Forest Foresight — AI Radar-Based Deforestation Prediction Protecting Gabon's Rainforest
Gabon
WWF and Wageningen University's Forest Foresight uses Sentinel-1 radar and ML to flag illegal-logging and mining risk months ahead. Gabonese …
Peru · Lima · See the Peru profile · See the Lima profile
Evidence: Randomised controlled trial Top 14% 80/100 · Ask Evidence Copilot about this practice
Peru's Ministry of Environment has run Geobosques, a weekly satellite-based deforestation early-warning system, since April 2017; a PNAS randomized controlled trial found community patrols using its alerts cut deforestation 37% over two years versus untreated villages.
Peru holds roughly 78 million hectares of primary tropical Amazon forest, and its remoteness has long made illegal logging and land-use conversion difficult to detect and enforce against. Peru's Ministry of Environment (MINAM) launched Geobosques in April 2017 as a national early-warning platform ("Alertas Tempranas") built on Landsat satellite imagery, later supplemented with Synthetic Aperture Radar data - with support from Japan's JICA to the national forest service, SERFOR - to see through the persistent cloud cover that limits optical imagery.
Distribute timely, satellite-based forest-disturbance alerts widely enough, and reliably enough, that they can be acted on before deforestation spreads.
The platform distributes weekly forest-disturbance alerts to more than 2,500 subscribers across government, civil society and the private sector. A randomized controlled trial by Slough, Kopas and Urpelainen, published in PNAS in 2021, tested what happens when these alerts are paired with trained, incentivized community patrols. The programme, implemented by the NGO Rainforest Foundation US in coordination with indigenous federations, was randomly assigned to 39 of 76 Amazon communities.
A peer-reviewed comparison published in IOP Science found Geobosques detected 137,143 hectares of forest-cover loss between March and December 2017, versus 99,958 hectares flagged by the global GLAD alert system over the same period. Communities that received the alerts and patrol support cut deforestation by 52% in year one (2018) and 21% in year two (2019) relative to control communities - a combined 37% reduction over the two years.
The RCT's patrol-and-incentive intervention was run by an NGO layered on top of the government's alert data, not directly by MINAM, so the rigorous causal evidence speaks to how the alerts are acted upon locally rather than to the government platform's enforcement follow-through in isolation; Global Forest Watch's own reporting on the programme is titled to acknowledge both "triumphs and challenges" in turning alerts into results on the ground.
LIFE Programme Horizon Europe National / regional programmes
Indicative funding routes for practices of this type — always check each programme's current calls and eligibility rules.
Reusable artefacts from this practice — as published by their sources.
Do you run this practice? Claim it — verified implementers get a public contact pathway and can propose corrections.
Where this practice's information was retrieved from, and when.
Gabon
WWF and Wageningen University's Forest Foresight uses Sentinel-1 radar and ML to flag illegal-logging and mining risk months ahead. Gabonese …
Brazil
INPE's DETER system issues near-real-time satellite alerts on forest clearing across the Amazon and three other biomes, letting IBAMA dispatch …
Peru
A University of Sheffield machine-learning system detects illegally logged trees from very-high-resolution satellite imagery, cross-checked against Peru's forestry registers. OSINFOR …
Romania
Romania's Environment Ministry piloted Google Vertex AI on its SUMAL 2.0 timber-permit system, scanning 400,000 monthly permits and 1.6M photos …
Open full copilot Grounded in cited practices — always check the sources.