Mapping the Future: How Can‑Peat researchers are building the next generation of peatland maps

What if we could see Canada’s peatlands, their carbon stocks, disturbances, and restoration potential, layered in a single interactive map? That vision is moving closer to reality through Can‑Peat’s rapidly expanding geospatial initiatives.

Across the Can‑Peat network, researchers are developing an impressive suite of geospatial products that will shape how Canada identifies, protects, and restores peatlands. Rather than working in isolation, these teams are building complementary maps that, together, create a powerful foundation for decision support, climate mitigation, and land‑use planning.

April Dalton, Postdoctoral Fellow at the Université du Québec à Montréal, is producing new national‑scale peat depth and carbon‑stock maps using machine‑learning approaches trained on almost a hundred thousand peat core measurements. By integrating carbon stocks and peat depths, her models will enable users to visualize carbon hotspots, uncertainty, and potential restoration gains.

Working alongside her, Sanghyeon Song, PhD Candidate at McGill University, is developing machine‑learning layers that map anthropogenic disturbance types across Canada’s peatlands. His models identify features such as extraction areas, roads, seismic lines, and wildfire footprints, filling a critical gap for land managers and policy teams evaluating cumulative impacts.

Nicholas Pontone, PhD Candidate at Carleton University, is advancing wildfire‑related mapping with a focus on long‑term carbon balance. He is creating an improved peatland vegetation and subclass map by expanding national coverage, strengthening classification accuracy especially in mountainous regions, and adding a swamp class. In parallel, Nicholas is developing a new Peat Moisture Code that will better represent peat fuel conditions within Canada’s Fire Danger Rating System. His 30‑metre resolution second‑edition boreal peatland map is expected before fall 2026.

Rebecca Edwards, Lead Remote Sensing Specialist at Ducks Unlimited Canada, is refining DUC’s Enhanced Wetland Classification, a multilayer machine‑learning inventory that distinguishes wetland classes using optical, radar, elevation, and canopy‑height data. With extensive helicopter and ground‑based validation, Rebecca’s updated workflows significantly improve the accuracy of bogs, fens, and swamps, offering a robust baseline for national peatland analyses.

In the United States, Erik Lilleskov, Research Scientist with the USDA Forest Service, is developing restoration‑targeting maps that intersect soil surveys, land‑cover datasets, and infrastructure layers under the PeatRestore project. His current peat map, built from remotely sensed soil surveys, opens new opportunities for cross‑border harmonization with Canadian datasets.