Occupancy and detectability modelling of vertebrates in northern Australia using multiple sampling methods

摘要: Understanding where species occur and how difficult they are to detect during surveys is crucial for designing evaluating monitoring programs, has broader applications conservation planning management. In this study, we modelled occupancy the effectiveness of six sampling methods at detecting vertebrates across Top End northern Australia. We fitted occupancy-detection models 136 (83 birds, 33 reptiles, 20 mammals) 242 recorded 333 sites in eight reserves between 2011 2016. For species, mean was highly variable: birds reptiles ranged from 0.01–0.81 0.01–0.49, respectively, whereas mammal lower, ranging 0.02–0.30. Of 11 environmental covariates considered as potential predictors occupancy, topographic ruggedness, elevation, maximum temperature, fire frequency were retained more readily top models. Using these models, predicted Australia (293,017 km2) generated richness maps each group. mammals high associated with rugged terrain, while bird highest coastal lowland woodlands. On average, detectability diurnal higher per day (0.33 ± 0.09) compared nocturnal night spotlighting (0.13 0.06). Detectability similar day/night pit trapping (0.30 (0.29 0.11). detectable using motion-sensor cameras a week (0.36 0.06), exception smaller-bodied species. One Elliott (0.20 0.06) (0.19 effective than cage (0.08 0.03) (0.05 0.04). Our estimates will help inform decisions about best redesign long-running vertebrate program