Estuarine Research Reserve (SINERR), defined as the

Sapelo field site, which was chosen initially as a reference site for comparison to the Brunswick field site.

There are four main tools that have been utilized to study the stock structure of coastal bottlenose dolphins. Photo-identification surveys have been used to identify abundance, distribution, and site-fidelity patterns of different groups of dolphins within a region (e.g. Balmer et al., 2008; Read et al., 2003;

Speakman et al., 2010; Urian et al., 2009; Williams et al., 1993; Wilson et al.,

1999). Satellite-linked and radio telemetry data have determined ranging patterns of individuals that have distributions outside of the boundaries of photo- identification surveys (e.g. NMFS, unpublished data, Balmer et al., 2008).

Genetic studies have provided insight into different groups of dolphins that reside within the estuaries and coastal waters of a region (reviewed in Rosel et al.,

2009). Anthropogenic contaminant concentrations have been used to identify different groups of dolphins on a broad (e.g. Hansen et al., 2004) and fine (e.g.

Litz et al., 2007) geographic scale. My dissertation focuses on combining three of these tools (photo-identification, telemetry, and contaminants) to enhance our knowledge of dolphin stock structure as well as provide insight into the influence of contaminants on bottlenose dolphins within the estuaries of southern Georgia.

My Master‟s thesis investigated the stock structure of bottlenose dolphins within the St. Joseph Bay region of Florida‟s northern Gulf of Mexico coast following three unusual mortality events (UMEs), which resulted in over 300 dolphin deaths (NOAA, 2004). I utilized two of the above mentioned tools, photo- identification and telemetry, to determine seasonal abundance, site-fidelity, and

v distribution patterns of dolphins within this region. Although the results of this study provided insight into different groups of dolphins utilizing the St. Joseph

Bay region, it is still unclear which dolphins were affected by the UMEs, since research had not been conducted prior to these mortality events.

The research for my Ph.D. dissertation provided the opportunity to utilize the tools that I implemented in my Master‟s research, as well as an additional stock structure tool, that of contaminant sampling, and a new telemetry methodology, a prototype, single-pin satellite linked tag. The results from my

Ph.D. research provided baseline data on bottlenose dolphins in a region with extremely high levels of anthropogenic, point-source contamination, and insight into the potential effects that these contaminants may be having on dolphins within this region. Currently, there have been no UMEs along the southern coast of Georgia, and hopefully there will not be any in the future. However, extremely high levels of contaminants have been identified as a localized stressor on dolphin communities (Schwacke et al., 2002). Thus, unlike in the St. Joseph Bay region where baseline data were collected following several mortality events, the results from my Ph.D. dissertation will provide resource managers with baseline data prior to any mortalities that may occur along the southern coast of Georgia.

These data can be utilized to identify causative factors of mortality and determine which groups of dolphins may have been affected.

There are two common themes that will be addressed throughout my

Ph.D. dissertation. The first is the ability to identify baseline data (abundance, site-fidelity, habitat use, and ranging patterns) for bottlenose dolphins in a

vi complex salt marsh ecosystem, in which the defined geographic boundaries of a survey region may or may not be representative of the distribution pattern(s) of a given dolphin or group of dolphins. The second is utilizing differences in contaminant levels between individuals and groups to provide insight into the effects of these contaminant burdens on dolphins and differentiate between dolphin groups by their respective contaminant loads. My research interests have always been focused on how many individuals can be supported within a particular region (i.e. abundance) and why animals live where they do (i.e. habitat use/distribution). The combination of these two themes within the context of my

Ph.D. dissertation research has expanded my interest to include how we measure these parameters when animals‟ distributions do not fit the boundaries of our study design. In addition, how do we identify stressors (i.e. contaminants) that may be influencing these parameters, when the animals we are studying spend the majority of their time underwater and conceal many of the effects associated with these anthropogenic stressors?

Dissertation structure

This dissertation has four chapters that are formatted as manuscripts to specific targeted journals. The first chapter focuses on characterizing the contaminant exposure of dolphins within the two field sites (Brunswick and

Sapelo), and examines contaminant profiles in relation to individual dolphin ranging patterns based upon photo-identification sighting histories. The second chapter identifies baseline stock structure data (abundance, site-fidelity, habitat

vii use, and ranging patterns) for both field sites using photo-identification surveys and telemetry data. The third chapter defines and utilizes dolphin residency patterns and variations in habitat use to differentiate between dolphins whose ranging patterns overlap within the estuaries of southern Georgia. The fourth chapter evaluates a new satellite-linked tag attachment design that was tested to identify the impacts of this tag on the dorsal fin and duration of tag attachment/transmission.

Chapter 1, which is currently in the journal Science of the Total

Environment, examines the relationship between anthropogenic contaminants and ranging patterns in bottlenose dolphins within the estuaries of southern

Georgia. In this chapter, I utilized contaminant levels measured from dolphin blubber samples and individual ranging patterns from photo-identification efforts within the region. Dolphins were classified into one of three distinct ranging patterns based upon the photo-identification data. Individuals sighted exclusively within one of the defined field sites were considered to have either Brunswick or

Sapelo ranging patterns. Individuals sighted in both field sites were classified as having a Mixed ranging pattern. I utilized a two-way analysis of variance

(ANOVA) including sex (male, female) and ranging pattern (Brunswick, Sapelo,

Mixed) as factors and a Tukey‟s Honestly Significant Difference (HSD) test for pairwise comparisons. Brunswick males had the highest concentrations of polychlorinated biphenyls (PCBs) reported for any marine mammal. Sapelo males had lower PCB levels than in Brunswick males, but comparable to the highest levels measured in other dolphin populations along the southeastern U.S.

viii

The pattern of PCB congeners was consistent with Aroclor 1268, a highly chlorinated PCB mixture used by, LCP Chemicals, a Superfund site located within the Brunswick field site. I performed a linear regression to identify relationships between Aroclor 1268 proportion and mean sighting distance from the point source for each biopsy sampled individual. For both male and female dolphins, there was a negative relationship between the mean sighting distance from the point source and Aroclor 1268 proportion. The results of this chapter were the first to identify the geographic range of point source contaminants in bottlenose dolphins within the estuaries of southern Georgia and provide insight into the pathway(s) in which these contaminants may be accumulating in dolphins within this region.

Chapter 2, which is currently in review by co-authors to be submitted to the Journal of Wildlife Management, determined seasonal abundance, site- fidelity, habitat use, and ranging patterns of bottlenose dolphins within the estuaries of southern Georgia. In this chapter, I utilized capture-recapture (CR) photo-identification surveys to estimate bottlenose dolphin abundance within the

Brunswick and Sapelo field sites across seasons, and short-term radio tracking and satellite-linked telemetry data, in addition to photo-identification data, to determine dolphin ranging patterns, habitat use, and site-fidelity. Strahler Stream

Order (SSO) was utilized as a novel, quantitative technique to classify habitat use across and within each field site. Despite similar survey areas, dolphin abundance in the Sapelo field site was higher in almost every season than in the

Brunswick field site. Dolphin density, measured as the number of dolphins per

ix kilometer of survey effort, was also significantly higher in the Sapelo field site than Brunswick field site across all seasons. Dolphins also utilized habitat differently between the two sites. Dolphin densities were similar across tributary sizes in the Brunswick field site, while dolphin density increased with larger tributary size in the Sapelo field site. Within both field sites, there were seasonal fluctuations in abundance, with highest numbers observed in summer and/or fall.

The majority of dolphins sighted during these peak abundance periods had low site-fidelity, and were sighted within larger tributaries, suggesting that these individuals may be visitors to the region. During seasons with lower abundance, most dolphins had moderate to high site-fidelity and were sighted across all tributary sizes, suggesting that these individuals may be estuarine residents.

Telemetry data collected for four months during summer and fall identified that tagged individuals remained exclusively within the estuarine waters of one or both field sites, or had ranges that extended slightly outside of the field sites‟ boundaries into the adjacent estuarine and coastal waters. These results suggest that all of the tagged individuals are likely residents within the southern

Georgia estuaries. However, the geographic boundaries of the field sites utilized in this study may be insufficient to define the ranging patterns of all dolphins within this region. This chapter provided the first baseline data on bottlenose dolphins within the estuaries of southern Georgia and suggested that additional parameters such as seasonality and habitat use patterns should be utilized, in addition to the current geographic boundaries to differentiate estuarine and coastal stocks within the Western North Atlantic and Gulf of Mexico.

Chapter 3, which is in preparation to be submitted to the Journal of

Cetacean Research and Management, defined and utilized dolphin residency patterns and variations in habitat use to identify and differentiate between dolphins whose ranging patterns overlap within the estuaries of southern

Georgia. Three residency patterns, based upon previous studies, were defined utilizing the total number of years and seasons that individual dolphins were sighted throughout all photo-identification effort in the region. Transients were defined as individuals that were sighted in only one season for one year, seasonal visitors were sighted in one or two seasons for greater than one year, and residents were sighted in three or four seasons for greater than one year.

For each residency pattern, the total number of sightings within each SSO tributary was determined to identify differences in habitat use among residency patterns. Residents were primarily sighted across all seasons and all SSO tributaries. These individuals are likely members of the estuarine stock(s) for this region. Seasonal visitors and transients were sighted predominantly in summer or seasons adjacent to summer and in larger SSO tributaries. The majority of these individuals are likely members of the South Carolina/Georgia Coastal

Stock that enter the larger estuarine tributaries as they travel along the coast.

The results of this chapter provided insight into the appropriate survey effort required to assist with future stock assessments in other estuaries along the U.S.

Atlantic and Gulf of Mexico coasts. A minimum of two years of seasonal CR photo-identification surveys were required to group individuals into the residency patterns defined in this chapter. This survey methodology initially requires an

xi intensive amount of survey effort within a given region. However, after these residency patterns have been determined, photo-identification surveys can be performed on a yearly or multi-year basis, during the appropriate season, to determine abundance for a targeted stock.

Chapter 4, which is currently in press in the journal Aquatic Mammals, evaluated a new satellite-linked tag attachment design, with the goals of minimizing negative impacts to the dorsal fin while maximizing transmitter longevity. Three prototype satellite-linked tags were attached to bottlenose dolphins along with radio transmitters that enabled follow-up monitoring to assess the impact of the prototype tags and to identify the likely causes of tag loss. The 43% weight reduction from earlier side-mount tags, reduction in the number of attachment pins from three to one, and re-positioning of the tag attachment from the middle-upper third to the lower trailing edge of the dorsal fin reduced potential damage to major venous regions in the dorsal fin and minimized long-term effects to the tagged individuals. In addition, these satellite- linked tags provided location data for a mean of 62 + 9 S.D. days, which is comparable to other previous satellite-linked tag transmission durations for small cetaceans. The results of this chapter suggest that the new satellite-linked tag attachment design was a significant improvement in tagging small cetaceans over the previous multi-pin, side-mount designs.

In summary, these chapters offer us insights into the bottlenose dolphins found within the estuaries of southern Georgia and evaluate the appropriate sampling methods to further enhance our understanding of bottlenose dolphin

xii stock structure and the potential effects of contaminant exposure on bottlenose dolphins within this region.

xiii

ACKNOWLEDGEMENTS

There are many people I wish to thank throughout my dissertation. My three “co-advisors,” Dr. Ann Pabst, Dr. Lori Schwacke, and Dr. Randy Wells, have provided an amazing amount of support and mentorship as I have developed through my academic studies as a student at the University of North

Carolina Wilmington, a “contractor/student” with NOAA Charleston, and a researcher with the Chicago Zoological Society. I will always be appreciative of their abilities to keep me focused on class work and writing while in Wilmington,

NC, and ambition to keep me going in sometimes arduous field conditions, wherever I may be. I also wish to thank Dr. Fred Scharf for not only his vast statistical knowledge, but his ability to transfer some portion of that knowledge to me through his rigorous (but well worth it) class or by always being available for an additional random question. I am extremely appreciative of my committee,

Drs. Doug Nowacek, Amanda Southwood, and Andrew Westgate, who continually kept me thinking about different ways to apply my data, such as “how does a dolphin navigate in a swamp,” “how do contaminants disrupt reproductive physiology,” and “when a tag falls off a dolphin is it lost or shed”?

The field work required for my dissertation was some of the most challenging I have experienced to date, and would not have been possible without the hard work and dedication of many. I wish to thank Jeff Adams, Penn

Clarke, Larry Fulford, Ryan McAlarney, Bill McLellan, Peter Nilsson, Dr. John

Schwacke, Todd Speakman, Dr. Forrest Townsend, and Eric Zolman for their assistance during the field components of this project as well as their detailed

xiv involvement in the logistics and data analysis required throughout this project. I particularly wish to thank Barbie Danielson, Clay George, Suzanne Lane, and

Kate Sparks who were not only involved with all of the requirements above, but kept me sane and grounded during many of the tough times throughout the field work required during this project.

I wish to thank my funding sources: NOAA‟s Ocean and Human Health

Initiative, NOAA‟s Marine Mammal Health and Stranding Response Program, the

Chicago Zoological Society, University of North Carolina Wilmington, The

Dolphin Project, and Georgia Department of Natural Resources. Data were collected under Scientific Research Permit Number 932-1905/MA-009526 issued by NOAA Fisheries and IACUC permit numbers HQ-2009-001 and UNCW 2007-

016.

I also would like to deeply thank my friends and family, whose encouragement throughout this process I will always be appreciate of. My father,

Richard Balmer, and brother, Stan Balmer, have been tremendously supportive of me throughout my dissertation. My dad was one of the first to spark my biological interests by taking me fishing every Sunday of my youth. My brother, who has always been there for me, even made it down to help during the

Georgia health assessment and was on the water the day after he passed kidney stones, certainly not an easy feat. My friends Alley Ejlali, Henry Luciano, Jason

“Smiley” Marley, Jack Morris, Stephanie Nowacek, Mike Nystrom, Nick

Robinson, and Bryan Spaulding have continually been there when I have needed them most.

Lastly, I thank my girlfriend Jenn Yordy, who not only was a huge help to me in understanding how to interpret and analyze contaminant data, but most importantly has kept me on the right path throughout this entire process with continual love, support, and encouragement.

xvi

DEDICATION

I would like to dedicate this dissertation to the memory of Dr. Nelio Barros; an amazing mentor, colleague, and most importantly friend. The first time I had the utmost pleasure of meeting Nelio was on the first Sarasota dolphin health assessment I had ever been a part of as a Mote Marine Laboratory intern in

2000. Coming from a background in chasing black bears through the hills of southwestern Virginia, and the fact I had never even seen a bottlenose dolphin up-close, if not for Nelio explaining every intricate detail of the health assessment as well as answering every question I had that day, I would have been truly lost.

Throughout the years of our friendship, he was always someone I could count on and feel instantly better after talking to. I will never forget that during one of the

Sarasota health assessments, he brought me an actual picnic basket of food every night while I worked in lab processing samples. He was by far one of the most genuine and caring people I have ever had the privilege to call a friend.

Nelio, you will be missed and never forgotten.

xvii

LIST OF TABLES

Table Page

CHAPTER ONE

1. Photographic records and biopsy samples obtained from 2004-2009 for all survey effort in the southern Georgia study area (SGA)………………………………………………………………………...... 12

2. Geometric mean persistent organic pollutant (POP) contaminant values and 95% confidence intervals in (µg/g) lipid weight mass from Brunswick, Mixed, and Sapelo bottlenose dolphins sampled in the SGA. Significant P-values are indicated in bold……………...... 18

CHAPTER TWO

1. Capture-recapture photo-identification survey effort for each primary session and Strahler Stream Order (SSO) tributaries within each field site.…………………………..…………………………………………………. 50

2. Program MARK‟s results summarizing Akaike's Information Criterion (AICc), delta AICc, and number of parameters, for “Markovian Emigration” and “Random Emigration” robust models……………………. 51

3. Bottlenose dolphin population data utilizing the “Random Emigration” and “Markovian Emigration” robust models for the Brunswick and Sapelo field sites.……………………………………………………………... 52

4. Effect tests from two-way analysis of variance (ANOVA)…………………56

5. Effect tests from one-way analysis of variance (ANOVA)…………………59

6. Ranging patterns for all individuals in the Southern Georgia Bottlenose Dolphin Photo-identification Catalog………………………….. 61

7. Site-fidelity classification and ranging pattern for each satellite- linked and radio tagged individual dolphin…………………………………. 63

CHAPTER FOUR

1. Satellite-linked and radio tracking summaries for three dolphins tagged along the Georgia coast…………………………………………… 120

xviii

LIST OF FIGURES

Figure Page Page

CHAPTER ONE

1. Brunswick and Sapelo field sites located in the southern Georgia study area (SGA) …………………………………………………………...... 7

2. Relationship between the proportions of Aroclor 1268 congeners found in the blubber of each biopsy sampled individual and its calculated mean sighting distance from LCP Chemicals.……………...... 20

CHAPTER TWO

1. Southern Georgia photo-identification study area (SGA), defined by Balmer et al. (2011). …………………………………………………………..34

2. Schematic representation of Strahler Stream Order (SSO) classification utilized in the Brunswick and Sapelo field sites.…………….45

3. Transect survey design by Strahler Stream Order (SSO) for the (a) Sapelo and (b) Brunswick field sites …………………………………… 46

4. Bottlenose dolphin total abundance and 95% confidence interval (CI) for each primary session in the Brunswick (generated using the “Random Emigration” robust model) and Sapelo (generated using the “Markovian Emigration” robust model) field sites.……………………….… 53

5. Frequency of individuals sighted by site-fidelity classification and year in the (a) Brunswick and (b) Sapelo field sites …………...... 55

6. Dolphins per kilometer of survey effort within the Brunswick and Sapelo field sites, grouped by Strahler Stream Order (SSO) classification…………………………………………………………………… 57

7. Dolphins per kilometer of survey effort within the (a) Brunswick and (b) Sapelo field sites, grouped by site-fidelity classification and Strahler Stream Order (SSO)…………………………………………....60

8. Ranging patterns of (a) radio-tagged and (b) satellite-linked tagged individuals that were sighted outside of the Brunswick field site……….... 64

9. Ranging patterns of (a) radio-tagged individuals that were sighted outside of the Sapelo field site and (b) satellite-linked tagged individual with ranging pattern exclusively within Sapelo field site…….… 65

xix

10. Ranging patterns of satellite-linked tagged individual, Z04, that included estuarine and coastal waters of both the Brunswick and Sapelo field sites………………………………………………………… 67

CHAPTER THREE

1. Southern Georgia photo-identification study area (SGA), defined by Balmer et al. (2011).………………...... 91

2. Frequency sighted of individuals grouped by site-fidelity classification and total number of year sighted ……………………………. 93

3. Frequency sighted of individuals grouped by site-fidelity classification and total numbers of seasons sighted in (a) 1 year and (b) greater than 1 year……………………………………….94

4. Frequency sighted of individuals grouped by site-fidelity classification and (a) in only 1 year and 1 season, (b) greater than 1 year and 1 or 2 seasons, and (c) greater than 1 year and 3 or 4 seasons ………………………………………………………………………... 95

5. Frequency sighted of (a) transients grouped by season, (b) seasonal visitors grouped by site-fidelity classification, and (c) residents grouped by site-fidelity classification in different Strahler Stream Order (SSO) tributaries……………………………………………… 96

6. Highlighted subareas for sightings of dolphins considered to be (a) transients, (b) seasonal visitors, and (c) residents of the SGA………. 97

CHAPTER FOUR

1. SirTrack Kiwisat 202 Cetacean Fin Tag (model K2F161) (Havelock North, NZ)………………………………………………………... 110

2. Turtle/Brunswick River Estuary (TBRE) and Sapelo Island National Estuarine Research Reserve (SINERR) field sites…………………….... 112

3. (a) Z04 with VHF radio transmitter (Top) and Kiwisat 202 Cetacean Fin Tag (Bottom), (b) Z04 with slight migration and slight biogrowth of satellite-linked transmitter; day 39, (c) Z04 without satellite-linked transmitter; day 62………………………………... 114

4. (a) Z08 with VHF radio transmitter (Top) and Kiwisat 202 Cetacean Fin Tag (Bottom), (b) Z08 with severe migration and heavy biogrowth of satellite-linked transmitter; day 64, and (c) Z08 without satellite-linked transmitter; day 77………………………. 115

5. (a) Z22 with VHF radio transmitter (Top) and Kiwisat 202 Cetacean Fin Tag (Bottom), (b) Z22 with no migration and no biogrowth of satellite-linked transmitter; day 20, and (c) Z22 with satellite-linked transmitter; day 71……………………………………. 116

xxi

CHAPTER ONE: RELATIONSHIP BETWEEN PERSISTENT ORGANIC POLLUTANTS (POPS) AND RANGING PATTERNS IN COMMON BOTTLENOSE DOLPHINS (TURSIOPS TRUNCATUS) FROM COASTAL GEORGIA, USA

ABSTRACT

Bottlenose dolphins (Tursiops truncatus) are apex predators in coastal southeastern U.S. waters; as such they are indicators of persistent organic pollutants (POPs) in coastal ecosystems. POP concentrations measured in a dolphin‟s blubber are influenced by a number of factors, including the animal‟s sex and ranging pattern in relation to POP point sources. This study examined

POP concentrations measured in bottlenose dolphin blubber samples (n = 102) from the Georgia, USA coast in relation to individual ranging patterns and specifically, distance of sightings from a polychlorinated biphenyl (PCB) point source near Brunswick, Georgia. Dolphin ranging patterns were determined based upon 5 years of photo-identification data from two field sites approximately

40 kilometers apart: (1) the Brunswick field site, which included the Turtle

Brunswick River Estuary (TBRE), and (2) the Sapelo field site, which included the

Sapelo Island National Estuarine Research Reserve (SINERR). Dolphins were categorized into one of three ranging patterns from photo-identification data.

Individuals with sighting histories exclusively within one of the defined field sites were considered to have either Brunswick or Sapelo ranging patterns.

Individuals sighted in both field sites were classified as having a Mixed ranging pattern. Brunswick males had the highest concentrations of PCBs reported for any marine mammal. The pattern of PCB congeners was consistent with Aroclor

1268, a highly chlorinated PCB mixture associated with a Superfund site in

Brunswick. PCB levels in Sapelo males were lower than in Brunswick males, but comparable to the highest levels measured in other dolphin populations along the

2 southeastern U.S. Female dolphins had higher Aroclor 1268 proportions than males, suggesting that the highly chlorinated congeners associated with Aroclor

1268 may not be offloaded through parturition and lactation, as easily as less halogenated POPs. Individuals sighted farther from the Superfund point source had lower Aroclor 1268 proportions.

INTRODUCTION

Bottlenose dolphins (Tursiops truncatus) are top-level predators and long- lived residents of bays, estuaries, and tidal marshes along the southeastern

United States (reviewed in Shane et al., 1986; Wells and Scott, 1999). Lipophilic persistent organic pollutants (POPs), which are biomagnified in organisms at higher trophic levels, are stored in their lipid-rich blubber, making the bottlenose dolphin a sensitive indicator for POPs in coastal ecosystems (Kucklick et al., in review).

Contamination of the Turtle/Brunswick River Estuary (TBRE) in southern coastal Georgia (Figure 1) by the highly chlorinated (> 5 chlorines) polychlorinated biphenyls (PCBs) mixture Aroclor 1268 has been well documented (Kannan et al., 1997; Kannan et al., 1998; Maruya and Lee, 1998;

Maruya et al., 2001). The primary PCB congeners found in the TBRE are those that comprise Aroclor 1268, a highly chlorinated (>5 chlorines) mixture of PCBs.

This mixture was used extensively at a chlor-alkali plant that operated in the

TBRE from 1955 to 1994. The site, referred to as LCP Chemicals, was designated a National Priority List (i.e. Superfund) site in 1996 due to extensive environmental contamination from mercury, lead, PCBs, dioxin, and other organic compounds (EPA, 2007; Kannan et al., 1997). Understanding the long- term impacts of these contaminants requires knowledge of the extent to which they contaminate the adjacent environment and food web.

Kannan et al. (1997) measured PCB levels in sediments within the TBRE and determined that sediments sampled from the LCP Chemicals site had PCB

4 concentrations 50 times higher than those measured 500 m from the site. Fish species, including spotted sea trout (Cynoscion nebulosus) and striped mullet

(Mugil cephalus), sampled in the TBRE had PCB concentrations that were three times higher than PCB levels measured in fish from the Skidaway River, approximately 100 km north of the TBRE (Maruya and Lee, 1998). High concentrations of PCBs, specifically those with the Aroclor 1268 congener pattern, were also reported from a pilot study which sampled bottlenose dolphins in the TBRE (Pulster et al., 2009). Pulster et al. (2009) compared PCB levels from blubber of live dolphins sampled via remote biopsy in St. Simons Sound and the adjacent Back River in the TBRE with blubber samples from stranded dolphins collected approximately 90 km to the north, near Savannah, Georgia.

Even with a small sample size of only four male TBRE dolphins, the study was able to discern a congener pattern indicative of an Aroclor 1268 source and similar to the congener profile documented in prey fish from the area (Pulster et al., 2009; Pulster et al., 2005). In addition, Rosel (unpublished NOAA data) reported that mitochondrial DNA control region sequences and microsatellite markers from dolphins remotely biopsied in the TBRE were significantly different from those of dolphins sampled in Savannah, Georgia, and Charleston, South

Carolina. Thus, it has been hypothesized that the dolphins in the TBRE and surrounding waters may be long-term residents to this region (Pulster et al.,

2009). However, to date, this hypothesis has not been tested and no previous data have been published on ranging patterns of dolphins along this region of the

Georgia coast.

This study builds on the previous research of Pulster et al. (2009) by expanding the sampling of dolphins within and outside of the TBRE to examine the relationship between measured POP concentrations and individual dolphin ranging patterns. Biopsy sampling was extended 40 km northeast of the TBRE to the waters in and around the Sapelo Island National Estuarine Research

Reserve (SINERR) (Figure 1). The SINERR is a federal- and state-managed protected area and is the focus of long-term ecological research projects such as water quality monitoring, primary productivity assessment, and fisheries sampling

(e.g. Dresser and Kneib, 2007; Hanson and Synder, 1979; Owen and White,

2005). The area surrounding Sapelo Island, including the SINERR, is relatively undeveloped and was chosen with the intent that dolphins in this area could potentially act as a reference group for comparison with dolphins inhabiting the more contaminated TBRE. However, nothing was known about the ranging patterns of bottlenose dolphins within and between the TBRE and SINERR regions. Thus, if dolphins in the SINERR region were found to have elevated

POP levels, it would be unclear whether such findings were due to contaminant transport or movement of dolphins between the two regions.

Photo-identification of dorsal fins has proven to be a very effective method of identifying individual dolphins and determining their ranging patterns (e.g. Irvine et al., 1981; Scott et al., 1990a; Wells and Scott, 1990). Photo-identification surveys were initiated within the TBRE and SINERR regions to document the presence of individual dolphins and their potential movement between the sites.

The goals of this study were to characterize the POP, and specifically PCB,

Figure 1. Brunswick and Sapelo field sites located in the southern Georgia study area (SGA). The Sapelo Island National Estuarine Research Reserve (SINERR) is the area within the rectangular box located within the Sapelo field site. The Turtle/Brunswick River Estuary (TBRE), is the area within the square box located in the Brunswick field site. The brackets define the SGA boundaries including the division between the Brunswick and Sapelo field sites and 15 km upriver of the major tributaries.

7 exposure of dolphins in the TBRE and SINERR regions and examine patterns of

PCB congeners in relation to individual dolphin ranging patterns based upon photo-identification sighting histories.

METHODS

Study Area

The southern Georgia photo-identification study area (SGA) included the estuarine waters from Sapelo Sound south to St. Simons Sound, representing approximately 60 km of north-south estuarine shoreline (Figure 1). The study area‟s eastern boundaries were defined as the mouths of Sapelo, Doboy,

Altamaha, and St. Simons Sounds. The western boundaries were defined as 15 km upriver of the Sapelo, Altamaha, and Turtle rivers. The SGA was divided into two field sites based upon the location of major sounds within each site. The

Brunswick field site included the TBRE and all estuarine waters from St. Simons

Sound north to and including Altamaha Sound. The Sapelo field site excluded

Altamaha Sound and covered all estuarine waters north to, and including Sapelo

Sound.

Biopsy sample collection

Biopsy samples from individual bottlenose dolphins were collected during both remote biopsy sampling surveys and a capture-release health assessment.

Remote biopsy sampling was conducted in the Brunswick field site in August

2006 and March 2007 and in the Sapelo field site during August 2007, March

2008, and August 2008 utilizing standard techniques demonstrated to be safe and effective in numerous studies of small cetaceans (Kiszka et al., 2010; Sellas et al., 2005; Wells and Scott, 1990). The remote biopsy samples were obtained using a 0.3 m long carbon fiber dart with a 25 mm stainless steel cutterhead, which was propelled by a 0.22 blank charge from a modified 0.22 caliber rifle.

The rifle was equipped with a holosight (Bushnell Corporation, Overland Park,

KS) to improve sampling accuracy and a digital video camera and/or digital still camera to identify the dorsal fin of the sampled individual and to document the animal‟s reactions post-sampling. Dolphins were sampled within a range of 2-6 m from the vessel. Sampling location was typically along the animal‟s flank, approximately 10 cm below and 10 cm behind the caudal insertion of the dorsal fin. The sample collected from the biopsy cutterhead included a superficial layer of epidermis in addition to a full thickness section of blubber approximately 10 mm in diameter and 0.5-1.0 g in weight. Once a sample was obtained, the epidermis was removed from the blubber using latex gloves and sterile instruments. The blubber samples were placed in Teflon jars and frozen in a liquid N2 dry shipper to be analyzed for persistent organic pollutant (POP) concentrations. Only full thickness blubber samples were utilized to determine

POP concentrations in this study. The epidermis, which was stored in 20%

DMSO/saturated NaCl, was used to identify the sex of the sampled individual using molecular methods (Rosel, 2003).

In addition to the remote biopsy samples, surgical biopsy wedges were collected during a health assessment of bottlenose dolphins in August 2009

(Schwacke et al., in review). Dolphins were captured through encirclement with a seine net and brought aboard a specially designed veterinary examination and sampling vessel. Biopsy wedge samples were collected by a veterinarian at a site 10 cm below and 10 cm behind the caudal insertion of the dorsal fin. A chlorohexiderm and ethanol scrub was used to sterilize the sampling region and lidocaine hydrochloride with epinephrine was administered as a local anesthetic.

Sterilized instruments that were hexane and acetone washed as well as autoclaved were used to surgically remove the biopsy wedge sample. For POP analysis, a 0.7 - 1.0 g, full-depth, subsection of the biopsy wedge sample was placed into a 15 ml Teflon jar and frozen in a liquid N2 dry shipper on the sample processing vessel. Following sampling, the dolphins were radio-tagged and released at the capture site. At the lab, the sample was stored frozen at -80oC until analysis. Epidermal samples were also collected and utilized to identify sex as described above.

Biopsy sample analysis

Blubber samples were analyzed for POPs as described previously (Litz et al., 2007). Briefly, approximately 1 g of blubber was minced, dried with sodium sulfate and extracted by pressurized fluid extraction using dichloromethane.

Samples were cleaned up by size exclusion chromatography and aluminum solid phase extraction prior to analysis by gas chromatography mass spectrometry.

Lipid content was calculated gravimetrically from a weighed portion of the PFE

10 extract. POP concentrations were determined using a gas chromatograph-mass spectrometer (GC/MS; Agilent 6890/5973, Palo Alto, CA).

A five to seven point calibration curve of compounds was determined from National Institute of Standards and Technology (NIST) Standard Reference

Material (SRM) solutions and utilized to quantify all analytes and calibrants.

Samples were extracted, cleaned, and analyzed by GC/MS in lots of 30-40 with a minimum of one blank and 1-3 aliquots of NIST SRM 1945 Organics in Whale

Blubber (Kucklick et al., 2010). POP concentrations identified within each aliquot of SRM 1945 were within 7.5% + 3.5% (mean + standard deviation) of the certified values. The limit of detection (LOD) for each analyte was defined as the greater of (a) the mass of the analyte in the lowest detectable calibration solution divided by the sample mass, or (b) the average mass of the analyte detected in blanks plus three times the standard deviation. The limits of detection ranged from 0.089 ng/g wet mass to 16.9 ng/g wet mass for all measured analytes.

Photo-identification

The photographic records for this study were from three efforts of varying duration and scope, totaling 238 surveys from 2004-2009 (Table 1). All efforts were included in this analysis to establish the broadest record possible for each individual dolphin‟s sighting history.

Dorsal fin images were obtained from remote biopsy sampling surveys conducted in 1-2 week sessions in the TBRE during December 2004, August

# of # of remote # of surgical Date Field site Survey type individuals biopsy samples biopsy samples sighted obtained obtained

14 - 17 Dec. 2004 Brunswick Remote Biopsy 11

21 - 30 Aug. 2006 Brunswick Remote Biopsy 130 13

12 - 23 Mar. 2007 Brunswick Remote Biopsy 114 19

20 - 31 Aug. 2007 Sapelo Remote Biopsy 169 20

Brunswick and 04 -16 Feb. 2008 Abundance 146 Sapelo

17 - 27 Mar. 2008 Sapelo Remote Biopsy 77 10

Brunswick and 01 - 11 Apr. 2008 Abundance 146 Sapelo Brunswick and 29 Jul. - 9 Aug. 2008 Abundance 222 Sapelo

18 - 28 Aug. 2008 Sapelo Remote Biopsy 106 14

Brunswick and 06 - 16 Oct. 2008 Abundance 100 Sapelo Brunswick and 29 Jan. - 9 Feb. 2009 Abundance 131 Sapelo Brunswick and 31 Mar. - 11 Apr. 2009 Abundance 159 Sapelo Brunswick and 06 - 16 Jul. 2009 Abundance 196 Sapelo Brunswick and Health 03 - 14 Aug. 2009 26 26 Sapelo Assessment Brunswick and 15 Aug. - 9 Oct. 2009 Radio tracking 224 Sapelo Brunswick and 13 - 24 Oct. 2009 Abundance 179 Sapelo Brunswick and 25 Oct. - 20 Nov. 2009 Radio tracking 69 Sapelo

Table 1. Photographic records and biopsy samples obtained from 2004-2009 for all survey effort in the southern Georgia study area (SGA).

2006, and March 2007 and in and around the SINERR during August 2007,

March 2008 and August 2008 (Table 1). Contaminant results of biopsy samples from the December 2004 TBRE surveys were previously reported (Pulster et al.,

2009) and are not included in this analysis. However, photographic images obtained during the 2004 surveys were included for analysis of individual sighting histories.

Abundance surveys utilizing photo-identification of individuals‟ dorsal fins were conducted during every season for 2008 and 2009 in both the Brunswick and Sapelo field sites. During this effort, a 6 -7 m, center console vessel with three observers surveyed both field sites to obtain photographs of every individual dolphin‟s dorsal fin. Mark-recapture analyses were then performed to determine seasonal abundance (methods reviewed in Balmer et al., 2008) in both the Brunswick and Sapelo field sites.

Radio-tracking was used to identify ranging patterns during summer/fall

2009, following the capture-release health assessment. The two goals of the health assessment were to (1) perform detailed health examinations of bottlenose dolphins from the Brunswick and Sapelo field sites including collection of a surgical wedge biopsy sample for contaminant analysis and (2) attach radio transmitters on bottlenose dolphins to determine short-term ranging patterns.

Balmer et al. (2008) have previously described the methodology for radio transmitter attachment and follow-up tracking. Briefly, bottlenose dolphins in both the Brunswick and Sapelo field sites were temporarily captured and restrained utilizing practices similar to those implemented by the Chicago

Zoological Society‟s Sarasota Dolphin Research Program (Wells et al., 2004).

Radio transmitters were deployed on 28 dolphins (14 male, 14 female) and subsequently tracked by vessel for over 100 days with GPS positions recorded for the visual locations of all tagged individuals.

For all three survey efforts, dorsal fin images were graded on both distinctiveness of the dorsal fin, and photographic quality, following the methods of Urian et al. (1999). A catalog of all fins was created with each individual receiving a unique number based on its distinctive markings. Currently, the SGA photo-identification catalog consists of 646 individual bottlenose dolphins. The photo-identification records from the remote biopsy, abundance, and radio- tracking surveys were used to analyze individuals‟ sighting histories and classify each biopsy sampled individual into one of three ranging patterns. In this study, a ranging pattern is defined as the photo-identification sighting history for an individual dolphin within the SGA region. If all photo-identification sightings of a biopsy sampled individual were in either the defined Brunswick or Sapelo field site, they were identified as having a “Brunswick” or “Sapelo” ranging pattern, respectively. Biopsy sampled individuals that were sighted in both field sites were identified as having a “Mixed” ranging pattern.

Data Analysis

Blubber samples in this study were analyzed for PCB congeners (IUPAC

PCB numbers 18, 28+31, 44, 49, 52, 56, 66, 70, 74, 87, 92, 95, 99, 101, 105,

110, 118, 119, 128, 130, 137,138, 146, 149, 153+132, 151, 154, 156, 157, 158,

163, 170, 172, 174, 176, 177, 178, 180, 183, 185, 187, 189, 194, 195, 197, 199,

200, 201, 202, 203+196, 206, 207,208, and 209), polybrominated diphenyl ether

(PBDE) congeners (47, 99, 100, 153, and 154), dichlorodiphenyl-dichloroethanes

(DDTs) (2,4‟-DDD DDE, and DDT; and 4,4‟- DDD, DDE, and DDT), chlordanes

(CHLs) (cis- and trans-chlordane and nonachlor, oxychlordane and heptachlor epoxide), hexachlorobenzene (HCB), dieldrin, and mirex.  PCBs was defined as the sum of the 54 PCB congeners.  Aroclor 1268 was defined as the sum of the following congeners identified by Maruya and Lee (1998) as indicative of

Aroclor 1268 (174, 180,183, 187, 194, 196, 199, 200, 201, 202, 206, 207, 208, and 209). Aroclor 1268 proportion was calculated as  Aroclor 1268/  PCBs.

To control for lipid content variability between individuals and sampling seasons,

POP concentrations for all samples were calculated on a lipid-weight basis and log transformed to meet the assumptions of normality.

Because mothers transfer much of their accumulated lipophilic contaminant loads to their offspring during each pregnancy and associated lactation period (Aguilar et al., 1999; Wells et al., 2005; Yordy et al., 2010), all biopsied individuals were separated based upon sex. Each sampled individual was classified into its respective ranging pattern (Brunswick, Sapelo, or Mixed) based upon its photo-identification sighting history from all survey efforts. If a sampled individual had a non-distinctive fin or had not been sighted pre- or post- biopsy sampling (i.e. its ranging pattern could not be identified), it was excluded from these analyses. The proportion of Aroclor 1268 congeners was arcsine transformed to meet the assumption of normality. A two-way analysis of variance

(ANOVA) including sex (male, female) and ranging pattern (Brunswick, Sapelo,

Mixed) as factors was performed. When the F-statistic was significant for ranging pattern, pairwise comparisons were made using Tukey‟s Honestly

Significant Difference (HSD) test.

The location of the LCP Chemicals site (31.189440 N, 81.508330 W)

(EPA, 2002), the likely point source for Aroclor 1268 contamination, was used as a reference point and photo-identification sighting histories for each biopsy sampled individual were utilized to calculate the distance of each sighting from this point. Distance for each photo-identification sighting was calculated as the closest on-water distance between the sighting and the reference point using the

“Measure” tool in ArcMap 9.2 (ESRI, Redlands, CA). For each individual dolphin, the mean distance to point source was determined from that dolphin‟s entire sighting history. Linear regression analysis was performed to examine any relationships between the proportions of Aroclor 1268 congeners, and mean sighting distance from point source. A test for homogeneity of slopes was used to determine interactions between sex and distance from point source.

RESULTS

A total of 105 blubber samples were collected via remote biopsy from dolphins in the Brunswick and Sapelo field sites. Of these, 29 remote biopsy samples were excluded because individuals had non-distinctive fins or were not sighted pre- or post-sampling. In addition, 26 samples were collected via surgical biopsy during the capture-release health assessment bringing the total

16 number of samples utilized in this study to 102. Sampled individuals, which were sighted a mean number of 14 + 12 (+ standard deviation) times, were separated by sex and grouped into one of three ranging patterns; Brunswick (♀ = 10, ♂ =

24), Mixed (♀ = 4, ♂ = 18), and Sapelo (♀ = 14, ♂ = 32).

Male dolphins had significantly higher mean concentrations for all POP classes than did females (Table 2). Mean percent lipid was significantly higher in female dolphins than male dolphins (P = 0.0022). PCB and Aroclor 1268 differed significantly across all ranging patterns. There were no significant differences in mean percent lipid and all other POP classes, across male ranging patterns. The highest PCB concentrations in male dolphins were 2870 µg/g

(Brunswick), 756 µg/g (Mixed), and 333 µg/g (Sapelo). Brunswick males had significantly higher mean PCB and  Aroclor 1268 concentrations than did

Sapelo males (P < 0.0001 and P < 0.0001, respectively). Mean PCB and 

Aroclor 1268 concentrations for Mixed males were significantly lower than

Brunswick males (P = 0.0036 and P = 0.0024, respectively) and significantly higher than Sapelo males (P = 0.0028 and P = 0.0090, respectively). The highest PCB concentrations measured in female dolphins were 339 µg/g

(Brunswick), 154 µg/g (Mixed), and 279 µg/g (Sapelo). There were no significant differences in mean percent lipid, PCB,  Aroclor 1268, and all other POP classes between females across ranging patterns. However, the low sample size

(n = 4) for Mixed females limits interpretation of contaminant data associated with this ranging pattern in comparison to the other female ranging patterns.

POP Aroclor 1268 Lipid (%) Σ Aroclor 1268 HCB Dieldrin Mirex Class: proportion

Brunswick 25.12 A A A 3.85 36.77 6.30 0.04 0.16 2.65 Males 509.56 407.78 0.77 (n = 24) (13.17 - 37.07) (369.04 - 703.59) (290.30 - 572.78) (0.74 - 0.80) (2.79 - 5.32) (21.93 - 61.65) (4.31 - 9.22) (0.03 - 0.06) (0.06 - 0.39) (1.86 - 3.78)

B B B Mixed 27.90 253.57 170.71 0. 68 5.12 28.55 5.75 0.05 0.32 2.17 (n = 18) (17.02 - 38.77) (177.89 - 361.45) (119.14 - 244.61) (0.65 - 0.71) (3.78 - 6.95) (16.87 - 48.32) (3.68 - 9.01) (0.04 - 0.07) (0.19 - 0.55) (1.53 - 3.08)

C C C Sapelo 23.57 115.73 69.10 0.60 2.48 20.49 3.83 0.04 0.15 1.69 (n = 32) (14.39 - 32.74) (91.66 - 146.13) (54.97 - 86.86) (0.58 - 0.62) (1.95 - 3.17) (14.03 - 29.93) (2.76 - 5.34) (0.03 - 0.04) (0.11 - 0.21) (1.30 - 2.20)

Brunswick 32.80 a a a 0.63 15.68 0.63 0.02 0.16 0.45 Females 116.47 94.87 0.85 (n = 10) (12.71 - 52.90) (78.14 - 173.60) (64.41 - 139.72) (0.79 - 0.84) (0.22 - 1.82) (2.79 - 88.10) (0.24 - 1.63) (0.01 - 0.04) (0.03 - 0.72) (0.27 - 0.76)

a a a Mixed 28.61 45.94 35.15 0.78 0.38 1.59 0.49 0.01 0.22 0.46 (n = 4) (17.18 - 40.03) (20.75 - 101.72) (19.43 - 63.60) (0.55 - 1.00) (0.05 - 2.57) (0.23 - 10.99) (0.08 - 3.05) (0.00 - 0.03) (0.04 - 1.30) (0.12 - 1.74)

a a b Sapelo 36.44 48.27 30.60 0.63 1.27 10.03 1.31 0.03 0.09 0.77 (n = 14) (19.04 - 53.84) (27.25 - 85.50) (17.72 - 52.86) (0.59 - 0.67) (0.63 - 2.55) (3.98 - 25.32) (0.37 - 4.74) (0.02 - 0.04) (0.03 - 0.26) (0.42 - 1.41)

P -value P = 0.8960 P < 0.0001 P < 0.0001 P < 0.0001 P = 0.7237 P = 0.0674 P = 0.7384 P = 0.3640 P = 0.8094 P = 0.8948 (ranging pattern):

P-value P = 0.0022 P < 0.0001 P < 0.0001 P < 0.0001 P < 0.0001 P = 0.0006 P < 0.0001 P < 0.0001 P = 0.0132 P < 0.0001 (sex):

Table 2. Geometric mean persistent organic pollutant (POP) contaminant values and 95% confidence intervals in (µg/g) lipid weight mass from Brunswick, Mixed, and Sapelo bottlenose dolphins sampled in the SGA. Significant P-values are indicated in bold. Note: For each POP class, statistical differences were determined utilizing a two-way ANOVA with sex and ranging pattern as factors. When the F-statistic was significant for ranging pattern, pairwise comparisons for ranging patterns within each sex were made using Tukey‟s Honestly Significant Difference (HSD) test. Homogeneous groups are indicated by capital (male) or lower case (female) letter subscripts.

Aroclor 1268 proportion in male dolphins differed significantly between all three ranging patterns (P < 0.0001 for all pairwise comparisons), with Brunswick males having the highest proportion followed by Mixed, and Sapelo males (Table

2). Brunswick and Mixed females had a significantly higher proportion of Aroclor

1268 (P < 0.0001 and P = 0.0009, respectively) than did Sapelo females. Aroclor

1268 proportion did not differ significantly between Brunswick and Mixed females

(P = 0.9288).

Linear regression analysis was performed to identify relationships between Aroclor 1268 proportion and mean sighting distance from the point source for each biopsy sampled individual (Figure 2). For both male and female dolphins, there was a negative relationship between the proportion of Aroclor

1268 congeners and mean sighting distance from the point source (males: R2 =

0.6842, P < 0.0001; females: R2 = 0.7137, P < 0.0001). The slopes of the regression lines did not differ between males and females (P = 0.4020).

DISCUSSION

This study confirms that dolphins utilizing the TBRE are exposed to extraordinarily high levels of PCBs. The maximum PCB concentration measured in a Brunswick male was over 1.5 times greater than the maximum PCB level measured in transient, male Pacific killer whales (Orcincus orca), which were previously reported to have the highest PCB levels of any cetacean (Krahn et al.,

2007; Ross et al., 2000). Biomagnification of contaminant concentrations has

Figure 2. Relationship between the proportions of Aroclor 1268 congeners found in the blubber of each biopsy sampled individual and its calculated mean sighting distance from LCP Chemicals.

20 been extensively documented in marine mammal species (reviewed in Houde et al., 2005). Transient killer whales, at the top of the northeastern Pacific marine food web, primarily feed on other marine mammal species (Ford et al., 1998), therefore high contaminant levels would be expected in these individuals through biomagnification. Bottlenose dolphins along the southeastern U.S. are also considered top-level, marine predators (reviewed in Wells et al., 2005).

However, bottlenose dolphin prey is primarily based on lower trophic levels such as pinfish (Lagodon rhomboides), mullet (Mugil spp.), and a variety of soniferous fish species (Barros and Odell, 1990; Barros and Wells, 1998; Berens McCabe et al., 2010; Gannon and Waples, 2004). Thus, based solely on trophic level differences, it would be expected that bottlenose dolphin contaminant concentrations should typically be lower than those of transient killer whales.

The higher levels of PCBs measured in Brunswick male dolphins compared to male transient killer whales is related to the proximity of this population to a major

PCB point source and the exposure to these contaminants within their localized environment due to their ranging patterns.

 PCB concentrations measured in male dolphins that were only sighted in the Sapelo field site were lower than in Brunswick males, but were comparable to those measured for male bottlenose dolphins in northern Biscayne Bay, Florida

(Litz et al., 2007). These males were previously reported to have the highest

PCB concentrations for bottlenose dolphins in the southeastern U.S. The Sapelo

Island National Estuarine Research Reserve (SINERR) has been identified in numerous studies as a “pristine” reference site based upon the minimal amount

21 of urbanization in the region (e.g. Alberts et al., 1990; Chalmers et al., 1985;

Plumley et al., 1980). The elevated levels of PCBs and high Aroclor 1268 proportion in Sapelo male dolphins suggest otherwise. Although there are limited industrial influences surrounding the SINERR, dolphins that have been sighted exclusively in this region have elevated PCB levels associated with a point source located 40 km southwest of their observed ranging pattern. Future research is necessary to identify the pathways leading to Aroclor 1268 contamination in Sapelo dolphins, such as determining contaminant levels and movement patterns of key bottlenose dolphin prey fish species. Contaminated prey or sediments are the most likely routes leading to dolphin exposure as the

Aroclor 1268 mixture is extremely hydrophobic (mean log Kow = 7.9 L/kg)

(Maruya and Lee, 1998) and water transport is unlikely.

For each ranging pattern within the southern Georgia study area (SGA), female dolphins had significantly lower mean PCB and  Aroclor 1268 concentrations, but significantly higher proportions of Aroclor 1268 than males.

Female cetaceans, upon reaching sexually maturity, offload the majority of their contaminants to their first born offspring, primarily through lactation (reviewed in

Aguilar et al., 1999). For example, PCB concentrations measured in adult female bottlenose dolphins from Sarasota are much lower than those of juvenile females from the same community (Wells et al., 2005; Yordy et al., 2010). Yordy et al.

(2010) identified significant changes in POP profiles of female bottlenose dolphins at sexual maturity, where the smallest, least lipophilic contaminants were offloaded through lactation to their first offspring. The predominant Aroclor

1268 congeners are highly chlorinated and therefore may not partition to the milk during lactation, making them resistant to offloading (Kannan et al., 1997;

Kannan et al., 1998; Yordy et al., 2010). Thus, the proportion of Aroclor 1268 in female dolphins would be expected to be higher than in males, as females offload the less lipophilic contaminants and retain the most lipophilic contaminants. The results of this study suggest that SGA female bottlenose dolphins either continue to be exposed to PCBs, or are not offloading contaminants at the same rate as dolphins in other regions, or some combination of these two processes.

Schwacke et al. (2002) suggested that risk of reproductive failure, such as neonate mortality, would be highest for primiparous female bottlenose dolphins, but that following a successful birth and lactation, the risk of reproductive failure would be reduced with a lower contaminant load. The high PCB levels in SGA females, maintained over the course of a reproductive lifetime, may also maintain the high risk for reproductive failure, even for subsequent reproductive events.

Photo-identification data from the 2008 survey effort identified six neonates within the SGA, only one of which survived until the following year (B. Balmer, unpublished data), yielding an annual neonate survival rate of 0.167. For comparison, Speakman et al. (2010) calculated an annual neonatal survival rate of 0.754 (95% CI=0.647-0.878) for bottlenose dolphins in the Charleston

Estuarine Stock. In Sarasota Bay, Florida, the average annual overall neonatal survival is approximately 80%, with about 50% of first-born calves surviving the first year (Wells and Scott, 1990; Wells et al., 2005). Although our SGA estimate

23 is only for a single year, and survival rates often vary greatly across years, these results suggest that dolphin reproductive potential in the SGA may be limited in comparison to other estuarine areas. Knowledge of life history parameters from stranding data is necessary to improve the accuracy of neonatal survivorship estimates. However, collection of high quality stranded carcasses in the SGA has been hampered by geographic remoteness, high tidal flux, and other logistical constraints in the region. Enhanced stranding response, stranding reporting and continuation of photo-identification surveys in the SGA are all needed in order for survival estimates to be calculated and compared with other dolphin populations.

The PCB congeners that comprise Aroclor 1268 have been identified as a point source pollutant from the LCP Chemicals Superfund site (Kannan et al.,

1997; Kucklick et al., in review; Maruya and Lee, 1998; Pulster and Maruya,

2008). There was a significant negative relationship between the proportion of

Aroclor 1268 congeners and mean sighting distance from the LCP Superfund site, indicating that the exposure of a SGA dolphin is directly associated to its proximity to this site. Although PCBs are ubiquitous contaminants and there is potentially some background exposure resultant from long-range environmental transport, the high levels and proportion of Aroclor 1268 congeners indicate that

PCB exposure of the sampled dolphins was predominantly from this single point source. Other studies along the southeastern U.S. have reported elevated levels of highly chlorinated PCB congeners in bottlenose dolphins (Hansen et al., 2004;

Houde et al., 2006; Kucklick et al., in review; Watanabe et al., 2000). Watanabe

24 et al. (2000) determined that over 60% of the PCB profile measured in liver samples from stranded bottlenose dolphins consisted of six (hexa) and seven

(hepta) chlorobiphenyls. Similarly, in blood plasma samples from bottlenose dolphins obtained during capture-release health assessments, the predominant

PCB homolog groups measured were those that contained between five (penta) and seven (hepta) chlorines (Yordy et al., 2010). However, the specific PCB profile of the highly chlorinated congeners associated with Aroclor 1268 have only been identified along the southern coast of Georgia (Kucklick et al., in review). Although our study has identified SGA dolphins with localized ranging patterns exclusively within the Brunswick and Sapelo field sites, future research is necessary to determine if other groups of dolphins are entering the SGA as well as prey species‟ movements into and out of the region.

Kucklick et al. (in review) utilized POP concentrations measured in bottlenose dolphins at 14 locations along the southeastern U.S. and Gulf of

Mexico coasts, to identify geographic differences in POPs. The contaminant levels measured in the Brunswick and Sapelo field sites for this study were two of the locations included in this analysis. Kuckilck et al. (in review) confirmed the results of this study, which identified that Brunswick dolphins had the highest

PCB concentrations measured along the southeastern U.S. and Gulf of Mexico coasts. PBDE concentrations in SGA dolphins were comparable to dolphins sampled in Charleston, SC, and Mississippi Sound, and higher than dolphins sampled in all other sampling locations. Mirex concentrations in SGA dolphins were comparable to dolphins sampled in Sarasota Bay, FL, Tampa Bay, FL, and

Mississippi Sound, and higher than all other sampling locations.  DDT,  CHL,

HCB, and dieldrin concentrations were intermediate in SGA dolphins, in comparison to all other sampling locations. The geographic differences in POP concentrations provide an additional tool to identify bottlenose dolphin stock delineations.

NOAA has defined five coastal and nine estuarine North Western Atlantic

(NWA) bottlenose dolphin stocks, based upon photo-identification, telemetry, and genetic studies at multiple locations along the southeastern U.S. coast (reviewed in Waring et al., 2009). Numerous NWA bottlenose dolphin stocks overlap with each other and the precise delineations of these stocks, and movements of individuals between these stocks, are currently not well understood. On a broad- scale, Hansen et al. (2004) identified differences in POP concentrations between individual dolphins biopsy sampled in multiple states along the southeastern U.S.

Similarly, Litz et al. (2007) identified significant differences in PCB exposure of different bottlenose dolphin communities in the localized estuary of Biscayne

Bay, Florida. The results of this study suggest that the elevated POP levels and patterns may provide insight into Georgia bottlenose dolphin population structure.

The two NOAA defined stocks in this region are the South Carolina/Georgia

Coastal Stock (SCGCS) and the Southern Georgia Estuarine Stock (SGES)

(Waring et al., 2009). The SCGCS includes all of the coastal waters of South

Carolina and Georgia out to 25 m in depth. The SGES includes all of the estuarine waters from Altamaha Sound south to the Cumberland Sound

(Georgia/Florida border). The spatial extent, ranging patterns, and overlap

26 between these two stocks are not well understood. Dolphins that live in the estuarine waters to the north of the SGES, including Sapelo Island and the

SINERR, are not classified into any stock at this time. The results from the photo-identification data and measured contaminant concentrations from this study suggest that Brunswick and Sapelo bottlenose dolphins may be part of separate estuarine stocks; SGES and a previously undefined stock beginning at the Altamaha Sound and extending northward, respectively. Recent studies determining seasonal abundance estimates, as well as ranging and movement patterns of bottlenose dolphins within the Brunswick and Sapelo field sites will augment this study and enhance these proposed changes in current SGA stock delineations.

Results of this study suggest that POP, and specifically Aroclor 1268, contamination extends farther outside of the TBRE than previously documented.

Elevated levels of POPs, such as PCBs, have been identified as potential stressors to marine mammals (reviewed in Houde et al., 2005). Numerous studies have linked high tissue levels of PCBs to deleterious effects on reproduction and immune function (Aguilar and Borrell, 1998; DeLong et al.,

1973; Helle et al., 1976; Martineau et al., 1987). However, identifying POPs as a causative factor of reproductive failure and immune suppression has proven difficult due to the logistical, political, and ethical constraints involved with marine mammals (reviewed in Schwacke et al., 2002). SGA bottlenose dolphins have extremely high levels of PCBs, specifically the highly chlorinated congeners associated with Aroclor 1268, which have been suggested to be resistant to

27 offloading. Individual dolphins within the SGA have relatively localized distribution patterns facilitating routine follow up monitoring. Thus, the bottlenose dolphins within the SGA provide a unique opportunity to identify possible deleterious effects associated with chronic PCB exposure.

CHAPTER TWO: SEASONAL ABUNDANCE, SITE-FIDELITY, HABITAT USE, AND RANGING PATTERNS OF BOTTLENOSE DOLPHINS (TURSIOPS TRUNCATUS) WITHIN THE ESTUARIES OF SOUTHERN GEORGIA, USA

ABSTRACT

Bottlenose dolphins (Tursiops truncatus) within southern Georgia estuaries have been exposed to extremely high levels of persistent organic pollutants (POPs). Dolphins in this region have the highest polychlorinated biphenyl (PCB) levels recorded for any marine mammal and these levels are related to the distance from a known EPA Superfund point-source in the

Turtle/Brunswick River Estuary (TBRE). Currently little is known about the population structure of dolphins in this region. This study provides baseline data on abundance, site-fidelity, habitat use, and ranging patterns of dolphins across two adjacent field sites defined as the Brunswick field site, which included the

TBRE, and the Sapelo field site, which included the Sapelo Island National

Estuarine Research Reserve. The Sapelo field site is relatively undeveloped and was selected for comparison to the more contaminated TBRE. Despite similar survey areas, total dolphin abundance, calculated using photo-identification surveys and capture-recapture (CR) techniques, in the Sapelo field site was higher in almost every season surveyed than in the Brunswick field site. Dolphin density, measured as the number of dolphins per kilometer of survey effort, was also significantly higher in the Sapelo field site than Brunswick field site across all seasons. Dolphins utilized habitat differently between the two sites. Dolphin densities were similar across tributary sizes in the Brunswick field site, while dolphin density increased with larger tributary size in the Sapelo field site. Within both field sites, there were seasonal fluctuations in abundance, with highest numbers observed in summer and/or fall. The majority of dolphins sighted during

30 these peak abundance periods had low site-fidelity, and were sighted within larger tributaries, suggesting that these individuals may be visitors to the region.

During seasons with lower abundance, most dolphins had moderate to high site- fidelity and were sighted across all tributary sizes, suggesting that these individuals may be estuarine residents. Telemetry data collected for four months during summer and fall identified that tagged individuals (n = 28) remained exclusively within the estuarine waters of one or both field sites, or had ranges that extended slightly outside of their boundaries into the adjacent estuarine and coastal waters. These results suggest that all of the tagged individuals are likely residents within the southern Georgia estuaries. However, the geographic boundaries of the field sites utilized in this study may be insufficient to define the ranging patterns of all dolphins within this region.

INTRODUCTION

The coastal morphotype of the bottlenose dolphin (Tursiops truncatus) is distributed along the east coast of the United States throughout the coastal and estuarine waters from Long Island, New York south to the Florida Peninsula

(Waring et al., 2009). Seven coastal stocks and nine estuarine stocks of bottlenose dolphins are currently defined based upon photo-identification, telemetry, and genetic studies that have been carried out at multiple locations along the southeastern U.S. coast (reviewed in Waring et al., 2009). These stocks are defined, in principle, to ensure that any required conservation effort to mitigate a particular human activity is aimed at the correct management unit

(Waring et al., 2009).

The Turtle/Brunswick River Estuary (TBRE), located along the southern coast of Georgia, has been the focus of numerous contaminant studies due to the presence of four heavily contaminated hazardous waste sites which have been designated as National Priority List (NPL) sites by the U.S. Environmental

Protection Agency (EPA, 2007). Extremely elevated levels of persistent organic pollutants (POPs) and metals have been identified in the marsh and surrounding biota within the TBRE (Kannan et al., 1997; Maruya and Lee, 1998; Maruya et al., 2001). Initial studies involving a small number of samples (n = 4) found that bottlenose dolphins in the TBRE had polychlorinated biphenyl (PCB) concentrations that were 10 times higher than those of dolphins sampled 90 km to the north in Savannah, Georgia (Pulster et al., 2009). Pulster et al. (2009)

32 identified a PCB congener pattern indicative of a point-source contaminant,

Aroclor 1268, used by LCP Chemicals, one of the four NPL sites in the TBRE.

Recently, bottlenose dolphins within the TBRE were the focus of an intensive effort to identify the relationship between individual ranging patterns and POP concentrations (Balmer et al., 2011). Photo-identification sighting histories and blubber contaminant levels (measured from remote and surgical biopsy samples) were compared for 102 individual dolphins across two field sites: (1) Brunswick, which included the waters surrounding the city of Brunswick,

Georgia and the TBRE, and (2) Sapelo, which included the Sapelo Island

National Estuarine Research Reserve (SINERR) (Fig. 1). The SINERR, which is jointly managed by NOAA and the Georgia Department of Natural Resources, is the focus of long-term ecological studies including water quality monitoring, habitat restoration, and fisheries sampling (e.g. Dresser and Kneib, 2007;

Hanson and Synder, 1979; Owen and White, 2005). The area surrounding

Sapelo Island, including the SINERR, is relatively undeveloped and was selected for comparison to the more contaminated TBRE. Male bottlenose dolphins sighted exclusively within the Brunswick field site had the highest PCB concentrations yet reported for any marine mammal, which were also significantly higher than PCB concentrations measured in males sighted exclusively within the Sapelo field site. For both male and female dolphins, there was a negative relationship between their mean sighting distance from the known point source of PCB contamination within the TBRE, and the proportion of their total blubber PCB load formed by Aroclor 1268 congeners (Balmer et al., 2011).

Figure 1. Southern Georgia photo-identification study area (SGA), defined by Balmer et al. (2011).

Currently, bottlenose dolphins within the TBRE are considered to be part of the Southern Georgia Estuarine System (SGES) stock, whose geographic boundaries include all of the estuarine waters from Altamaha Sound south to the

Cumberland Sound (Georgia/Florida border) (Waring et al., 2009). Mitochondrial

DNA control region sequences and microsatellite markers from dolphins in the

TBRE are significantly different from those of dolphins sampled within the defined geographic boundaries of two more northern estuarine stocks, the Northern

Georgia/Southern South Carolina Estuarine System and Charleston Estuarine

System stocks (P. E. Rosel, NOAA, personal communication). Bottlenose dolphins in the waters surrounding the SINERR, though, are not yet included in any estuarine system stock. Ranging patterns and spatial overlap of dolphins primarily inhabiting the TBRE and SINERR regions, as well as other estuarine and coastal stocks, are unknown. Understanding stock delineations in this region is of particular importance given the high levels of contaminants identified in these dolphins.

Identifying stocks and estimating the number of individuals belonging to those stocks are necessary steps for establishing effective management plans

(Taylor and Gerrodette, 1993). Capture-recapture (CR) methods have been used intensively in ecological studies to estimate population numbers (Le Cren,

1965). This technique has been modified for bottlenose dolphins by defining a

„capture‟ as a photograph of an individual‟s dorsal fin (i.e. Balmer et al., 2008;

Read et al., 2003; Wells et al., 1996; Williams et al., 1993; Wilson et al., 1999).

CR techniques can be used to model both open and closed populations. Open

35 population models are appealing because they allow for changes in movements of animals in and out of the study area, mortality, and recruitment (Pine et al.,

2003). Closed CR models rely upon the following assumptions: (1) the population is demographically and geographically closed, i.e. no births, death, immigration, or emigration; (2) there is homogeneity of capture probabilities, i.e. all animals have the same likelihood of being sampled; (3) marks are recognized on recapture; and (4) marks are not lost during the study (Seber, 1982).

However, in many wildlife populations, the assumptions of geographic closure and capture homogeneity are violated when performing CR studies (Lancia et al.,

1996). Pollock (1982) combined the strengths of both open and closed population models into a synthetic, robust design that allows for estimation of population size during short-term studies, with closed population parameters, and variations in survival and recruitment between these studies with an open population model (Lancia et al., 1996; Pine et al., 2003).

Previous photo-identification studies of coastal bottlenose dolphins have demonstrated that residency patterns may differ among individuals. Three general patterns have been identified: year-round residents, sighted in the same region for multiple years; seasonal residents, sighted during the same season(s) for multiple years; and transients or visitors, sighted only during a single season

(e.g. Balmer et al., 2008; Zolman, 2002). Thus, it is known that bottlenose dolphins may violate the closed model assumptions of capture homogeneity and geographic closure depending on the year or season of survey effort. However,

Pollock‟s robust design model (1982)can be utilized to estimate seasonal

36 abundance in a survey region by performing multiple, short-term photo- identification surveys (secondary sessions) throughout the year and to account for variation in capture probabilities and recruitment rates across survey periods

(primary sessions) (e.g. Balmer et al., 2008; Speakman et al., 2010).

The goals of this study were to gain insight into the stock structure of bottlenose dolphins in the southern Georgia estuaries, within the Brunswick and

Sapelo field sites. Photo-identification surveys and CR techniques (e.g. Balmer et al., 2008; Read et al., 2003; Williams et al., 1993; Wilson et al., 1999) were used to estimate bottlenose dolphin abundance at both field sites across seasons. Short-term radio tracking and satellite-linked telemetry data (e.g.

Balmer et al., 2008; Evans, 1971; Read and Gaskin, 1985), in addition to photo- identification data, were utilized to identify dolphin ranging patterns, habitat use, and site-fidelity (reviewed in Switzer, 1993; Yochem et al., 1987).

METHODS

Study area

The boundaries of the southern Georgia photo-identification study area

(SGA), defined by Balmer et al. (2011), were chosen based upon the location of the TBRE point source and the SINERR reference site (Fig. 1). Field sites of roughly equal size were selected surrounding these two core features that included approximately 60 km of coastline from St. Simons Sound north to

Sapelo Sound. Altamaha Sound, which delivers the third largest input of freshwater into the Atlantic Ocean from North America (Frangiamore and

Gibbons, 2004), was selected as the boundary between the two field sites, as this geographic feature was located at the median distance of north-south coastline between field sites. The Brunswick field site included all estuarine waters from St. Simons Sound north to and including Altamaha Sound. The

Sapelo field site excluded Altamaha Sound and covered all estuarine waters north to, and including, Sapelo Sound. Bottlenose dolphin seasonal abundance, site-fidelity, ranging patterns, and habitat use were determined for each field site.

Capture-recapture (CR) photo-identification surveys

Capture-recapture (CR) photo-identification surveys were conducted at both field sites, during each season in 2008 and 2009 [winter (January/February), spring (March/April), summer (July/August), and fall (October)]. All tributaries greater than 1 km in length and with a mean low-tide depth of at least 1 m were included in this study. Two contour transect designs (i.e. transects that follow a particular geographic feature) were implemented in this study. For all tributaries less than 1 km across, transects were completed by traveling in the middle of the tributary from beginning to end waypoints. For tributaries greater than 1 km across, the vessel surveyed 500 m off each side of the tributary‟s banks.

Survey effort was temporally divided into primary and secondary sessions.

A primary session was defined as a season of survey effort; thus, there were eight primary sessions for each field site. Within each primary session, three secondary sessions of survey effort were completed. Because of the high number of tributaries in both field sites, it was not possible to survey all transects

38 during each secondary session. During each secondary session, greater than

75% of all transects were completed at random within each field site. Completion of one secondary session took approximately two days of survey effort. Once a secondary session was completed in one field site, survey effort alternated to the other field site, which allowed for sufficient population mixing (> 2 days) between secondary sessions to meet the assumptions of CR modeling. All transects were completed in a Beaufort Sea State of 3 or less to optimize sightability.

Three observers participated in the surveys, each covering 60 degrees of the 180 degrees at the bow of the vessel. The vessel used for surveying was a 6 m center-console, outboard-powered boat. During each CR survey, a sighting was recorded when any dolphin was encountered. Digital photographs of the dorsal fins of all sighted individuals, geographic location, and other environmental data (i.e. salinity, water temperature, and dissolved oxygen) were recorded during each sighting.

Photo-identification data analysis

At the end of each survey day, digital photographs were downloaded to a field laptop and individually labeled with their respective survey and sighting number. Images were then sorted utilizing Adobe Photoshop 7.0 (Adobe

Systems Inc., San Jose, California, USA) and ACDSee 7.0 (ACD Systems,

Victoria, British Columbia, Canada) to identify the best left and/or right side of each individual (reviewed in Speakman et al., 2010). Photographic quality of each image was graded based upon the focus, contrast, angle, dorsal fin

39 visibility, and proportion of the dorsal fin within the frame of the image (Urian et al., 1999). Digital photographs with a Q-1 (excellent) or Q-2 (good) quality grade were included in data analyses; images with a Q-3 (poor) grade were excluded.

Two independent investigators searched for all Q-1 and Q-2 images in the

Southern Georgia Bottlenose Dolphin Photo-identification Catalog, maintained in

Finbase (Adams et al., 2006), a Microsoft Access (Microsoft Corporation,

Redmond, Washington, USA) database that contains all photo-identification records and associated field data for each identified individual sighted in the SGA

(Balmer et al., 2011). The Southern Georgia Bottlenose Dolphin Photo- identification Catalog was created in 2004 and includes images and associated sighting data of dolphins that were identified during remote biopsy, CR, and radio telemetry surveys carried out by NOAA since 2004 (see Balmer et al., 2011). If an individual was not matched in the catalog, it received a new four digit numeric code based upon the location of the most prominent dorsal fin feature (reviewed in Speakman et al., 2010).

A distinctiveness rating (D1-very distinctive, D2-moderately distinctive, D3- not distinctive) was given to each identified individual, which was agreed upon by two separate investigators (Urian et al., 1999). Sighting histories for each D-1 and D-2 individual encountered during the two years of CR photo-identification surveys were exported to the program MARK (White and Burnham, 1999), which was used to estimate dolphin seasonal abundance for each field site (see below).

CR data analysis

CR photo-identification data in this study were analyzed in the program

MARK (White and Burnham, 1999) utilizing the robust design with closed captures. Two different robust models were utilized to estimate dolphin abundance in each field site. The “Markovian Emigration” model sets immigration and emigration as different probability variables across primary sessions, in which individuals return to the study area based upon time dependent functions (Kendall et al., 1997). The “Random Emigration” model sets immigration and emigration as equal probability variables, where individuals can leave the study area and randomly return during other primary sessions (Kendall et al., 1997). Model fitness was based on Akaike's Information Criterion (AICc)

(Akaike, 1974), which is a measure of the goodness of fit for each of the tested population models (Burnham and Anderson, 1992). For each field site, abundance estimates were calculated utilizing the “Markovian Emigration” and

“Random Emigration” robust models. The models with the lowest AICc values were selected to compare abundance across field sites.

Abundance estimates generated in the program MARK were based only upon data from distinctive individuals (D-1 and D-2) in the Brunswick and Sapelo field sites. Thus, the total abundance estimate [distinctive and not distinctive (D-

3) individuals] was calculated as:

(1) Ntotal = Nmodel / Ө where Ntotal = total abundance estimate, Nmodel = robust model abundance estimate, and Ө = mean proportion of distinctive (D-1 and D-2) individuals in

41 each primary session (Wilson et al., 1999). The proportion of distinctive fins was calculated by determining the total number of distinctive and not distinctive individuals identified using photo-identification during a given sighting. Only sightings in which all individual dolphins were photographed (full photo coverage) were utilized in this calculation. Full photo coverage was determined for each sighting by comparing the number of individuals identified from photo- identification analysis and the estimate of total dolphins sighted in the field. If these two numbers were the same, then the sighting would be utilized for calculating the proportion of distinctive individuals. For each primary session, the mean proportion of distinctive individuals (Ө) was calculated. Standard error and

95% confidence intervals for each primary session within each field site of the total abundance estimates were calculated using the delta method as described in Wilson et al. (1999).

Site-fidelity data analysis

Three sources of data were included for analysis to examine site-fidelity:

1) sighting data from photo-identification surveys conducted in 2008-2009 as previously described, 2) existing sighting data from the Southern Georgia

Bottlenose Dolphin Photo-identification Catalog, and 3) sighting data and GPS positions determined from radio and satellite-linked telemetry studies, respectively, conducted in 2009.

Satellite-linked and radio telemetry were conducted over a three month period beginning in August 2009 (Scientific Research Permit Number 932-

1905/MA-009526 issued by NOAA Fisheries and IACUC permit numbers HQ-

2009-001 and UNCW 2007-016) (reviewed in Balmer et al., 2011). The methodology for satellite-linked and radio telemetry utilized in this study has been described in Balmer et al. (2011). Briefly, 28 bottlenose dolphins (14 male, 14 female) across the Brunswick and Sapelo field sites were temporarily captured and restrained utilizing practices similar to those developed and implemented by the Chicago Zoological Society‟s Sarasota Dolphin Research Program (Wells et al., 2004). Each individual was freeze-branded on the dorsal fin with a letter (“Z”) and two digit number (“01, 02, 03,” etc.). Even numbers were assigned to males and odd numbers to females. All dolphins received radio transmitters. Three male dolphins (Z04, Z08, and Z22) also received satellite-linked transmitters.

The total number of sightings for each cataloged individual throughout all photo-identification effort was calculated. For each primary session, individuals were then grouped into a sighting bin based upon the total number of times they were sighted, which describes a relative site-fidelity index. The size of the sighting bins was based upon:

(2) BIN SIZE = (2*IQR)/ n where IQR = the interquartile range of the total number of sightings, and n = the total number of animals sighted. This estimator generates histograms that reliably represent the underlying density distribution of the data (Freedman and

Diaconis 1981). In this study, three site-fidelity bins were identified; individuals with 1-6 sightings (low), 7-12 sightings (moderate), and 13-18 sightings (high) throughout all photo-identification effort.

Satellite-linked and radio telemetry permitted short-term (approximately 3 months), intensive monitoring of tagged dolphins. Satellite-linked locations and radio tracking sightings of tagged individuals were excluded from site-fidelity classifications to prevent an overestimation in the total number of sightings for an individual. However, telemetry data were utilized in identifying habitat use and ranging patterns for tagged individuals. Due to the complexity of the salt marsh estuaries along the southern coast of Georgia, only the highest quality satellite- linked location data were utilized in these analyses (i.e. Class 3- greater than five uplinks received in one satellite pass, position accuracy of approximately 150 m).

Habitat use

Strahler Stream Order (SSO) was used to classify the complex salt marsh estuary habitat in the SGA. SSO is a quantitative technique used to define habitat based upon the number of upstream tributaries (Strahler, 1952).

Tributaries measured with SSO range from headwaters (first order) to progressively larger tributaries (second, third, and fourth order) (Fig. 2). The

Sapelo field site included one 4th order tributary, two 3rd order tributaries, sixteen

2nd order tributaries, and ten 1st order tributaries (Fig. 3a). The Brunswick field site included one 4th order tributary, three 3rd order tributaries, twelve 2nd order tributaries, and ten 1st order tributaries (Fig. 3b).

SSO has been used to quantitatively describe habitat use of invertebrate, amphibian, bird, fish, and mammal species (e.g. Duncan and Ward 1985, Manel et al. 1999, Waite et al. 2004, Hartman and Tornlov 2006). This technique has

Figure 2. Schematic representation of Strahler Stream Order (SSO) classification utilized in the Brunswick and Sapelo field sites.

Figure 3. Transect survey design by Strahler Stream Order (SSO) for the (a) Sapelo and (b) Brunswick field sites.

46 not previously been used to quantify habitat use of marine mammal species. To identify if dolphins were differentially utilizing different sized tributaries across field sites, all photo-identification sightings and satellite-linked locations for each cataloged individual were plotted in ArcMap 9.2 (ESRI, Redlands, California,

USA). SSO classifications for the locations of all sightings were then determined.

Dolphin density, defined here as the total number of dolphins sighted per kilometer of survey effort, within each SSO was then calculated. A two-way analysis of variance (ANOVA) including SSO (1, 2, 3, 4) and field site

(Brunswick, Sapelo) as factors was performed in JMP 7.0 (SAS Institute Inc.,

Cary, North Carolina, USA). When the F-statistic was significant for the interaction between SSO and field site, pairwise comparisons were made using

Tukey‟s Honestly Significant Difference (HSD) test.

Within each field site, to identify if dolphins with different site-fidelity classifications were differentially utilizing different sized tributaries, the total number of dolphins per kilometer within each SSO and site-fidelity classification were determined. For each site-fidelity classification, a one-way ANOVA with

SSO (1, 2, 3, 4) as a factor was performed in JMP 7.0. When the F-statistic was significant for SSO, pairwise comparisons were made using Tukey‟s Honestly

Significant Difference (HSD) test.

Ranging patterns

In this study, all of the CR photo-identification surveys and the majority of remote biopsy effort were conducted exclusively within the boundaries of the

SGA. Thus, it was unclear to what extent individual dolphins may be ranging outside of these boundaries. Satellite-linked and radio telemetry permitted short- term monitoring of individual dolphins to determine if they ranged outside of the

SGA. When radio-tagged individuals were tracked outside of the SGA‟s boundaries, photos were also obtained of any dolphins sighted with tagged individuals. Thus, extended ranging patterns of some non-tagged individuals could also be determined.

Dolphin sighting histories from all survey effort (remote biopsy, CR, and telemetry) were utilized to determine the ranging patterns of all cataloged individuals. Sighting locations for all cataloged individuals were plotted in

ArcMap 9.2 and compared to the boundaries of the Brunswick and Sapelo field sites. For individuals sighted exclusively within the SGA boundaries, three ranging pattern classifications were used: “Brunswick”, individuals sighted exclusively within the Brunswick field site; “Sapelo”, individuals sighted exclusively within the Sapelo field site; and “Mixed”, individuals sighted within both field sites. For individuals sighted outside of the SGA, ranging patterns were defined based upon their cardinal direction in relation to the closest field site, and included whether they were sighted in coastal or estuarine waters (e.g.

“Brunswick and estuarine waters south of Brunswick”).

RESULTS

The Southern Georgia Bottlenose Dolphin Photo-identification Catalog currently consists of 646 distinctive individuals, including 293 dolphins that were

48 first sighted during the two years of photo-identification surveys conducted for this study. From the beginning of the CR surveys in February 2008 until the end of the radio telemetry in November 2009, 154 surveys were completed representing 22,402 km travelled by vessel, 1,006 on-water hours, and 4,185 dolphins sighted.

Seasonal abundance estimates

CR photo-identification survey effort was generally comparable among primary sessions (Table 1). For the Brunswick and Sapelo field sites, the

“Random Emigration” and “Markovian Emigration” robust models had the lowest

Akaike's Information Criterion (AICc), respectively (Table 2). For each field site, abundance estimates generated by both models were generally comparable

(Table 3). Total dolphin abundance (Ntotal) was higher during all primary sessions in the Sapelo field site, except during the winter of 2008 (Table 3).

Brunswick field site.- The number of distinctive individuals (Ndistinct) and total abundance (Ntotal) were highest during the summer primary sessions (Table

3). Total abundance was similar in winter and spring, increased in summer, and decreased to its lowest levels in fall for both years (Fig. 4).

Sapelo field site.- The number of distinctive individuals (Ndistinct) and total abundance (Ntotal) were highest during the 2008 summer and 2009 fall primary sessions (Table 3). Total abundance increased from winter to summer in 2008, decreased in fall of 2008 and increased from winter to fall of 2009 (Fig. 4).

Brunswick field site

Primary SSO 4 SSO 3 SSO 2 SSO 1 Total session (km) (km) (km) (km) km Feb-08 70.86 137.30 144.54 68.46 421.16 Apr-08 70.86 137.30 186.55 72.34 467.05 Jul/Aug-08 70.86 137.30 188.34 75.57 472.08 Oct-08 47.24 122.60 199.93 105.78 475.54 Total: 1835.83

Primary SSO 4 SSO 3 SSO 2 SSO 1 Total session (km) (km) (km) (km) km Jan/Feb-09 70.86 110.88 209.87 46.72 438.33 Mar/Apr-09 70.86 137.30 203.51 46.56 458.23 Jul-09 70.86 137.30 223.31 44.82 476.29 Oct-09 70.86 137.30 198.69 52.41 459.26 Total: 1832.11

Sapelo field site

Primary SSO 4 SSO 3 SSO 2 SSO 1 Total session (km) (km) (km) (km) km Feb-08 39.50 50.39 260.42 47.70 398.00 Apr-08 39.50 57.41 242.35 66.10 405.36 Jul/Aug-08 59.25 57.41 266.41 67.96 451.03 Oct-08 39.50 57.41 281.90 68.70 447.51 Total: 1701.90

Primary SSO 4 SSO 3 SSO 2 SSO 1 Total session (km) (km) (km) (km) km Jan/Feb-09 59.23 57.41 242.91 69.29 428.84 Mar/Apr-09 39.50 57.41 234.62 64.33 395.86 Jul-09 59.23 57.41 287.97 63.15 467.76 Oct-09 39.50 57.41 282.12 62.80 441.83 Total: 1734.29

Table 1. Capture-recapture photo-identification survey effort for each primary session and Strahler Stream Order (SSO) tributaries within each field site.

Brunswick field site

Model AICc Delta AICc # of parameters “Random Emigration” -270.699 0.000 45 “Markovian Emigration” -266.151 4.547 51

Sapelo field site

Model AICc Delta AICc # of parameters “Markovian Emigration” -1127.07 0.000 51 “Random Emigration” -1124.32 2.757 45

Table 2. Program MARK‟s results summarizing Akaike's Information Criterion (AICc), delta AICc, and number of parameters, for “Markovian Emigration” and “Random Emigration” robust models.

Brunswick field site Random Emigration Markovian Emigration Primary Year N N N SE (N ) 95% CI N N SE (N ) 95% CI Session distinct θ model total total model total total 2008 Winter 66 0.50 111 221 17.0 195-246 111 221 17.0 195-246 2008 Spring 47 0.57 117 207 21.7 179-234 109 193 20.0 165-221 2008 Summer 97 0.58 168 288 21.9 255-320 168 288 17.4 256-321 2008 Fall 29 0.51 35 67 6.9 55-80 35 68 7.2 56-81 2009 Winter 61 0.55 92 166 12.3 147-186 95 173 10.9 153-191 2009 Spring 59 0.60 99 164 15.7 139-187 96 159 13.9 135-183 2009 Summer 90 0.55 129 236 12.9 215-257 126 230 11.8 209-252 2009 Fall 36 0.54 52 96 12.7 75-117 52 96 12.7 75-117 Sapelo field site Random Emigration Markovian Emigration Primary Year N N N SE (N ) 95% CI N N SE (N ) 95% CI Session distinct θ model total total model total total 2008 Winter 79 0.61 109 180 11.1 161-198 109 180 11.1 161-198 2008 Spring 99 0.59 171 289 17.5 257-322 171 289 21.6 257-321 2008 Summer 136 0.65 223 344 23.5 308-380 223 344 23.5 308-380 2008 Fall 69 0.61 102 168 12.8 147-188 103 168 12.9 147-189 2009 Winter 89 0.67 130 195 13.5 175-215 136 204 12.2 184-223 2009 Spring 83 0.58 128 219 15.6 197-241 124 213 13.9 191-235 2009 Summer 108 0.59 164 276 16.7 251-300 160 269 15.2 245-294 2009 Fall 70 0.56 240 426 77.9 387-465 180 318 37.0 279-364

Table 3. Bottlenose dolphin population data utilizing the “Random Emigration” and “Markovian Emigration” robust models for the Brunswick and Sapelo field sites. Notes: Ndistinct = the number of distinctive individuals identified during each survey season. θ = the proportion of distinctive (D-1 and D-2) individuals in each survey period. Nmodel = robust model abundance estimate. Ntotal = total abundance estimate adjusting for distinctive and not distinctive individuals. SE = standard error of the total abundance estimate (Ntotal). 95% CI = 95% confidence interval for the total abundance estimate (Ntotal).

Figure 4. Bottlenose dolphin total abundance and 95% confidence interval (CI) for each primary session in the Brunswick (generated using the “Random Emigration” robust model) and Sapelo (generated using the “Markovian Emigration” robust model) field sites.

Site-fidelity

Brunswick field site.- During winter, spring, and fall, the majority of individuals sighted in the Brunswick field site had moderate site-fidelity (Fig. 5a).

During summer, the majority of individuals sighted were those with low site- fidelity (2008) or low and moderate site-fidelity (2009). The percentage of individuals with high site-fidelity remained relatively constant across seasons and years.

Sapelo field site.- During winter, spring, and fall, the majority of individuals sighted in the Sapelo field site had high site-fidelity (Fig. 5b). During summer, the majority of individuals sighted were those with low site-fidelity. The percentage of individuals with moderate site-fidelity remained relatively constant throughout seasons and years.

Habitat use

Overall, dolphin density, defined here as the total number of dolphins per kilometer of survey effort, was significantly higher in the Sapelo field site than the

Brunswick field site (P < 0.0001), and at both field sites, dolphin density increased with Strahler Stream Order (SSO) (P < 0.0001) (Table 4, Fig. 6).

Dolphins utilized habitat differently across the two field sites (P = 0.0058). In the

Brunswick field site, dolphin densities were generally similar across SSO tributaries, differing significantly only across SSO 1 and SSO 4. In the Sapelo field site, dolphin densities differed significantly across all SSO tributaries, except

SSO 1 and SSO 2.

Figure 5. Frequency of individuals sighted by site-fidelity classification and year in the (a) Brunswick and (b) Sapelo field sites.

Dolphins per kilometer by field site and Strahler Stream Order (SSO)

Factor Nparm DF Sum of Squares F Ratio Prob > F Field site 1 1 0.20255025 23.3042 <.0001 SSO 3 3 0.66811133 25.6229 <.0001 Field site*SSO 3 3 0.11215858 4.3014 0.0058

Table 4. Effect tests from two-way analysis of variance (ANOVA).

Figure 6. Dolphins per kilometer of survey effort within the Brunswick and Sapelo field sites, grouped by Strahler Stream Order (SSO) classification. Note: Statistical differences were determined utilizing a two-way ANOVA with field site and SSO as factors. Pairwise comparisons for the interaction of field site and SSO were made using Tukey‟s Honestly Significant Difference (HSD) test. Points that share the same letter are not significantly different from each other.

Brunswick field site.- Dolphins that displayed low site-fidelity were sighted in significantly higher densities in SSO 4 tributaries than all other SSO tributaries

(Table 5, Fig. 7a). Dolphins that displayed moderate site-fidelity were also sighted at significantly higher densities in SSO 4 tributaries than all other SSO tributaries. Moderate site-fidelity dolphin densities were also significantly higher in SSO3 tributaries than SSO 1 tributaries. Dolphins that displayed high site- fidelity were sighted in equal densities across all SSO tributaries.

Sapelo field site.- Dolphins that displayed low site-fidelity were sighted in significantly higher densities in SSO 4 tributaries than all other SSO tributaries

(Table 5, Fig. 7b). Dolphins that displayed moderate site-fidelity were sighted at significantly higher densities in SSO 4 tributaries than SSO 1 tributaries.

Dolphins that displayed high site-fidelity were sighted in significantly higher densities in SSO 3 tributaries than SSO 1 and SSO 2 tributaries.

Ranging patterns- photo-identification

Ranging patterns were determined for all 646 individuals in the Southern

Georgia Bottlenose Dolphin Photo-identification Catalog (Table 6). The majority of dolphins sighted (n = 575, 89%) were defined as having either Brunswick,

Sapelo, or Mixed ranging patterns. Note, though, that the majority of survey effort was conducted within the SGA boundaries. The two predominant patterns for individuals that ranged outside of the SGA were those that included one field site and the adjacent estuarine waters (e.g. Sapelo and estuarine waters north of

Sapelo; Brunswick and estuarine waters south of Brunswick). Only one male

Dolphins per kilometer by Strahler Stream Order (SSO)

Brunswick field site

High site-fidelity

Source Nparm DF Sum of Squares F Ratio Prob > F SSO 3 3 0.00477148 2.1184 0.1204

Moderate site-fidelity

Source Nparm DF Sum of Squares F Ratio Prob > F SSO 3 3 0.67881073 20.7624 <.0001

Low site-fidelity

Source Nparm DF Sum of Squares F Ratio Prob > F SSO 3 3 0.80049992 14.3921 <.0001

Sapelo field site

High site-fidelity

Source Nparm DF Sum of Squares F Ratio Prob > F SSO 3 3 0.12916867 9.7222 0.0001

Moderate site-fidelity

Source Nparm DF Sum of Squares F Ratio Prob > F SSO 3 3 0.07787921 5.7598 0.0034

Low site-fidelity

Source Nparm DF Sum of Squares F Ratio Prob > F SSO 3 3 0.74464458 12.5154 <.0001

Table 5. Effect tests from one-way analysis of variance (ANOVA).

Figure 7. Dolphins per kilometer of survey effort within the (a) Brunswick and (b) Sapelo field sites, grouped by site-fidelity classification and Strahler Stream Order (SSO). Note: For each site-fidelity classification, statistical differences were determined utilizing a one-way ANOVA with SSO as a factor. Pairwise comparisons for SSO, within each site-fidelity classification, were made using Tukey‟s Honestly Significant Difference (HSD) test. Points that share the same letter (A, B, C; low site-fidelity), (a, b, c; moderate site-fidelity), and (X, Y, Z; high site-fidelity) are not significantly different from each other. 60

Ranging pattern # of individuals % Brunswick 186 28.79 Brunswick and estuaries south of Brunswick 26 4.02 Brunswick and coastal Brunswick 10 1.55 Brunswick, coastal Brunswick, and estuaries and coastal south of Brunswick 1 0.15 Estuaries south of Brunswick 8 1.24 Coastal Brunswick 14 2.17 Sapelo 296 45.82 Sapelo and estuaries north of Sapelo 10 1.55 Estuaries north of Sapelo 1 0.15 Mixed 93 14.40 Brunswick and Sapelo, and coastal Brunswick and Sapelo 1 0.15 Total 646

Table 6. Ranging patterns for all individuals in the Southern Georgia Bottlenose Dolphin Photo-identification Catalog.

61 individual, Z04, who was satellite-linked and radio tagged, had a ranging pattern across the estuarine and coastal waters of both field sites.

Ranging patterns- satellite-linked and radio telemetry

The majority of tagged individuals sighted in the Brunswick field site had ranging patterns that extended outside of this field site‟s boundaries (Table 7).

Four radio-tagged individuals, including one mother/calf pair (Z25/Z27) were sighted in the estuarine waters to the south of the Brunswick field site (Fig. 8a).

One individual, Z20, was sighted in the Brunswick field site and adjacent coastal waters (Fig. 8a). Z22 was sighted exclusively within the Brunswick field site during radio tracking surveys (Fig. 8b). However, Z22‟s satellite-linked telemetry identified that this individual had an extended ranging pattern which included the coastal waters of the Brunswick field site as well as the estuarine and coastal waters to the south of the Brunswick field site.

In contrast, the majority of tagged individuals sighted in the Sapelo field site had ranging patterns exclusively within this field site‟s boundaries (Table 7).

Two radio-tagged individuals (Z00 and Z11) were sighted in the estuarine waters north of the Sapelo field site (Fig. 9a). Satellite-linked telemetry and radio tracking location data for Z08, a dolphin with high site-fidelity, though, were exclusively within the Sapelo field site (Fig. 9b).

Three individuals had Mixed ranging patterns that overlapped the estuarine waters of both the Brunswick and Sapelo field sites (Table 7). The ranges of one satellite-linked and radio-tagged individual, Z04, overlapped the

Brunswick field site

FB Sex Site-fidelity Ranging pattern Z14 M High Brunswick Z15 F High Brunswick Z18 M High Brunswick Z23 F High Brunswick Z22 M High Brunswick, coastal Brunswick, and estuaries south of Brunswick Z16 M Moderate Mixed Z04 M Moderate Brunswick and Sapelo, and coastal Brunswick and Sapelo Z25 F Moderate Brunswick and estuaries south of Brunswick Z24 M Low Mixed Z26 M Low Mixed Z20 M Low Brunswick and coastal Brunswick Z27 F N/A Brunswick and estuaries south of Brunswick Z29 F N/A Brunswick Z19 F N/A Brunswick and estuaries south of Brunswick Z21 F N/A Brunswick and estuaries south of Brunswick Sapelo field site FB Sex Site-fidelity Ranging pattern Z01 F High Sapelo Z02 M High Sapelo Z03 F High Sapelo Z07 F High Sapelo Z08 M High Sapelo Z10 M High Sapelo Z12 M High Sapelo Z06 M Moderate Sapelo Z24 M Moderate Mixed Z26 M Moderate Mixed Z04 M Moderate Brunswick and Sapelo, and coastal Brunswick and Sapelo Z11 F Moderate Sapelo and estuaries north of Sapelo Z16 M Low Mixed Z00 M Low Sapelo and estuaries north of Sapelo Z05 F N/A Sapelo Z09 F N/A Sapelo Z13 F N/A Sapelo

Table 7. Site-fidelity classification and ranging pattern for each satellite-linked and radio tagged individual dolphin. Note: N/A refers to individual dolphins that were not distinctive prior to tag attachment.

Figure 8. Ranging patterns of (a) radio-tagged and (b) satellite-linked tagged individuals that were sighted outside of the Brunswick field site. Note: Z25 and Z27 were a mother/calf pair observed together in all radio-tracking sightings. 64

Figure 9. Ranging patterns of (a) radio-tagged individuals that were sighted outside of the Sapelo field site and (b) satellite-linked tagged individual with ranging pattern exclusively within Sapelo field site.

65 coastal waters in addition to the estuarine waters of both field sites (Fig. 10).

Similar to Z22 (Fig. 8b), satellite-linked telemetry identified an extended ranging pattern for Z04 along the coastal waters of the SGA that was not identified with radio telemetry alone.

Site-fidelity- satellite-linked and radio telemetry

Brunswick field site.- All individuals with high site-fidelity, except Z22, were sighted exclusively within the boundaries of the Brunswick field site (Table 7).

Individuals with moderate site-fidelity had ranging patterns that were within the estuaries and along the coast of both field sites (Z04), extended north into the

Sapelo field site (Z16), or south of the Brunswick field site (Z25). Individuals with low site-fidelity had ranging patterns that extended along the coastal waters of

Brunswick (Z20) or north into the Sapelo field site (Z24 and Z26). Four dolphins were not known prior to radio tracking, and, thus could not be included in this analysis.

Sapelo field site.- All individuals with high site-fidelity were sighted exclusively within the boundaries of the Sapelo field site (Table 7). Individuals with moderate site-fidelity had ranging patterns that were within the estuaries and along the coast of both field sites (Z04), exclusively within the Sapelo field site

(Z06), north of the Sapelo field site (Z11), or extended south into the Brunswick field site (Z24, and Z26). Individuals with low site-fidelity had ranging patterns that extended north of the Sapelo field site (Z00) or extended south into the

Figure 10. Ranging patterns of satellite-linked tagged individual, Z04, that included estuarine and coastal waters of both the Brunswick and Sapelo field sites.

Brunswick field site (Z16). Three dolphins were not known prior to radio tracking, and, thus could not be included in this analysis.

DISCUSSION

The goals of this study were to estimate seasonal abundance, identify site-fidelity, determine ranging patterns, and classify habitat use for bottlenose dolphins in an estuarine system that has experienced extremely high levels of

POP contamination.

Abundance, site-fidelity, and habitat use across the Brunswick and Sapelo field sites

Bottlenose dolphin abundance in the Sapelo field site was higher in almost every season than in the Brunswick field site (Fig. 4). Dolphin density in the larger tributaries (SSO 3 and SSO 4) was also approximately three to four times higher in the Sapelo field site than Brunswick field site (Fig. 6). There are a number of factors that may be contributing to these differences in dolphin abundance and density. The Brunswick field site may have a lower carrying capacity for bottlenose dolphins than does the Sapelo field site. Young and

Phillips (2002)developed a model to identify the amount of primary productivity and prey abundance required to support bottlenose dolphins in a salt marsh estuary in South Carolina. The results of this model suggest that variation in productivity and prey species abundance between field sites may contribute to differences in dolphin abundance. The development of salt marsh estuaries into

68 high density suburban housing, shopping centers, and industrial sites has been suggested to significantly reduce the productivity of adjacent coastal ecosystems

(Holland et al., 2004). Primary productivity within the relatively undeveloped salt marshes of the SINERR has been extensively studied since 1953 (Schelske and

Odum, 1961). Smooth cordgrass (Spartina alterniflora) has been identified as the most significant primary producer, which accounts for approximately 75% of the primary production within this estuary (Schelske and Odum, 1961). Sanger et al. (2008) classified the salt marshes within the SINERR as a mix of forested and marsh land use classes, based upon the percentage of impervious cover within this region utilizing National Land Cover Data in ArcGIS 9 (ESRI,

Redlands, California, USA). Currently, land use classifications for the TBRE have not been identified. However, the four NPL sites and numerous other industrial influences in the region (EPA, 2007) suggest that land use within the

TBRE differs dramatically from that of the SINERR. Future research investigating land use classifications and determining primary productivity in the

TBRE would be useful to determine whether differences in primary productivity may be contributing to the differences in dolphin abundance between field sites.

POP contamination can negatively affect reproduction in known prey species of bottlenose dolphins and, thus, also influence carrying capacity.

Thomas (1989) determined that increased PCB concentrations in female Atlantic croaker (Micropogonias undulates), an important prey species of bottlenose dolphins along the southeastern U.S. (Gannon and Waples, 2004), impaired their ovarian growth and disrupted sex hormone production and secretion. Within the

TBRE, Maruya and Lee (1998) identified high levels of PCBs in spotted sea trout

(Cynoscion nebulosus) and striped mullet (Mugil cephalus), two additional prey species of bottlenose dolphins (Barros and Wells, 1998; Berens McCabe et al.,

2010). Thus, the high PCB levels identified in fish from the Brunswick field site may be negatively impacting their reproduction, which could lower the abundance of dolphin prey in this region. Contaminant analysis of dolphin prey species, size, and age class, as conducted by Maruya and Lee (1998), in both field sites is required to test this hypothesis.

POP contamination may also be negatively impacting nursery habitats important to dolphin prey species. Rogers et al. (1984) identified that the upper waters of the Ogeechee River Estuary, just north of the defined boundaries of the

Sapelo field site, were essential habitat for recruitment of larval and juvenile

Atlantic croaker. The four NPL sites in the TBRE, including LCP Chemicals, are located along smaller tributaries that empty into the Turtle River (SSO 3) and St.

Simons Sound (SSO 4). Thus, the high levels of POPs from these NPL sites may be influencing the survival and recruitment of fish utilizing the TBRE as a nursery. The lower dolphin densities in the larger tributaries (SSO 3 and SSO 4) in the Brunswick field site (Fig. 6), as compared to the Sapelo field site, may be due to decreased numbers of dolphin prey in the Turtle River and St. Simons

Sound, which are adjacent to these NPL sites. Dolphin distribution has been linked to prey distribution in other field sites. For example, Barros and Wells

(1998), utilizing stomach contents from stranded bottlenose dolphins and direct observations of dolphin foraging, demonstrated that the primary habitat of dolphin

70 prey, sea grass, was also a main habitat type utilized by dolphins in Sarasota

Bay, Florida. Systematic dolphin prey sampling across both field sites as conducted in Sarasota Bay (e.g. Berens McCabe et al., 2010) is necessary to address this hypothesis further.

Increased localized stressors that impact dolphin demographic parameters, such as birth and survival rates, could also result in abundance differences between the Brunswick and Sapelo field sites. The Brunswick field site includes the city of Brunswick, Georgia, which is the sixth-busiest automobile port along the east coast of the U.S. (Morris, 2007), the four NPL sites, and numerous other industrial influences (EPA, 2007). In contrast, the Sapelo field site, which includes the SINEER, is relatively undeveloped, experiences less vessel disturbance, and is the focus of long-term salt marsh habitat restoration projects (Owen and White, 2005). In Shark Bay, Australia, Bejder et al. (2006) determined that dolphin abundance was lower in areas with higher vessel activity. Focal follows of individual dolphins across both field sites, as performed in other regions (e.g. Nowacek et al., 2001), would be useful to identify if the differences in abundance between the Brunswick and Sapelo field sites may be a result of different levels of human activity between the two field sites.

Extremely high levels of PCBs have also been identified as potential stressors to marine mammals (reviewed in Schwacke et al., 2002). Upon reaching sexual maturity, female dolphins offload the majority of their contaminants to their first born calves, primarily through lactation (Cockcroft et al., 1989). Schwacke et al. (2002) suggested that risk of reproductive failure is

71 highest for primiparous female bottlenose dolphins, but that following a successful birth and lactation that risk is reduced. The high PCB levels identified in Brunswick female dolphins may promote an increased risk for reproductive failure, even for subsequent reproductive events (Balmer et al., 2011). Thus, reproductive effects associated with a PCB point-source in the TBRE could be contributing to the differences in dolphin abundance between field sites.

However, bottlenose dolphin survival rates, which were not explicitly investigated in this study, often vary greatly from year to year (Wells et al., 2005). Thus, continued photo-identification surveys in the SGA are necessary to determine if dolphin survivorship differs between the Brunswick and Sapelo field sites.

Broad-scale comparisons to other studies

Currently, the geographic boundaries of the Brunswick field site are within the defined Southern Georgia Estuarine System (SGES) stock, which includes all of the estuarine waters from Altamaha Sound south to the Cumberland Sound

(Georgia/Florida border) (Waring et al., 2009). Dolphins that live in the estuarine waters to the north of the SGES, including the Sapelo field site, are not classified into any stock at this time. This study, based upon 6 years of photo-identification data, cataloged 307 individuals (47.5% of all animals in the SGA catalog) who were sighted exclusively within the Sapelo field site or estuarine waters to the north (Table 6). Ninety-three SGA cataloged individuals (14%) had ranging patterns that crossed Altamaha Sound and included the estuarine waters of both the Brunswick and Sapelo field sites (Table 6). The majority of SGA cataloged

72 individuals (n = 482, 75%), were sighted exclusively within either the Brunswick or Sapelo field sites. These data suggest that there are limited overlapping ranging patterns of dolphins between the two field sites and that the Altamaha

Sound may be an appropriate geographic boundary to utilize for stock delineations.

In both the Brunswick and Sapelo field sites, bottlenose dolphin abundance was highest during the summer of 2008 (288, 95% CI = 255-320;

344, 95% CI = 308-380, respectively) and decreased approximately threefold during the fall of 2008 (67, 95% CI = 55-80; 168, 95% CI = 147-189, respectively)

(Table 3). Seasonal fluctuations in dolphin abundance have been identified in other field sites along the Gulf of Mexico and Western North Atlantic. In the St.

Joseph Bay region along Florida‟s northern Gulf of Mexico coast, Balmer et al.

(2008) identified a threefold increase in dolphin abundance during spring and fall

(313-410) and lowest abundance during winter and summer (78-152). Similarly, in Charleston, SC, Speakman et al. (2010) identified a threefold increase in abundance during fall (910, 95% CI = 819-1018) and lowest abundance during winter (364, 95% CI = 305-442). For both of these studies, photo-identification surveys were conducted along the coast in addition to the estuarine waters. At each field site, the seasonal increases in abundance were attributed to an influx of animals hypothesized to be members of the adjacent coastal stock. The photo-identification surveys in the SGA did not extend along the coast; however, the majority of dolphins sighted during these peaks in abundance had low site- fidelity and were within the larger tributaries (SSO 4) (Fig. 5). Thus, these

73 individuals may be considered visitors to the SGA and may be members of the

South Carolina/Georgia Coastal Stock (Waring et al., 2009). In contrast, dolphins with high site-fidelity were sighted across all SSO tributaries, suggesting that these individuals are estuarine residents that utilize the habitat differently than dolphins identified as visitors to this region.

The causative factors underlying these seasonal changes in abundance are not yet known. Barco et al. (1999) identified seasonal changes in bottlenose dolphin distribution from the Chesapeake Bay to the ocean in the nearshore waters of Virginia Beach, Virginia. These seasonal shifts in distribution were hypothesized to be related to movement patterns of dolphin prey species that utilized the Chesapeake Bay in the spring and summer and migrated to deeper waters along the coast for spawning in the fall. Thus, shifts in dolphin distribution from the estuarine to coastal waters may contribute to the seasonal abundance patterns observed in the SGA. Gannon and Waples (2004) determined that the diet of North Carolina estuarine dolphins primarily consisted of Atlantic croaker, while North Carolina coastal dolphins primarily fed upon weakfish (Cynoscion regalis). Thus, prey species selection and distribution could be factors contributing to the differences identified in habitat selection between estuarine resident and visitor dolphins in the SGA. Continued CR survey effort, extending along the coast of the Brunswick and Sapelo field sites, coupled with prey sampling, would be required to better understand these seasonal abundance patterns.

The variations in site-fidelity and habitat use patterns displayed by dolphins within this study suggest that contaminant profiles may also differ between these dolphins. Previous biopsy sampling has focused primarily on dolphins with high site-fidelity within the SGA, since these individuals are continually exposed to high levels of PCBs from the TBRE point-source (Balmer et al., 2011). Additional biopsy sampling is necessary during seasons of peak abundance within the larger tributaries to compare contaminant levels of dolphins that visit the SGA and those that may be resident to the estuaries of this region.

Contaminant data (e.g. Balmer et al., 2011; Pulster et al., 2009) may provide an additional tool towards differentiating bottlenose dolphin stocks along the Georgia coast.

Photo-identification effort within the SGA could only identify the ranging patterns of dolphins within the defined study area‟s geographic boundaries.

Satellite-linked and radio telemetry data were utilized to augment the photo- identification data and investigate potential ranging patterns outside of these boundaries. The majority of tagged dolphins with high site-fidelity remained within the boundaries of their respective field site, suggesting that these individuals are likely estuarine residents with localized ranging patterns largely exclusive to the Brunswick or Sapelo field sites. However, most tagged individuals with moderate and low site-fidelity had ranging patterns that extended outside of their respective field site boundaries. These patterns affected their site-fidelity classification for a given field site. For example, three tagged individuals (Z16, Z24, and Z26) had ranging patterns that included both field

75 sites, but each was identified as having a different site-fidelity classification

(moderate or low) within each field site (Table 4). Based upon photo- identification results exclusively within the Brunswick or Sapelo field site, these individuals may have been considered visitors to the SGA. However, photo- identification effort in both field sites and the telemetry data, which extended outside of the SGA, suggest that these individuals are estuarine residents with ranging patterns that simply do not fit the defined field site boundaries. Thus, the use of geographic boundaries to differentiate estuarine stocks is useful, however, additional parameters, such as seasonality and habitat use, should be incorporated into future stock assessments.

Satellite-linked telemetry enhanced the radio tracking data in this study by providing additional locations of animals that were not possible via vessel-based radio tracking. For example, based upon radio tracking alone, Z04‟s ranging pattern included the estuarine waters of both field sites (Fig. 10). Satellite-linked telemetry, though, identified that this individual ranged into the coastal as well as estuarine waters of both field sites. Additional surveys that include the coastal waters surrounding the SGA would be useful in determining ranging patterns of dolphins in this region.

Based upon the lowest Akaike's Information Criterion, the best fit robust population model for abundance estimates differed between the Brunswick and

Sapelo field sites (Table 2). The photo-identification and telemetry data provide insight into why two different robust models were selected for these adjacent field sites. Seasonal fluctuations in abundance and individual ranging patterns of

76 dolphins differed between the Brunswick and Sapelo field sites. The majority of tagged dolphins sighted in the Brunswick field site had ranging patterns that extended outside of the field site‟s boundaries (Table 5), and seasonal abundance fluctuations that were similar across years (Fig. 4). In contrast, the majority of tagged dolphins sighted in the Sapelo field site had ranging patterns within this field site‟s boundaries and seasonal abundance fluctuations that were not similar between years. The two robust models utilized in this study apply different probability variables to immigration and emigration rates. Thus, the difference in abundance fluctuations and individual ranging patterns between field sites may be contributing factors to the best fit of the different models utilized to estimates abundance in the Brunswick and Sapelo field sites.

Collaborations between other researchers that are utilizing CR photo- identification methods are necessary to identify and standardize the best fit models for a region and to allow for comparisons of abundance between regions for the purposes of stock management.

Management implications

The Marine Mammal Protection Act (MMPA) was established in response to the threat of depletion or extinction of marine mammal species by human activities (reviewed in Coulston, 1990). Bottlenose dolphins within southern

Georgia estuaries are of particular concern due to the high levels of anthropogenic contaminants in their blubber. There are numerous factors that may be contributing to the differences in abundance and habitat use between

77 these two field sites, including the high levels of contaminants in this region.

Thus, management agencies should focus on additional research projects, such as determining contaminant levels of dolphin prey species and investigating dolphin survivorship across field sites.

Currently, NOAA has divided Western North Atlantic bottlenose dolphins into coastal and estuarine stocks, based upon geographic boundaries. This study identified two different groups of dolphins in southern Georgia estuaries; apparent residents that utilized all tributaries throughout the estuaries, and visitors that were sighted primarily in larger tributaries during summer and/or fall.

Although there was limited overlap of dolphin ranging patterns between the two field sites, individual dolphins did range across both field sites as well as outside of the field sites‟ boundaries. These results suggest that additional parameters such as seasonality and habitat use patterns should be utilized, in addition to the current geographic boundaries to differentiate estuarine and coastal stocks. This approach is currently in place for the Northern and Southern Migratory Coastal

Stocks of bottlenose dolphins, which are delineated by geographic boundaries that shift across seasons (Waring et al., 2009). Within the estuaries of southern

Georgia, the hypothesized overlap of estuarine and coastal stocks during summer and fall could be indicators for which stock(s) to target for a specific management plan.

The high levels of contaminants in southern Georgia estuaries likely have different effects on the two groups of bottlenose dolphins identified in this region; estuarine residents and visitors. Differences in contaminant profiles between

78 these two groups may also provide insight into stock identification. The distribution of contaminants from the TBRE may be greater than originally anticipated, depending on the ranging patterns of visitor dolphins outside of southern Georgia estuaries. Future research, including biopsy sampling of visitor dolphins, and extended photo-identification surveys along the coast, are necessary to address this hypothesis further.

CHAPTER THREE: DEFINING RESIDENCY PATTERNS FOR COMMON BOTTLENOSE DOLPHINS (TURSIOPS TRUNCATUS) WITHIN THE ESTUARIES OF SOUTHERN GEORGIA, USA

ABSTRACT

Marine mammal stocks are defined to ensure that appropriate conservation efforts are directed towards the correct management unit. Along the U.S. Atlantic and Gulf of Mexico coasts, bottlenose dolphins (Tursiops truncatus Montagu 1821) have been classified into multiple, separate estuarine stocks. Several of these defined stocks have been the focus of long-term photo- identification studies, which permit accurate abundance estimates for dolphins within a given study area utilizing capture-recapture (CR) modeling. However, often the geographic boundaries of these stocks are not well understood as stocks overlap and individuals cross stock boundaries. Thus, it can be unclear which stock(s) are included within these abundance estimates. In addition, for the majority of currently identified estuarine stocks, few baseline data yet exist on dolphins within the stock‟s defined geographic boundaries. The goals of this study were to define and utilize dolphin residency patterns and variations in habitat use to differentiate between dolphins whose ranging patterns overlap within the estuaries of southern Georgia. Three residency patterns were defined based individual dolphin sighting histories collected during all photo-identification effort in the region (2004-2009). Dolphins were identified as either transients

(sighted in only one season for one year), seasonal visitors (sighted in one or two seasons for greater than one year), or residents (sighted in three or four seasons for greater than one year). Strahler Stream Order (SSO) was utilized to quantitatively classify all tributaries within the study area. For each residency pattern, the total number of sightings within each SSO was determined to identify

81 differences in habitat use among residency patterns. Residents (n = 188) were primarily sighted across all seasons and in all SSO tributaries. These individuals are likely members of the estuarine stock(s) for this region. Seasonal visitors (n

= 129) and transients (n = 272) were sighted predominantly in summer or seasons adjacent to summer and in larger SSO tributaries. The majority of these individuals are likely members of the South Carolina/Georgia Coastal Stock that enter the larger estuarine tributaries as they travel along the coast. The results of this study may also lend insight into the appropriate survey effort required to assist with future stock assessments in other estuaries along the U.S. Atlantic and Gulf of Mexico coasts. A minimum of two years of seasonal CR photo- identification surveys and data on SSO distribution were required to group individuals into the residency patterns defined in this study. This survey methodology initially requires an intensive amount of focused survey effort within a given region. However, after these residency patterns have been determined, we suggest that CR photo-identification surveys can be performed on a yearly or multi-year basis, and effort targeted during the appropriate season and SSO, to determine abundance for a targeted stock.

INTRODUCTION

The Marine Mammal Protection Act of 1972 defines a stock as “a group of marine mammals of the same species in a common spatial arrangement that interbreed when mature” (reviewed in Coulston, 1990). Stocks are defined to ensure that any required conservation effort to mitigate a particular human activity is aimed at the correct management unit (Waring et al., 2009). The

National Oceanic and Atmospheric Administration (NOAA) has classified bottlenose dolphins (Tursiops truncatus Montagu 1821) into distinct coastal and estuarine stocks throughout the Western North Atlantic and Gulf of Mexico based in part upon photo-identification, telemetry, and genetic studies that have been carried out at multiple field sites along the southeastern U.S. (reviewed in Waring et al., 2009). However, the geographic delineations of these stocks are often not well understood as stocks overlap and individuals move across stock boundaries

(e.g. Read et al., 2003).

Estimating the number of individuals belonging to a stock is a necessary step for establishing effective management plans (Taylor and Gerrodette, 1993).

Within a number of inshore and estuarine regions along the U.S. Atlantic and

Gulf of Mexico coasts, bottlenose dolphin abundance has been determined utilizing photo-identification surveys and capture-recapture (CR) models (e.g.

Balmer et al., 2008; Read et al., 2003; Speakman et al., 2010). Abundance estimates can vary within a geographic region, though, dependent upon the spatial and temporal pattern of stock overlap. For example, Read et al. (2003) utilized photo-identification CR surveys to determine abundance of bottlenose

83 dolphins in the inshore waters of North Carolina during summer. This study provided the first abundance estimates for dolphins within this region, information critical to managers. The authors acknowledged, though, that the stock identities of dolphins represented within this sample were unknown. Currently, NOAA defines two migratory coastal stocks, one resident coastal stock, and two estuarine stocks, which overlap the geographic boundaries of this study‟s survey area (Waring et al., 2009). Thus, it was unclear which stock(s) were included within this abundance estimate.

Previous investigators have defined a number of dolphin residency patterns to differentiate between individuals whose ranging patterns overlap within a given geographic region. Zolman (2002) defined residency patterns of bottlenose dolphins in the Charleston, SC area, based upon seasonal presence of individuals sighted during a consecutive 15 month (October 1994 – January

1996) photo-identification study. These patterns were (1) year-round residents, which were sighted during all four seasons, regardless of year, (2) seasonal residents, which were sighted in the same season for consecutive years, and (3) transients, which were sighted in only one season or two consecutive seasons, within one year. Speakman et al. (2006) utilized these residency patterns, and additional photo-identification surveys conducted during 1997-2003, to classify dolphins as members of different stocks, whose ranges overlapped the waters surrounding Charleston, SC. Dolphins classified as year-round residents were considered to be members of the Charleston Estuary System (CES) Stock.

Seasonal residents and transients were hypothesized to be members of the

South Carolina/Georgia Coastal Stock, which entered into the defined geographic boundaries of the CES Stock typically during summer. The Mid-

Atlantic Bottlenose Dolphin Catalog (MABDC), which includes images and associated sighting data from multiple photo-identification studies along the southeastern U.S. (Urian et al., 1999), was subsequently utilized to determine whether individual dolphins extended their ranges outside of the Charleston, SC survey area (Speakman et al., 2010). Seven individuals classified as transients by Speakman et al. (2010) had ranges that extended as far south as

Jacksonville, FL and as far north as Wilmington, NC (K. Urian, pers. comm.), suggesting that these individuals may be members of adjacent coastal stocks that enter the boundaries of the CES Stock as they travel along the coast

(Speakman et al., 2010). Thus, the residency patterns identified by Zolman

(2002) provided insight into abundance of the CES Stock, in addition to abundance of other stocks such as the South Carolina/Georgia Coastal Stock that may be utilizing the waters within the Charleston, SC region (Speakman et al., 2010).

Along Florida‟s northern Gulf of Mexico coast, Balmer et al. (2008) utilized photo-identification surveys and telemetry data during 17 non-consecutive months (April 2004 – July 2007) to identify two groups of dolphins, residents and seasonal visitors, in the St. Joseph Bay region. Each identified dolphin in this study was grouped into a statistically defined sighting bin (Freedman and

Diaconis, 1981) based upon the total number of times they were sighted, which was used to define a relative site-fidelity index. During spring and fall, the St.

Joseph Bay region experienced an influx of dolphins, measured as an increase in abundance using CR models. The majority of dolphins sighted during these seasons were individuals classified in the lowest site-fidelity bin. A subset of dolphins that were tagged during these seasons also displayed ranges that extended outside of the St. Joseph Bay region. The combined results of the abundance estimates, site-fidelity, and telemetry data suggested that these individuals were seasonal visitors to St. Joseph Bay region and were hypothesized to be members of the Northern Gulf of Mexico Coastal Stock.

During winter and summer, abundance was lowest, the majority of dolphins sighted were within the higher site-fidelity bins, and individuals tagged in these seasons had localized movements within the St. Joseph Bay region. These individuals were considered to be residents, and members of the St. Joseph Bay

Estuarine Stock (Balmer et al., 2008).

These studies point out the difficulty of obtaining accurate stock abundance estimates, if ranging patterns of dolphins overlap within a given geographic region. Robust CR models (Pollock, 1982), though, can be utilized to account for overlapping ranges (i.e. variation in immigration and emigration rates), if these patterns are well understood. These models can generate accurate abundance estimates for a region, at any given point and time, but they cannot assign stock identification to any individual dolphin. Another long standing problem in the use of CR models to estimate abundance are transient individuals that may be travelling through a study area during a given survey period but are not resighted in subsequent survey periods (Hines et al., 2003).

The occurrence of transients in a study area results in inflated abundance estimates that may not be representative of the targeted stock (Pradel et al.,

1997). Transients are operationally defined in CR modeling as individuals that have a zero probability of survival after their initial capture (Hines et al., 2003;

Pradel et al., 1997). However, differentiating transients from individuals with other residency patterns is extremely difficult. Transients, for example, may be residents that were sighted only during the final capture session, or residents that died following the first capture session. Alternatively, transients may be seasonal residents or visitors, which return to the study area in subsequent years, but which may not have been resighted because of inadequate survey effort in subsequent survey periods.

Recently, Balmer et al. (in review) identified an additional parameter that could be used to differentiate between different dolphin residency patterns within the estuaries of southern Georgia, that of habitat use. Each identified dolphin in this study was grouped into a statistically defined sighting bin (methods of Balmer et al. 2008), based upon the total number of times they were sighted during a non-consecutive 19 month (December 2004 – November 2009) photo- identification study. Strahler Stream Order (SSO), a quantitative technique used to define habitat based upon the number of upstream tributaries (Strahler, 1952), was utilized to classify all tributaries within the study area. Individual dolphins were grouped by their site-fidelity classification and all individuals‟ sightings were mapped to determine their respective SSO. These data were used, in addition to seasonal abundance estimates calculated utilizing CR models and individual

87 ranging patterns based upon telemetry data, to provide insight into different dolphin residency patterns within the estuaries of southern Georgia. Dolphin abundance varied across seasons, with the highest number of individuals observed during summer and fall. During these seasons, the majority of dolphins sighted had low site-fidelity and were within larger SSO tributaries- sounds opening into the coastal environment. Balmer et al. (in review) hypothesized that these dolphins were members of the South Carolina/Georgia Coastal Stock.

During seasons with lower abundance, the majority of dolphins sighted had moderate or high site-fidelity, and were sighted across all SSO tributaries, suggesting that these individuals were estuarine residents, utilizing more fully the estuarine habitat.

The site-fidelity classifications utilized by Balmer et al. (in review) were aimed at comparing dolphins across two adjacent field sites within the estuaries of southern Georgia. In addition to identifying bottlenose dolphin stock structure within this region, the goals of this study were to determine the abundance and distribution of dolphins that were affected by high levels of contaminants in a highly polluted site, verses an adjacent, comparison site. Thus, the design of this study did not utilize residency patterns as defined by others.

The goals of this current study were to expand on the work of Balmer et al.

(in review) to identify residency patterns of dolphins within the estuaries of southern Georgia. Residency patterns defined in this study are similar to, but more conservative than those of Zolman (2002), due to in large part access to longer term data sets. In Charleston, SC, Zolman (2002) defined residency

88 patterns based upon photo-identification surveys conducted during 15 consecutive months over a 1.25 year period. This study will define residency patterns based upon photo-identification surveys conducted during 19 non- consecutive months over a 6 year period. Thus, the more conservative residency pattern definitions in this study will be based upon multi-year photo identification surveys. For each residency pattern, the total number of sightings within each SSO tributary was determined to identify differences in habitat use among residency patterns. These results will provide insight into the different stocks of dolphins within the estuaries of southern Georgia based upon both temporal and spatial data. These results may also be useful in identifying appropriate survey effort required to assist with future stock assessments in other estuaries along the U.S. Atlantic and Gulf of Mexico coasts.

METHODS

The southern Georgia photo-identification study area (SGA), defined by

Balmer et al. (2011), represents approximately 60 km of north-south estuarine shoreline (Fig.1). The city of Brunswick, GA is located in the southwest corner of the SGA. The study area‟s eastern boundaries were defined as the mouths of

Sapelo, Doboy, Altamaha, and St. Simons Sounds. The western boundaries were defined as 15 km upriver of the Sapelo, Altamaha, and Turtle Rivers.

The Southern Georgia Bottlenose Dolphin Photo-identification Catalog was created in 2004 and includes images and associated sighting data of individual dolphins that were identified during remote biopsy, capture-recapture, and radio

89 telemetry surveys carried out by NOAA and other researchers from 2004-2009 in the SGA (Balmer et al., 2011). The methodologies for these survey techniques have been discussed in detail in other studies (e.g. Balmer et al., 2008; Litz et al.,

2007; Speakman et al., 2010). Currently, the SGA photo-identification catalog consists of 646 individual bottlenose dolphins (Balmer et al., 2011).

Balmer et al. (in review) grouped each cataloged individual into a statistically defined site-fidelity classification based upon the methodology of

Balmer et al. (2008). Three site-fidelity classifications were identified- low

(individuals with 1-6 sightings), moderate (7-12 sightings), and high (13-18 sightings).

Three residency patterns were defined based upon the total number of years and seasons that individual dolphins were sighted- transients, seasonal visitors, and residents. Transients were sighted in only one season for one year, seasonal visitors were sighted in one or two seasons for greater than one year, and residents were sighted in three or four seasons for greater than one year.

Each individual was then classified into one of these three defined residency patterns.

SGA cataloged individuals were grouped by the total number of years that each individual was sighted. Individuals sighted in only one year were then grouped by the number of seasons that each individual was sighted. Individuals sighted in only one season, which were identified as SGA transients, were then separated into the specific season that each was sighted. The sightings for all

Figure 1. Southern Georgia photo-identification study area (SGA), defined by Balmer et al. (2011).

SGA transients were plotted in ArcMap 9.2 (ESRI, Redlands, CA) and the total number of sightings within each SSO tributary was calculated.

The above steps were then repeated for dolphins that were sighted across multiple years and in one or two seasons, which were identified as SGA seasonal visitors, and in three or four seasons, which were identified as SGA residents.

To illustrate differences in habitat use between residency patterns, all sightings for each pattern were plotted in ArcMap 9.2. If a sighting occurred anywhere along a given SSO transect, that entire SSO tributary would be shaded in (see Balmer et al. (in review) for a detailed methodology for transect surveying in the SGA). This technique permits a visual interpretation of habitat use differences among residency patterns. However, it is an overestimation of habitat use, because individuals within a given residency pattern may not have been sighted throughout the entire shaded SSO tributary.

RESULTS

Although not a surprising result, dolphins with low site-fidelity (i.e. individuals sighted only 1-6 times throughout all photo-identification effort) were sighted more frequently in just one year than individuals with moderate or high site-fidelity (Fig. 2). Transients (n = 272) (sighted in only one season for one year) were exclusively individuals with low site-fidelity (Fig. 3a). The majority of transients were sighted during summer (Fig. 4a). In all seasons, transients were primarily sighted within SSO 4 tributaries (Fig. 5a and 6a).

Figure 2. Frequency sighted of individuals grouped by site-fidelity classification and total number of year sighted. Note: Frequency sighted is the percentage of individuals within each site-fidelity classification that were sighted during the given number of year(s).

Figure 3. Frequency sighted of individuals grouped by site-fidelity classification and total numbers of seasons sighted in (a) 1 year and (b) greater than 1 year. Note: Frequency sighted is the percentage of individuals within each site-fidelity classification that were sighted during the given number of season(s).

Figure 4. Frequency sighted of individuals grouped by site-fidelity classification and (a) in only 1 year and 1 season, (b) greater than 1 year and 1 or 2 seasons, and (c) greater than 1 year and 3 or 4 seasons. Note: Frequency sighted is the percentage of individuals within each site-fidelity classification that were sighted during the given season(s). 95

Figure 5.- Frequency sighted of (a) transients grouped by season, (b) seasonal visitors grouped by site-fidelity classification, and (c) residents grouped by site- fidelity classification in different Strahler Stream Order (SSO) tributaries. Note (Fig. 5a): Frequency sighted is the percentage of individual sightings within each season that were sighted within each Strahler Stream Order (SSO) classification. Note (Fig. 5b and 5c): Frequency sighted is the percentage of individual sightings within each site-fidelity classification that were sighted within each Strahler Stream Order (SSO) classification.

Figure 6. Highlighted subareas for sightings of dolphins considered to be (a) transients, (b) seasonal visitors, and (c) residents of the SGA.

Seasonal visitors (n = 129) (sighted in one or two seasons for greater than one year) were primarily individuals with low site-fidelity (Fig. 3b). The majority of seasonal visitors were sighted during summer or seasons adjacent to summer

(spring or fall) (Fig. 4b). Seasonal visitors with low site-fidelity were primarily sighted within SSO 4 tributaries, while those with moderate site-fidelity were primarily sighted within SSO 3 tributaries (Fig. 5b and Fig. 6b).

Residents (n = 188) (sighted in three or four seasons for greater than one year) were primarily individuals with moderate and high site-fidelity (Fig. 3b). The majority of residents were sighted across all seasons (Fig. 4c). Residents with low site-fidelity were sighted similarly across SSO 2, 3, and 4 tributaries (Fig. 5c and Fig. 6c). Residents with moderate and high site-fidelity were sighted predominantly in SSO 2, followed by SSO 3 and SSO 4 tributaries.

DISCUSSION

Balmer et al. (in review) utilized capture-recapture (CR) modeling, total number of individual sightings, and geographic location of sightings to determine seasonal abundance, site-fidelity, and habitat use of bottlenose dolphins in the southern Georgia photo-identification study area (SGA). This study, which defined residency patterns based upon the total number of years and seasons that individuals were sighted, provides additional insight into the different dolphin groups that may be utilizing the SGA.

This study identified 188 photographically distinct individual dolphins as residents to the approximate 60 km of estuarine coastline within the SGA (Fig.

4c). This number is similar to those for bottlenose dolphin communities that have been studied in other regions, which tend to include between 100 and 200 individuals (Wells 1991, Williams et al. 1993, Wilson et al. 1999, Balmer et al.

2008). It is important to note, though, that this estimate excludes non-distinctive individuals, which could not be identified utilizing photo-identification. However, these results suggest that the distinctive individuals identified as SGA residents are likely members of the estuarine stock(s) for this region.

Balmer et al. (in review) determined that the highest number of individuals observed in the SGA occurred during summer and fall, when the majority of dolphins sighted had low site-fidelity and were found within larger SSO tributaries. Transients and most seasonal visitors were sighted predominantly in the summer or seasons adjacent to summer (Fig. 4a and 4b, respectively), and within larger SSO tributaries (Fig. 5a and 5b, respectively). These results suggest that the majority of transients and seasonal visitors are likely members of the South Carolina/Georgia Coastal Stock that enter the larger tributaries (i.e. sounds) of the SGA. Seasonal visitors are utilizing the SGA on a multi-year basis and may be entering the larger tributaries of the SGA as they travel along the coast. Transients are not resighted in subsequent years, thus, it is unclear if these individuals are returning to the SGA. One hypothesis to why these individuals are not resighted is bias in photo-identification coverage. During periods of high abundance, complete photo-identification coverage of all individuals in a sighting can be extremely difficult. Transients may be utilizing the

SGA on a multi-year, seasonal basis and, thus, would be defined as seasonal

99 visitors, but these individuals are missed due to incomplete photo-identification coverage.

Alternatively, transients may be residents that became newly distinctive during the final capture session or died following the first capture session. For example, transients sighted in the SGA during winter (n = 20) (Fig. 4a), when the majority of dolphins have moderate or high site-fidelity, are likely residents that are newly distinctive. Transients may also be members of adjacent estuarine stocks, and have ranging patterns that only occasionally overlap the boundaries of the SGA and therefore could be missed during any given survey period.

Extending photo-identification surveys along the coast and within adjacent estuaries may provide insight into the residency patterns for these individuals.

Transients could also be individuals that fit the definition by Hines et al. (2003), which are dolphins sighted during one season and never resighted again. These individuals could be members of adjacent coastal or estuarine stocks that do not return to the SGA. Photo-identification searches of all SGA transients in the Mid-

Atlantic Bottlenose Dolphin Catalog (MABDC) (Urian et al. 1999) would be useful in identifying extended ranging patterns of these individuals.

A small number of individuals (n = 37), which were sighted in only one year and in greater than one season, did not fit into any of the residency patterns defined in this study (Fig. 3a). These individuals are likely seasonal visitors or residents that have not yet been sighted in subsequent years. Continued photo- identification surveys within the SGA would likely provide insight into the appropriate residency patterns for these individuals.

100

Estuarine stocks of bottlenose dolphins in the Gulf of Mexico and Western

North Atlantic are defined by geographic boundaries surrounding individual dolphin communities (Waring et al. 2009). A community is defined as a group of resident animals that share home ranges, display similar genetic features, and interact more frequently with each other than with dolphins in adjacent waters

(Wells et al. 1987). Currently, NOAA has defined 33 Gulf of Mexico estuarine stocks and nine Western North Atlantic estuarine stocks (Waring et al. 2009).

There are numerous other estuaries in both basins that have not yet been included within the boundaries of any stock identification. Several of these defined stocks have been the focus of intensive, multi-year research projects

(e.g. Charleston Estuarine System, Indian River Lagoon Estuarine Stock,

Sarasota Bay Estuarine Stock, and St. Joseph Bay Estuarine Stock), which permit accurate abundance estimates and stock assessments (e.g. Balmer et al.

2008, Speakman et al. 2010). Currently, though, few such baseline data exist for the majority of estuarine stocks. Although it is not logistically possible to perform long-term research projects for all estuarine stocks, the development and identification of a survey methodology that efficiently gathers baseline data on abundance and residency patterns is necessary.

The results of this study suggest such a methodology to estimate the abundance of estuarine stocks. Line transect methods, by vessel or plane, have been used to estimate abundance of offshore cetaceans (e.g. Hammond 1986,

Buckland et al. 1993). However, the use of these survey techniques to determine inshore cetacean abundance are more difficult due to more complex

101 coastal topography (Wells et al. 1980, Wilson et al. 1997). These difficulties are made clear when one considers that aerial line transect surveys yielded a stock abundance estimate of zero for Saint Joseph Bay, Florida (Waring et al. 2000).

CR photo-identification surveys conducted within the same region identified seasonal fluctuations in abundance with a threefold increase during spring and fall (313-410) and lowest abundance during winter and summer (78-152) (Balmer et al. 2008). Thus, CR photo-identification surveys are likely the appropriate survey methodology for determining estuarine stock abundance.

Spatial patterns of dolphin sightings are also useful towards identifying residency patterns for a targeted estuary. In this study, SGA transients were sighted primarily within the larger tributaries, while residents were sighted similarly across all sized tributaries. Thus, CR photo-identification surveys within all SSO tributaries would provide insight into the different groups of dolphins utilizing a given estuary.

CR photo-identification surveys across seasons would provide insight into the different groups of dolphins within a targeted estuary. Seasonal fluctuations in dolphin abundance have been identified within the SGA (Balmer et al. in review), and in other field sites along the Gulf of Mexico and Western North

Atlantic (e.g. Balmer et al. 2008, Speakman et al. 2010). For example, CR photo-identification surveys performed during winter in SGA, when there are fewer seasonal visitors and transients (Fig. 4), would provide the appropriate abundance estimates for SGA residents that are likely members of the estuarine stock(s) for this region. CR photo-identification surveys during summer and

102 seasons adjacent to summer would generate abundance estimates that include residents and individuals with residency patterns suggestive of the South

Carolina/Georgia Coastal Stock (seasonal visitors and transients). The season(s) in which these fluctuations occur may vary among field sites, thus initial sampling across seasons would be required.

CR photo-identification surveys across years would assist in determining residency patterns with more confidence than one year and identify seasonal changes in abundance that may vary across years. In this study, two years

(2008-2009) of seasonal CR photo-identification surveys were necessary to identify residency patterns of dolphins within the SGA. Initially, this survey methodology requires an intensive amount of survey effort within a given estuary.

However, after residency patterns have been identified, CR photo-identification surveys can be performed on a yearly or multi-year basis, during the appropriate season, to determine abundance for a targeted estuary.

After baseline data on abundance estimates and residency patterns for a given estuary have been identified, there are several other tools that could be implemented to provide additional insight into stock structure. Satellite-linked and radio telemetry targeting individuals during a specific season and tributary

(SSO) size could identify ranging patterns of individuals with different residency patterns. For example, along Florida‟s northern Gulf of Mexico coast, Balmer et al. (2008) radio-tagged dolphins during the spring, when there was an influx of animals into the St. Joseph Bay region. Based upon the radio telemetry and photo-identification data, individuals were classified into one of two groups;

103 visitors that had extended ranging patterns along the coast, and residents that remained within the St. Joseph Bay region. Intensive investigation of stranding patterns within the region could also provide insight into stock structure. For example, McLellan et al. (2002) utilized 25 years of temporal stranding data to suggest that coastal bottlenose dolphin stock structure in the Western North

Atlantic was more complex than originally proposed. Periodic CR photo- identification surveys of sites adjacent to a given estuary would be useful to identify overlap of individuals from other estuaries.

104

CHAPTER FOUR: EVALUATION OF A SINGLE-PIN, SATELLITE-LINKED TRANSMITTER DEPLOYED ON BOTTLENOSE DOLPHINS (TURSIOPS TRUNCATUS) ALONG THE COAST OF GEORGIA, USA

105

ABSTRACT

Satellite-linked telemetry is a useful tool for determining small cetacean movement patterns. While transmitter design and success rates have varied, problems associated with transmitter size, attachment location on the dorsal fin, and the number of attachment pins have shortened the expected attachment duration or caused negative impacts to the dorsal fins of individual animals. The goals of this study were to develop a new satellite-linked transmitter attachment design that would minimize negative impacts to the dorsal fin while maximizing transmitter longevity. Three adult male bottlenose dolphins (Tusiops truncatus) received prototype satellite-linked transmitters, Kiwisat 202 Cetacean Fin Tags, during a health assessment conducted in August 2009 within the estuaries of southern Georgia. This transmitter design had a mass in air of 37g and was attached to the trailing edge of the dorsal fin via a modified plastic housing and two semi-rigid, plastic flanges. The tag was secured to the fin with a single Delrin pin with each end threaded for non-stainless steel lock nuts. In addition to the prototype satellite-linked transmitters, each of the three dolphins also received a radio transmitter that permitted routine follow-up observations of both animal and tag condition using vessel-based surveys. The 43% weight reduction from earlier side-mount tags, reduction in the number of attachment pins, and re-positioning of the tag attachment from the middle-upper third to the lower trailing edge of the dorsal fin reduced damage to major venous regions in the dorsal fin and minimized long-term effects to the tagged individuals. These satellite-linked transmitters provided location data for a mean of 62 + 9 S.D. days, which is

106 comparable to other previous satellite-linked tag transmission durations for small cetaceans. These results suggest that the new satellite-linked transmitter attachment design was a significant improvement in tagging small cetaceans over the previous multi-pin, side-mount designs.

107

INTRODUCTION

Electronic tags, which use the ARGOS system (CLS, 2008), have proven to be valuable tools in assessing small cetacean movement patterns and habitat use (e.g. Balmer et al., 2008; Corkeron and Martin, 2004; Klatsky et al., 2007;

Read and Westgate, 1997; Wells et al., 2008; Wells et al., 2009). While tag design and success rates have varied, problems associated with package size, attachment position on the dorsal fin, and number of attachment pins have, in some cases, shortened the predicted attachment duration or caused adverse impacts to the dorsal fins of the animals (Balmer et al., 2010; Read and

Westgate, 1997; Scott et al., 1990b). One of the most recent iterations in the evolution of smaller satellite-linked tags for dolphins used a 65g, side-mounted tag, which attached to the upper third of the dorsal fin with three plastic pins

(Balmer et al., 2010). This tag design has been used to determine movement patterns and dive durations in several species of small cetacean including: bottlenose dolphins (Tursiops truncatus) off Bermuda (Klatsky et al., 2007),

Risso‟s dolphins (Grampus griseus) in the Gulf of Mexico and Atlantic Ocean

(Wells et al., 2009; R. Wells, pers. obs.), rough-toothed dolphins (Steno bredanensis) in the Atlantic Ocean (Wells et al., 2008; R. Wells, pers. obs.), and

Franciscana dolphins (Pontoporia blainvillei) in the Atlantic coastal waters off

Argentina (R. Wells, pers. obs.). While this tag design appears to be relatively robust, a recent deployment of the tag demonstrated that the design is not well suited for coastal bottlenose dolphins (Balmer et al., 2010). It is generally believed that most dorsal fin packages release after galvanic corrosion of the

108 attachment fasteners allow the attachment pins to fall out, resulting in minimal dorsal fin damage. However, there have been few studies that have provided detailed follow-up monitoring from tag attachment to tag failure. In this case,

Balmer et al. (2010) identified that one of the attachment pins sheared prior to fastener corrosion, resulting in tag migration and damage to the dorsal fin as well as changes in the animal‟s dive behavior. These results motivated the development of a new satellite-linked tag attachment design that would minimize negative impacts to the dorsal fin while maximizing transmitter longevity.

A new prototype satellite-linked transmitter (Kiwisat 202 Cetacean Fin Tag model K2F161), was developed by SirTrack (Havelock North, NZ) (Figure 1). It had a mass in air of 37g and was attached to the trailing edge of the dorsal fin via a modified plastic housing and two semi-rigid, plastic flanges. The tag was secured to the fin with a single, 0.64 cm (1/4”) Delrin pin with each end threaded for M6 x 1 (1/4”-20) non-stainless steel (corrodible) lock nuts. The nylon ring inside the steel lock-nuts was scored to facilitate pin slide out (and tag loss) once the steel had rusted away. The nuts were tightened to a level in which there was a 1-2 mm gap between each plastic flange and the dorsal fin surface. This allowed for limited tag movement and reduced risk of pressure necrosis. The conventional single-pin VHF radio bullet tags (Trac Pac, Ft. Walton Beach, FL) that have been used to successfully monitor individual dolphins (e.g. Balmer et al., 2008). However, the overall size and weight of the satellite-linked transmitter was larger than that of the radio bullet tag (37g, 75 mm x 20 mm x 25 mm; 16g,

109

Figure 1. SirTrack Kiwisat 202 Cetacean Fin Tag (model K2F161) (Havelock

North, NZ)

110

52mm x 15 mm x 15 mm; respectively). Thus, it was unclear the effects that this larger, single-pin transmitter would have on the dorsal fin.

The goals of this new satellite-linked transmitter design aimed to minimize detrimental effects on tagged individuals while maximizing satellite-linked transmission duration. Specifically, we wanted a tag that would provide daily locations of the dolphins for periods up to 50 days and would either detach from the animal via fastener corrosion or pin breakage. In the event that tag migration occurred, it was reasoned that the attachment location, within 3-4 cm of the trailing edge of the dorsal fin, would minimize potential tissue damage. It is worth noting that many species of small odontocetes and bottlenose dolphins in particular, frequently bear natural notches along the trailing edges of their dorsal fins (reviewed in Scott et al., 1990b) and can therefore tolerate minor tissue damage in this region of the dorsal fin. This study assessed the efficacy of the

Kiwisat 202 Cetacean Fin Tag, and its impact on instrumented dorsal fins.

METHODS

In August 2009, the National Oceanic and Atmospheric Administration

(NOAA), with multiple partner organizations and agencies, conducted ten days of bottlenose dolphin health assessments at two sites along the coast of Georgia; the Turtle/Brunswick River Estuary (TBRE) near the city of Brunswick, Georgia, and the waters surrounding the Sapelo Island National Estuarine Research

Reserve (SINERR) near the city of Darien, Georgia (Figure 2). These sites were chosen to examine potential health impacts on bottlenose dolphins from elevated

111

Figure 2. Turtle/Brunswick River Estuary (TBRE) and Sapelo Island National

Estuarine Research Reserve (SINERR) field sites.

112 environmental contaminant exposures (Kannan et al., 1997; Kannan et al., 1998;

Maruya and Lee, 1998; Maruya et al., 2001; Pulster et al., 2005) and identify fine- scale movement patterns of individual dolphins in these polluted waters. Dolphins were captured, sampled, examined, tagged, and released using practices similar to those pioneered for health assessment and monitoring of bottlenose dolphins in Sarasota Bay, Florida (Wells et al., 2004). Each individual was freeze-branded on the dorsal fin with a letter (“Z”) and two digit number (“01, 02, 03,” etc.). Even numbers were given to males and odd numbers to females. A tooth sample was also collected from individual dolphins to estimate dolphin age (Hohn et al.,

1989).

To determine movement patterns in this region, 28 bottlenose dolphins were tagged with conventional VHF radio transmitters (MM130, Backmount

Transmitter, Advanced Telemetry Systems, Inc., Isanti, MN) using a single-pin attachment along the trailing edge of the dorsal fin (Balmer et al., 2008). Three adult male dolphins (Z04, Z08, and Z22) also received prototype satellite-linked transmitters, Kiwisat 202 Cetacean Fin Tags (Figures 3a, 4a, and 5a). The VHF radio transmitters were attached to the upper third of the dorsal fin to maximize signal strength for the follow-up radio tracking portion of the study. The satellite- linked transmitters were attached to the lower third of the dorsal fin to optimize tag stability and attachment duration. The Kiwisat 202 Cetacean Fin Tags provided location only data for a defined duty cycle (6 h/d) for the life of the tags.

Transmitter battery life for this duty cycle was estimated by SirTrack to be at least

50 days and supplemented by a saltwater switch which turned the transmitter off

113

Figure 3. (a) Z04 with VHF radio transmitter (Top) and Kiwisat 202 Cetacean Fin Tag (Bottom), (b) Z04 with slight migration and slight biogrowth of satellite- linked transmitter; day 39, and (c) Z04 without satellite-linked transmitter; day 62.

114

Figure 4. (a) Z08 with VHF radio transmitter (Top) and Kiwisat 202 Cetacean Fin Tag (Bottom), (b) Z08 with severe migration and heavy biogrowth of satellite- linked transmitter; day 64, and (c) Z08 without satellite-linked transmitter; day 77.

115

Figure 5. (a) Z22 with VHF radio transmitter (Top) and Kiwisat 202 Cetacean Fin Tag (Bottom), (b) Z22 with no migration and no biogrowth of satellite-linked transmitter; day 20, and (c) Z22 with satellite-linked transmitter; day 71.

116 when submerged to conserve battery life. In addition to testing the suitability of the new design, the satellite-linked tags provided locations of animals outside of the area defined for VHF radio tracking surveys. The VHF radio transmitters allowed routine follow-up observations of both animal and tag condition using radio telemetry during vessel surveys, while the satellite-linked transmitters provided animal locations both within and outside of the VHF radio survey area for all three individuals.

Kiwisat 202 Cetacean Fin Tags were attached along the lower third of each animal‟s dorsal fin, approximately 3.5 cm cranial to the trailing edge. Prior to tag attachment, the dorsal fin was scrubbed with chlorhexiderm and rinsed with ethanol. At the attachment position, a local anesthetic (lidocaine 2% with epinephrine) was injected with a Miltex 76-50 N-tralig injector (Miltex Inc., York,

PA) prior to drilling the pin attachment hole using a cordless power drill with a sterilized 0.6 cm thin walled brass bore (see Balmer et al., 2008; Wells et al.,

2009). Delrin pins and steel lock-nuts were used for attachment of both VHF radio and satellite-linked transmitters.

RESULTS

Z04‟s satellite-linked tag transmitted for a total of 57 days after deployment while the VHF radio tag transmitted for 65 days. This allowed for comprehensive monitoring of the condition of the satellite-linked tag from attachment to release. Slight caudal migration in the dorsal fin and minor biofouling (accumulation of algae) were first observed on the satellite-linked tag

117 beginning on day 39 (Figure 3b); the tag remained in this condition for the remainder of the life of the satellite-linked tag (through day 56). Satellite-linked location data transmissions ceased on day 57 and the animal was not resighted using VHF radio telemetry until day 62. At this time, the animal was observed without the satellite-linked transmitter but with a small, well-healed hole evident at the tag attachment location (Figure 3c). These observations suggest that either the Delrin pin sheared or a nut was lost which allowed the pin to slide out of the attachment hole. Based on the relatively good condition of the nut on the

VHF radio tag at day 62 (Figure 3c), we reason that the former is much more likely.

Z08‟s satellite-linked tag transmitted for 71 days while the VHF radio tag transmitted for 28 days. Although Z08‟s VHF radio tag failed before the satellite- linked tag, Z08‟s male associate, Z10, had a VHF radio tag that transmitted for

103 days. The close associations of these two males permitted a complete follow-up of Z08‟s satellite-linked tag. Between days 37-64, Z08‟s satellite-linked tag was observed progressively migrating out of the dorsal fin while an increasing amount of biofouling was evident on the satellite-linked tag (Figure 4b). Satellite- linked transmissions ceased on day 71 and the animal was not resighted again until day 77 by which time the satellite-linked tag was no longer present. The animal did have an open notch in the lower part of the dorsal fin (Figure 4c) suggesting that the satellite-linked tag migrated out of the dorsal fin. Z08 was resighted on day 103 and it was documented to have a fully healed dorsal fin notch at the location of the former satellite-linked tag.

118

Z22‟s satellite-linked tag transmitted for 55 days while the VHF radio tag transmitted for 20 days. Z22 was observed four times between day 1 and day 20.

On day 20, Z22‟s satellite-linked transmitter showed no migration or biofouling

(Figure 5b). Z22 was resighted on day 71, at which time the satellite-linked tag showed only slight migration and no biofouling (Figure 5c). This result suggests that battery or electronics failure caused the tag to fail prior to tag loss. Thus, the attachment design was well matched to the battery life in this configuration, but longer deployments may need to consider a stronger attachment mechanism.

Due to logistical constraints, we were unable to provide any additional follow-up monitoring of Z22‟s satellite-linked tag.

The attachment of satellite-linked and VHF radio transmitters enabled follow-up monitoring to assess the impact of the prototype Kiwisat 202 Cetacean

Fin Tags and to identify the likely causes of tag loss. All three satellite-linked tags transmitted longer than the 50 day estimate of battery life provided by

SirTrack (Table 1). The modes of failure in all three documented tags were different, with Z22‟s satellite-linked transmitter failing prior to tag loss. Z04‟s satellite-linked transmitter was probably lost due to shearing of the Delrin pin and

Z08‟s satellite-linked tag was apparently lost after it migrated caudally out of the dorsal fin. In both cases, healing at the former tag attachment site was documented; in the case of Z04 only eight days after tag loss and in Z08, 39 days post tag loss. Similar modes of tag loss have been observed in VHF radio- tagged individuals during the long-term study of bottlenose dolphins in Sarasota

119

Reason # of days # of days Reason for Dolphin Length Tagging for tag Sex Age transmitting transmitting tag failure ID (cm) Date failure (satellite) (radio) (satellite) (radio) Delrin pin Z04 M 241 16 5-Aug-09 57 65 sheering or Migration nut loss Migration Z08 M 257 27 6-Aug-09 71 28 and Migration biofouling Battery or Delrin pin Z22 M 251 32 11-Aug-09 55 20 electronic sheering or failure nut loss Mean + 61 + 9 37 + 24 S.D.

Table 1. Satellite-linked and radio tracking summaries for three dolphins tagged along the Georgia coast.

120

Bay (Scott et al., 1990b). Previously tagged Sarasota dolphins have been observed for over 20 years post tagging, suggesting no serious long-term effects.

DISCUSSION

Although our data are limited, tag loss via pin shear seems to be less traumatic to the dorsal fin tissues than pin migration, at least in terms of healing time. Despite this, the dolphin that did experience pin migration had completely healed in just over 1 month. While it would be prudent to design tags that could not migrate, given how little we understand about the physical and mechanical interactions between dorsal fins and the things we attach to them, this does not seem likely in the near future, making the present design more appealing. Future studies are necessary to evaluate the effectiveness of different diameter Delrin pins and their affect on tag attachment duration. The results from this study suggest that the single-pin attachment design of the Kiwisat 202 Cetacean Fin

Tag is a significant improvement in tagging small cetaceans over the previous multi-pin, side-mount designs (e.g. Balmer et al., 2010). The 43% weight reduction from earlier side-mount tags, reduction in the number of attachment pins from three to one, and re-positioning of the tag attachment from the middle- upper third to the lower trailing edge of the dorsal fin reduced potential damage to major venous regions in the dorsal fin and minimized long-term effects to the tagged individuals. In addition, these satellite-linked transmitters provided location data comparable to other previous satellite-linked tag transmission durations (e.g. Klatsky et al., 2007). With a more relaxed duty cycle and higher

121 repetition rate, the current battery configuration could easily be extended out past

100 days. Investigating different tag construction materials (i.e. more pliable plastics and epoxy molds) and streamline shaping of the tag would also likely contribute to improved retention times. Anti-fouling paints on the exterior surface of transmitters may be advantageous to reduce drag and improve tag retention.

However, caution must be taken to ensure that the chemicals utilized in anti- fouling paints do not have negative effects on the tagged animal. Future studies are needed to identify the optimal tag attachment location on the trailing edge of the dorsal fin to minimize drag and maximize tag retention.

The capability to provide direct observational monitoring for the life of a satellite-linked transmitter not only provides additional details of movement patterns for the individual animal, but also increases our knowledge of how to improve current tag designs (Hays et al., 2007). Future research is necessary to refine this new design and determine its success on other small cetacean species as well as bottlenose dolphins in different habitats.

122

LITERATURE CITED

Adams J, Speakman T, Zolman E, Schwacke LH. Automating image matching, cataloging, and analysis for photo-identification research. Aquatic Mammals 2006; 32: 374-384.

Aguilar A, Borrell A. Abnormally high polychlorinated biphenyl levels in striped dolphins (Stenella coeruleoalba) affected by the 1990–92 Mediterranean epizootic. Science of the Total Environment 1998; 154: 237-247.

Aguilar A, Borrell A, Pastor T. Biological factors affecting variability of persistent pollutant levels in cetaceans. Journal of Cetacean Research and Management 1999: 83-116.

Akaike H. A new look at the statistical model identification. IEEE Transactions on Automatic Control 1974; 19: 716-723.

Alberts JJ, Price MT, Kania M. Metal concentrations in tissues of Spartina alterniflora (Loisel.) and sediments of Georgia salt marshes. Estuaries, Coastal, and Shelf Science 1990; 30: 47-58.

Balmer B, Wells R, Nowacek S, Nowacek D, Schwacke L, McLellan W, et al. Seasonal abundance and distribution patterns of common bottlenose dolphins (Tursiops truncatus) near St. Joseph Bay, Florida, USA. Journal of Cetacean Research and Management 2008; 10: 157-167.

Balmer BC, Schwacke LH, Wells RS. Linking dive behavior to satellite-linked tag condition for a bottlenose dolphin (Tursiops truncatus) along Florida‟s northern Gulf of Mexico coast. Aquatic Mammals 2010; 36: 1-8.

Balmer BC, Schwacke LH, Wells RS, Adams J, George RC, Lane SM, et al. Seasonal abundance, site-fidelity, habitat use, and ranging patterns of bottlenose dolphins within the estuaries of southern Georgia, U.S.A. Journal of Wildlife Management in review.

Balmer BC, Schwacke LH, Wells RS, George RC, Hoguet J, Kucklick JR, et al. Relationship between persistent organic pollutants (POPs) and ranging patterns in common bottlenose dolphins (Tursiops truncatus) from coastal Georgia, USA. Science of the Total Environment 2011; 409: 2094-2101.

123

Barco SG, Swingle WM, McLellan WA, Harris RN, Pabst DA. Local abundance and distribution of bottlenose dolphins (Tursiops truncatus) in the nearshore waters of Virginia Beach, Virginia. Marine Mammal Science 1999; 15: 394-408.

Barros NB, Odell DK. Food habits of bottlenose dolphins in the Southeastern United States. In: Leatherwood JS, Reeves RR, editors. The bottlenose dolphin. Academic Press, San Diego, CA, 1990, pp. 309-328.

Barros NB, Wells RS. Prey and feeding patterns of resident bottlenose dolphins (Tursiops truncatus) in Sarasota Bay, Florida. Journal of Mammalogy 1998; 79: 1045-1059.

Bejder L, Samuels A, Whitehead H, Gales N, Mann J, Connor R, et al. Decline in relative abundance of bottlenose dolphins exposed to long-term disturbance. Conservation Biology 2006; 20: 1791-1798.

Berens McCabe EJ, Gannon DP, Barros NB, Wells RS. Prey selection by resident common bottlenose dolphins (Tursiops truncatus) in Sarasota Bay, Florida. Marine Biology 2010; 157: 931-942.

Burnham KP, Anderson DR. Data-based selection of an appropriate biological model: the key to modern data analysis. In: McCullough DR, Barrett RH, editors. Wildlife 2001: Populations. Elsevier Science Publishers, London, 1992, pp. 16-30.

Chalmers AG, Wiegert RG, Wolf PL. Carbon balance in a salt marsh: Interactions of diffusive export, tidal deposition and rainfall-caused erosion. Estuarine, Coastal and Shelf Science 1985; 21: 757-771.

CLS. Argos User's Manual 2007-2008. Landover, MD, 2008.

Cockcroft VG, De Kock AC, Lord DA, Ross GJB. Organochlorines in bottlenose dolphins Tursiops truncatus from the east coast of South Africa. South African Journal of Marine Science 1989; 8: 207-217.

Corkeron PJ, Martin AR. Ranging and diving behaviour of two 'offshore' bottlenose dolphins, Tursiops sp., off eastern Australia. J. Mar. Biol. Ass. U.K. 2004; 84: 465-468.

124

Coulston C. Flipper caught in the net of commerce: reauthorization of the marine mammal protection act and its effect on dolphin. Journal of Energy, Natural Resources, and Environmental Law 1990; 11: 97-125.

DeLong RL, Gilmartin WG, Simpson JG. Premature births in California sea lions: Associations with organochlorine pollutant residue levels. Science 1973; 181: 1168-1170.

Dresser BK, Kneib RT. Site fidelity and movement patterns of wild subadult red drum, Sciaenops ocellatus (Linnaeus), within a salt marsh-dominated estuarine landscape. Fisheries Management and Ecology 2007; 14: 183- 190.

EPA. U.S. Environmental Protection Agency Remedial Investigation Fact Sheet: LCP Chemicals, Brunswick, Georgia. 2002.

EPA. National Priorities List, Final Rule; 40 CFR Part 300. Federal Register 2007; 72: 53463-53470.

Evans WE. Orientation behavior of delphinids: Radio telemetric studies. Annals New York Academy of Sciences 1971; 188: 142-160.

Ford JKB, Ellis GM, Barrett-Lennard LG, Morton AB, Palm RS, Balcomb KC. Dietary specialization in two sympatric populations of killer whales (Orcinus orca) in coastal British Columbia and adjacent waters. Canadian Journal of Zoology 1998; 76: 1456-1471.

Frangiamore CS, Gibbons W. The New Georgia Encyclopedia. Atlanta: Georgia Humanities Council, 2004.

Freedman D, Diaconis P. On the histogram as a density estimator: L2 theory. Probability Theory and Related Fields 1981; 57: 453-476.

Gannon DP, Waples DM. Diets of coastal bottlenose dolphins from the U.S. mid- Atlantic coast differ by habitat. Marine Mammal Science 2004; 20: 527- 545.

Hansen LJ, Schwacke LH, Mitchum GB, Hohn AA, Wells RS, Zolman ES, et al. Geographic variation in polychorinated biphenyl and organochlorine

125

pesticide concentrations in the blubber of bottlenose dolphins from the US Atlantic coast. Science of the Total Environment 2004; 319: 147-172.

Hanson RB, Synder J. Microheterotrophic activity in a salt-marsh estuary, Sapelo Island, Georgia. Ecology 1979; 60: 99-107.

Hays GC, Bradshaw CJA, James MC, Lovell P, Sims DW. Why do Argos satellite tags deployed on marine mammals stop transmitting. Journal of Experimental Marine Biology and Ecology 2007; 349: 52-60.

Helle E, Olsson M, Jensen S. PCB levels correlated with pathological changes in seal uteri. Ambio 1976; 5: 261-263.

Hines JE, Kendall WL, Nichols JD. On the use of the robust design with transient capture-recapture models. The Auk 2003; 120: 1151-1158.

Hohn A, Scott M, Wells R, Sweeney J, Irvine A. Growth layers in teeth from free- ranging, known-age bottlenose dolphins. Marine Mammal Science 1989; 5: 315-342.

Holland AF, Sanger DM, Gawle CP, Lerberg SB, Santiago MS, Riekerk GH, et al. Linkages between tidal creek ecosystems and the landscape and demographic attributes of their watershed. Journal of Experimental Marine Biology and Ecology 2004; 298: 151-178.

Houde M, Hoekstra PF, Solomon KR, Muir DCG. Organohalogen contaminants in Delphinoid Cetaceans. Reviews of Environmental Contamination & Toxicology 2005; 184: 1-57.

Houde M, Pacepavicius G, Wells RS, Fair PA, Letcher RJ, Alaee M, et al. Polychlorinated biphenyls and hydroxylated polychlorinated biphenyls in plasma of bottlenose dolphins (Tursiops truncatus) from the Western Atlantic and the Gulf of Mexico. Environmental Science & Technology 2006; 40: 5860-66.

Irvine AB, Scott MD, Wells RS, Kaufmann JH. Movements and activities of the Atlantic bottlenose dolphin, Tursiops truncatus, near Sarasota, Florida. Fishery Bulletin 1981; 79: 671-688.

126

Kannan K, Maruya KA, Tanabe S. Distribution and characterization of polychlorinated biphenyl congeners in soil and sediments from a Superfund site contaminated with Aroclor 1268. Environmental Science and Technology 1997; 31: 1483-1488.

Kannan K, Nakata H, Stafford R, Masson G, Tanabe S, Giesy JP. Bioaccumulation and toxic potential of extremely hydrophobic polychlorinated biphenyl congeners in biota collected at a Superfund site contaminated with Aroclor 1268. Environmental Science and Technology 1998; 32: 1214-1221.

Kendall WL, Nichols JD, Hines JE. Estimating temporary emigration using capture-recapture data with Pollock's robust design. Ecology 1997; 78: 563-578.

Kiszka JJ, Simon-Bouhet B, Charlier F, Pusineri C, Ridoux V. Individual and group behavioural reactions of small delphinids to remote biopsy sampling. Animal Welfare 2010; 19: 411-417.

Klatsky LJ, Wells RS, Sweeney JC. Offshore bottlenose dolphins (Tursiops truncatus): Movement and dive behavior near the Bermuda Pedestal. Journal of Mammalogy 2007; 88: 59-66.

Krahn MM, Herman DP, Matkin CO, Durban JW, Barrett-Lennard LG, Burrows DG, et al. Use of chemical tracers in assessing the diet and foraging regions of eastern North Pacific killer whales. Marine Environmental Research 2007; 63: 91-114.

Kucklick J, Schwacke L, Wells R, Hohn A, Guichard A, Yordy J, et al. Bottlenose dolphins as indicators of persistent organic pollutants in waters along the US East and Gulf of Mexico coasts. Environmental Science & Technology in review.

Kucklick JR, Schantz MM, Pugh RS, Porter BJ, Poster DL, Becker PR, et al. Marine mammal blubber reference and control materials for use in the determination of halogenated organic compounds and fatty acids. Anal Bioanal Chem 2010; 397: 423-32.

Lancia RA, Nichols JD, Pollock KH. Estimating the number of animals in wildlife populations. In: Bookhout TA, editor. Research and Management

127

Techniques for Wildlife and Habitats. The Wildlife Society, Bethesda, 1996, pp. 215-253.

Le Cren ED. A note on the history of mark-recapture population estimates. Journal of Animal Ecology 1965; 34: 453-454.

Litz JA, Garrison LP, Fieber LA, Martinez A, Contillo JP, Kucklick JR. Fine-scale spatial variation of persistent organic pollutants in bottlenose dolphins (Tursiops truncatus) in Biscayne Bay, Florida. Environmental Science and Technology 2007; 41: 7222-7228.

Martineau D, Beland P, Desjardins C, Lagace A. Levels of organochlorine chemicals in tissues of beluga whales (Delphinapterus leucas) from the St. Lawrence Estuary, Quebec, Canada. Archives of Environmental Contamination and Toxicology 1987; 16: 137-147.

Maruya K, Lee R. Aroclor 1268 and toxaphene in fish from a southeastern U.S. estuary. Environmental Science and Technology 1998; 32: 1069-1075.

Maruya K, Walters T, Manning R. Residues of toxaphene in fin- and shellfish from Terry/Dupree Creek, Georgia, U.S.A. Estuaries 2001; 24: 585-596.

Morris RC. Redefining the port of Brunswick. Georgia Ports Authority 2007; 47: 1-40.

NOAA. Interim report on the bottlenose dolphin (Tursiops truncatus) Unusual Mortality Event along the Panhandle of Florida March-April 2004. NOAA Fisheries Office of Protected Resources, 2004, pp. 36.

Nowacek SM, Wells RS, Solow AR. Short-term effects of boat traffic on bottlenose dolphins, Tursiops truncatus, in Sarasota Bay, Florida. Marine Mammal Science 2001; 17: 673-688.

Owen BO, White S. National Estuarine Research Reserve System 10th Anniversary Report on the System-wide Monitoring Program (SWMP) Data Applications (1995-2005). National Estuarine Research Reserve System, NOAA, Silver Spring, MD. 32 pp. 2005.

128

Pine WE, Pollock KH, Hightower JE, Kwak TJ, Rice JA. A review of tagging methods for estimating fish population size and components of mortality. Fisheries 2003; 28: 10-23.

Plumley FG, Davis DE, McEnerney JT, Everest JW. Effects of a photosynthesis inhibitor, atrazine, on the salt-marsh fiddler crab, Uca pugnax (Smith). Estuaries and Coasts 1980; 3: 217-223.

Pollock KH. A capture-recapture design robust to unequal probability of capture. Journal of Wildlife Management 1982; 46: 757-760.

Pradel R, Hines JE, Lebretion JD, Nichols JD. Capture-recapture survival models taking account of transients. Biometrics 1997; 53: 60-72.

Pulster E, Smalling K, Zolman E, Schwacke L, Maruya K. Persistent organochlorine pollutants in the blubber of bottlenose dolphins(Tursiops truncatus) along the Georgia coast (USA). Environmental Toxicology and Chemistry 2009; 28: 1390-1399.

Pulster EL, Maruya KA. Geographic specificity of Aroclor 1268 in bottlenose dolphins (Tursiops truncatus) frequenting the Turtle/Brunswick River Estuary, Georgia (USA). Science of the Total Environment 2008; 393: 367-75.

Pulster EL, Smalling KL, Maruya KA. Polychlorinated biphenyls and toxaphene in preferred prey fish of coastal southeastern US bottlenose dolphins (Tursiops truncatus). Environmental Toxicology and Chemistry 2005; 24: 3128-3136.

Read AJ, Gaskin DE. Radio tracking the movements and activities of harbor porpoises, Phocoena phocoena (L.), in the Bay of Fundy, Canada. Fishery Bulletin 1985; 83: 543-552.

Read AJ, Urian KW, Wilson B, Waples DM. Abundance of bottlenose dolphins in the bays, sounds, and estuaries of North Carolina. Marine Mammal Science 2003; 19: 59-73.

Read AJ, Westgate AJ. Monitoring the movements of harbour porpoises (Phocoena phocoena) with satellite telemetry. Marine Biology 1997; 130: 315-322.

129

Rogers SG, Targett TE, Van Sant SB. Fish-nursery use in Georgia salt-marsh estuaries: the influence of springtime freshwater conditions. Transactions of the American Fisheries Society 1984; 113: 595-606.

Rosel PE. PCR-based sexing in Odontocete cetaceans. Conservation Genetics 2003; 4: 647-649.

Rosel PE, Hansen LJ, Hohn AA. Restricted dispersal in a continuously distributed marine species: common bottlenose dolphins Tursiops truncatus in coastal waters of the western North Atlantic. Molecular Ecology 2009; 18: 5030-5045.

Ross PS, Ellis GM, Ikonomou MG, Barrett-Lennard LG, Addison RF. PCB concentrations in free ranging Pacific Killer whales, Orcinus orca: Effects of age, sex and dietary preference. Marine Pollution Bulletin 2000; 40: 504:515.

Sanger D, Blair A, DiDonato G, Washburn T, Jones S, Chapman R, et al. Support for integrated ecosystem assessments of NOAA's National Estuarine Research Reserve System (NERRS), Volume I: the impacts of coastal development on the ecology and human well-being of tidal creek ecosystems of the US Southeast. NOAA Technical Memorandium NOS NCCOS 82 2008; Charleston, SC, USA.

Schelske CL, Odum EP. Mechanisms maintaining high productivity in Georgia estuaries. Proceedings of Gulf and Caribbean Fishery Institute 1961; 14: 75-80.

Schwacke LH, Voit EO, Hansen LJ, Wells RS, Mitchum GB, Hohn AA, et al. Probabilistic risk assessment of reproductive effects of polychlorinated biphenyls on bottlenose dolphins (Tursiops truncatus) from the southeast United States coast. Environmental Toxicology and Chemistry 2002; 21: 2752-2764.

Schwacke LH, Zolman ES, Townsend FI, Balmer BC, Litz J, DeGuise S, et al. Anemia, hypothyroidism and suppressed immune response associated with extreme polychlorinated biphenyl exposure in common bottlenose dolphins (Tursiops truncatus) from the Georgia, USA, coast. Proceedings of the Royal Society B-Biological Sciences in review.

130

Scott MD, Wells RS, Irvine AB. A long-term study of bottlenose dolphins on the west coast of Florida. In: Leatherwood S, Reeves RR, editors. The bottlenose dolphin. Academic Press, San Diego, 1990a, pp. 235-244.

Scott MD, Wells RS, Irvine AB, Mate BR. Tagging and marking studies on small cetaceans. In: Leatherwood JS, Reeves RR, editors. The Bottlenose Dolphin. Academic Press, San Diego, CA, 1990b, pp. 489-514.

Seber GAF. The estimation of animal abundance and related parameters. New York: MacMillan, 1982.

Sellas AB, Wells RS, Rosel PE. Mitochondrial and nuclear DNA analyses reveal fine scale geographic structure in bottlenose dolphins (Tursiops truncatus) in the Gulf of Mexico. Conservation Genetics 2005; 6: 715-728.

Shane S, Wells RS, Würsig B. Ecology, behavior, and social organization of the bottlenose dolphin: a review. Marine Mammal Science 1986; 2: 34-63.

Speakman T, Zolman ES, Adams J, Defran RH, Laska D, Schwacke LH, et al. Temporal and spatial aspects of bottlenose dolphin occurrence in coastal and estuarine waters near Charleston, South Carolina. NOAA Technical Memorandum NOS-NCCOS-37, 243 pp 2006.

Speakman TR, Lane SM, Schwacke LH, Fair PA, Zolman ES. Mark-recapture estimates of seasonal abundance and survivorship for bottlenose dolphins (Tursiops truncatus) near Charleston, South Carolina, USA. Journal of Cetacean Research and Management 2010; 11: 153-162.

Strahler AN. Dynamic basis of geomorphology. Geological Society of America Bulletin 1952; 63: 923-938.

Switzer PV. Site fidelity in predictable and unpredictable habitats. Evolutionary Ecology 1993; 7: 535-555.

Taylor BL, Gerrodette T. The uses of statistical power in conservation biology: The vaquita and Northern spotted owl. Conservation Biology 1993; 7: 489- 500.

131

Thomas P. Effects of Aroclor 1254 and cadmium on reproductive endocrine function and ovarian growth in atlantic croaker. Marine Environmental Research 1989; 28: 499-503.

Urian KW, Hofmann S, Wells RS, Read AJ. Fine-scale population structure of bottlenose dolphins (Tursiops truncatus) in Tampa Bay, Florida. Marine Mammal Science 2009; 25: 619-638.

Urian KW, Hohn AA, Hansen LJ. Status of the photo-identification catalog of coastal bottlenose dolphins of the western North Atlantic: Report of a workshop of catalog contributors. NOAA Technical Memorandum NMFS- SEFSC-425 1999.

Waring GT, Josephson E, Fairfield CP, Maze-Foley K. U.S. Atlantic and Gulf of Mexico marine mammal stock assessments. NOAA Technical Memorandum NMFS-NE 210 2009: 440.

Watanabe M, Kannan K, Takahashi A, Loganathan BG, Odell DK, Shinsuke T, et al. Polychlorinated biphenyls, organochlorine pesticides, tris (4- chlorophenyl) methane, and tris (4-chlorophenyl) methanol in livers of small cetaceans stranded along Florida coastal waters, USA. Environmental Toxicology and Chemistry 2000; 19: 1566-1574.

Wells RS, Early GA, Gannon JG, Lingenfelser RG, Sweeney P. Tagging and tracking of rough-toothed dolphins (Steno bredanensis) from the March 2005 mass stranding in the Florida Keys. NOAA Technical Memorandum NMFS-SEFSC-574 2008: 44.

Wells RS, Manire CA, Smith D, Gannon JG, Fauquier D, Mullin KD. Movements and dive patterns of a rehabilitated Risso‟s dolphin, Grampus griseus, in the Gulf of Mexico and Atlantic Ocean. Marine Mammal Science 2009; 25: 420-429.

Wells RS, Rhinehart HL, Hansen LJ, Sweeney JC, Townsend FI, Stone CR, et al. Bottlenose dolphins as marine ecosystem sentinels: developing a health monitoring system. EcoHealth 2004; 1: 246:254.

Wells RS, Scott MD. Estimating bottlenose dolphin population parameters from individual identification and capture release techniques. . In: Hammond PS, Mizroch SA, Donovan GP, editors. Individual Recognition of Cetaceans: Use of Photo-identification and Other Techniques to Estimate 132

Population Parameters. IWC Special Issue 12. International Whaling Commission, Cambridge, 1990, pp. 407:415.

Wells RS, Scott MD. Bottlenose dolphin, Tursiops truncatus (Montagu, 1821). In: Ridgway SH, Harrison R, editors. Handbook of marine mammals; Volume 6: The second book of dolphins and the porpoises. Academic Press, San Diego, 1999, pp. 137-182.

Wells RS, Tornero V, Borrell A, Aguilar A, Rowles TK, Rhinehart HL, et al. Integrating life-history and reproductive success data to examine potential relationships with organochlorine compounds for bottlenose dolphin (Tursiops truncatus) in Sarasota Bay, FL. Science of the Total Environment 2005; 349: 106:19.

Wells RS, Urian KW, Read AJ, Bassos MK, Carr WJ, Scott MD. Low-level monitoring of bottlenose dolphins, Tursiops truncatus, in Tampa Bay, Florida: 1988-1993. NOAA Technical Memorandum NMFS-SEFSC-385 1996: 25.

White GC, Burnham KP. Program MARK: Survival estimation from populations of marked animals. Bird Study 46 Supplement 1999: 120-138.

Williams JA, Dawson SM, Slooten E. The abundance and distribution of bottlenose dolphins (Tursiops truncatus) in Doubtful Sound, New Zealand. Canadian Journal of Zoology 1993; 71: 2080-2088.

Wilson B, Hammond PS, Thompson PM. Estimating size and assessing trends in a coastal bottlenose dolphin population. Ecological Applications 1999; 9: 288-300.

Yochem PK, Stewart BS, DeLong RL, DeMaster DP. Diel haul-out patterns and site fidelity of harbor seals (Phoca vitulina richardsi) on San Miguel Island, California, in autumn. Marine Mammal Science 1987; 3: 323-332.

Yordy JE, Wells RS, Balmer BC, Schwacke LH, Rowles TK, Kucklick JR. Life history as a source of variation for persistent organic pollutant (POP) patterns in a community of common bottlenose dolphins (Tursiops truncatus) resident to Sarasota Bay, FL. Science of the Total Environment 2010; 408: 2163-2172.

133

Young RF, Phillips HD. Primary production required to support bottlenose dolphins in a salt marsh estuarine creek system. Marine Mammal Science 2002; 18: 358-373.

Zolman ES. Residence patters of bottlenose dolphins (Tursiops truncatus) in the Stono River Estuary, Charleston County, South Carolina, USA. Marine Mammal Science 2002; 18: 879-892.

134