Using Occupancy Modeling to Estimate American Marten Distribution in Michigan’s Lower Peninsula

The current distribution American marten (Martes americana) in the Lower Peninsula of Michigan, USA is largely unknown. We used detection-nondetection data from a large-scale camera trap study encompassing 27,371km2 of public lands on within the 1836 Ceded Territory to estimate marten occurrence across Michigan’s Northern Lower Peninsula. We modeled marten detection probability and occupancy as a function of covariates from two summer season that took place in 2019 and 2020. Model averaged predictions indicate high occupancy probabilities (ψ \u3e 90%) when agricultural land was low (\u3c 12% per 8km2) and canopy cover was high (\u3e 85% per 8 km2). Additionally, we conducted multiple single season occupancy models to compare basic detection probabilities when marten were known to be present (ψ = 1). Across both years, summer detection probabilities were drastically lower than winter with an average detection rate of 2% - 46% in the summer, and 10% - 85% in the winter. Furthermore, we did not find any significant relationship between detection probabilities and weather data, however, Julian date of the survey was found to be significant in the winter of 2020. In this study, we demonstrate the use of occupancy modeling as an aid to determining species distribution and improving management decisions for this imperiled species throughout its range in Michigan