Fires in the Peruvian Amazon: the impact of El Niño and government inaction
A dozen fires are raging across the Andes and the Peruvian Amazon, worsened by an exceptional drought linked to El Niño. In the absence of an official response, local communities are mobilizing to contain the flames.
Incendies en Amazonie et dans les Andes péruviennes : le feu qui se propage sous l'effet d'El Niño
A dozen fires are raging simultaneously in the Andes and the Peruvian Amazon, fueled by a record drought linked to El Niño. Copernicus satellite images show isolated but highly active fire outbreaks, while riverside villages fight with buckets and improvised chainsaws.
Current fires: location, extent, and satellite data
The flames have spread across more than 1,200 km², from the south of the Cusco region to the humid forest of Madre de Dios. Satellite data from the Sentinel‑2 mission, processed by neural networks, allows estimation of the burned area with a forecast uncertainty below 5% according to the GraphCast predictive model. This precision, made possible by machine learning applied to multispectral imagery, helps ground teams prioritize the most critical areas.
The ECMWF system, coupled with Pangu‑Weather AI models, anticipated the rise in fire risk as early as July, but the alerts were not relayed to local authorities. The lack of coordination between civil protection services and forecasters created an information gap that is reflected today in the extent of the fire.
El Niño and drought mechanism: why the Peruvian climate is shifting
El Niño, an oceanic phenomenon characterized by abnormal warming of equatorial Pacific waters, alters atmospheric currents and reduces precipitation in the western part of South America. Copernicus climate models indicate a precipitation anomaly of -30% over the Amazon basin over the past three months, which has dried out soils and made vegetation highly flammable.
The evapotranspiration mechanism, amplified by temperatures 2°C above the seasonal average, has increased the soil's energy load, promoting spontaneous combustion of organic matter. Neural networks integrated into ECMWF forecasts reproduce these non-linear interactions, but uncertainty remains high due to El Niño's natural variability.
Consequences for populations and the role of AI models in fire forecasting
Indigenous communities, already vulnerable, are suffering the loss of agricultural land, degradation of air quality, and forced displacement. According to ONI reports, fine particulate matter (PM2.5) concentrations exceeded WHO safety thresholds for several consecutive days, increasing respiratory risks.
