Peatlands play an integral role in the global carbon and hydrologic cycles and make up 3% of the Earth’s total landscape. Despite their importance, there is a lack of accurate information on the global distribution of peatlands. With more precise information, Earth system models (ESMs) are able to simulate the effects of climate change to better understand its effects on the global carbon and hydrologic cycles.
To fill this gap, Melton et al. (2022) utilized machine learning (ML) to develop a global peatland fractional coverage map, referred to as Peat-ML, with an intended use in ESM applications. The model utilized information on climate, geomorphological and soil data, vegetation, and terrain, known drivers of peatland formation to detect peatland coverage. When compared against Ducks Unlimited Canada’s Enhanced Wetland Classification for the Boreal Plains Ecozone, an extensively ground-truthed map, Peat-ML was found to be of comparable quality for boreal peatlands. To learn more about Peat-ML's process click here.