Assessment of surface water at the Sobradinho reservoir under the effects of drought using multi-temporal Landsat Images

Abstract

The matrix of energy generation in Brazil is predominantly hydroelectric power. Consequently, the reservoirs need constant monitoring due to the large volume of artificially dammed water. Images from remote sensing can provide reliable information concerning water bodies. In this paper, we use remote sensing imagery to monitor the Sobradinho dam in three different epochs. The objective was to verify quantitatively the area of the dam's surface reduced due to the drought of 2015, which was considered the worst in history. The approach used water surface area estimations from bands of Landsat5 and Landsat8 satellites which highlight water bodies better from other features present on surface of the Earth. Through the techniques of growth region and normalized difference water index (NDWI), we determined the surface area of the reservoir in 2011 and calculated the decrease caused by the drought. By analyzing the numbers provided by the results it is possible to estimate how the Sobradinho reservoir has been affected by the drastic drought. The results show that the Landsat images enable the monitoring of large reservoirs. Bearing in mind that monitoring is a primary and indispensable tool, not only for technical study, but also for economic and environmental research, it can help establish planning projects and water administration strategies for future decisions about the hydrical resource priority.