Marine mammals and sea turtles listed under the U.S. Endangered Species Act are recovering

The U.S. Endangered Species Act (ESA) is a powerful environmental law protecting imperiled plants and animals, and a growing number of marine species have been protected under this law as extinction risk in the oceans has increased. Marine mammals and sea turtles comprise 38% of the 163 ESA-listed marine “species”, which includes subspecies and distinct population segments, yet analyses of recovery trends after listing are lacking. Here we gathered the best available annual abundance estimates for geographically delimited populations of all 62 marine mammal and sea turtle species listed under the ESA. Of these, we chose representative populations of species that were listed before 2012, occur and reproduce in U.S. waters, and have data of sufficient quality and timespan for trend analyses. Thus, we quantitatively analyzed population trends, magnitude of population change, and recovery status for 23 and 8 representative populations of 14 marine mammal and 5 sea turtle species, respectively. Using generalized linear and non-linear models, we found that 18 marine mammal (78%) and 6 sea turtle (75%) populations significantly increased after listing; 3 marine mammal (13%) and 2 sea turtle (25%) populations showed non-significant changes; while 2 marine mammal (9%), but no sea turtle populations declined after ESA protection. Overall, the 24 populations that increased in abundance were from species listed for 20 years or more (e.g., large whales, manatees, and sea turtles). Conservation measures triggered by ESA listing such as ending exploitation, tailored species management, and fishery regulations, and other national and international measures, appear to have been largely successful in promoting species recovery, leading to the delisting of some species and to increases in most populations. These findings underscore the capacity of marine mammal and sea turtle species to recover from substantial geographical population declines when conservation actions are implemented in a timely and effective manner.


INTRODUCTION
We collected information and population-level abundance estimates for ESA-listed marine 147 mammals and sea turtles from published papers and government reports. Main data sources 148 included NMFS and USFWS technical memorandum and administrative reports, U.S. marine 149 mammal stock assessment reports, species recovery plans, five-year status reviews, and primary 150 sources from peer-reviewed scientific journals (see data in supporting information). When  Estimate bias and errors in population abundance obtained from data sources were variable 165 among species and even within the same species over time. For example, survey effort and 166 methodologies changed over time and population estimates have been calculated using different 167 approaches over the years for the same population (e.g., traditional population abundance 168 models, Bayesian population models, or habitat-based density models). Thus, where available, each data point was accompanied with information on data collection methodology, error 170 information (e.g., coefficient of variation), and data estimation reliability (see data in supporting 171 information). Time-series of population abundance for each species were carefully constructed to 172 ensure all annual data points were derived from adequate and quantitative methodologies.  (Table 2 and S2 Table). Population trajectories were classified as 179 significantly increasing, non-significant change (non-significant slope), or significantly 180 decreasing as in Magera et al. [43]. Recovering populations were defined as those that 181 significantly increased in abundance after ESA listing, independently of whether or not they 182 were on track to meet the recovery criteria for downlisting or delisting found in recovery plans.
To explore population trajectories after listing we used several models including linear models 191 (lm), generalized linear models (glm), generalized least square models (gls), or generalized 192 additive models (gam) in which population abundance estimates were modeled by running time 193 in years (S2 Table). Because population trends were species specific, we used several family 194 distributions and error links for each of the population-level models (S2 Table). For each 195 population, we performed a comprehensive exploratory analysis using all models and possible 196 combinations of families and links with and without a log transformation of the abundance 197 estimates. In several gls models we added correlation and variance structures to account for 198 potential temporal autocorrelation among years and variation in the data (S2 Table).  241 Overall, approximately 78% of marine mammals (18 out of 23) and 78% of sea turtles (7 out of 242 9) analyzed that met our selection criteria significantly increased in abundance after ESA listing 243 ( Fig 1A). Representative populations of three marine mammals (~13%) and two sea turtles 244 (~22%) experienced non-significant change. Only two marine mammals (~9%), but no sea 245 turtles significantly declined after ESA protection ( Fig 1A). Marine mammals and sea turtles 246 with populations that significantly increased were listed between two to five decades and  Table 2). Out of the 25 species with populations that 250 significantly increased, 52% were listed as endangered, 32% as threatened, and 16% were 251 delisted, indicating that population increases occurred independent of whether a species was 252 classified as threatened or endangered (Tables 1 and 2).    (Table 3). For example, all four DPSs of humpback whales analyzed in our study showed   Table 3 and S2 Table). At least 17 individuals died in 2017, representing nearly 4% of the entire 280 population (Table 3 and S2 Table).    protection status see Table 1; and for results of fitting models see S2 townsendi) increased about nine times at ~15% per year since listing in 1985 (Fig 3; Table 3).

303
The California population of the Southern sea otter (Enhydra lutris nereis) approximately 304 doubled in numbers and it is likely to reach the demographic recovery criteria in the coming 305 years (Fig 3; Table 3). The Eastern DPS of Steller sea lion (Eumetopias jubatus) tripled its 306 population at ~6% per year since 1990, reaching its recovery criteria of ~60,000 individuals in 307 2013, and was subsequently delisted from the ESA (Fig 3; Table 3). Also, both the Florida and 308 Antillean manatee subspecies increased approximately eight and three times (~17% and ~5% per 309 year), respectively, for the past 40 years (Fig 3; Table 3); and the USFWS downlisted them from 310 endangered to threatened in 2017 (Table 1).   (Fig 2, Table 3). This population suffered major declines after a  Table 3).

333
However, the population has increased to 968 seals by 2016 (Table 3).

335
Nearly all sea turtle species analyzed in our study significantly increased after ESA listing ( Fig   336   4; Table 3 and S2 Table). Estimates of the number of individuals, nesting females, and number Atlantic leatherback sea turtles showed that these species increased at considerably high growth 339 rates (~13% to ~284% per year) for several decades, depending on initial estimates (Fig 4; Table   340 3 and S2 Table). For example, the number of nesting females of green sea turtle at East Island of   significantly increased after ESA listing (Fig 4; Table 3). protection status see Table 1; and for results of fitting models see S2 Among the sea turtles analyzed in this study, our models were not able to detect significant  Table 3 and S2 Table). For example, the best models for the number of nests of   indirect harm [62].

428
Conservation measures for these two species were developed relatively recently and ongoing threats have not been mitigated [83,84]. It is likely that these species will require more time

446
These species face several similar regional anthropogenic threats including prey reduction due to 447 fishing, habitat degradation, toxic pollutants, disturbance from boat traffic and marine noise, 448 fishery interactions, as well as global threats associated with climate change and ocean regime 449 shifts that affect food availability [88][89][90][91][92]