The Deforestation Driver & Carbon Emission (DeDuCE) model aims to identify deforestation—the permanent replacement of natural forests by other land-uses—across the globe to expanding croplands, pastures, and forest plantations. It then links this deforestation to the commodities produced on the deforested land and estimates the carbon dioxide emissions resulting from this land-use change.
The model does so by overlaying satellite data on forest loss with maps of specific crops (e.g., soybeans, oil palm, cocoa, and rubber) or of broader land-uses (e.g., croplands, forest plantations, and pastures) or deforestation drivers. Through a procedure that prioritizes data with higher spatio-temporal accuracy and detail, the model identifies where deforestation occurs and attributes this directly to a commodity using spatial data (e.g., soybeans) or to a broader land-use (e.g., agriculture or cropland). Where deforestation cannot be spatially attributed to a specific commodity, the model uses non-spatial agricultural and forestry statistics to assess commodity-driven deforestation in a two-step procedure: first, deforestation attributed to broad land-uses (e.g., agriculture or commodity production) is further subdivided between cropland, pastures, and forest plantations based on their relative (gross) expansion in a region (typically at country-level); second, deforestation attributed to cropland expansion (either based on cropland maps or statistics) is further allocated between different crop commodities in proportion to their respective increase in harvested area.
Finally, the model estimates carbon losses due to deforestation using maps of forest carbon stocks—in above- and below-ground biomass, dead wood, litter, and soils—and accounts for the carbon sequestered in the replacing land use. Furthermore, carbon dioxide emissions from peatland drainage are estimated by overlaying the identified deforestation data with a map of the global extent of peatlands.