The North American beaver (Castor Canadensis) is an ecosystem engineer and keystone species that widely influences stream hydrology and ecosystems within the Hudson Bay Lowlands (HBL) by constructing dams across stream channels. Many of these dams can also span the surrounding floodplain valleys and extend for over several hundred meters. Previous studies have shown that changes in the extent of beaver ponds and wetlands mapped using aerial photographs can serve as an indicator of shifting regional abundance of beavers. In this research we are investigating the use of freely available satellite data to map long-term changes in beaver ponding in the HBL. Inter-annual variability in beaver pond extents since 1985 is being measured using sub-pixel surface water mapping methods applied to the 30-m resolution Landsat satellite archive. We are also investigating the long-term effects of fire disturbance on beaver activity and developing a machine learning model to predict beaver pond locations across the HBL. An initial application of the beaver pond mapping method to a 5127 km2 study region in the Coastal Hudson Bay Lowlands indicated that ponding can demonstrate large, multi-year fluctuations. In a second HBL study region, we found that fire disturbance in mature conifer forest leads to rejuvenation of broadleaf beaver forage and expansion of beaver ponding by 140% after several decades. A preliminary satellite-based predictive model of beaver pond locations across the entire HBL indicates a patchy distribution, with relatively higher density in the southern and coastal regions. We will apply the sub-pixel mapping method to measure long-term changes in the extent of all beaver ponds identified using a final model.