Over 20 years ago, ecologists were introduced to the site occupancy model (SOM) for estimating occupancy rates from detection‐nondetection data. In the ensuing decades, the SOM and its hierarchical modeling extensions have become mainstays of quantitative ecology, and estimating occupancy rates has become one of the most common applications of ecological field data. Here, we review 364 peer‐reviewed articles published between 2019–2021 that estimated occupancy. We first document broad patterns in study design and statistical methods to provide educators, developers of methodology and software, and ecologists with a clear picture of the landscape of methodologies used to estimate animal occupancy. Second, we conduct a focused review of a subset of 98 papers that applied the hierarchical SOM, drawing from methodological literature to identify discrepancies between SOM applications and methodological best practices. We discuss limits to statistical power, issues with model checking and model selection procedures, potential problems arising from unmodeled non‐independence, and reproducibility. We highlight areas of rapid advancement in interpreting animal occupancy related to animal movement, imperfect detection, and the occupancy–density relationship. We aim to help readers understand the landscape of methods available, motivate shifts toward robust and reproducible science, and inspire new software and methodological research.

中文翻译：

---

生态学家在实践中如何估计占用率？


20 多年前，生态学家被引入场地占用模型 （SOM），用于根据检测-非检测数据估计占用率。在随后的几十年里，SOM 及其分层建模扩展已成为定量生态学的支柱，估计入住率已成为生态田间数据最常见的应用之一。在这里，我们回顾了 2019 年至 2021 年间发表的 364 篇同行评审文章，这些文章估计了入住率。我们首先记录了研究设计和统计方法的广泛模式，为教育工作者、方法和软件开发人员以及生态学家提供了用于估计动物占用率的方法的清晰图景。其次，我们对应用分层 SOM 的 98 篇论文子集进行了重点审查，从方法论文献中汲取灵感，以确定 SOM 应用与方法学最佳实践之间的差异。我们讨论了统计功效的限制、模型检查和模型选择程序的问题、未建模的非独立性引起的潜在问题以及可重复性。我们强调了在解释与动物运动、不完美检测和占位-密度关系相关的动物占用率方面取得快速进展的领域。我们的目标是帮助读者了解可用方法的前景，推动向稳健和可重复的科学转变，并激发新的软件和方法研究。