Physically getting to beaver ponds and taking high quality data on them in the field is hard. Beaver ponds are muddy, mucky, sensitive ecosystems teeming with wildlife. They’re also enormous. Instead of tromping around the field, I develop methods to find them and collect data from them using satellite and aerial imagery. 

This crowdsourced effort will expand the training dataset and increase both image diversity and ecological variation, improving the machine-learning image recognition model EEAGER’s precision, recall, and generalizability.

This project is a collaboration with the National Parks Service on Wenabozho ominisan (the Apostle Islands) in Lake Superior. This study will utilize field observations, camera traps, and GIS analysis to map beaver colonies, document dam and lodge construction, and assess the resulting hydrological and vegetation changes. Particular attention will be given to how beaver-modified landscapes contribute to ecosystem resilience, biodiversity support, and potential wildfire mitigation within the unique island ecosystems of Lake Superior.