A new data-driven map of peat distribution predicts substantial unknown peatland areas in Amazonia.

Tropical peatlands, the most carbon dense terrestrial ecosystems in the world, are a substantial but highly uncertain component of the global carbon cycle with wide-ranging estimates of their total area (441,025 -1,700,000 km²), and total below-ground carbon (105-215 Pg C). Protecting and restoring tropical peatlands can make a significant contribution to limiting CO₂ emissions and global warming, but policy instruments such as REDD+ and Nationally Determined Contributions to the Paris Agreement must be informed by high resolution maps of peatland distribution if they are to be effective.

In regions such as South East Asia, we have relatively high confidence in existing peatland estimates, supported by decades of research and field data. There have also been recent intensive field efforts in regions such as the Congo basin and Peruvian Amazonia, which have driven the development of more accurate regional peatland and below-ground carbon maps. However, substantial gaps remain in our understanding of the distribution of peatlands and their carbon store in some key tropical regions, including most of tropical South America.

Here we collate around 2,000 ground reference points in and around Amazonian peatlands, including data from Peru, Colombia, Brazil and Ecuador. These data are used to drive a high-resolution machine learning model predicting peatland distribution across the Amazon basin. We estimate a total peatland extent of approximately 200,000 km², greater for example than that of the Congo basin, with Brazil hosting the greatest predicted area. However, a lack of field data, particularly across some regions of Brazil, Bolivia and Venezuela, means that substantial uncertainties remain. We predict large areas of peatland in the Rio Negro basin which overlap with predicted maps of Campinarana or white sand forest habitats, but we do not know what proportion of these ecosystems represent true peatlands and how deep and carbon dense their peat could be. We discuss remaining uncertainties and call for expanding sampling in these ecosystems to better understand these potentially extensive peatlands. The past response of tropical peatlands to climate and land-use change, particularly in terms of carbon accumulation, also requires urgent investigation given that the central and eastern Amazon is projected to undergo more significant climatic change than the western basin.