Wildlife managers often need to estimate population abundance to make well‐informed decisions. However, obtaining such estimates can be difficult and costly, particularly for species with small populations, wide distributions, and spatial clustering of individuals. For this reason, DNA surveys and capture–recapture modeling has become increasingly common where direct observation is consistently difficult or counts are small or variable. We compared the precision, as indicated by the coefficient of variation (CV), and cost‐effectiveness of 2 methods to estimate abundance of desert bighorn sheep (Ovis canadensis nelsoni) populations: traditional ground‐based mark–resight and fecal DNA capture–recapture. In the Marble Mountains in the Mojave Desert of southeastern California, USA, we conducted annual ground‐based mark–resight surveys and collected fecal samples at water sources concurrently during the dry seasons (Jun–Jul) of 2016 and 2017. Fecal DNA samples were genotyped to identify unique individuals. The Lincoln–Peterson bias‐corrected estimator and Huggins closed‐capture recapture models were used to estimate abundance for the ground‐based mark resight and fecal DNA capture–recapture, respectively. We compared costs between the 2 methods for our study and used simulations to estimate costs for a variety of possible sampling scenarios for our study system based on field‐based estimates. Population abundance estimates from fecal DNA capture–recapture achieved much greater precision (CV = 5–7%) than estimates derived from ground‐based mark–resight (CV = 21–56%). Our simulations indicated that for a population of 100, 2 sampling occasions, and resight probability of 0.20, the lowest CV obtained by mark–resight was approximately 12%. We predict the cost of abundance estimates for this level of precision (CV = 12%) from fecal DNA capture–recapture would be 28% of the cost of ground‐based mark–resight (i.e., a 72% cost reduction). We conclude that fecal DNA capture–recapture is a highly cost‐effective alternative for estimating abundance of relatively small populations (≤300) of desert bighorn sheep. More broadly, integrating simulated study designs with cost analyses provides a tool to identify the most effective method for estimating abundance over a wide variety of sampling scenarios. © 2020 The Wildlife Society.

中文翻译：

---

粪便DNA标记回收与传统标记回收的成本和精度

野生动物管理者通常需要估计种群数量以做出明智的决定。然而，获得这样的估计可能是困难且昂贵的，特别是对于种群少，分布广泛和个体空间聚集的物种。因此，在直接观察始终困难或计数很小或可变的情况下，DNA调查和捕获-捕获模型变得越来越普遍。我们比较了变异系数（CV）和两种估算沙漠大角羊（Ovis canadensis nelsoni）丰度的成本效益方法的精度）人群：传统的基于地面的标记检查和粪便DNA捕获-捕获。在美国东南部加利福尼亚州莫哈韦沙漠的大理石山中，我们进行了年度地基标记检验调查，并在2016年和2017年的干旱季节（6月至7月）同时收集了水源中的粪便样本。粪便DNA样本为基因型鉴定独特的个体。使用Lincoln-Peterson偏差校正估计器和Huggins封闭式捕获再捕获模型分别估计基于地面标记的捕获和粪便DNA捕获的捕​​获量。我们比较了我们的研究的两种方法之间的成本，并使用了模拟方法，以基于实地的估算为基础，为我们的研究系统估算了各种可能的采样方案的成本。粪便DNA捕获-再捕获的种群丰度估算值（CV = 5–7％）比基于地面标记监测的估算值（CV = 21–56％）高得多。我们的模拟结果表明，对于100个人口而言，有2个采样时机，且观察概率为0.20，通过标记观察获得的最低CV约为12％。我们预测，从粪便DNA捕获到再捕获中，达到此精确度水平（CV = 12％）的丰度估算成本将是基于地面标记检查的成本的28％（即，成本降低72％）。我们得出的结论是，粪便DNA捕获再捕获是一种高成本效益的替代方法，可用于估计相对较小的种群（≤300）沙漠大角羊的数量。更广泛地，将模拟研究设计与成本分析相结合，提供了一种工具，可用于确定用于评估各种采样方案中的丰度的最有效方法。©2020野生动物协会。