https://orcid.org/0000-0002-5095-4334
Figures
Abstract
Common bottlenose dolphins (Tursiops truncatus truncatus) inhabiting the Indian River Lagoon (IRL) estuarine system along the east coast of Florida are impacted by anthropogenic activities and have had multiple unexplained mortality events. Given this, managers need precise estimates of demographic and abundance parameters. Mark-recapture photo-identification boat-based surveys following a Robust Design were used to estimate abundance, adult survival, and temporary emigration for the IRL estuarine system stock of bottlenose dolphins. Models allowed for temporary emigration and included a parameter (time since first capture) to assess evidence for transient individuals. Surveys (n = 135) were conducted along predetermined contour and transect lines throughout the entire IRL (2016–2017). The best fitting model allowed survival to differ for residents and transients and to vary by primary period, detection to vary by secondary session, and did not include temporary emigration. Dolphin abundance was estimated from 981 (95% CI: 882–1,090) in winter to 1,078 (95% CI: 968–1,201) in summer with a mean of 1,032 (95% CI: 969–1,098). Model averaged seasonal survival rate for marked residents was 0.85–1.00. Capture probability was 0.20 to 0.42 during secondary sessions and the transient rate was estimated as 0.06 to 0.07. This study is the first Robust Design mark-recapture survey to estimate abundance for IRL dolphins and provides population estimates to improve future survey design, as well as an example of data simulation to validate and optimize sampling design. Transients likely included individuals with home ranges extending north of the IRL requiring further assessment of stock delineation. Results were similar to prior abundance estimates from line-transect aerial surveys suggesting population stability over the last decade. These results will enable managers to evaluate the impact of fisheries-related takes and provide baseline demographic parameters for the IRL dolphin population which contends with anthropogenic impacts and repeated mortality events.
Citation: Durden WN, Stolen ED, Jablonski T, Moreland L, Howells E, Sleeman A, et al. (2021) Abundance and demography of common bottlenose dolphins (Tursiops truncatus truncatus) in the Indian River Lagoon, Florida: A robust design capture-recapture analysis. PLoS ONE 16(4): e0250657. https://doi.org/10.1371/journal.pone.0250657
Editor: Michael E. Douglas, University of Arkansas Fayetteville, UNITED STATES
Received: December 22, 2020; Accepted: April 9, 2021; Published: April 28, 2021
Copyright: © 2021 Durden et al. This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
Data Availability: All relevant data are within the manuscript and its Supporting Information files.
Funding: This study was supported by a grant (2016-16-DMM-01) received by WND from North Carolina SeaGrant (https://ncseagrant.ncsu.edu/). Hubbs-SeaWorld Research Institute staff also received support from a grant received by WND from SeaWorld Busch Gardens Conservation Fund (2016-Session 1) (https://swbg-conservationfund.org/) as well funding for the project that was provided from proceeds of Discover Florida’s Oceans specialty license plate (http://discoverfloridasoceans.com). The funders provided support in the form of salaries for authors [WND, EDS, TJ, LM, EH, AS, MD, GB, MM], equipment, supplies and travel, but did not have any additional role in the study design, data collection and analysis, decision to publish, or preparation of the manuscript. The specific roles of these authors are articulated in the “Author Contributions” section.
Competing interests: We note that one or more of the authors are currently employed by a commercial company: GeoSan Services LLC, Centersoft Technologies, and Dolphin Census, Inc. This does not alter our adherence to PLOS ONE policies on sharing data and materials.
Introduction
Anthropogenic impacts continue to threaten both coastal and estuarine cetacean populations [1–9], rendering the accurate evaluation of abundance and demographic parameters critical to population management. Various methods have been employed to estimate the abundance of cetacean species, most based on line-transect or mark-recapture methodology [10–15]. Both approaches can provide accurate estimates. While line-transect surveys (aerial and vessel based) are the most utilized methods for coastal species, they only estimate density (abundance) and cannot estimate survival or distinguish between resident and transient animals. In contrast, mark-recapture methods, which rely on patterns of individuals resighted over time, can estimate abundance, survival, and temporary immigration [16,17]. Identifying resident animals within a population is critical, as those animals are most vulnerable to the impacts of repeated anthropogenic activities as well as ecological deterioration in the region.
Common bottlenose dolphins (Tursiops truncatus truncatus) inhabiting the Indian River Lagoon (IRL) estuarine system along the east coast of Florida between Ponce Inlet and Jupiter Inlet have been studied for decades and are considered long-term residents comprising the IRL estuarine system dolphin stock [18–20]. The expansive range (~250 km) of this dolphin stock has made vessel-based abundance estimation difficult and prior studies have thus employed line-transect aerial surveys [21–23]. However, more recent studies support the occurrence of transients as well as movements of some individuals that extend beyond the northern boundary of the IRL [21,24–27]. Many cetacean species, including bottlenose dolphins, can be individually identified by naturally occurring markings on the trailing edge of the dorsal fin [28–30]. Photo-identification, or the identification of individuals based on these unique markings, is widely used to study cetaceans [30]. It is commonly combined with mark-recapture methods where marked individuals are “captured” (first identified) and subsequently “recaptured” (resighted) during survey efforts [31–36].
Accurate estimates of abundance and demographic parameters are essential to the management and conservation of the IRL dolphin stock and have become increasingly important as IRL dolphins have experienced multiple Unusual Mortality Events (UMEs) (2001, 2008, 2013; 2013–2015) [37]. During the largest mortality event (2013 UME), a minimum of 77 dolphin mortalities occurred. Based on the mean abundance estimate prior to the event (1,032 dolphins) [21], mortalities represented ~7.5% of the population (19 marked individual mortalities in 2013). Concurrent with this event, the Mid Atlantic UME (2013–2015) coincided and further impacted IRL dolphins [37]. Reoccurring mortality events could indicate serious ecological pressures that may lead to the decline of this stock. Indian River Lagoon dolphins are listed as a strategic stock since anthropogenic mortality likely exceeds Potential Biological Removal set by managers (PBR; the maximum number of mortalities, excluding natural mortalities, that can be removed annually while still allowing the stock to reach or maintain an optimal sustainable population level) [37]. In recent years, the IRL has undergone several large-scale ecosystem changes, most notably phytoplankton blooms that yielded catastrophic seagrass loss [38]. Seagrass meadows have been found to provide critical habitat to prey consumed by estuarine dolphins [39]; therefore, these significant ecological changes are likely to jeopardize the health of the already vulnerable IRL dolphin stock. Recent studies have documented diminished health in IRL dolphins including high concentrations of mercury [40], lingual and genital papillomas [41], and skin disease (lacaziosis) [42,43]. Moreover, interactions with both commercial and recreational fisheries account for up to 12% of the annual mortality [8,44].
The Marine Mammal Protection Act requires dolphin stock assessment, and abundance estimates that are necessary to manage stocks and to calculate the level of sustainable anthropogenic mortality (PBR). Data from aerial surveys and observations of movements of IRL dolphins through inlets have suggested that temporary emigration (movement of resident individuals out of the study area during a portion of the study), and transience (movement of non-resident dolphins through the study area) may have contributed to fluctuating abundance estimates. This is most obvious in portions of the lagoon near inlet access, particularly in response to dramatic declines in water temperature [21,23] which may influence dolphin movement [45–48]. Satellite telemetry has further corroborated oceanic habitat usage in resident IRL dolphins [49]. Photo-identification surveys further support transience occurrence, documenting dolphin movements between the northern portion of the IRL (Mosquito Lagoon) and the St. Johns River (Jacksonville Estuarine system stock-JES), [24,25,50] with a 13% exchange in the individuals examined [25]. Furthermore, genetic research suggests that dolphins inhabiting the Mosquito Lagoon sub-basin may be a disjunct community from the IRL as these animals are genetically distinct from the rest of the IRL and most closely associated with the JES stock [50], suggesting genetic exchange [26,27]. Evaluating transience in IRL dolphins is imperative as parameter estimates will be biased if transients are not considered.
The objectives of this study were to utilize dolphin photo-identification surveys and mark-recapture methodology as components of a Robust Design survey [51] to estimate abundance, survival, occurrence of transients, and temporary emigration of IRL dolphins. Prior to survey initiation, a simulation study was conducted to validate the study design (S1 Text, S1 Fig). We measured transient rates by incorporating a time since first capture parameter into our survival sub-models. Closed population capture-recapture models were used to estimate abundance by sub-basin and primary period (= season).
Methods
Ethics statement
Data collection, vessel-based photo-identification surveys of free-ranging bottlenose dolphins, was conducted under permits issued by NOAA Fisheries under General Authorization Letter of Confirmation No.: 16522, 18182, and 20377–01 in tandem with permits issued by Canaveral National Seashore: CANA-2015-SCI-0010 and the U.S. Fish and Wildlife Service: MI-2016-207R. This allowed data collection on the protected/privately owned lands including national wildlife refuges (Cape Canaveral National Seashore, Merritt Island National Wildlife Refuge) and otherwise secured properties (Canaveral Air Force Station, Kennedy Space Center). Data collection did not involve animal handling and thus did not require additional IACUC authorization.
Study area
The IRL is a shallow and diverse estuarine system located along the east coast of central Florida. It opens to the Atlantic Ocean at four inlets and consists of three interconnected basins: the Indian River, Banana River and Mosquito Lagoon [52–54] (Fig 1). The 902 km2 estuary spans 220 km with a width of 0.93 to 9.30 km [22] extending from Ponce Inlet to Jupiter Inlet [54]. Although most of the estuary is shallow (<1 m at high tide), depths > 5 m occur in the dredged basins and channels of the Intracoastal Waterway (ICW) [52], which encompasses approximately 2.2% of the lagoon. The IRL in its entirety was divided into four regions (= sub-basins). The Banana River (BR) (202 km2) and the Mosquito Lagoon (ML) (140 km2) included each sub-basin in its entirety (Fig 1). The Indian River basin was divided into two sub-basins: the northern Indian River (NIR) (378 km2, previously defined as north of Eau Gallie Causeway [55], with little tidal and non-tidal flushing [56]), and the southern Indian River (SIR) (182 km2) which consisted of three previously defined basins [55] and includes three of the four inlets (Fig 1). Due to a lack of tidal flushing, the BR and NIR have decreased water quality compared to the majority of ML and SIR [56–58], and were the common epicenter of prior IRL UMEs [37,59]. By portioning the lagoon, we were able to compare our results with previous abundance studies [21,23] and with community estimates peripheral to the basins [60].
Sub-basins aligned N-S are: Mosquito Lagoon; Banana River; northern Indian River; southern Indian River. Contour lines and alternating saw-tooth transects were utilized throughout the lagoon as illustrated in the inset portion of the Banana River (upper right). The map was generated using ERSI software, ArcGIS 10.8 (http://www.esri.com/software/argis), using geographic data from Florida Fish and Wildlife Conservation Commission. Reprinted from [Florida Fish and Wildlife Conservation Commission. Florida Marine Research Institute. Atlas of Marine Resources, Version 1.3] under a CC BY license, with permission from [Richard Flamm, Florida Fish and Wildlife Conservation Commission], original copyright [2000].
Capture-recapture photo-identification surveys
We used a Robust Design photo-identification survey [16,61], which is the preferred method for estimating estuarine dolphin abundance [12,13,51]. The design grouped photo-identification surveys into four short-duration (less than three weeks) primary periods, each containing three replicate surveys (secondary sessions), during which population closure was assumed. The four primary periods were separated by longer duration time intervals (~three months) during which an open population was assumed. The Robust Design estimates abundance using closed mark-recapture models based on recaptures of individuals between the secondary sessions, and survival and temporary immigration/emigration using open mark-recapture models based on recaptures of individuals between the primary periods. Vessel based capture-recapture surveys were conducted in the IRL between August 2016 and May 2017, with the following assumptions: 1) Population closure within a primary period and emigration/immigration between primary sampling periods being temporary; 2) Dorsal fin marks were unique, permanent, and correctly identified; 3) Individual capture probabilities were equal within a secondary session, and 4) Individual survival probabilities were equal among individuals within a primary period [16,61,62]. To account for the potential occurrence of non-resident individuals, we included a transient parameter by modeling survival for individuals during the interval following the primary period of their first sighting differently than for subsequent intervals. Additional assumptions were that marked and unmarked animals did not differ in detection, movement, or survival parameters, and mixed randomly [16,61,62]. We defined four primary periods (summer = June-August, fall = September-November, winter = December-February, and spring = March-May) [63]. Primary periods contained three secondary sessions that were completed under optimal conditions with a Beaufort Sea State ≤3 (conditions = glassy to crests of some large wavelets breaking), in the shortest time period to meet the assumption of closure (target: ≤3 weeks). Secondary sessions (complete survey of the IRL) were separated by at least one day to allow for population mixing [13]. Existing dorsal fin catalogs were utilized and established protocols were closely followed [51]. The survey design used both depth contour lines and alternating saw-tooth transects (total length ~ 743 km) to minimize capture heterogeneity (Fig 1). Alternating saw tooth transects (2.5 km apart) were traversed in areas where lagoon width exceeded 1.25 km in either direction from the N-S contour line (based on 95% of prior sightings) [64]. Predetermined routes were downloaded into a Global Positioning System unit to ensure transect adherence. Each secondary survey covering the IRL bottlenose dolphin stock (Ponce Inlet-Jupiter Inlet) [65] was completed within one to three days (Fig 1).
Vessels (5–7 m) containing researchers (n ≥3) traveled at 10–12 knots to search along pre-determined tracks. Dorsal fins of encountered dolphins were photographed using a Canon EOS digital camera with a 100–400 mm telephoto lens (Canon USA, Inc., Melville, NY, USA). A dolphin group (= sighting) was defined as all dolphins within 100 m that were engaged in similar behavior with the same general heading [66]. Calves were identified as <75% adult size and swimming in adult echelon (i.e., in proximity to adult mid-lateral flank). Young-of-year (YOY) were <50% adult size, swimming in echelon, and characterized by: (a) dark coloration; (b) floppy dorsal fin; (c) presence of fetal lines; (d) extreme buoyancy; and (e) rostrum-first surfacing [67–69]. Animals not identified as calves or young of the year were considered adults. Additional data included: (a) time-location-GPS readings; (b) lat-long in decimal degrees; (c) behavior; (d) estimated water depth; (e) estimated group size/composition; (f) environmental covariates; and (g) sighting conditions (Beaufort Sea state, chop height, glare).
Photo-identification analyses and cumulative curve
Dorsal fin image analyses followed established protocols [70]. Briefly, marked dorsal fins were sorted by notch patterns, with the best photograph serving as the ‘type’ for each dolphin. Subsequently, unambiguous matches were accepted as re-identifications if a minimum of two experienced personnel agreed. A new individual was added to the catalog if distinctly marked but not matched with an existing photo.
To minimize false matches, images were graded for photographic quality based on a weighted scale of five characteristics [71]. Images were then assigned a quality score as follows: Q1 = excellent; Q2 = average; Q3 = poor. Dorsal fin distinctiveness was assigned a rating as follows: D1 = very distinct, D2 = moderately distinct, at least two features or one major feature; and D3 = not distinct, few to no features [71,72]. To avoid potential identification errors, only Q1 and Q2 quality photographs were used as individual detections in capture-recapture analyses. Thus mark-recapture estimates only applied to the identifiable individuals (D1 and D2).Unidentifiable individuals (D3; not meeting D1, D2 criteria), were used to estimate the proportion of marked individuals and, as discussed below, the total population size of an area. Because of the potential for changes in distinctiveness, non-independence from mothers, and other causes of capture heterogeneity, calves and YOY were excluded from capture-recapture analyses [51]. Dolphin survival can vary greatly by age class [73]; therefore, focusing on only adults ensured less capture heterogeneity and more reliable survival estimates.
To informally evaluate population closure, a cumulative curve of marked individuals (i.e., discovery curve) [74] was plotted based on the cumulative total number of distinct (D1 andD2) dolphins across each secondary session. Lastly, the percent of exchange of marked dolphins between sub-basins was evaluated.
Robust design models
We used program MARK [75] via package RMark [76] in R [77] to assess various Robust Design capture-recapture models. For each primary period, parameters estimated included dolphin abundance (N), the probability of apparent survival (φ), the probability of detection (p), and two temporary emigration parameters: the conditional probability of an animal not being available for capture (e.g., outside of the study area or within the study area but not available for detection) given that it was available (ϒ’’) or not available (ϒ’) during the previous primary period. Some models also included a time since initial capture (“transient”) parameter to allow separate estimates of survival for animals captured in more than one primary period-season (= residents) [78].
A total of 36 models were fit with combinations of structural covariates for detection, survival, and temporary emigration parameters. Detection models allowed detection to vary between primary periods (p(season)), secondary sessions (p(season*session)), or to be equal across all sampling occasions (p(.)). The probability of first detection and the probability of recapture was assumed equal for all models (no behavioral response). Survival was modeled as varying by primary period (ϕ(season)) or as a constant (ϕ(.)). A transient covariate was also included in some models to allow survival during interval following the primary period of an individual’s initial capture to be estimated separately from survival for all subsequent primary periods [78]. For those models, survival was allowed to vary by primary period (ϕ(transient*season)) or to be constant (ϕ(transient)). We assumed that transients were likely to be sighted in only one primary period and the set of dolphins captured for the first time in any primary period was a mixture of transients and residents. Therefore, survival estimates for the first primary period after initial capture (φ1) included apparent mortality (due to permanent emigration) and were likely biased low [79]. The survival estimates for the remaining observations (φ2) more accurately reflected resident survival since animals seen in more than one primary period (season) were likely residents. Assuming that the transients and residents had the same detection probability within the relatively short primary periods (and thus that the transients left between the primary periods), then the proportion of transients among the newly-marked animals (distinct dolphins seen for the first time) can be calculated as τi = φ i2/φ i1 where the i subscript refers to the primary period at the beginning of the survival interval. Although the number of transients could be computed as Ti = Ni/(Ni + mi), where Ni = the number of newly-marked individuals and mi = the number of previously marked individuals captured within time period i [79,80], we did not adjust abundance estimates by removing estimated transients. Temporary emigration was modeled as: 1) Markovian movement in which the probability of availability was dependent on the previous state (available or unavailable) (ϒ’’(.) ϒ’(.)); 2) random movement in which availability probability did not differ based on the previous state (ϒ’’ = ϒ’(.)); or 3) no movement models (ϒ” = ϒ’ = 0) with no temporary emigration. For Markovian and random movement models, the temporary emigration parameters were constrained to be constant throughout the study. Attempts made to model temporary emigration as varying over time resulted in some parameters not being identifiable when the transient covariate was included, potentially due to the number of primary periods (n = 4) [81]. We chose to constrain movement parameters to be equal over time for all models with temporary emigration as a conservative approach to comparing models.
The robust design lacks an overall goodness-of-fit test, therefore to evaluate model fit we used Fletcher’s generic goodness-of-fit statistic (c-hat) [82,83] calculated by program MARK for the two most parameterized models. This included both time (season) and “transient” (time since initial capture) effects on survival, time-specific (season and session) detection parameters, and random temporary emigration parameters. The assumption of population closure during primary periods was evaluated using goodness-of-fit statistics implemented in program Close Test version 3 [84]. Model comparisons were based on the small-sample adjusted Akaike Information Criterion AICc [85], calculated with the theoretical number of estimable model parameters (i.e., the column rank of the design matrix), rather than the estimated rank of the variance–covariance matrix. Models with ΔAICc < 10 were evaluated for evidence of numerical estimation errors (indicating a lack of parameter identifiability). Akaike weights were reported and convey the relative support compared to all candidate models for each model on a scale of zero to one [85]. Akaike weights and final abundance estimation were based on model sets including only similar movement sub-model types [85], excluding models with numerical estimation problems for some parameters (S1 Table).
To obtain abundance estimates for the total population (identifiable, D1+D2 and unidentifiable, D3 dolphins) an ad hoc adjustment was applied as follows [78]. The proportion of identified individuals (Ɵ) was calculated as the number of adult individuals that were marked (D1 +D2) divided by the total number of adult individuals (D1 + D2 + D3) observed during each primary period (season). Abundance estimates for the total population in each primary period (season) were then calculated by dividing the population estimates for identified individuals
by the proportion marked (Ɵ). The variance of each total population estimate was calculated as:
where n was the number of dolphins from which Ɵ was estimated. Log-normal 95% confidence intervals for total population size was calculated with a lower limit of
/C and an upper limit of
·C, where
in which z is the normal deviate, CV is the coefficient of variation, and α = 0.05. Density was estimated as dolphins/km2 by dividing total abundance as calculated above by IRL area: (902 km2).
Closed sub-basin abundance models
To obtain seasonal abundance estimates () for each of the four sub-basins, closed population capture-recapture abundance models were fit with program MARK [75] via package RMARK [76] in R [77]. These capture-recapture abundance models made the following assumptions: population closure (no births, deaths, immigration or emigration), unique marks that were permanent and correctly read, and equal capture probability for marked and unmarked animals (random mixing after first capture) [86,87]. A separate model was fit to each sub-basin by primary period (season), which allowed each secondary session to have a unique capture probability [88]. Multi-state robust design capture-recapture models could potentially have provided sub-basin-specific abundance estimates (and movement probabilities), but these models require many additional parameters. Given that data were limited to four primary periods [81], such complex models were prohibited (too many parameters for the available data). Although potentially losing some statistical power, fitting separate closed models produced reasonable abundance estimates with good precision and provided valuable baseline estimates for resource managers. Closed population capture-recapture abundance estimates were adjusted for unidentifiable individuals following the methods described above for Robust Design abundance estimates. Density by sub-basin was estimated as dolphins/km2 by dividing the total abundance estimates by the area of each sub-basin.
Results
Field effort and photo-identification
From August 2016 through May 2017, 135 vessel surveys (25 survey days) were conducted throughout the IRL during four capture-recapture primary periods (12 secondary sessions). Each secondary session (complete IRL survey replicate) employed 11–13 vessels (11.42 ± 1.00 SD) over 1–3 d (2.3 ± 0.65 SD). Each primary period was completed in 13–36 d (20.0 ± 10.9 SD). Vessel surveys ranged from 3.28–14.25 h (8.72 ± 2.21 SD; total field hours: 1,177.58 h). Over 159,000 photographs were taken of 1,465 dolphin groups, totaling 5,973 dolphins (S2 Table). Dorsal fin images used in analyses were Q1 = 87.3% or Q2 = 8.4%, while Q3 = 4.3% were excluded. Calves and YOYs comprised 19.2% of animals sighted. Mean group size was 4.08 (± 4.24 SD) (Table 1).
Cumulative curve of marked individuals and distribution patterns
Distinctively marked individuals (n = 503) were recorded, with 369 (73.4%) observed in one sub-basin. The southern Indian River had the greatest number with 124 (33.6%) followed by 115 (31.2%) in Mosquito Lagoon, 73 (19.8%) in the northern Indian River and 57 (15.4%) in the Banana River. A total of 112 marked individuals (22%) were seen in two sub-basins. Movements between three sub-basins were only observed between the NIR, BR and SIR, 22 individuals (4%). Of the individuals seen in two sub-basins, the greatest exchange occurred between the northern Indian River and the Banana River, followed by the northern Indian River and the southern Indian River (Table 2). Eighty-seven percent of the distinct individuals in Mosquito Lagoon were only observed there, while the remaining 13% were also observed in the northern Indian River (Table 2). Of the 503 distinctly marked animals, 84 (16.7%) were sighted during a single survey only (Fig 2). These 84 animals were distributed across all primary periods (fall = 17; spring = 18, summer = 29, and winter = 20) and sub-basins (Mosquito Lagoon = 23, Banana River = 17, northern Indian River = 22, southern Indian River = 22). Frequency of sightings of marked animals over the four seasons was one (n = 105; 20.9%), two (n = 155, 30.8%), three (n = 168, 33.4%) and four seasons (n = 75, 14.9%), and over the 12 secondary sessions frequency ranged from one to nine (Fig 2). Over the 12 secondary sessions, cumulative captures of marked animals increased steadily and plateaued over the final sessions (Fig 3). Within primary periods, a sharp decline in the number of new marked animals between sessions two and three was evident for nearly all periods (Fig 3).
Number of sightings (= number of secondary sessions out of 12) in which an individual was photographed.
Superimposed onto the plot is the cumulative curve of distinct (D1) and moderately distinct (D2) individuals.
Robust design and model selection
Fletcher c-hat statistic was 1.03 for the two most parameterized models, indicating no significant overdispersion in our data, thus models were not adjusted for overdispersion. The best supported models (12 of the 36) had AICc values <4 (S1 and S3 Tables). Of these, five had ΔAICc < 2 (= high support). The remaining seven had 2 < ΔAICc < 4 (= moderate support) [89]. Population closure during primary periods was supported except during spring when potential losses between secondary sessions one and two were indicated (S4 Table). The best supported model allowed survival to vary by primary period (season) and by time since initial capture (“transient”), and no temporary emigration (S3 Table). The second-best model also allowed survival to vary by time since initial capture but had random temporary emigration. All of the best supported models allowed a unique detection probability estimate for each secondary session within seasons. Model average estimates of detectability ranged from 0.20 to 0.42 between secondary sessions (Table 3, S2 Fig). The model average estimates of seasonal (three month) marked dolphin survival for residents ranged from 0.85–1.00 (Table 4).
Temporary emigration models with random movement (combined Akaike model weight = 0.42) had more support than those with no movement (combined Akaike model weight = 0.34) or Markovian movement (combined Akaike model weight = 0.22) (S1 Table). Based on model average estimates of Markovian models, the probability that a dolphin became unavailable if it had been available in the previous primary period (ϒ”) was quite low with the model average estimate being 0.05 (SE = 0.05). The probability was greater that a dolphin remained unavailable for observation (outside the study area) given that it had been unavailable in the previous primary period (ϒ’), but estimates lacked precision 0.48 (SE = 0.44). The estimated proportion of transients (transient rate) among the marked population ranged from 0.06 in winter to 0.07 in fall, the only seasons for which the proportion of transients could be estimated. The proportion of marked individuals ranged from 0.40–0.45 between primary periods (Table 5). Abundance for the IRL Estuarine System stock ranged from 981 (95% CI: 882–1,090) dolphins in the winter season to 1,078 (95% CI: 968–1,201) dolphins in the summer season (Table 5, S3 Fig). Mean estimated dolphin abundance was 1,032 (95% CI: 969–1,098) (Table 5). IRL density ranged from 1.09–1.20 dolphins/km2 (1.15 ± 0.05 SD) (Table 5).
Abundance by sub-basin (closed models)
The greatest mean abundance was in the southern Indian River, followed by the Banana River and northern Indian River; while the lowest mean abundance was observed in Mosquito Lagoon (Table 6, S4 Fig). Mean estimates of detectability were greatest for the Banana River (0.40) and Mosquito Lagoon (0.32), followed by the Northern Indian River (0.23) and the Southern Indian River (0.20) (S5 Table). Dolphin density (dolphins/km2) varied by sub-basin with the largest mean density in the southern Indian River (2.00 ± 0.60 SD), followed by the Banana River (1.72 ± 0.57 SD), Mosquito Lagoon (1.23 ± 0.32 SD), and the northern Indian River (0.91 ± 0.27 SD) (Table 6).
The mean percent of identified individuals 43 ± 0.12 SD varied between sub-basins with the largest occurring in the southern Indian River (58.8%), followed by Mosquito Lagoon (47.0%), the Northern Indian River (38.0%) and the Banana River (29.4%) (Table 6).
Discussion
Estimating the abundance and demography of bay, sound and estuarine bottlenose dolphin populations pose challenges due to shallow water inaccessibility, the uncertainty of population boundaries, and the movement of non-residents within the system [1–3,13,90,91]. Here we implemented a Robust Design photo-identification survey to estimate bottlenose dolphin abundance for the Indian River Lagoon estuarine system stock (Florida) while accounting for both temporary emigration by residents and occurrence of transient (non-resident) dolphins. Our estimates achieved good precision and were comparable to those derived from recent IRL studies [21], suggesting that the population size has been stable over the last decade. Additionally, our study provided estimates of high seasonal survival, low transient rate and moderate temporary emigration rate. Model selection preference for models without temporary emigration (the ’no movement’ model) and the low estimated transient rate supported a resident IRL dolphin stock, in agreement with results from prior studies [18,19]. While the observed transient rate was low, by excluding potential transients we were able to more accurately estimate resident IRL dolphin survival by separating dolphins sighted in only one season (= transients) from those sighted in more than one season. These data provide critical guidance for a population whose diminishing health, coupled with anthropogenic impacts, likely promotes a mortality rate that exceeds potential biological removal [37].
Demographic and abundance parameters
Our capture-recapture study estimated a transient rate of 7%, which was less than the observed rate of single sightings (e.g., 84 of 503 marked dolphins were sighted only once, with 27% of these in Mosquito Lagoon). This illustrates the utility of the model allowing for transients to be separated from residents that were potentially only seen once. The low estimate of transient rate further supported the use of a closed Robust Design. Within primary periods, a sharp decline in the number of new marked animals between sessions two and three was evident for nearly all periods (exception was a slight winter increase; 13 to 19 new marked animals between sessions), supporting population closure within the primary periods [74]. Furthermore, the majority of those that were sighted only once in Mosquito Lagoon (57%) also inhabited the adjacent Halifax River estuary to the north, and a few ranged further north into the JES (S6 Table) [25]. However, while this low rate suggests transients did not play a large role in the population biology of IRL dolphins, results should be interpreted with the understanding that rates depend on how study boundaries were defined which likely did not perfectly align with biological factors. Furthermore, transient rates may be driven by environmental conditions, as found in a recent study [24]. Although rare (4.3%), it is possible that excluded poor quality images contributed to some resident dolphins being categorized as transients. Since apparent survival is a product of true survival and site fidelity [92], it was expected that the inclusion of a transient parameter would increase survival rates for marked IRL adults by reducing the effect of non-residents on apparent survival (i.e., reducing negative bias). However, including the transient parameter may have over corrected and biased resident survival upward by removing individuals from those estimates with the lowest survival [93]. This could be reduced in future studies by including covariates that identify dolphins known to range beyond the study area.
Evidence was mixed for the occurrence of temporary emigration, but model selection generally supported non-structured movement (i.e., random), suggesting that some dolphins were, at times, unavailable for re-sighting. Although less supported, the Markovian movement model indicated that while IRL dolphins rarely temporarily emigrate from the study area, those that do may remain absent for more than one season. Likewise, we found the probability of an individual not being available during a primary period if it was not available in the prior session was greater (48%) than if it was available in the prior session (5%). Temporary emigration, as measured by the robust design, can be caused by true absence from the study area, or by individuals present but unobservable during a primary period. Individuals inhabiting the periphery of the study region, with home ranges extending beyond the IRL border, may have contributed to such temporary emigration. Resident dolphins have also been documented utilizing oceanic habitat [49]. Availability bias [94] may also occur due to dolphins utilizing the labyrinth of canals, freshwater creeks, and shallow waters surrounding islands throughout the IRL [49,64]. This bias could be potentially mitigated by including covariates that model detection heterogeneity such as distance from shallow labyrinth or inlets.
Survival rates for identified residents varied widely across movement models. A trend of initially high resident survival, which decreased between the first (fall-winter) and the second (winter-spring) interval, was evident across all models that included both transient and season effects. While variation in survival between seasons may have occurred, the reduction may instead be due to the limited number of primary periods employed. Low-biased survival at the end of a study (the last survival estimates in a time series) has been noted in prior studies, and is a potential result of constraints or biased estimates at the end of the time series [81]. During our study, cases of documented marked adult mortality were low (n = 3, S7 Table), although may be underrepresented since even in the well-studied Sarasota Bay estuary, only one-third of dolphin carcasses were recovered [95]. Marked resident survival in the first (fall-winter) time period, range 0.98–1.00 more closely matched observed mortality. Thus, the first time period likely provides the most reliable estimate of survival for utilization by resource managers. Longer-term studies with increased primary periods should improve survival estimation.
Detection probability between secondary sessions ranged from 0.20–0.42, values which fell within recommended bounds of effective capture-recapture survey designs (0.2–0.3; [51]). Detection was least in fall, and likely influenced by an increased sea state (waves) common to that season [96]. The mean proportion of marked dolphins (0.43 ± 0.12) was less than other study systems with mean rates of 0.79 and 0.72 respectively [13,97]. Differences in dorsal fin marking rates between populations may be influenced by ecosystems [98]. Since extensive portions of the IRL are relatively isolated [56] in contrast to more open bays [2,13,90,91], heterogeneity between regions (e.g., anthropogenic activities, conspecific interactions) could contribute to differences in marked ratios. Furthermore, in other study regions, rates of dorsal fin marking have a sex-bias, with significantly higher rates of dorsal fin nicks in males [99]. Potential heterogeneity in capture probability may occur as larger groups contain more calves [100], and by extension female dolphins with less propensity to be marked [99]. Interestingly, the marked ratio varied by sub-basin, with the greatest in the southern Indian River (0.588) and the least in the Banana River (0.294). Dolphins in the southern lagoon have the highest prevalence of boat-injuries that yield dorsal fin disfigurement, in relation to other sub-basins [101], and this may contribute to variability in marks among sub-basins. Much of the northern Banana River prohibits motorized vessels and has no public usage [102], thereby limiting entanglement, vessel strikes, and capture-release activities for dolphin health assessment that may influence dorsal fin marking [8,101,103].
Abundance and distribution patterns
Our mean dolphin abundance and density estimates for the Indian River Lagoon stock were similar to those derived from multi-year line transect aerial surveys [21]. Both the current study and prior aerial efforts [21] observed the greatest mean abundance in the southern Indian River sub-basin and the lowest in Mosquito Lagoon. Only slight changes in abundance were observed between seasons, in contrast to the aerial surveys, in which dolphin abundance was greater in winter than summer [21]. Winter aerial surveys were conducted during several unusually cold periods, which may have contributed to increased abundance in the Mosquito Lagoon and the southern Indian River [21]. While aerial survey estimates of dolphin detectability did not vary seasonally [21], these extreme cold events possibly influenced abundance estimates through heterogeneity in availability due to dolphin behavior. Longer-term mark-recapture studies in northeastern Florida (2011–2016), have found evidence for a winter influx of transients into Mosquito Lagoon from adjacent northern estuaries [24] which also may influence abundance estimates. Seasonal variance in abundance for the southern Indian River (i.e., lower summer abundance and increased winter abundance) was similar to that observed in previous studies [21], although the cause of this variation has yet to be determined. Studies aimed at estimating IRL dolphin abundance will ultimately need to address movements of animals in and out of the study area which may be more common near the perimeters.
Although we found limited evidence for transience and temporary emigration, a prior study identified two separate dolphin communities in Mosquito Lagoon, with the northern community extending beyond the IRL boundary [60]. These findings, coupled with studies that indicate movements of IRL dolphins beyond the northern boundary [24,25], suggest this sub-basin may have higher temporary emigration and transient rates, causing greater fluctuations in abundance [21]. The majority of dolphins inhabiting the Mosquito Lagoon sub-basin have been found to exhibit strong site fidelity (71% exclusive to ML) [104], which was also observed in this study (87% of Mosquito Lagoon individuals were only observed therein). These results support Mosquito Lagoon dolphins being relatively isolated from the rest of the IRL. The current study found the greatest sub-basin exchange between the northern Indian River and the Banana River. Results are similar to a prior study which found four communities within the Indian and Banana Rivers, with one occupying a portion of the NIR and BR, and three in the SIR (this study); all overlapping adjacent basins [60]. Movements observed within the IRL coincided with previously described IRL dolphin communities, further supporting the lack of connectively between ML dolphins and the other IRL sub-basins.
Dolphin group size may be influenced by environmental factors, with smaller groups predominating in shallow estuarine waters [66,105–108]. Slightly larger mean dolphin group sizes were observed in this study than was recorded in the aerial surveys [21]. The proportion of calves (including YOY) in this study (19.2% of observations) was significantly larger than the 5.42% estimated from the aerial survey [21]. However, differences were likely due to the conservative methods utilized to define calves during aerial surveys (half the size of the adult) which would be expected to under-count calves. Our results were comparable to prior vessel-based IRL studies where calves constituted 24% of encounters [100], and to Sarasota Bay where calves constituted 21.5% [109].
Management implications
A goal of future capture-recapture studies should be to improve stock delineation and community structure and to better quantify transient movements. Depending on management objectives of future studies, our parameter estimates could be used in simulations to identify opportunities to reduce sampling effort or increase precision. For example, variability of survival estimates increased during the current study, possibly influenced by the low number of primary periods (n = 4). Conducting future survey efforts over a longer time period, with an increased number of primary periods, could provide more precision in estimating temporal parameters and thus improve survival estimates. Incorporating long-term monitoring designs and comprehensive ecosystem monitoring approaches [110] may also help increase efficiencies. Lastly, the northern portion of the study area did not incorporate the full extent of the home range for dolphins in Mosquito Lagoon [21,24,25]. This could potentially cause availability bias due to temporary emigration; therefore, future studies should give further consideration to stock delineation in this area.
Effective management of the IRL dolphin stock requires data on distribution and abundance. Estimates derived herein demonstrate the feasibility of utilizing a robust design methodology and provide the first estimates of temporary emigration, transient rate, and survival for this population. The most reliable estimate of marked resident survival for IRL dolphins in this study was similar to that recorded in other bays and estuaries [31,111]. However, IRL dolphins have experienced recurring mortality peaks (2001, 2008, 2013, 2013–2015) where annual mortality peaked at 77 dolphins [37,59]. Variability in annual mortality is quite high for this population [37,59], and diminished ecosystem health [38], fisheries-related takes, and other anthropogenic activities continue to threaten this compromised dolphin population. Thus, our parameter estimates will provide critical guidance to stock management.
Supporting information
S1 Fig. Parameter estimates from robust design analysis of 1000 simulated data sets.
Dolphin populations were set with initial size 1000, a detection parameter of 0.3, survival at 0.95, and the rate of gamma prime (ϒ’ = probability of a dolphin being unavailable for observation if unavailable in the prior primary period) and gamma double prime (ϒ” = probability of a dolphin being unavailable if available in the prior primary period) were both set at 0.1. *Mean parameter estimate = dot, median = triangle, vertical line = true parameter value.
https://doi.org/10.1371/journal.pone.0250657.s001
(TIF)
S2 Fig. Model averaged estimates (95%CI) of detection of marked bottlenose dolphins during Indian River Lagoon surveys (Florida east-central coast) by secondary session.
Detection was calculated using Robust Design models for capture-recapture in the Indian River Lagoon Estuarine System.
https://doi.org/10.1371/journal.pone.0250657.s002
(TIF)
S3 Fig. Estimated abundance (total and marked) derived for bottlenose dolphins from the Indian River Lagoon, Florida (2016–2017).
Marked abundance estimates include marked animals only. Total abundance estimates were adjusted for the ratio of marked: Unmarked individuals.
https://doi.org/10.1371/journal.pone.0250657.s003
(TIF)
S4 Fig. Estimated dolphin abundance estimated by sub-basin and primary period using closed population capture-recapture abundance models.
Total dolphin abundance estimates were obtained by adjusting for the ratio of marked: Unmarked individuals observed in each sub-basin of the Indian River Lagoon (2016–2017).
https://doi.org/10.1371/journal.pone.0250657.s004
(TIF)
S1 Table. Results of robust design model selection.
Parameters include: Apparent survival (φ), capture probability (p), the probability being unavailable for capture if the individual was available in the prior period (ϒ’’) or unavailable in the prior period (ϒ’) and were constrained constant (.) or allowed to vary by primary period (season) and/or secondary session. Models allowed apparent transient survival to be constant (φ (transient)), or to vary by primary period (φ(transient*season)). Three models (italics) had numerical estimation problems and were not included in inferential procedures.
https://doi.org/10.1371/journal.pone.0250657.s005
(XLSX)
S2 Table. Observed and photographed bottlenose dolphin groups (sightings) in the Indian River Lagoon, compiled by sub-basin and primary period.
Primary periods as follows: Summer = 10–23 Aug. 2016, Fall = 1–5 Nov. 2016, Winter = 18 Jan.-3 Feb. 3, 2017, Spring = 17 Apr.-23 May 2017.
https://doi.org/10.1371/journal.pone.0250657.s006
(XLSX)
S3 Table. Estimated survival and movement parameters (95% CI) for the 12 best supported models for Indian River Lagoon bottlenose dolphins.
Standard errors were calculated using profile likelihood estimation. To aid estimate comparisons, models with (a) and without (b) the transient parameter (time since first capture) are grouped. Time-specific adult survival rates (S) are presented for transients and residents, and survival estimates for time constant models are listed in the first time-specific row. Parameters include apparent survival (φ), model averaged estimates of adult survival rates (S) between primary periods, and temporary emigration: Conditional probability of an animal not being available for capture if it was available during the previous primary period (ϒ’’) or not available (ϒ’).
https://doi.org/10.1371/journal.pone.0250657.s007
(XLSX)
S4 Table. Population closure test for Indian River Lagoon dolphins suggesting population closure for all primary periods except for spring.
Test subcomponent statistics indicated lack of closure in spring was due to losses between secondary sessions one and two.
https://doi.org/10.1371/journal.pone.0250657.s008
(XLSX)
S5 Table. Estimated detection (p) (95% CI) of marked bottlenose dolphins by sub-basin and seasonal replicate (= secondary session) during Indian River Lagoon surveys (Florida east-central coast).
Detection was estimated using closed capture-recapture models for Indian River Lagoon dolphins (2016–2017).
https://doi.org/10.1371/journal.pone.0250657.s009
(XLSX)
S6 Table. Sighting history for thirteen bottlenose dolphins sighted only once in Mosquito Lagoon (considered potential transients).
Sighting data were collected between 2008 and the end of the current study. All animals listed had prior sightings north of the Indian River Lagoon in the Halifax River. *Indicate animals known to range north into the Jacksonville Estuarine System stock.
https://doi.org/10.1371/journal.pone.0250657.s010
(XLSX)
S7 Table. Marked adult dolphins sighted during secondary sessions and subsequently recovered deceased during the study.
https://doi.org/10.1371/journal.pone.0250657.s011
(XLSX)
S1 Text. Simulation study methods and results.
https://doi.org/10.1371/journal.pone.0250657.s012
(DOCX)
Acknowledgments
We are very grateful to staff from the Kennedy Space Center Ecological Program Office, Kennedy Space Center, Cape Canaveral Air Force Station, and U.S. Fish and Wildlife Service for their assistance with entry into portions of the study area. We thank Megan Stolen for assistance with stranding response and data collection. We sincerely thank our colleagues, Debbie Wingfield (Volusia County Environmental Management), Lisa Gemma, and St. Johns River Water Management District for providing critical field assistance and vessel support which made the project feasible. Likewise, we are grateful to those that were essential in the field as well as with photographic analyses: Amy Brossard, Heidy Clifford, Mackenzie Daniel, Agatha Fabry, Brandy Nelson, Samantha Nekolny, Jessie Stevens, and Elizabeth Titcomb. We sincerely appreciate Nicole Mader (Dolphin Ecology Project) enabling the study by providing essential coverage and photographic analyses in the southern end of the study area. Furthermore, we are grateful to the countless volunteers and interns who contributed to data collection and management. This study would not have been possible without our colleagues’ perseverance, talent and the dedication, for which we have the utmost gratitude. We thank anonymous reviewers and Michael Douglas whose suggestions significantly improved this manuscript. Lastly, we thank Lance Garrison for his assistance with the design of the project and continued guidance throughout the study.
References
- 1. McDonald TL, Hornsby FE, Speakman TR, Zolman ES, Mullin KD, Sinclair C, et al. Survival, density, and abundance of common bottlenose dolphins in Barataria Bay (USA) following the Deepwater Horizon oil spill. Endanger Species Res. 2017; 33: 193–209.
- 2. Mullin KD, McDonald T, Wells RS, Balmer BC, Speakman T, Sinclair C, et al. Density, abundance, survival, and ranging patterns in common bottlenose dolphins (Tursiops truncatus) in Mississippi Sound following Deepwater Horizon oil spill. PLos ONE, 2017; 12(10):e0186265. pmid:29053728
- 3. Lane SM, Smith CR, Mitchell J, Balmer BC, Barry KP, McDonald T, et al. Reproductive outcome and survival of common bottlenose dolphins sampled in Barataria Bay, Louisiana, USA, following the Deepwater Horizon oil spill. Proc. Biol Sci. 2015; 282(1818): 20151944. pmid:26538595
- 4. Read AJ, Drinker P, Northridge S. Bycatch of marine mammals in U.S. and global fisheries. Conserv Biol. 2006; 20(1): 163–169. pmid:16909669
- 5. 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. Conserv Biol. 2006; 20(6): 1791–1798. pmid:17181814
- 6. Weilgart LS. The impacts of anthropogenic ocean noise on cetaceans and implications for management. Can J Zool. 2007; 85(11): 1091–1116.
- 7. Taylor BL, Rojas-Bracho L, Moore J, Jaramillo-Legorreta A, Ver Hoef JM, Cardenas-Hinojosa G, et al. Extinction is imminent for Mexico’s endemic porpoise unless fishery bycatch is eliminated. Conserv Lett. 2016; 10(5): 588–595.
- 8. Stolen M, Durden WN, Mazza T, Barros N, St. Leger J. Effects of fishing gear on bottlenose dolphins (Tursiops truncatus) in the Indian River Lagoon system, Florida. Mar. Mamm. Sci. 2013;
- 9. Baulch S, Perry C. Evaluating the impacts of marine debris on cetaceans. Mar Pollut Bull. 2014; 80(1–2): 210–221. pmid:24525134
- 10. Hiby AR, Hammond PS. Survey techniques for estimating abundance of cetaceans. Rep Int Whal Commn. 1989; 11(2): 47–80.
- 11.
Buckland ST, Anderson DR, Burnham KP, Laake JL. Distance sampling: Estimation of biological populations. New York: Chapman and Hill; 1993.
- 12. Read A.J, Urian KW, Wilson B, Waples DM. Abundance of bottlenose dolphins in the bays, sounds, and estuaries of North Carolina. Mar Mamm Sci. 2003; 19: 59–73.
- 13. Balmer BC, Wells RS, Nowacek SM, Nowacek DP, Schwacke LH, McLellan WA. et al. Seasonal abundance and distribution patterns of common bottlenose dolphins (Tursiops truncatus) near St. Joseph Bay, Florida, USA. J Cetacean Res Manag. 2008; 10: 157–167.
- 14. Dawson S, Wade P, Slooten E, Barlow J. Design and field methods for sighting surveys of cetaceans in coastal and riverine habitats. Mamm Rev. 2008; 38(1): 19–49.
- 15. Williamson LD, Brookes KL, Scott BE, Graham IM, Bradbury G, Hammond PS. et al. Echolocation detections and digital video surveys provide reliable estimates of relative density of harbor porpoises. Methods Ecol Evolut. 2016; 7(7): 762–769.
- 16. Kendall WL, Pollock KH, Brownie C. A likelihood-based approach to capture-recapture estimation of demographic parameters under the robust design. Biometrics. 1995; 51: 293–308. pmid:7766783
- 17. Pollock KH, Nichols JD, Brownie C, Hines JE. Statistical inference for capture-recapture experiments. Wildlife Monogr. 1990; 107: 1–97.
- 18.
Odell DK, Asper ED. Distribution and movements of freeze-branded bottlenose dolphins in the Indian and Banana Rivers, Florida. In Leatherwood S, Reeves R, editors. The bottlenose dolphin. San Diego: Academic Press; 1990. pp. 515–540.
- 19. Mazzoil M, Mc Culloch S, Defran RH. Observations of the site fidelity of bottlenose dolphins (Tursiops truncatus) in the Indian River Lagoon, Florida. Fla Sci. 2005; 68: 217–226.
- 20.
NOAA. Marine mammal stock assessment reports (SARs) by species/stock: Bottlenose dolphin (Tursiops truncatus) Indian River Lagoon estuarine system stock. NOAA Fisheries, Office of Protected Resources. 2009 [cited 2019 30 October]; https://www.fisheries.noaa.gov/webdam/download/70874369.
- 21. Durden WN, Stolen ED, Jablonski TA, Puckett SA, Stolen MK. Monitoring seasonal abundance of Indian River Lagoon bottlenose dolphins (Tursiops truncatus) using aerial surveys. Aquat Mamm. 2017; 43(1): 90–112.
- 22. Leatherwood S. Aerial survey of the bottlenosed dolphin, Tursiops truncatus, and the West Indian manatee, Trichechus manatus, in the Indian and Banana Rivers, Florida. Fish Bull. 1979; 77: 47–59.
- 23. Durden WN, Stolen ED, Stolen MK. Abundance, distribution, and group composition of Indian River Lagoon bottlenose dolphins (Tursiops truncatus). Aquat Mamm 2011; 37(2): 175–186,
- 24. Mazzoil M, Gibson Q, Durden WN, Borkowski R, Biedenbach G, McKenna Z, et al. Spatiotemporal movements of common bottlenose dolphins (Tursiops truncutus truncatus) in Northeast Florida, USA. Aquat Mamm. 2020; 46(3): 285–300.
- 25. Nekolny SR, Denny M, Biedenbach G, Howells EM, Mazzoil MM, Durden WN, et al. The effects of study area size on home range estimates of bottlenose dolphins (Tursiops truncatus). Curr Zool. 2017; 63(6): 693–701. pmid:29492031
- 26. Richards VP, Greig TW, Fair PA, McCulloch SD, Politz C, Natoli A, et al. Patterns of population structure for inshore bottlenose dolphins along the eastern United States. J Hered. 2013; 104(6): 765–778. pmid:24129993
- 27.
Rodgers, SE. Population structure and dispersal of bottlenose dolphins (Tursiops truncatus) of the Indian River Lagoon estuary, Florida and adjacent Atlantic waters. M.Sc. Thesis, Florida Atlantic University. 2013. Available from: https://media.proquest.com/media/pq/classic/doc/3142735451/fmt/ai/rep/NPDF?_s=BdYlL7JZZzKu6T4d42rKywX%2Fmso%3D..
- 28. Würsig B, Würsig M. The photographic determination of group size, composition, and stability of coastal porpoises (Tursiops truncatus). Science. 1977; 18: 755–756.
- 29. Würsig B, Jefferson TA. Methods of photo-identification for small cetaceans. Rep Int Whal Commn. 1990; 12: 43–52.
- 30. Hammond PS, Mizroch SA, Donovan GP. Individual Recognition of cetaceans: Use of photo-identification and other techniques to estimate population parameters. Rep Int Whal Commn. 1990; 12: 440 pp.
- 31. 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. J Cetacean Res Manag. 2010; 11(2): 153–162.
- 32. Cantor M, Wedekin LL, Daura-Jorge FG, Rossi-Santos, MR. Assessing population parameters and trends of Guiana dolphins (Sotalia guianensis): An eight-year mark-recapture study. Mar Mamm Sci. 2011; 28(1): 63–83.
- 33. Félix F, Castro C, Laake JL, Haase B, Scheidat M. Abundance and survival estimates of the southeastern Pacific humpback whale stock from 1991–2016 photo-identification surveys in Ecuador. J Cetacean Res Manag. 2011; 3: 301–307.
- 34. Nicholson K, Bejder L, Allen SJ, Krützen M, Pollock KH. Abundance, survival and temporary emigration of bottlenose dolphins (Tursiops sp.) off Useless Loop in the western gulf of Shark Bay, Western Australia. Mar Freshw Res. 2012; 63(11): 1059–1068.
- 35. Santostasi NL, Bonizzoni S, Bearzi G, Eddy L, Gimenez O. A robust-design capture-recapture analysis of abundance, survival and temporary emigration of three odontocete species in the Gulf of Corinth, Greece. PLoS ONE. 2016; 11(12): e0166650. pmid:27926926
- 36. Sprogis KR, Pollock KH, Raudino HC, Allen SJ, Kopps AM, Manlik, O. et al. Sex-specific patterns in abundance, temporary emigration and survival of Indo-Pacific bottlenose dolphins (Tursiops aduncus) in coastal and estuarine waters. Front Mar Sci. 2016; 3 (12): 1–14.
- 37.
NOAA. Marine mammal stock assessment reports (SARs) by species/stock: Bottlenose dolphin (Tursiops truncatus) Indian River Lagoon estuarine system stock. NOAA Fisheries Office of Protected Resources. 2015 [cited 2020 7 January]; Available from: https://www.fisheries.noaa.gov/webdam/download/70874365.
- 38.
Morris LJ, Chamberlain RC, Jacoby CA. Summary report for the northern Indian River Lagoon. In: Yarbro L, Carlson PR, editors. Seagrass integrated mapping and monitoring report no. 2. Fish and Wildlife Research Institute Technical Report, TR-17 version 2. Florida: Fish and Wildlife Research Institute; 2016. pp. 266-.281.
- 39. Barros NB, Wells RS. Prey and feeding patterns of resident bottlenose dolphins (Tursiops truncatus) in Sarasota Bay, Florida. J Mammal. 1998; 79(3): 1045–1059.
- 40. Durden WN, Stolen MK, Adams DH, Stolen ED. Mercury and selenium concentrations in stranded bottlenose dolphins from the Indian River Lagoon system, Florida. Bull Mar Sci. 2007; 81(1): 37–54.
- 41. Bossart GD. Emerging diseases in marine mammals: from dolphins to manatees. Microbe. 2007; 2: 544–547.
- 42. Murdoch ME, Reif JS, Mazzoil M, McCulloch SD, Fair PA, Bossart GD. Lobomycosis in bottlenose dolphins (Tursiops truncatus) from the Indian River Lagoon, Florida: Estimation of prevalence, temporal trends and spatial distribution. EcoHealth. 2008; 5(3) 289–297; pmid:18712442
- 43. Durden WN, St. Leger J, Stolen M, Mazza T, Londono C. Lacaziosis in bottlenose dolphins (Tursiops truncatus) in the Indian River Lagoon, Florida, USA. J Wildl Dis. 2009; 45: 849–856. pmid:19617500
- 44. Noke WD, Odell DK. Interactions between the Indian River Lagoon blue crab fishery and the bottlenose dolphin Tursiops truncatus. Mar Mamm Sci. 2002; 18: 819–832.
- 45.
Kenney RD. Bottlenose dolphins off the northeastern United States. In Leatherwood S, Reeves R, editors. The bottlenose dolphin. San Diego: Academic Press; 1990. pp. 369–386. https://doi.org/10.1016/B978-0-12-440280-5.50025–1
- 46. 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. Mar Mamm Sci. 1999, 15: 394–408.
- 47. Gubbins CM, Caldwell M, Barco S, Rittmaster K, Bowles N, Thayer V. Abundance and sighting patterns of bottlenose dolphins (Tursiops truncatus) at four northwest Atlantic coastal sites. J Cetacean Res Manag, 2003, 5(2): 141–147.
- 48. Torres LG, McLellan WA, Meagher E, Pabst DA. Seasonal distribution and relative abundance of bottlenose dolphins, Tursiops truncatus, along the U.S. mid-Atlantic coast. J Cetacean Res Manag. 2005, 7(2): 153–161.
- 49. Hartel EF, Durden WN, O’Corry-Crowe G. Testing satellite telemetry within narrow ecosystems: nocturnal movements and habitat use of bottlenose dolphins within a convoluted estuarine system. Anim Biotelemetry. 2020; 8(13): 2020.
- 50.
NOAA. Marine mammal stock assessment reports (SARs) by species/stock: Bottlenose dolphin (Tursiops truncatus) Jacksonville estuarine system stock. NOAA Fisheries Office of Protected Resources. 2015 [cited 2020 7 January]; Available from: https://www.fisheries.noaa.gov/webdam/download/71600796.
- 51.
Rosel PE, Mullin KD, Garrison L, Schwacke L, Adams J, Balmer B, et al. Photo-identification capture-mark-recapture techniques for estimating abundance of bay, sound and estuary populations of bottlenose dolphins along the U.S. East coast and Gulf of Mexico: A Workshop Report. NOAA Technical Memorandum, National Marine Fisheries Service, NMFS-SEFSC-621. 2011. pp. 1–30.
- 52. Gilmore RG Jr. Fishes of the Indian River Lagoon and adjacent waters, Florida. Bulletin of the Florida State Museum Biological Sciences. 1977; 22: 103–147.
- 53. Mulligan TJ, Snelson FF Jr., Summer season population of epibenthic marine fishes in the Indian River Lagoon system, Florida. Fla Sci. 1983; 46: 250–276.
- 54. National Estuary Program. National Estuary Program, Indian River Lagoon Comprehensive Conservation and Management Plan (IRLCCMP). 1996 [cited 2019 30 October]; Available from: https://onelagoon.org/wp-content/uploads/irl_1996_ccmp.pdf.
- 55.
Woodward-Clyde Consultants. Biological resources of the Indian River Lagoon, National Estuary Program Final Technical Report, 92F274C. Florida: Woodward-Clyde Consultants; 1994.
- 56. Smith NP. Tidal and nontidal flushing of Florida’s Indian River Lagoon. Estuaries. 1993; 16(4): 739–746.
- 57. Sigua GC, Steward JS, Tweedale WA. Water-quality monitoring and biological integrity assessment in the Indian River Lagoon, Florida: Status, trends, and loadings (1988–1994). Environ Manage. 2000; 25(2): 199–209. pmid:10594193
- 58. Sigua GC, Tweedale WA. Watershed scale assessment of nitrogen and phosphorus loadings in the Indian River Lagoon basin, Florida. J Environ Manage. 2003; 67: 363–372. pmid:12710924
- 59. Stolen MK, Durden WN, Odell DK. Historical synthesis of bottlenose dolphin (Tursiops truncatus) stranding data in the Indian River Lagoon system, Florida, from 1977–2005. Fla. Sci. 2007; 70(1): 45–54.
- 60. Titcomb EM, O’Corry-Crowe G, Hartel E, Mazzoil MS. Social communities and spatiotemporal dynamics of association patterns in estuarine bottlenose dolphins. Mar Mamm Sci. 2015; 31(4): 1314–1337,
- 61. Pollock KH. A capture-recapture design robust to unequal probabilities of capture. J Wildl Manage. 1982; 46: 752–757.
- 62. Pollock KH, Nichols JD, Brownie C, Hines JE. Statistical inference for capture-recapture experiments. Wildlife Monogr. 1990; 107: 1–97.
- 63.
Shane SH. Behavior and ecology of the bottlenose dolphin at Sanibel Island, Florida. In: Leatherwood S, Reeves RR, editors. The bottlenose dolphin San Diego: Academic Press; 1990. pp. 245–265.
- 64. Durden WN, O’Corry-Crowe G, Shippee S, Jablonski T, Rodgers S, Mazzoil M, et al. Small-scale movement patterns, activity budgets, and association patterns of radio-tagged Indian River Lagoon bottlenose dolphins (Tursiops truncatus). Aquat Mamm. 2019; 45(1): 66–87.
- 65.
NOAA. Marine mammal stock assessment reports (SARs) by species/stock: Bottlenose dolphin (Tursiops truncatus) Indian River Lagoon estuarine system stock. NOAA Fisheries Office of Protected Resources. 2013 [cited 2019 30 October]; Available from: https://www.fisheries.noaa.gov/webdam/download/70874367.
- 66.
Wells RS, Irvine AB, Scott MD. The social ecology of inshore odontocetes. In: Herman LM, editor. Cetacean behavior: Mechanisms and functions. New York: Wiley & Sons; 1980. pp. 263–317.
- 67.
Cockcroft VG, Ross JB. Early development of a captive calf. In: Leatherwood S, Reeves RR, editors. The bottlenose dolphin. San Diego: Academic Press; 1990. pp. 461–478.
- 68.
Mead JG, Potter CW. Natural history of bottlenose dolphins along the central Atlantic coast of the United States. In: Leatherwood S, Reeves RR, editors. The bottlenose dolphin. San Diego: Academic Press; 1990. pp. 165–195.
- 69. Mann J, Smuts BB. Natal attraction: Allomaternal care and mother-infant separations in wild bottlenose dolphins. Anim Behav. 1998; 55: 1097–1113. pmid:9632497
- 70. Mazzoil M, McCulloch SD, Defran RH, Murdoch E. Use of digital photography and analysis of dorsal fins for photo-identification of bottlenose dolphins. Aquat Mamm. 2004; 30: 209–219.
- 71.
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 catalogue contributors. National Oceanic and Atmospheric Administration Report, NMFS-SEFSC-425. North Carolina: National Oceanic and Atmospheric Administration; 1999. pp. 298–321. https://doi.org/10.1098/rspb.1999.0746 pmid:10380684
- 72. Urian KW, Gorgone A, Read A, Balmer B, Wells RS, Berggren P, et al. Recommendations for photo-identification methods used in capture-recapture models with cetaceans. Mar Mamm Sci. 2015; 31 (1): 298–321.
- 73. Stolen MK, Barlow J. A model life table for bottlenose dolphins (Tursiops truncatus) from the Indian River Lagoon system, Florida, U.S.A. Mar Mamm Sci. 2003; 19(4): 630–649.
- 74. Williams JA, Dawson SM, Slooten E. The abundance and distribution of bottlenosed dolphins (Tursiops truncatus) in Doubtful Sound, New Zealand. Can J Zool. 1993; 71: 2080–2088.
- 75. White GC, Burnham KP. Program MARK: survival estimation from populations of marked animals. Bird Study. 1999; 46: S120–S139.
- 76.
Laake JL. RMark: An R interface for analysis of capture-recapture data with MARK. AFSC Processed Rep. Alaska Fish. Sci. Cent. Seattle: NOAA, 2013. pp. 25.
- 77.
R Core Team. R: A language and environment for statistical computing. 2018 [cited 30 October 2019]. R Foundation for Statistical Computing, Vienna, Austria. Available from: https://www.R-project.org/.
- 78. Silva MA, Magalhães S, Prieto R, Santos RS, Hammond PS. Estimating survival and abundance in a bottlenose dolphin population taking into account transience and temporary emigration. Mar Ecol Prog Ser. 2009; 392: 263–276.
- 79. Pradel R, Hines JE, Lebreton JD, Nichols JD. Capture-recapture survival models taking account of transients. Biometrics. 1997; 53(1): 60–72.
- 80. Madon B, Garrigue C, Pradel R, Gimenez O. Transience in the humpback whale population of New Caledonia and implications for abundance estimation. Mar Mamm Sci. 2013; 29: 669–678.
- 81.
Langtimm CA. Non-random temporary emigration and the robust design: Conditions for bias at the end of a time series. In Thomson DL, Cooch EG, Conroy MJ, editors. Modeling demographic processes in marked populations. Boston: Springer US; 2009. pp. 745–761.
- 82. Kendall WL, Stapleton S, White GC, Richardson JI, Pearson KN, Mason P. A multistate open robust design: population dynamics, reproductive effort, and phenology of sea turtles from tagging data. Ecol Monogr. 2019; 89 (1): e01329.
- 83. Fletcher D. Estimating overdispersion when fitting a generalized linear model to sparse data. Biometrika. 2012; 99: 230–237.
- 84. Stanley TR, Burnham KP. A closure test for time-specific capture-recapture data. Environ Ecol Stat. 1999; 6:197–209.
- 85.
Burnham KP, Anderson DR. Model selection and multimodel inference: A practical information-theoretic approach. 2nd ed. New York: Springer-Verlag; 2002.
- 86. Pollock KH. Capture-recapture models. J Am Stat Assoc. 2000; 95(449): 293–296. pmid:10877301
- 87.
Williams BK, Nichols JD, Conroy MJ. Analysis and management of animal populations. San Diego: Academic Press; 2002.
- 88. Otis DL, Burnham P, White GC, Anderson DR. Statistical Inference from capture data on closed animal populations. Wildlife Monographs. 1978; 62et al. 1978–80: 3–135.
- 89.
Anderson DR. Model based inference in the life sciences: A primer on evidence. New York: Springer; 2008.
- 90. Conn PB, Gorgone AM, Jugovich AR, Byrd , Hansen LJ. Accounting for transients when estimating abundance of bottlenose dolphins in Choctawhatchee Bay, Florida. J Wildl Manage. 2011; 75(3): 569–579.
- 91. Litz JA, Ronje EL, Whitehead HR, Garrison LP. Updated abundance estimates for common bottlenose dolphins (Tursiops truncatus) inhabiting West Bay, Texas. Aquat Conserv. 2019; 29(11): 1959–1970.
- 92. Lebreton JD, Burnham KP, Clobert J, Anderson DR. Modeling survival and testing biological hypothesis using marked animals: a unified approach with case studies. Ecol Monogr. 1992; 62(1): 67–118.
- 93. Hines JE, Kendall WL, Nichols JD. On the use of the robust design with transient capture-recapture models. Auk. 2003; 120 (4): 1151–1158.
- 94. Marsh H, Sinclair DF. Correcting for visibility bias in strip transect aerial surveys of aquatic fauna. J Wildlife Manage. 1989; 53: 1017–1024.
- 95. Wells RS, Allen JB, Lovewell G, Gorzelany J, Delynn RE, Fauquier DA, et al. Carcass-recovery rates for resident bottlenose dolphins in Sarasota Bay, Florida. Mar Mamm Sci. 2015; 31(1): 355–368.
- 96.
Noke WD. Interactions between the Indian River Lagoon blue crab fishery and the bottlenose dolphin Tursiops truncatus. M.Sc. Thesis, University of Central Florida. 1999. Available from: https://stars.library.ucf.edu/rtd/2214/.
- 97.
Tyson RB, Wells RS. Sarasota Bay/Little Sarasota Bay bottlenose dolphin abundance estimates: 2015. Southeast Fisheries Science Center Reference Document PRBD-2016-02. 7 pp. 2016. Available from: https://repository.library.noaa.gov/view/noaa/12664 https://doi.org/10.7289/V5/RD-PRBD-2016-02
- 98. Oudejans MG, Visser F, Englund A, Rogan E, Ingram SN. Evidence for distinct coastal and offshore communities of bottlenose dolphins in the north east Atlantic. PLos ONE. 2015; 10(4): e0122668. pmid:25853823
- 99. Rowe LE, Dawson SM. Determining the sex of bottlenose dolphins from Doubtful Sound using dorsal fin photographs. Mar Mamm Sci. 2009; 25(1): 19–34.
- 100. Kent EE, Mazzoil M, McCulloch SD, Defran RH. Group characteristics and social affiliation patterns of bottlenose dolphins (Tursiops truncatus) in the Indian River Lagoon, Florida. Fla Sci. 2008; 71(2): 149–168.
- 101. Bechdel SE, Mazzoil M, Murdoch E, Howells EM, Reif JS, McCulloch SD, et al. Prevalence and impacts of motorized vessels on bottlenose dolphins (Tursiops truncatus) in the Indian River Lagoon, Florida. Aquat Mamm. 2009; 35(3): 367–37.
- 102.
Provancha JA, Scheidt DM. Long-term trends in seagrass beds in Mosquito Lagoon and northern Banana River. In Bortone SA, editor. Seagrasses: monitoring ecology, physiology, and management. Boca Roton: CRC Press; 2000. pp. 177–193.
- 103. Bossart GD, Fair P, Shaefer A, Reif JS. 2017. Health and Environmental Risk Assessment Project for bottlenose dolphins Tursiops truncatus from the southeastern USA. I. Infectious diseases. Dis Aquat Organ. 2017; 125: 141–153. pmid:28737159
- 104. Mazzoil M, Reif JS, Youngbluth M, Murdoch ME, Bechdel SE, Howells E, et al. Home ranges of bottlenose dolphins (Tursiops truncatus) in the Indian River Lagoon, Florida: Environmental correlates and implications for management strategies. EcoHealth. 2008; 5(3): 278–288. pmid:18841416
- 105. Shane SH, Wells RS, Würsig B. Ecology, behavior and social organization of the bottlenose dolphin: A review. Mar Mamm Sci. 1986; 2: 34–63.
- 106.
Wells RS, Scott MD, Irvine AB. The social structure of free-ranging bottlenose dolphins. In Genoways H, editor. Current Mammalogy, vol 1. New York: Plenum Press; 1987. pp. 247–305.
- 107. Zolman ES. Residence patterns of bottlenose dolphins (Tursiops truncatus) in the Stono River Estuary, Charleston County, South Carolina, U.S.A. Mar Mamm Sci. 2002; 18: 879–889.
- 108. Toth JL, Hohn AA, Able KW, Gorgone AM. Patterns of seasonal occurrence, distribution and site fidelity of coastal bottlenose dolphins (Tursiops truncatus) in southern New Jersey, U.S.A. Mar Mamm Sci. 2011; 27: 94–110.
- 109. Bejarano AC, Wells RS, Costa DP. Development of a bioenergetics model for estimating energy requirements and prey biomass consumption of the bottlenose dolphin Tursiops truncatus. Ecol Model. 2017; 356: 162–172.
- 110. Haase P, Tonkin JD, Stoll S, Burkhard B, Frenzel M, Geijzendorffer IR, et al. The next generation of site-based long-term ecological monitoring: Linking essential biodiversity variables and ecosystem integrity. Sci Total Environ. 2018; 613–614: 1376–1384. pmid:29898505
- 111. Fruet PF, Daura-Jorge FG, Moller LM, Genoves RC, Secchi ES. Abundance and demography of bottlenose dolphins inhabiting a subtropical estuary in the Southwestern Atlantic Ocean. J Mammal. 2015; 96(2): 332–343.