Sampling the nonexistent? Evaluating the presence/absence of wolverines (Gulo gulo) in the Sierra Nevada

Detecting species that are both rare and range widely across the landscape presents a difficult challenge.  Wolverines have disappeared from almost half of their range in North America.  We present a survey designed to have a high probability to determine if a viable population persists within Sequoia-Kings Canyon National Parks, California.  We placed baited camera traps at high elevation sites throughout the parks from late January through mid-May 2006.  No wolverines were detected during our survey, although we did record other mesocarnivore species at 17 of the 18 bait stations.  We conducted power analyses to determine the effort required to have a high probability of detecting wolverines before setting up the survey and to interpret the absence of wolverine detections after the survey.  Power analyses were based on trap efficiency estimates from studies in extant wolverine populations and movement simulations for a hypothetical wolverine population in our study area.  We estimated that our survey had a low chance of missing a wolverine population under a wide range of assumptions.  We further evaluate the power of our survey considering that smaller populations have both a lower chance of being detected but a higher chance of extinction in the near future.  Weighting detection probability against extinction risk under reasonable assumptions about home range size and bait attraction distances, our survey had a less than 5% chance of missing a persisting population of wolverines.