Open-Earth-Monitor Global Workshop 2025

Tropical forests are biodiversity hotspots, providing critical ecosystem services that sustain millions of plant and animal species. However, these forests are increasingly threatened by human activities, through the expansion of commodity crops such as soy, oil palm, rubber, cocoa, coffee, corn, logging, avocado, and pasture (Masolele et al., 2022, 2024). While significant efforts have been made to monitor deforestation using satellite imagery, most initiatives stop at detecting forest loss without tracking the land use that follows (Hansen et al., 2013). Understanding post-deforestation land use is crucial for addressing deforestation's root causes and mitigating its impacts (Masolele et al., 2022, 2024).
Currently, there is no global monitoring system capable of providing annual, spatially detailed updates on the land use that follows after deforestation. Existing datasets and methods frequently lack the spatial, thematic, and temporal resolution necessary to accurately map post-deforestation land uses (Curtis et al., 2018), limiting their utility for targeted rapid policy response and regulatory compliance, such as the European Union’s Deforestation Regulation (EUDR) (European Commission., 2024). This gap poses challenges for ensuring EUDR compliance, limiting the capacity to detect and mitigate deforestation linked to commodity production. Here, we present the first high-resolution (10 m) maps of land use following deforestation covering the entire pan-tropics. We utilize an extensive reference database containing 23 different land use types (including, soy, oil palm, rubber, cocoa, coffee, corn, logging, avocado, mining, cashew, corn, sugar, rice, and pasture), and employ Sentinel-1 and Sentinel-2 data combined with deep learning algorithms, to map land use following tropical deforestation from 2001 to 2023 with an F1-score of 83%. Our approach incorporates location encodings and environmental variables, such as elevation, temperature, and precipitation, to enhance the model’s ability to distinguish various land uses across diverse geographies. In general our results shows increased deforestation as a result of expansion of key commodity crops such as cocoa in Liberia, Cameroon, Ivory Coast, Ghana, Ecuador, Peru, Papua New Guinea; oil palm, in Indonesia, Malaysia; rubber in Malyasia, Thailand, Laos, Indonesia; coffee in Central America (Guatemala, Nicaragua, Costa rica), Peru, Ethiopia, Colombia, Vietnam; soy in Brazil; pasture in Paraguay, Bolivia, Mexico, Brazil, Cashew in in Cambodia, Tanzania, Mozambique, Benin and, logging in Suriname, Guyana, Papua New Guinea, Equatorial Guinea, Gabon, Republic of Congo, and Cameroon.
This work directly supports the European Union’s Deforestation Regulation (EUDR), aimed at curbing the EU market’s contribution to global deforestation (European Commission., 2024). Our research offers crucial insights for monitoring land use following deforestation, aiding environmental conservation initiatives and advancing carbon neutrality goals by providing detailed, high-resolution maps on land use that follows after deforestation events.