Forecasting Severe Meteorological Droughts in the Central Amazon Basin: A Supervised Learning Approach

The Amazon is the largest tropical forest in the world and plays a central role in the Earth's climate system, both influencing and responding to various climatic processes. Its primary basin, which shares its name, is closely coupled with global atmospheric circulation patterns. However, the increasing occurrence of droughts in recent decades is pushing this ecosystem toward a tipping point. This study develops an early warning system based on observational data to anticipate meteorological droughts in the Central Amazon Basin at seasonal lead times. We propose the use of machine learning algorithms to identify relationships between precipitation deficits in the region and sea surface temperature (SST) anomalies. This methodology, rooted in a supervised learning approach for time series forecasting, leverages signals from relevant SST modes to predict a regional meteorological drought index. The approach yielded promising results, demonstrating that observational data-driven methodologies can contribute to the anticipation of extreme events in this critical climatic hotspot.