Integrating targeted surveys during community-wide occupancy assessments improves detection rates for a pollinator species of concern

Conservation practitioners can often struggle to achieve multiple goals—such as monitoring community diversity, tracking species of concern, and evaluating community and species' responses to on-the-ground activities—within limited time and budgets. This study aimed to identify the most efficient field and analytical methods for achieving monitoring goals simultaneously in pollinator surveys. We evaluated whether integrating targeted visual surveys with standard community netting surveys could improve detection efficiency for Bombus occidentalis (western bumble bee), a declining species of concern, while still capturing broader community data. In 2023, we conducted netting surveys for bumble bees (Bombus spp.) at 22 sites in Glacier National Park and added visual surveys specifically for western bumble bees if not detected during netting. We estimated western bumble bee occupancy and detection using two site-occupancy models: (1) traditional detection/non-detection (DND) models and (2) time-to-detection (TTD) models. We also assessed model performance with reduced (subsampled) survey effort. Models that integrated net and visual surveys produced similar occupancy estimates but higher cumulative detection probability than net-only surveys, with visual surveys requiring less effort than additional net surveys. Notably, 17% of visits (11 of 66) yielded western bumble bee detections during visual surveys only. Although full datasets yielded comparable results between TTD and DND models, subsampling introduced large bias and decreased precision, particularly for net-only and single-survey TTD models. For conservation practitioners with limited time and budgets for monitoring, incorporating targeted visual surveys into community-level surveys can increase detection of species of concern without increasing the number of visits.