Highlights

  • Many companies have begun assessing not only whether their supply chains are free from deforestation, but also whether they are avoiding the conversion of other (non-forest) ecosystems to agriculture—ensuring their supply chains are truly deforestation- and conversion-free (DCF).
  • This working paper compares widely used, open-access land use and cover change (LUCC) data sources that track soy-linked deforestation across Latin America. These datasets are commonly used to verify DCF soy supply chains but can produce different results for a variety of reasons explored in this paper.
  • Hansen Tree Cover Loss (Hansen TCL), INPE PRODES, and MapBiomas datasets are shown to produce different estimates at aggregate jurisdictional levels and at the pixel level depending on the forest and land use change definition used to produce the dataset, the minimum mapping unit (MMU), and forest canopy cover and canopy height applied.
  • When intersecting forest loss with a high-resolution soy production extent, the scale of the soy-linked forest conversion ratios is an order of magnitude smaller than general forest loss identified by the datasets.
  • Understanding differences in datasets monitoring forest loss provides a methodological foundation for a future analysis of non-forest loss and conversion.