Experimentally quantifying how long recreation noise displaces mammals: Recovery over weeks to months varies by disturbance type and species tolerance

Outdoor recreation is expanding across public lands, yet managers lack evidence for how long wildlife responses to recreation persist and whether response duration depends on disturbance type and frequency. Using a field-based acoustic playback experiment in the Bridger–Teton National Forest, Wyoming, USA, we isolated recreation noise from other components of human presence and quantified how long mammal detections took to return to baseline. Across 12 sites over two field seasons, we broadcast motorized and non-motorized recreation sounds at rates mimicking high- and low-use trails, paired with nature-sound controls. We quantified treatment effects as deviations from expected weekly detections at our control sites, defining time-to-return as the first week when deviations overlapped zero. Response duration differed between disturbance types. High-rate motorized playbacks produced the longest suppression: all species except mule deer did not return to baseline within our 15-week season, whereas mule deer recovered by six weeks. For high-rate non-motorized playbacks, all species except for mule deer had suppressed detections until seven weeks while mule deer detections were not different from baseline. For low-rate non-motorized playbacks, mule deer detections returned to baseline at nine weeks while all other species returned to baseline at seven weeks. Species differences indicate that apparent overall recovery can mask prolonged avoidance by less tolerant taxa. Recreation effects persisted for weeks to months and varied with disturbance type and exposure frequency, highlighting how temporal patterns of use influence wildlife responses and where recreational planning of spatial refugia or seasonal restrictions could reduce impacts on sensitive species.