Google Earth Engine aplicado na Análise Espaço-Temporal da Seca Meteorológica e nas Projeções CMIP6

Abstract

Meteorological drought is a recurrent phenomenon across the Brazilian territory, representing one of the country's major environmental and socioeconomic challenges. It directly affects water security, agricultural production, energy generation, and the availability of natural resources. Droughts are characterized as slow-onset natural disasters whose complexity lies in their spatio-temporal variability and lack of defined geographical boundaries. In Brazil, the irregularity of precipitation, coupled with the progression of global climate change, has aggravated the intensity and frequency of these events, increasing climate risk in vulnerable biomes such as the Caatinga and the Cerrado. Given this problematic, understanding drought dynamics and their standardized quantification is essential for subsidizing public mitigation and adaptation policies. This study aimed to analyze the spatio-temporal variability of meteorological drought in Brazil using the Standardized Precipitation Index (SPI). It integrates a historical assessment (1961–2024) and future climate projections from the Coupled Model Intercomparison Project Phase 6 (CMIP6), specifically under the medium and high forcing scenarios (SSP2-4.5 and SSP5-8.5). The methodology was implemented on the Google Earth Engine (GEE) platform, an environment that enables the cloud-based processing of large volumes of geospatial data, fostering the scalability and replicability of the analyses. Moving windows of 6, 12, 24, and 48 months were applied to represent different durations of dry and wet events. We observed that the Northeast and Midwest regions exhibit greater susceptibility to prolonged drought events, resulting from the combination of rainfall variability and local climatic fragility. The projections indicated an intensification of droughts under the highemissions scenario, revealing spatial shifts of critical areas and increased persistence of arid periods. In contrast, the sensitivity of short-term temporal scales tends to reflect seasonal events more accurately, reinforcing the importance of multiscale analysis. The present study establishes a scope that integrates the spatial, temporal, and operational dimensions of research by characterizing meteorological drought in Brazil, covering the national scale for the historical period of 1961–2024 and in projections up to the year 2100. The work consists of a standardized evaluation of meteorological drought based on an internationally recognized climate index, developed through an automated, geospatial, and reproducible processing method that objectively quantifies the frequency and severity of dry events. The proposed approach demonstrates high potential for applicability and scalability, constituting a technical methodology that can be incorporated into territorial planning systems, water risk management policies, and national and regional climate change adaptation strategies, providing scientific and operational support for understanding droughts in Brazil within a scenario of continuous global warming.