Journal Article

Can Distance Sampling be Used to Estimate the Abundance of Mountain Ungulates? A Comparison of Distance Sampling and Double‐Observer Survey Methods

Corlatti, Luca, Lu, Zhi, Weckworth, Byron, Xiao, Lingyun, Liu, Yanlin

2025

Overview

Robust population surveys and monitoring of ungulates in mountainous areas are hindered by complex terrain and limited access to remote regions. Distance sampling, a widely used method for surveying ungulates in non‐mountainous habitats, faces challenges in such environments due to violations of key assumptions. In this study, we examine the assumptions underlying distance sampling and explore how they can be addressed in mountainous terrain. We then applied distance sampling to estimate the abundance of bharal ( Pseudois nayaur ) on the Tibetan Plateau and compared the results with estimates from the double‐observer survey method, which is commonly used for mountain ungulates. Both methods produced highly comparable estimates—2189 from distance sampling and 2038 from the double‐observer survey—approximately twice as high as the total count (963). Distance sampling yielded a wider confidence interval (95% CI: 1207–3972) than the double‐observer survey (95% CI: 1405–2670). We recommend using distance sampling when human resources are limited or when the closed population assumption is difficult to meet. Otherwise, the double‐observer method is preferable due to its narrower confidence intervals and greater ability to detect population changes. 山地有蹄类的可靠种群调查和监测因地形复杂和偏远地区难以进入而受阻。距离样线法作为非山地地区常用的种群调查方法, 在山地地形中往往难以满足其假设条件。本文讨论了距离采样的关键假设及其在山地地形中的应对方法。随后, 我们应用距离样线法在青藏高原一条山谷中估算了岩羊的数量, 并与常用于山地有蹄类动物的双观察者法调查结果进行了比较。两种方法得出高度相似的估计值 (距离样线法2189只, 双观察者法2038只), 均为直接计数法 (963只) 的两倍。距离样线法得出的置信区间较宽 (95% CI: 1207‐3972), 而双观察者较窄 (95% CI: 1405‐2670) 。建议在人力有限或难以满足种群闭合假设时采用距离采样, 否则双观察者调查因置信区间较窄更适合检测到种群变化。 This study addresses the challenge of accurately estimating populations of mountain‐dwelling ungulates such as the bharal ( Pseudois nayaur ), which are difficult to survey due to rugged terrain and remote locations. Traditional methods like total counts often underestimate true population sizes because animals can be missed. We tested and compared two advanced survey methods—distance sampling and double‐observer surveys—in a valley on the Tibetan Plateau. Distance sampling estimates population size by recording distances to detected groups along transects, but its assumptions are difficult to meet in mountainous areas. The double‐observer method uses two independent teams to improve detection estimates but requires more personnel and careful timing to avoid disturbing animals. Both methods produced similar population estimates—around 2100 bharal—double the number found by total counts (963). Distance sampling had wider confidence intervals, indicating lower precision, while double‐observer surveys were more precise and better at detecting population changes. We adapted distance sampling by dividing the area into watersheds and designing transects to maximize visual coverage. We recommend using double‐observer surveys when a sufficient number of surveyors are available and the population remains closed during surveys. Distance sampling is more appropriate when resources are limited or when the population may move in and out of the survey areas. With appropriate adjustments, both methods can provide reliable estimates to support the conservation of mountain ungulates, which are key prey for predators such as the snow leopard. Distance sampling and double‐observer surveys produced very similar population estimates (2189 and 2038, respectively), both about twice as high as the total count (963). Distance sampling yielded wider confidence intervals (95% CI: 1207–3972) compared to the double‐observer survey (95% CI: 1405–2670). Key assumptions of distance sampling can be met in rugged terrain by dividing the study area into smaller sampling units (e.g., watersheds) and designing transects to maximize visual coverage rather than adhering strictly to straight‐line or random placement. Distance sampling is recommended when human resources are limited or when maintaining a closed population during the survey is difficult. The double‐observer survey is preferable when sufficient personnel are available, as its narrower confidence intervals improve sensitivity to population changes. 本研究针对山地有蹄类动物 (如岩羊 Pseudois nayaur ) 数量准确估算的难题展开, 这类动物因地形复杂和栖息地偏远, 调查十分困难。传统的总数调查方法常因动物未被发现而低估真实数量。我们在青藏高原一条河谷中测试并比较了两种先进的调查方法——距离样线法和双观察者法调查。距离样线法通过沿样线记录观察到的动物群体距离来估算数量, 但其核心假设在山地环境中往往难以满足。双观察者法则由两组独立调查队伍同时进行, 以提高检测率, 但需要更多人力和合理安排间隔时间以避免干扰动物。两种方法得出的数量估计结果相近, 均约为2100只岩羊, 是直接计数法 (963只) 的两倍。距离样线法的置信区间较宽, 表明其精度较低, 而双观察者法调查置信区间较窄, 更适合检测种群变化。研究调整了距离样线法方法以应对山地的复杂地形, 通过将调查区域划分为多个小流域并最大化视野覆盖来保证核心假设的满足。建议在人力充足且种群在调查期间相对闭合时采用双观察者法；在人力有限或难以保证种群封闭时采用距离样线法。总体而言, 经过合理调整, 两种方法均能提供可靠的种群估计, 为保护雪豹等捕食者关键猎物的山地有蹄类动物提供支持。 距离样线法和双观察者法调查得出的种群估计非常接近 (距离样线法2189只, 双观察者法2038只), 均为直接计数法 (963只) 的两倍。距离样线法的置信区间较宽 (95% CI: 1207‐3972), 而双观察者法较窄 (95% CI: 1405‐2670) 。 通过将调查区域划分为较小的采样单元 (如流域), 并设计样线以最大化视觉覆盖, 而非严格遵循直线或随机布置, 可以满足距离样线法在复杂地形中的关键假设。 建议在人力有限或难以保证种群闭合时使用距离样线法；在人力充足且种群相对稳定时优先采用双观察者法调查, 因其置信区间较窄, 更能灵敏地检测种群变化。