Reproductive outcome and survival of common bottlenose dolphins sampled in Barataria Bay, Louisiana, USA, following the Deepwater Horizon oil spill

Common bottlenose dolphins (Tursiops truncatus) inhabit bays, sounds and estuaries across the Gulf of Mexico. Following the Deepwater Horizon oil spill, studies were initiated to assess potential effects on these ecologically important apex predators. A previous study reported disease conditions, including lung disease and impaired stress response, for 32 dolphins that were temporarily captured and given health assessments in Barataria Bay, Louisiana, USA. Ten of the sampled dolphins were determined to be pregnant, with expected due dates the following spring or summer. Here, we report findings after 47 months of follow-up monitoring of those sampled dolphins. Only 20% (95% CI: 2.50–55.6%) of the pregnant dolphins produced viable calves, as compared with a previously reported pregnancy success rate of 83% in a reference population. Fifty-seven per cent of pregnant females that did not successfully produce a calf had been previously diagnosed with moderate–severe lung disease. In addition, the estimated annual survival rate of the sampled cohort was low (86.8%, 95% CI: 80.0–92.7%) as compared with survival rates of 95.1% and 96.2% from two other previously studied bottlenose dolphin populations. Our findings confirm low reproductive success and high mortality in dolphins from a heavily oiled estuary when compared with other populations. Follow-up studies are needed to better understand the potential recovery of dolphins in Barataria Bay and, by extension, other Gulf coastal regions impacted by the spill.


Photographic-identification surveys for mark-recapture analysis
In June 2010, intensive photo-ID surveys for mark-recapture analysis were initiated in Barataria Bay. Mark-recapture photo-ID surveys have been used extensively to estimate abundance and determine survivorship of small cetaceans within a given region [3][4][5]. The mark-recapture photo-ID surveys in Barataria Bay were based on a robust design [6], and each primary session included three traversals of a defined survey route [4] (Figure 1). In all, ten primary sessions were conducted from June 2010 through April 2014, six of which followed the 2011 health evaluations.

Radio-tracking and freeze brand monitoring surveys
Surveys were initiated in September 2011 specifically to re-sight dolphins sampled during the preceding health evaluations. The objective of these surveys was to maximize the chance of encountering one or more of the 32 sampled dolphins by focusing survey effort in areas utilized by the sampled animals. Ad hoc survey routes were determined daily based on weather, sighting conditions, recent satellite-linked transmitter locations, and sightings of freeze-branded individuals from prior photo-ID surveys.
Surveys, each covering three days, were conducted at three, five, and ten weeks post capturerelease. Two, 4-element mast-mounted Yagi antennas [Advanced Telemetry Systems (ATS), Isanti, MN, USA] attached to ATS R2000 scientific receivers were used to scan all tagged animal frequencies at 60-second intervals. Once a tag signal was received, observers on the tracking vessel would attempt to visually locate the tagged animal, obtain photos of it along with any associated dolphins, and record a GPS position. The goal within each three-day tracking interval was to visually locate and obtain photos of all tagged dolphins.
Additional monitoring surveys were conducted following the expected transmission life of the VHF tags, with the primary objective to monitor tag migration and loss [1]. Animals were identified by the freeze brands applied during the health evaluations. These survey sessions, each lasting 3-5 days, were conducted in February, March, April, and May 2012.

Reproductive outcome surveys
Surveys specifically designed to document reproductive outcomes, utilizing photo-ID survey methods as described above and in greater detail by Melancon et al. [7], were conducted over one-week periods in June, July, and August 2012. The objective of these surveys was first to locate pregnant females from the August 2011 health evaluations (determined via ultrasound), and second to document the presence or absence of an accompanying neonate as they were expected to give birth in spring 2012. Survey routes were designed to provide coverage of areas encompassing previous sightings and recent satellite-linked locations of the targeted females.

R code to estimate CJS models
Data Preparation: The following code sets up data for later MCMC sampling, and is common to all models. Here, H32 is the 32 X 20 matrix of capture histories (Table 1) Vector of first encounters get.first <-function(x) min(which(x!=0)) f <-apply(ch,1,get.first) # Toss individuals first seen at last occasion (no information for CJS) Utility function to execute MCMC chains and compute DIC: The following function is a copy of the coda.samples function contained in rjags, with the addition of DIC computation and modification to return a list object.

JAGS code and execution:
The following is JAGS code used to fit the {constant,constant} model. # Write the modelString to a file, using R commands: .temp = file("model.txt","w") ; writeLines(modelString,con=.        Figure S1. Timing and types of surveys used in the survival and reproductive outcome analysis. Numbers above each symbol represent the number of surveys conducted in each month. Table S1. Summary information for individual dolphins monitored following the 2011 health evaluations in Barataria Bay, LA. Criteria for overall prognosis, lung disease, and hematological and serum biochemical panels were defined in Schwacke et al. [8]. A "1" indicates the individual was classified as out-of-range for the given panel, "0" indicates the individual was within normal range. *Indicates the serum aldosterone was below assay detection limit (5.5 pg/mL).