Análise temporal da dinâmica na UHE Santo Antônio por meio de imagens multiespectrais

Abstract

The Amazon hydrographic region, characterized as the most extensive terrestrial hydrographic network in the world, covers almost 50% of Brazilian territory, with great potential for hydroelectric power generation due to its large volume of water and variations in the region's topography. The main basis of the country's electricity generation system consists of hydroelectric power plants (HPPs), which account for approximately 60% of Brazilian energy production. However, the presence of HPPs in Brazilian territory results in changes to the area's water resources, accompanied by significant environmental impacts. Given the environmental sensitivity of this area, this study analyzes the water dynamics of the Santo Antônio Hydroelectric Power Plant (HPP) and strategic tributaries, such as the Madeira River and the Madre de Dios River, between 2015 and 2025. The central objective was to identify and map variations in the water body, focusing on the detection of drought and flooding areas through time series data. To this end, image processing from Landsat 8 and 9 satellites (OLI/TIRS sensors) and the Normalized Difference Vegetation Index (NDVI) were used. Supervised classification was performed using the Minimum Distance (MinDist) method, allowing the quantification of spatial changes. The results revealed that the "Dry" class showed the greatest variability, being significantly impacted by the historical droughts recently recorded in the Madeira River basin. In contrast, the "Flooded" class demonstrated stability throughout the period. It is concluded that the use of low-cost geotechnologies is indispensable for precision monitoring in remote regions, providing support for water resource management in the face of the challenges posed by climate change in the Amazon.