干旱区生态环境知识资源中心

We used sightability models with aerial counts to estimate regional abundance of desert bighorn sheep (Ovis canadensis) in southwestern Arizona during 1992-2012. Because of incomplete spatial and temporal coverage of surveys, we integrated sightability estimates using Bayesian state-space models. Although some game management units (GMUs) exhibited negative average growth rates during this period, temporal variability in growth was large. The combination of temporal variability and statistical uncertainty in the sightability estimates led to the selection of a regional model having zero mean growth from initial abundance. Hence, the model indicated a relatively stable population of desert bighorn sheep in southern Arizona, beginning with approximately 2,900 bighorn sheep in 1992 to the current number of approximately 2,300 sheep (2012). Our method provides statistically valid estimates of overall regional abundance not otherwise possible given the incomplete spatial coverage of surveys. Additionally, results have implications for other studies, common in wildlife management, where managers seek to integrate information across surveys with partial time-series. (c) 2014 The Wildlife Society.