Trail vs. thicket: Evaluating conservation detection dog performance in varied terrain

Abstract Use of detection dogs has emerged as an effective survey method for rare and cryptic species, outperforming traditional methods. However, there is little information available on how survey conditions may affect survey efforts. This study investigates the detection probability, effective sweep width, and search time of a detection dog working on‐lead in varying vegetation and weather conditions in Florida. We established 4 200‐m long transects in habitats with varying vegetation structure: open field, open canopy forest, closed canopy forest, and trail. American mink ( Neogale vison ) scat was set along each transect, in locations unknown to the detection dog handler, at distances ranging from 0 to 24 m from the transect. Vegetation and weather conditions were recorded for each survey. Between August 2021 and October 2022, we conducted 28 surveys (112 total transect surveys). Across all transects the detection dog located 66% of hidden scats ( n = 384). We calculated an effective sweep width of 49 m (24.5 m on either side of the transect) and detection probability decreased with distance from transect line. Contrary to expectations, vegetation structure and weather conditions did not influence detection probability, but vegetation structure did affect search time, an important variable to consider when designing surveys. Search time increased with increases in understory vegetation obstruction and basal area. Conducting surveys on trails showed promise as an effective search method, having similar detection probability to off‐trail transects and faster search times than the closed canopy forest transect. Trail surveys may be particularly useful for areas that would otherwise be impassable, but their effectiveness may vary based on the target species' behavior. Our study underscores the importance of understanding search metrics in conservation detection dog surveys to design effective and appropriate surveys for anticipated field conditions.