Recent decline of an endangered, endemic rodent: does exclusion of disturbance play a role for Hastings River mouse (Pseudomys oralis)?

Context The role of disturbance and its exclusion is fundamental to the conservation of threatened species. Aims We used the habitat accommodation model as a framework to investigate the importance of forest disturbance for the endangered Hastings River mouse, Pseudomys oralis, focusing on timber harvesting. Methods Our study comprised two separate surveys. We resurveyed old survey sites (n = 24) where the species was originally recorded as either present (logging excluded) or absent (subsequently logged). A second survey targeted trapping in high-quality habitat stratified by different times since logging. Finally, we analysed a 15-year trapping dataset targeting P. oralis to assess associations with co-occurring species. Key results The resurvey of old sites resulted in 12 P. oralis individuals being trapped, compared with 46 individuals in original surveys. Substantial declines were observed over time in transects where logging was excluded (60–82% decline), whereas there was little change at transects where P. oralis was not previously trapped and that were subsequently logged. The second survey yielded 27 P. oralis captures at post-logging sites assessed as high quality. Occupancy was very high (ψ = 0.93 ± 0.21) in transects logged 7–15 years ago and was 60% less in transects where logging was excluded for 35–45 years (ψ = 0.37 ± 0.22), whereas occupancy in transects logged 2–6 years ago was intermediate. This pattern of higher occupancy in logged areas was mirrored for the mean number of P. oralis trapped per transect. Ordination of habitat data showed an association of P. oralis with heath, mat-rushes and logs, whereas rats (Rattus and Melomys) were associated with ferns and shrubs. Camera traps revealed low background levels of predator presence. A negative exponential relationship was found between probability of occupancy of P. oralis and rat abundance from a 15-year trapping dataset (44 275 trap-nights), suggesting that rats may compete with P. oralis. Conclusions Our results supported the habitat accommodation model and suggested that disturbance is likely to influence the persistence of P. oralis. However, an interaction between predation and loss of cover from high-frequency disturbance (fire or intense grazing) cannot be excluded as a key threat. Implications Disturbance should be incorporated into the management of some species. Adaptive monitoring is recommended to assess alternative management regimes.