State-space model combining local camera data and regional administration data reveals population dynamics of wild boar
State-space model combining local camera data and regional administration data reveals population dynamics of wild boar
复制标题
结合本地相机数据和区域管理数据的状态空间模型揭示了野猪的种群动态
DOI：
10.1002/1438-390x.12138
复制
发表时间：
2023
影响因子：
1.7
通讯作者：
Tadashi Miyashita
中科院分区：
文献类型：
--
作者：
Minoru Kasada;Yoshihiro Nakashima;Keita Fukasawa;Gota Yajima;Hiroyuki Yokomizo;Tadashi Miyashita
最近野生动物的增加通过经济和生态破坏以及病原体的传播对人类造成负面影响。了解野生动物的种群动态对于制定有效的管理策略至关重要。然而，由于可用数据有限，很难在大时空尺度上准确估计人口规模。我们通过首先拟合基于相机陷阱数据的随机遭遇和停留时间（REST）模型来解决这些问题，以在基于收获的贝叶斯状态空间模型中为捕获率参数构建信息丰富的先验分布。我们构建了一个贝叶斯状态空间模型，将捕获野猪数量的管理数据与通过相机陷阱数据估计的捕获效率的先验分布相结合。具有来自 REST 模型的信息先验分布的模型成功地估计了种群动态，而仅使用管理数据的模型由于缺乏参数收敛而未能成功估计。我们确定了以下区域：(1) 野猪表现出较高的潜在种群增长率和较高的承载能力；(2) 当前的诱捕工作正在有效抑制当地种群；(3) 需要加强诱捕以控制整个地区的种群。该模型可用于在持续捕获压力的假设下预测种群的未来趋势。这将有助于确定空间上明确的诱捕工作，以实现目标种群水平。
Recent increases in wildlife cause negative impacts on humans through both economic and ecological damage, as well as the spread of pathogens. Understanding the population dynamics of wildlife is crucial to develop effective management strategies. However, it is difficult to estimate accurate and precise population size over large spatial and temporal scales because of the limited data availability. We addressed these issues by first fitting a random encounter and staying time (REST) model based on camera trap data to construct an informative prior distribution for a capture rate parameter in a harvest‐based Bayesian state‐space model. We constructed a Bayesian state‐space model that integrated administration data on the number of captured wild boar with the prior distribution of capture efficiency estimated by camera trap data. The model with informative prior distribution from the REST model successfully estimated population dynamics, whereas the model using only the administration data did not, owing to a lack of parameter convergence. We identified areas where (1) wild boars exhibit a high potential population growth rate and a high carrying capacity, (2) current trapping efforts are effectively suppressing local populations, and (3) trapping reinforcement is required to control populations in the whole region. The model could be used to predict future trends in populations under the assumptions of ongoing trapping pressure. This will help identify spatially explicit trapping efforts to achieve target population levels.