Comparing detection dog and livetrapping surveys for a cryptic rodent

We compared the effectiveness and cost of distribution surveys using livetrapping to those using detection dog-handler teams for a cryptic rodent (Franklin's ground squirrel [Poliocitellus franklinii]). We livetrapped at 62 sites in Illinois, Indiana, Iowa, Missouri, and Wisconsin for Franklin's ground squirrels in 2007-2009 and surveyed 40 of those sites using detection dogs in 2009. Independent surveys of a site by 2 dog-handler teams took <1 and yielded detection rates comparable to 2 daily livetrapping surveys (detection rate=83-84%). However, false presences are a potential problem when detection dogs are trained to scent of a species that leaves little visual sign to confirm its presence. Surveys by 2 dog-handler teams cost >2 daily livetrapping surveys conducted by 2 technicians but more and larger sites can be surveyed by dog-handler teams in a shorter time. For surveys covering large spatial scales or when time is a limiting factor, number of false presences, and study costs can be reduced by employing a 2-stage survey protocol in which livetrapping is conducted only at sites where detection dog surveys indicate presence. We conclude a 2-stage strategy could be used effectively in large-scale surveys for a variety of rare and cryptic species.