Cooperative Fish and Wildlife Research Units Program: Pennsylvania
Education, Research and Technical Assistance for Managing Our Natural Resources

Pennsylvania encompasses the southernmost range of snowshoe hares. Changes in climate or habitat conditions could result in hares becoming extirpated from Pennsylvania. Our study will be essential to the Pennsylvania Game Commission as it attempts to implement any management actions to offset potential effects of habitat changes and global climate change. Identifying if and how the distribution has changed and what habitat conditions impact those changes for hares in Pennsylvania will provide the PGC with the necessary information to make appropriate management decisions for harvest regulations, identify priority areas for conservation and management action, and to protect and manage existing habitat.

In year one, we will evaluate how distribution of snowshoe hares has changed by replicating the 2004 track and fecal pellet methods used by Diefenbach et al. (2016). We will delineate the geographic distribution and large-scale habitat associations of snowshoe hares across northern Pennsylvania by collecting fecal pellet groups deposited by cottontails and hares and then extract DNA from fecal pellets to discern the presence of snowshoe hares. If snow is present, tracks will also be used to identify presence of snowshoe hares and hares may be visually detected. Technicians will navigate to pre-selected sampling locations and sampling sites will be searched for sign of cottontails and hares, and fecal pellet groups will be collected and stored in ethanol. DNA will be extracted at the Pennsylvania Cooperative Fish and Wildlife Research Unit genetics lab. Genetic analyses will be conducted by the Nucleic Acid Facility, Life Sciences Consortium, Pennsylvania State University. Mitochondrial DNA will be extracted from epithelial cells found in fecal pellets and gene sequences unique to each species of lagomorph will be used to identify the presence or absence of snowshoe hare, cottontail rabbit, and Appalachian cottontail. Additional genetic analyses will use microsatellite markers to investigate the spatial structure of genetic diversity to estimate levels of gene flow and identify isolated populations.

To ensure a representative sample of hare habitats, we will use a probability-based sampling design to select areas to search for fecal pellets. Repeated samples will be taken from a portion of the sample sites to obtain estimates of detectability, and hence correct for errors attributed to the failure to detect hares when they are in fact present. For each sampled site, we will use a geographic information system (GIS) with vegetation cover types and other digital layers to identify habitat characteristics associated with the presence of snowshoe hares. To suggest a plausible model for the effects of habitat characteristics, a logistic regression will initially be fit to the presence/absence data for snowshoe hares. However, logistic regression does not take into account non-detection errors, and so can lead to under-estimates of the probability that hares are present. Logistic regression is based on the untenable assumption that observations are independent and not spatially correlated. Therefore, new geostatistical methods shall be developed to correct for non-detection errors, to account for spatial dependence in the data, and to predict whether or not hares are present at unsampled sites. Once we have identified habitat or landscape characteristics associated with the presence or absence of hares we can use a GIS, coupled with the newly developed geostatistical procedures, to map the distribution of hares across Pennsylvania. Estimates of uncertainty of model predictions will be an essential component of our analyses.

In year two, we will identify focal areas in McKean, Cameron, and Elk counties to evaluate fine scale habitat features influence on local snowshoe hare population density and potentially population connectivity. Features we will evaluate include stem density of mid and understory vegetation, height of midstory, deer browse, and coarse woody debris among others. One-mi² focal areas will be identified within this core Pennsylvania hare range that was identified by Diefenbach et al (2016). Specific areas will be selected for landscape features such as scale of timber sales, aspect, cover type, and stand age.