Contrasting wildlife detection trends without occupancy change: interpreting long-term camera-trap indicators in a protected woodland

•Human recreation is increasingly widespread in protected landscapes, raising concern about its potential effects on wildlife. At the same time, long-term camera-trap networks are becoming central tools for biodiversity monitoring, yet detection-based indices integrate abundance, movement and behavioural activity, complicating interpretation. Long duration, standardised datasets therefore offer valuable opportunities to distinguish behavioural disturbance from broader ecological change.
•We analysed 16 years of camera-trap data from five fixed woodland locations within a protected landscape to evaluate long-term wildlife trends and assess whether detection-based indicators reflect recreational disturbance or broader ecological dynamics. Species-specific temporal trends were estimated using generalised additive mixed models accounting for sampling effort and location. Dynamic occupancy models were used to distinguish changes in site use from changes in detection or behaviour, and short term disturbance models tested behavioural responses to daily variation in recreation.
•Wildlife detection trajectories differed among species. Deer species showed overall declines, whereas mesopredators and small mammals were stable or increased. Despite these contrasting detection trends, occupancy probabilities remained stable. Recreational activity also varied among user types: cyclist and dog-related detections increased, while detections of people without dogs or bicycles showed little long-term change. Short-term disturbance effects were weak and inconsistent, and strong diel segregation between humans and wildlife persisted across recreation levels.
These findings suggest routine recreation alone is unlikely to explain long-term changes in wildlife detections. Long-term camera-trap monitoring can reveal shifts in wildlife activity patterns and community structure, but detection-based indicators require cautious interpretation alongside occupancy estimates and broader ecological context.