Ranging patterns, spatial overlap, and association with dolphin morbillivirus exposure in common bottlenose dolphins (Tursiops truncatus) along the Georgia, USA coast

Abstract During 2013–2015, an outbreak of dolphin morbillivirus (DMV) occurred in the western North Atlantic, which resulted in the stranding of over 1,600 common bottlenose dolphins (Tursiops truncatus). There are currently five coastal and 10 bay, sound, and estuary dolphin stocks along the U.S. Atlantic coast, yet there is very limited understanding of which stocks were exposed to DMV during the recent outbreak, or how DMV was transmitted across stocks. In order to address these questions, information is needed on spatial overlap and stock interactions. The goals of this project were to determine ranging patterns, prevalence of DMV, and spatial overlap of the South Carolina‐Georgia (SC‐GA) Coastal Stock, and adjacent Southern Georgia Estuarine System (SGES) Stock. During September 2015, a health assessment and telemetry study was conducted in which 19 dolphins were captured, tested for antibodies to DMV, and satellite tagged. Dolphins were classified into one of three ranging patterns (Coastal, Sound, or Estuary) based upon telemetry data. Coastal dolphins (likely members of the SC‐GA Coastal Stock) had a significantly higher prevalence of positive DMV antibody titers (0.67; N = 2/3), than Sound and Estuary dolphins (likely members of the SGES Stock) (0.13; N = 2/16). These results suggest that the SC‐GA Coastal Stock may have experienced greater exposure to DMV as compared to the SGES Stock. However, due to the small size of the SGES Stock and its exposure to high levels of persistent contaminants, this stock may be particularly vulnerable to DMV infection in the future.


| INTRODUC TI ON
During a 10-month period between June 1987 and March 1988, an outbreak of cetacean morbillivirus in the western North Atlantic led to the stranding of at least 667 common bottlenose dolphins (Tursiops truncatus) along the U.S. Atlantic coast from New Jersey southward to Florida (Geraci, 1989;McLellan, Friedlaender, Mead, Potter, & Pabst, 2002). Recently, another major die-off occurred along the same span of the U.S. Atlantic coast in which over 1,600 common bottlenose dolphins stranded between July 2013 and March 2015 (NMFS unpub. data, Morris et al., 2015). The apparent cause of the significant increase in stranded common bottlenose dolphins, which represents an 8-fold increase over historical stranding rates, was determined to be dolphin morbillivirus (DMV) infection (Morris et al., 2015;Van Bressem et al., 2014). Currently, it is unclear why DMV outbreaks periodically impact common bottlenose dolphins in the western North Atlantic. Possible factors may include reintroduction of the virus associated with natural cyclic fluctuations of "herd immunity" (Duignan et al., 1996), changes in disease transmission associated with dolphin migrations, and spatial overlap between populations of small cetaceans. To assess factors associated with DMV transmission among common bottlenose dolphins in the western North Atlantic, a better understanding of movement patterns, spatiotemporal overlap of stocks, and assessment of exposure to DMV within stocks is needed.
The U.S. Marine Mammal Protection Act defines a stock as "…a group of marine mammals of the same species in a common spatial arrangement that interbreed when mature" (Marine Mammal Protection Act 16 U.S.C. 1,361 et seq.). Stock definitions ensure that conservation efforts to mitigate human activity are aimed at the appropriate management unit (reviewed in Conn, Gorgone, Jugovich, Byrd, & Hansen, 2011). Common bottlenose dolphin stock structure in the western North Atlantic is a complex mosaic of overlapping bay, sound, and estuary (BSE), and coastal stocks (Hayes, Josephson, Maze-Foley, & Rosel, 2017) (Figure 1). From North Carolina to Florida, there are 10 BSE stocks currently recognized (Hayes et al., 2017), characterized by year-round residency, high site fidelity, and localized ranging patterns (e.g., Zolman, 2002, Read, Urian, Wilson, & Waples, 2003, Mazzoil et al., 2008, Balmer et al., 2013. Concurrently, five coastal stocks (Northern and Southern Migratory, South Carolina-Georgia, Northern Florida, and Central Florida Coastal Stocks) have been designated (Hayes et al., 2017). Coastal Stock (Morris et al., 2015). Strandings from three BSE stocks, one in Northern North Carolina and two in Florida from Jacksonville and Indian River Lagoon, were also documented to be positive for DMV (NMFS unpublished data), but is not known whether infection occurred in other BSE stocks due to lack of data.
Aerial survey, genetic, small vessel photographic-identification (photo-ID), stranding, and telemetry data have provided insight into coastal dolphin ranging patterns and spatial overlap with BSE stocks (reviewed in Hayes et al., 2017). Several sites in South Carolina and Georgia have been relatively well-studied and found to experience seasonal increases in dolphin abundance during summer and fall suggestive of individuals from coastal stocks entering BSE waters (Balmer et al., 2013;Speakman, Lane, Schwacke, Fair, & Zolman, 2010). During the 1987-1988 and 2013-2015 mortality events, F I G U R E 1 Western North Atlantic common bottlenose dolphin bay, sound, and estuary (BSE), and coastal stock structure (adapted from Hayes et al., 2017). BSE stock boundaries include the estuarine and nearshore waters (≤1 km from shore). Coastal stock boundaries include primarily estuarine and inshore waters (≤20 m deep) with evidence for some coastal stocks to extend into continental shelf waters temporal stranding patterns (McLellan et al., 2002) and long-term photo-ID data (NMFS unpub. data, Urian, 2016) suggested that the Northern and Southern Migratory Coastal Stocks were the first stocks impacted by both of the DMV outbreaks. As these stocks migrated along the western North Atlantic coast, animals may have spread the disease to other coastal and/or BSE stocks causing the mortality events to extend spatially along the coast over time.
Migratory animals experience high energetic demands that may reduce immune function and increase susceptibility to disease (reviewed in Bowlin et al., 2010). These factors make migrating animals potential sources for infectious diseases that can in turn expand the geographic distribution of a pathogen or mortality event (Altizer, Bartel, & Han, 2011). Diseases, such as DMV, are thought to be transmitted between animals via inhalation or direct contact (Black, 1991;Van Bressem, Waerebeek, & Raga, 1999) and contact rates are generally higher in social species than solitary species (Craft, Hawthorne, Packer, & Dobson, 2008). Positive correlations between animal contact and spatial overlap of hosts have been identified in numerous species across many taxa and may affect disease transmissions among animals (reviewed in Robert, Garant, & Pelletier, 2012). Within a species, disease transmission rates may vary by pathogen (or strain) (Mideo, Alizon, & Day, 2008), age-class (Greig, Gulland, & Kreuder, 2005), group size (Côté & Poulinb, 1995), migration pattern (Altizer et al., 2011), season (Rogers et al., 1998), or sex (Creel & Creel, 1991).
A model for DMV transmission during the recent 2013-2015 mortality event was recently developed, and model analyses suggest that frequency-dependent transmission predominantly regulated the outbreak (Morris et al., 2015). Common bottlenose dolphins are highly sociable and have a fission-fusion society in which group composition can change quickly on an hourly to daily basis (reviewed in Connor, Wells, Mann, & Read, 2000). In the western North Atlantic, common bottlenose dolphin group size and cohesion tend to be greater in coastal as opposed to BSE environments (Speakman et al., 2006;Toth, Hohn, Able, & Gorgone, 2011), which may suggest different contact rates and potential differences in disease F I G U R E 2 Georgia health assessment and telemetry study area and capture locations for free-ranging common bottlenose dolphins (N = 19). Letter (Z) and two-digit number (♀, odd; ♂, even) are identifiers for individual tagged dolphins transmission rates within and across stocks. However, little is known about degree of interaction and possible spatiotemporal overlap between coastal and BSE stocks, both of which are essential for assessing contact rates and disease transmission.
A confluence of common bottlenose dolphin stocks occurs along the Georgia coast, where the ranges of five stocks, the Southern Migratory Coastal Stock, the South Carolina-Georgia (SC-GA) Coastal Stock, and adjacent Northern, Central, and Southern Georgia Estuarine System Stocks, overlap. The goals of the current study were to examine ranging patterns and spatial overlap of two of these stocks, the South Carolina-Georgia (SC-GA) Coastal Stock, and adjacent Southern Georgia Estuarine System (SGES) Stock, and to assess DMV antibody prevalence and pathogen presence through capture-release sampling and satellite telemetry. On field days in which conditions did not permit coastal captures, health assessments were carried out in BSE waters, initially targeting animals farther south and progressively working closer to the field base as the day progressed.

| Dolphin handling
The 2015 dolphin health assessment was conducted over 10 field days between 14 and 25 September. Capture-release methodologies have been detailed for previous studies (Asper, 1975;Norman, Hobbs, Foster, Schroeder, & Townsend, 2004). Briefly, one to two dolphins were encircled with a 366 m by 7 m deep seine net. For the majority of capture sets, water depth (>2 m) and bottom substrate type (heavy mud and oyster shell) prevented shallow-water capture protocols. Instead, once an animal became entangled in the net, it was handled from one of three 6.3 m, center-console, Zodiac (Zodiac Milpro International, Paris, France) rigid-hulled inflatable boats (RhIBs) with twin 90-hp four stroke outboard engines. Once restrained, the animal was moved onto a 3 m long, tri-fold floating mat. Sex was determined for all dolphins and any female dolphin greater than or equal to 220 cm was held on the floating mat until an ultrasound exam could be conducted to determine if the dolphin was pregnant (Smith et al., 2013). Abbreviated sampling was conducted for any diagnosed pregnant females, or dolphins that were becoming overly stressed (i.e., rapid respirations, greater than 8 breaths per minute, or arching), or if weather deteriorated to unworkable conditions. Samples were either collected while the animal was on the floating mat or on a specially designed processing vessel (R/V Megamouth, a 9.1 m Munson "Packman" monohull; William E. Munson Company, Burlington, WA, USA). Prior to release, dolphins were freeze-branded with a letter (Z) and two-digit number (♀, odd; ♂, even) to provide long-term identification (Scott, Wells, Irvine, & Mate, 1990) (Figure 3).

| Dolphin Morbillivirus (DMV)
Blood was collected from the ventral fluke vasculature using a 19 g × ¾" butterfly catheter. For serum samples, blood was spun in for DMV polymerase chain-reaction (PCR) testing. DMV titers were measured using the virus neutralization test (VNT) as previously described (Saliki & Lehenbauer, 2001). The Belfast strain of DMV was used. Twofold dilutions of serum (25 μL) were made in 96well microtiter plates with Dulbecco's minimum essential medium (DMEM). An equal volume of virus (25 μL) containing approximately 100 TCID50 was added to each well and plates were incubated at 35.5°C for 1 hr. Vero Dog Slam cells (1.5 × 104 cells in 150 μL) were added to each well and the plates incubated at 35.5°C in 5% CO 2 for 3 days, after which they were examined for cytopathic effects.
The antibody titer was defined as the highest dilution of serum that neutralized CPE. Results were expressed as positive if titers were ≥1:16 (Rowles et al., 2011). Frozen blowhole and rectal swabs were tested for cetacean morbillivirus by reverse transcriptase-PCR (RT-PCR). RNA was extracted using TRIzol reagent (Invitrogen Corp., Carlsbad, CA, USA), and cDNA was transcribed using the Superscript III First Strand kit (Invitrogen Corp.). DMV testing was performed using universal morbillivirus primers targeting a 429 base pair fragment of the phosphoprotein (P) gene (Barrett et al., 1993) followed by nested primers specific for dolphin morbillivirus. Bands of correct size were excised and purified PCR products were cloned (pCR4-TOPO vector; Invitrogen Corp.) and sequenced (ABI 3,730 Capillary Electrophoresis Genetic Analyzer; Applied Biosystems, Inc., Foster City, CA, USA). Partial sequences of the P gene were compared against morbilliviral sequences available in the GenBank database.
A tooth was extracted for age determination (Hohn, Scott, Wells, Sweeney, & Irvine, 1989;McFee, Adams, Fair, & Bossart, 2012;Ridgeway, Green, & Sweeney, 1975) as conditions permitted from dolphins that were not pregnant or identified by the veterinary team to be overly stressed.
Individual and cumulative ranging patterns were defined as 95% and 50% utilization distributions (UDs), which are probability distributions of an animal's or group of animals' movements (use) in the available habitat (plane) (Worton, 1989). UDs summarize ranging pattern data and provide insight into habitat use, spatiotemporal overlap, and short-term site fidelity (reviewed in Fieberg & Kochanny, 2005). Kernel density estimates (KDEs) are a quantitative method to determine UDs (Worton, 1989). The selection of bandwidth, or the smoothing parameter (h), is an important decision in which KDE distributions can be over-or under-estimated depending on this value (Horne & Garton, 2006;Kie et al., 2010). The methodology for bandwidth selection is dependent on the goals of the project, ranging patterns of the target species, and amount of data available for spatial analyses (Gitzen, Millspaugh, & Kernohan, 2006;Rayment et al., 2009). A rulebased ad hoc method (Kie, 2013) and Home Range Tools (HRT) for ArcGIS (Rodgers, Kie, Wright, Beyer, & Carr, 2015) were used to determine the appropriate bandwidth for KDEs of each individual and cumulative ranging pattern. UDs (95% and 50%) were then determined from these KDEs. Spatial overlap among ranging patterns was calculated following methods described in MacLeod (2013).

| Dolphin health assessment, dolphin morbillivirus (DMV) prevalence and satellite tagging
During  (Table 1) and titers ranged from 1:32 to 1:256. All dolphins were negative for DMV by PCR in blowhole and rectal swabs. had the largest UDs, followed by Sound and Estuary ranging patterns (Table 2; Figure 5).

| Satellite telemetry
F I G U R E 5 50% and 95% utilization distributions (UDs) and capture location with dolphin morbillivirus (DMV) titer for the Coastal, Sound, and Estuary ranging patterns Coastal dolphins had the highest prevalence of antibodies to DMV (0.67; N = 2/3), followed by Sound (0.20; N = 2/10) and Estuary (0.00; N = 0/6) dolphins (Table 1; Figure 5). All ranging patterns had some degree of 95% UD overlap with each other (Table 2; Figure 6). Simons and St. Andrew Sounds (Figure 6). ies have used comparable sample sizes to assess health and ranging patterns of large, marine vertebrates (e.g., Elwen et al., 2006, Meyer, Clark, Papastamatiou, Whitney, & Holland, 2009, Schwacke et al., 2012, Lane et al., 2015. This study provided some of the first insights into ranging patterns for dolphins in the BSE and coastal waters of Georgia, estimated DMV antibody prevalence following the 2013-2015 DMV outbreak, and quantified spatial overlap to assess differences between ranging patterns to assess animal movement and contact DMV transmission.

| Stock structure
Stocks of marine mammals have been primarily delimited using genetic analyses (Rosel, Forgetta, & Dewar, 2005), but, additional sampling techniques such as photo-ID and telemetry have been used to test and assist in classification of individuals into their respective stocks (Balmer, Wells, Schwacke, et al., 2014b;Sveegaard et al., 2015). In the present study, both Estuary and Sound dolphins had small to moderate UDs (Tables 1 and 2 habitat preferences (e.g., Lusseau et al., 2006, Wiszniewski, Allen, & Möller, 2009). In the southeastern U.S., similar sub-populations have been identified for dolphins in more interior estuarine waters and those in larger sounds and surrounding barrier islands (Urian, Hofmann, Wells, & Read, 2009;Wells et al., 2017). Estuary dolphins had UDs (34 ± 8 km 2 ; mean ± SD; mean ± SD; Table 1) several times larger than Estuary and other southeastern U.S. BSE UDs, and included coastal waters primarily within 5 km from shore (approximately 75% of satellite locations), but did have limited ranges extending over 10 km offshore.
Coastal dolphins were characterized by large UDs that included the coastal waters of Savannah, Georgia to Jacksonville, Florida primarily within 10 km from shore (approximately 75% of satellite locations), but did have limited ranges extending to over 15 km offshore (Tables 1 and 2; Figure 5a). None of the three Coastal dolphins were previously identified in the long-term photo-ID catalog for this region. However, historical photo-ID effort was primarily in BSE waters. The 2015 health assessment was conducted in September when the Southern Migratory Coastal Stock was hypothesized to be farther north off North Carolina (Hayes et al., 2017;Silva, 2016;Urian, 2016 coastal waters, photo-ID comparisons between projects/field sites (e.g., Balmer et al., 2016, Urian, 2016 and survey effort including both coastal and BSE waters (e.g., Laska, Speakman, & Fair, 2011, Silva, 2016 can provide data essential in determining ranging patterns, site fidelity, and stock discreteness of the SC-GA Coastal Stock.

| Dolphin morbillivirus (DMV) and spatial overlap
Determining spatial overlap of exposed and naive hosts is essential for predicting the spread of infectious diseases (Robert et al., 2012).
Preliminary modeling efforts for the recent western North Atlantic DMV outbreak concluded that information on movements and interactions among different stocks was one of the greatest needs for refining the model in order to better understand the dynamics of the outbreak (Morris et al., 2015). Here, we provide the first data on movement, spatial overlap, and DMV antibody titers for surviving dolphins following that outbreak. Rowles et al. (2011) previously documented that DMV antibodies decrease over time, but this trend may not always be consistently observed across all individuals. In this study, three of the sampled dolphins had relatively high DMV titers (1:256) and at least two of these were too young to have been exposed during the previous 1987-1988 outbreak; Z46 was only 12 years old, and Z28, with a length of only 211 cm, was likely less than 10 years old (McFee, Schwacke, Stolen, & Mullin, 2010). It is therefore likely that these dolphins were exposed during the more recent outbreak, although the potential for exposure outside of an outbreak is also possible. The fourth DMV positive has a much lower titer (1:32), and at 30 years old, could have potentially been exposed during the prior 1987-1988 outbreak.
Prevalence of DMV antibodies differed for the Estuary and  2003Bossart et al., 2010). In contrast, Duignan et al. (1996)  Atlantic coast (Hayes et al., 2017). In this study, Coastal dolphins not only had the highest prevalence of DMV antibodies (0.67; N = 2/3), but also the largest UDs (approximately 200 km of coastline) (Tables 1 and 2; Figure 5a). These results suggest that contact rates between coastal stocks could be relatively high, at least seasonally, with members of the SC-GA Coastal Stock exposed to DMV as the Southern Migratory Coastal Stock migrates through the region.
Within the SGES Stock, the prevalence of positive DMV titers was higher for dolphins of the Sound ranging pattern (0.20; N = 2/10) as compared to that of the Estuary ranging pattern and Estuary (0.00; N = 0/6), although once stratified by ranging pattern, sample sizes were small. Differences in spatial overlap may provide insight into the gradient of DMV exposure identified across ranging patterns. Although numerous studies across a variety of taxa have linked spatial overlap and disease transmission rates (reviewed in Robert et al., 2012), social barriers (i.e., group size and interaction rates) (Loehle, 1995) may be an additional factor influencing DMV prevalence. In the western North Atlantic, dolphin group size is higher in coastal waters than within BSEs (Speakman et al., 2006;Torres, Mclellan, Meagher, & Pabst, 2005;Toth et al., 2011) Sweeney, 2007) and novel methods to remotely tag individuals from these stocks will provide data necessary for a comprehensive assessment of contact rates and disease prevalence.

| CON CLUS IONS
This study provides the first data from dolphins that were exposed SGES should be considered a stock of concern with a small population estimate (N = 194; CV = 0.05) (Hayes et al., 2017), and extremely high PCB levels (Balmer et al., 2011;Kucklick et al., 2011), which have been documented to adversely affect the immune system and may facilitate the emergence of infectious disease (e.g., Ross, 2002). In fact, previous health assessments of the SGES stock documented a decrease in immune function associated with increasing PCB levels (Schwacke et al., 2012). Therefore, while exposure risk from a future outbreak may be lower for the SGES stock, the fact that the majority of the stock are naïve to DMV, and may additionally exhibit suppressed immune response, suggests that this population may be particularly vulnerable. Future health assessment and telemetry projects in the coastal waters are essential to better assess contact rates and disease prevalence among western North Atlantic coastal and BSE stocks.

| NOAA disclaimer
This publication does not constitute an endorsement of any commercial product or intend to be an opinion beyond scientific or other results obtained by the National Oceanic and Atmospheric Administration (NOAA). No reference shall be made to NOAA, or this publication furnished by NOAA, to any advertising or sales promotion which would indicate or imply that NOAA recommends or endorses any proprietary product mentioned herein, or which has as its purpose an interest to cause the advertised product to be used or purchased because of this publication. Lydia Staggs (Gulf World); and Rob Yordi (Sea World).

CO N FLI C T O F I NTE R E S T
None declared.

AUTH O R CO NTR I B UTI O N S
Dolphin health assessments require a great deal of funding, logistical support, and a multi-faceted team to link all of the data into a comprehensive manuscript. The authors are a mix of field and laboratory researchers that at a minimum did two of the following: 1) preparations and collection of the field data [Brian Balmer

DATA ACCE SS I B I LIT Y
All satellite telemetry, time depth, and DMV data are located on