Remote Sensing and Machine Learning for Drought Assessment: The Role of Aerodynamic Drivers in Ceará, Brazil
Year: 2026
Type: article
Abstract: This study aimed to create a spatiotemporal model of the Palmer Drought Severity Index in the state of Ceará, Brazil, using cloud computing and integrating a Random Forest machine learning model. Using the Google Earth Engine platform, the dataset selected for this work was TerraClimate, and the selected time interval was from 1994 to 2024. The model achieved a coefficient of determination (R²) of 0.72 and a Root Mean Squared Error of 1.27, demonstrating its ability to capture the non-linear dynamics of the region. The importance analysis of variables indicated 12-month accumulated precipitation as the main predictor (26.08%), followed by soil moisture (14.75%). Furthermore, wind speed (13.04%) stands out as a factor with greater influence than maximum temperature (8.19%), highlighting the relevance of the aerodynamic factor in the drought dynamics of the region. Spatially, the results estimate that approximately 8.7% of the state’s territory, particularly in the Inhamuns region and in the desertification nuclei of the municipality of Irauçuba, face extreme or severe drought vulnerability conditions. These findings suggest that data-driven models offer a robust and scalable alternative for assessing semi-arid areas with scarce data.
Institution: Universidade Federal do Ceará
Language: English
Subfield: Global and Planetary Change
Field: Environmental Science
Domain: Physical Sciences
SDGs: Climate action, Life on land
Open Access status: diamond
OpenAlex ID: W7223254215