Prediction of the spatial distribution and relative abundance of ground-dwelling mammals using remote sensing imagery and simulation models

摘要: We present an approach that allows current, retrospective and future relative abundances of mammal species to be predicted across landscapes. A spatial generalized regression model abundance based on habitat quality time since disturbance was combined with coverages the distribution derived from a simulation which predicts historical arrangement forest habitat. The strength this is input data can as part standard inventory mapping program addition high resolution remote sensing imagery. Furthermore, it operates at scale used for wildlife management in Australia, makes widely applicable. To demonstrate we use collected over 20 years long-nosed potoroo (Potorous tridactylus) large wallabies (red-necked wallaby, Macropus rufogriseus, swamp Wallabia bicolor) their habitats following wildfire. Results indicate has increased, initially sparse numbers less than 0.5 % plot-night occurrences close 3% approximately twenty after major fire event. by contrast decreased about 20% Presently modelled 2% plot-nights tracks very low. Predictions without additional were low, region likely unsuitable next 5 years. These models offer tools investigating current key provide managers thereby translating scientific understanding into suitable every-day managers.