Data from: Separating mortality and emigration: modelling space use, dispersal and survival with robust-design spatial-capture-recapture data

1. Capture-recapture (CR) techniques are commonly used to gain information about population dynamics, demography and life-history traits of populations. However, traditional CR models cannot separate mortality from emigration. Recently developed spatial-capture-recapture (SCR) models explicitly incorporate spatial information into traditional CR models, thus allowing for individuals’ movements to be modelled explicitly. 2. In this paper, we extend SCR models using robust-design data to allow for both processes in which individuals can disappear from the population, mortality and dispersal, to be estimated separately. We formulate a general robust-design spatial capture-recapture (RD-SCR) model, explore the properties of the model in a simulation study, and compare the results to a Cormack-Jolly-Seber model and a non-spatial robust-design model with temporary emigration. In the case study, we fit several versions of the general model to data on field voles (Microtus agrestis) and compare the results with those from the non-spatial models fitted to the same data. We also evaluate assumptions of the fitted models with a series of simulation-based posterior predictive goodness-of-fit checks that are applicable to SCR models in general and the RD-SCR model in particular. 3. The simulation results show that the model preforms well under a wide range of dispersal distances. Our model outperforms the traditional CR models in terms of both accuracy and precision for survival. The case study showed that adult females have an approximately 3.5 times higher mortality rate than adult males. Males have larger home-ranges and disperse longer distances than females, but both males and females mostly move their activity centres within their previous home-range between trapping sessions at three week intervals. 4. Our RD-SCR model has several advantages compared to other approaches to estimate “true” survival instead of only “apparent” survival. Additionally, the model extracts information about space use and dispersal distributions that are relevant for behavioural studies as well as studies of life-history variation, population dynamics and management. The model can be widely applied due to the flexible framework, and other variations of the model could easily be implemented.

1. 捕获-再捕获（Capture-recapture, CR）技术常被用于获取种群动态、种群统计学以及种群生活史特征的相关信息。然而传统CR模型无法区分死亡与迁出事件。近年来提出的空间捕获-再捕获（spatial-capture-recapture, SCR）模型将空间信息显式融入传统CR模型，从而能够对个体移动进行显式建模。2. 本文基于稳健设计数据对SCR模型进行拓展，以实现对种群消失的两类过程——死亡与扩散——的分别估算。本文构建了通用的稳健设计空间捕获-再捕获（RD-SCR）模型，通过模拟实验探究了该模型的性能，并将其结果与康马克-乔利-西伯（Cormack-Jolly-Seber）模型以及带临时迁出的非空间稳健设计模型进行对比。在案例研究中，我们将通用模型的多个变体应用于田鼠（Microtus agrestis）的观测数据，并将结果与基于同一数据拟合的非空间模型结果进行比较。此外，我们通过一系列基于模拟的后验预测拟合优度检验对拟合模型的假设进行评估，这类检验方法可普遍适用于SCR模型，尤其适配RD-SCR模型。3. 模拟实验结果表明，该模型在广泛的扩散距离范围内均表现良好。在存活率的估算精度与准确度两方面，我们提出的模型均优于传统CR模型。案例研究显示，成年雌性田鼠的死亡率约为成年雄性的3.5倍。雄性个体的家域范围更大，扩散距离也长于雌性，但在以三周为间隔的捕夹周期之间，雌雄个体的活动中心大多仍在其原有家域范围内移动。4. 相较于仅能估算“表观”存活率的其他方法，本文提出的RD-SCR模型具备多项优势。此外，该模型还能提取出与行为研究、生活史变异研究、种群动态研究以及种群管理相关的空间利用与扩散分布信息。由于其框架具备灵活性，该模型具备广泛的应用潜力，且其各类变体也易于实现。