MULTI-SPECIES CARNIVORE MONITORING: INVESTIGATING STATISTICAL POWER TO DETECT MARTEN AND FISHER POPULATION DECLINES USING SPATIALLY-EXPLICIT SIMULATIONS.
Jody M. Tucker; U.S Forest Service, Pacific Southwest Region; 24545 Highway 120, Groveland, CA, 95321; (559) 359-5888; jtucker@fs.fed.us; Katie M. Moriarty, Jessie D. Golding, Martha M. Ellis, Jody M. Tucker and Katie M. Moriarty
Surveying for multiple species simultaneously increases cost-effectiveness and allows insights into community ecology. However, multi-species monitoring may not be effective for rare species. We assessed the statistical power to detect population trends for marten and fisher using a single survey framework. We used a spatially-explicit, individually-based framework to simulate biologically realistic populations and then simulated population declines to assess statistical power to detect this decline while varying grid size, detection probability, number of visits, effective survey areas, sampling frequency, and sampling location. We (1) tested whether we could effectively detect declines for both species; (2) explored trade-offs in power between number of visits, detectability, effective survey area, and proportion of landscape surveyed, (3) evaluated power for various survey frequencies to inform cost-efficiency, and (4) assessed power of irregular sampling (variable sampling locations over time). Simulations with low detection probability or irregular sampling locations were unable to detect declines regardless of sampling intensity. But with moderate detectability and number of visits we were able to simultaneously detect declines in both species as long as survey devices encompassed an effective survey area >6.75 km2 and occurred in fixed locations over time.
Ecology and Conservation of Martens and Fishers