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.

No abstract

No abstract