Assessing the Natural History and Habitat Use of the Fisher in Eastern North Dakota (SWG T-12-R)

Fishers were extirpated from North Dakota by the early 1900s as a result of over-harvesting and habitat loss. However, within the past decade there has been an increase in the number of verified reports in northeastern North Dakota. Determining habitat associations is important to evaluate how fishers use different habitats within their range to identifying areas of conservation priority. Fishers have been documented to be associated with large contiguous forested tracts that have extensive canopy cover. However, the forest in North Dakota is highly fragmented and presents a unique opportunity to assess how fisher preferences in less than optimal habitat. The purpose of my study was to evaluate natural history information, assess if occupancy and visitation patterns at detection sites vary depending on size and isolation of forested patches and compare the efficacy of track-plates and remote cameras at detecting the species. I created habitat covariates defining patch-size and isolation. I then used the software PRESENCE (MacKenzie et al. 2002) to determine if the covariates I created to define patch-size and isolation had an impact on fisher occupancy (ψ) at a site. I assessed model-fit using a Pearson chi-square test with a parametric boot-strap of 1,000 simulations (MacKenzie and Bailey 2004). Also, I ran a Poisson regression using the site covariates that defined a sites category of patch-size and isolation to assess the impact that site patch-size and isolation had on Latency to Detection (LTD). Fishers were detected more frequently in the diurnal hours in 2008 and more often in the crepuscular hours in 2009. Fishers had similar rates of occupancy regardless of patch-size or isolation, demonstrating their adaptability to occupy non-preferred habitat. I compared track-plates and remote cameras in their number of false absences, percentage of sites with a detection, percentage of check periods (time from set-up to re-bait and re-bait to pull) with a detection, unit effort (number of unique detections by number of DDs), and number of functioning days to total detection days. Of 127 sites, track-plates had false absences at 11 of 41 (27%) visits to a site, and cameras only failed to detect a fisher visiting a site (based on detections at the track-plate) on 4 of 41 (10%) occasions. Fishers were detected at 30 (24%) sites by track-plates and 37 (28%) sites by cameras. Cameras outperformed track-plates in every category except for initial cost. Cameras had less false absences, more detections, provided a more thorough detection history, and captured natural history information that the track-plates could not. Advances in camera technology have increased their reliability and performance enabling them to outperform track-plates when sampling for fisher presence. Cameras now provide more detection information and require less surveyor effort than track-plates.