SITE OCCUPANCY AND ABUNDANCE MODELS FOR ANALYZING MULTIPLE-VISIT DETECTION/NONDETECTION DATA

• 成果类型: Article
• 署名作者: Huynh, Huu-dinh; Schofield, Matthew; Hwang, Wen-han
• 署名单位: National Chung Hsing University; National Tsing Hua University
• 刊物名称: ANNALS OF APPLIED STATISTICS
• ISSN/ISSBN: 1932-6157
• DOI: 10.1214/24-AOAS1888
• 发表日期: 2024
• 页码: 2424-2443
• 关键词: n-mixture models; population-size; trade-off; density; counts; rates

摘要

We propose an enhanced site occupancy model for analyzing ecological detection/nondetection data obtained from multiple visits. The model distinguishes between abundance, occupancy, and detection probabilities. We allow for transient individuals through a community parameter, c, that characterizes the proportion of individuals fixed across visits. This parameter seamlessly transitions from the standard occupancy model (c = 0) to the Nmixture model (c = 1), enabling a more accurate analysis of site occupancy data. Through theoretical developments and simulation studies, we demonstrate how this model effectively addresses biases inherent in conventional approaches, particularly for c is not at 0 or 1. We apply the model to various datasets of mammal and bird species and compare it to current approaches.