Data_Sheet_1_Environmental Factors May Drive the Post-release Movements of Surrogate-Reared Sea Otters.docx

Southern sea otters (Enhydra lutris nereis) are currently >80% depleted with respect to both abundance and range occupancy baselines, challenging the long-term persistence of the species and the ecosystem benefits their populations might provide. From 2001 to 2018, the Monterey Bay Aquarium rescued stranded sea otter pups and reared them in captivity through a surrogacy program using non-releasable adult females. We gave 11,396 days of captive care to 56 otters, reintroduced them into the wild, and observed them over 894 total field days after release. This study describes the post-release movements of the 42 successfully released otters, quantifying their dispersal patterns and modeling environmental, demographic, and animal care influences through a machine learning framework. This random forest model specifically considers predictor variable correlation, accounts for individual and joint variable impacts, and evaluates robustness through sensitivity analyses. Heavy tailed dispersal models best explained the (n = 641) daily movements of surrogate-reared otters, and the random forest outputs ranked population demography, population growth, and El Niño most significantly. Occasionally aided by recaptures, the scale of dispersals consistently declined after release, indicating successfully released otters stabilized their movements within 3 weeks in the wild. Our results show dispersal is an important metric for measuring the success of sea otter releases and suggest environmental factors (including climate) at release sites may determine the success of reintroduction programs.

南方海獭（Enhydra lutris nereis）当前的种群数量与分布范围均较基线水平减少80%以上，这对该物种的长期存续及其种群所能提供的生态系统服务功能均构成严峻挑战。2001年至2018年间，蒙特雷湾水族馆（Monterey Bay Aquarium）救助了搁浅的海獭幼崽，并通过使用无法放归野外的成年雌性海獭作为代养母的代育计划，在人工圈养环境下对幼崽进行抚育。本团队共计为56只海獭提供了总计11396天的人工圈养抚育，随后将其放归野外，并在放归后累计开展了894天的野外跟踪观测。本研究针对其中42只成功放归的海獭，描述其放归后的移动行为，量化其扩散模式，并通过机器学习框架构建环境、种群统计特征以及人工抚育因素对其移动行为的影响模型。该随机森林（random forest）模型专门考量预测变量间的相关性，兼顾单个变量与联合变量的影响，并通过敏感性分析评估模型的鲁棒性。重尾扩散模型能够最优拟合代育海獭的641次每日移动轨迹（n=641），且随机森林模型的输出结果显示，种群统计特征、种群增长以及厄尔尼诺（El Niño）现象是影响程度最高的因子。尽管偶尔通过重捕获取辅助数据，海獭的扩散规模在放归后持续下降，表明成功放归的海獭可在野外3周内稳定其移动范围。本研究结果表明，扩散行为是评估海獭放归成效的重要指标，同时提示放归地的环境因素（包括气候）或可决定海獭再引入项目的成败。