Statistical Methods and Computational Tools for Marine Animal Movement, Distribution and Population Size

海洋环境中正在收集越来越多的复杂时空数据。这使得海洋科学家(包括生态学家、养护学家和渔业科学家)能够提出比以前更精细的空间和/或时间尺度的问题，例如检查特定渔业管理区内的丰度变化，而不是区域之间的变化。这些新的数据来源，其中一些来自迅速发展的数字技术(例如，用于海洋动物追踪的全球定位系统)，另一些则来自迄今未得到充分利用的公民科学努力(例如，新斯科舍省海滩的夏季水母目击)，以及伴随它们而来的重要问题，这就要求在时空建模方面取得进展。管理和保护海洋动物是具有挑战性的，因为它们迁徙，随着时间的推移改变它们的分布，并且很难观察到。*我的研究计划专注于开发方法学和计算工具，以回答与海洋保护和管理相关的重要科学问题。我的研究的具体短期目标是：1)开发选择正确数量的行为状态并评估适用于动物运动的状态空间模型(SSM)和隐马尔可夫模型(HMM)框架的拟合优度的方法；2)将空间(和其他)信息直接纳入种群动力学模型，例如，以更准确地识别副渔获物热点；以及3)开发近亲标记重新捕获(CKMR)方法，以估计特定保护物种的种群数量。这些目标将通过推进标准方法(例如，动物运动的SSM)、通过模拟严格评估相互竞争的方法(例如，用于捕获和/或描述空间依赖的方法)、以及通过精心选择的案例研究面对尚处于起步阶段的具有挑战性的统计推断环境(例如，CKMR)来实现。将尽一切努力确保对离群值和其他偏离模型假设的小偏差具有稳健性。将要开发的方法将广泛适用，但我将重点介绍加拿大大西洋地区令人担忧的物种保护和管理。我的研究的长期目标是继续提供可用的统计方法和辅助计算工具，以确保加拿大能够谨慎地管理其海洋资源。**

Increasingly large amounts of complex spatiotemporal data are being collected in the marine environment. This has enabled ocean scientists (including ecologists, conservationists, and fisheries scientists) to ask questions at finer spatial and/or temporal scales than previously possible, for example to examine variation in abundance within a particular fisheries management area as opposed to between areas. These new data sources, some derived from rapidly advancing digital technologies (e.g., global positioning systems for marine animal tracking) and others from hitherto under-utilized citizen science efforts (e.g., summer jellyfish sightings on Nova Scotia beaches), and the important questions that accompany them demand advancements in spatiotemporal modelling. Managing and conserving marine animals is challenging because they migrate, change their distributions through time, and are difficult to observe. ***My research program focuses on developing methodology and computational tools to answer important scientific questions related to marine conservation and management. The specific short-term objectives of my research are to: 1) develop methods for choosing the correct number of behavioural states and assessing goodness of fit suitable for both state-space model (SSM) and hidden Markov model (HMM) frameworks for animal movement; 2) incorporate spatial (and other) information into population dynamics models directly, for example to more accurately identify by-catch hotspots; and 3) develop Close Kin Mark Recapture (CKMR) methods for estimating population size for particular species of conservation concern. These objectives will be achieved by advancing standard approaches (e.g., SSMs for animal movement), critically assessing competing approaches (e.g., those for capturing and/or describing spatial dependence) through simulation, and confronting challenging statistical inference settings that are in their infancy (e.g., CKMR) by way of carefully chosen case studies. Every effort will be made to ensure robustness to outliers and other small departures from model assumptions. The methods to be developed will be broadly applicable, but I will focus on species of conservation and management concern in Atlantic Canada. The long-term objective of my research is to continually make available statistical methods and supporting computational tools that ensure Canada can steward its ocean resources with care. **