Effects of forest management, weather, and landscape pattern on furbearer harvests at large-scales

Abstract

Over the past 50 years, Ontario’s forest landscape has changed due to ever increasing natural resource management. The natural vegetation pattern, forest composition, and the fire regime have been altered. Maintaining wildlife species diversity is an important goal of current forest management. However, little is understood about the impacts of large-scale land use and landscape scale processes that influence wildlife. This project used trapline harvest statistics from 1972-1990 to identify broad-scale effects of forest management, weather, and landscape structure on furbearers (marten, beaver, fisher, and lynx). Spatial variables for logging and fire disturbance, forest cover type, weather, spatial pattern, and road density were compiled in a geographic information system (GIS) and standardized by trapline. Regression models were created for each species and analysed at five spatial scales ranging from the Ontario Ministry of Natural Resources (OMNR) district (5000 sq. km) to the ‘provincial’ (800,000 sq. km) scales. The models were then compared temporally and spatially for consistency in variable contribution to the regression models. Forest cover type, weather, and spatial pattern variables accounted for the greatest variation in furbearer harvest, while disturbance and road density variables accounted for little variation. Model predictive capability ranged from 10 to 55% for all species. Marten models had the greatest predictive power (r2) at the ‘OMNR District’ scale, while fisher and beaver models had the highest r2 values at the ‘Hills site region’ and ‘provincial’ scales, respectively. Lynx models were inconsistent with relatively low predictive power at all scales. The models suggest that disturbance from forest management is not affecting furbearer harvests. Landscape scale variables such as forest cover type, weather, and landscape pattern account for a relatively high proportion of marten, beaver, and fisher harvests. These variables and the predictive power of the models reveal the influence that broad landscape factors have on wildlife.