Uso da regressão linear múltipla na predição do nível dos reservatórios a partir de imagens de satélite.

Abstract

The present study focused on the application of the Google Earth Engine (GEE) platform for delineating flooded areas in seven reservoirs located in the state of Ceará. This analysis covered various dates throughout the operational cycle of these reservoirs. By using water level data, an approach based on multiple linear regression was adopted for each of the reservoirs in question. This methodology, complemented by the contribution of the Scilab programming environment, facilitated the establishment of links between the reservoir water level, the pixels composing the internal area of the reservoir, and the boundary pixels (perimeter). These relationships aimed at creating a management tool with significant potential. This developed tool, in turn, has the ability to estimate current water levels in reservoirs based on the analysis of satellite images. The outcomes of the analysis led to the formulation of multiple linear regression equations that exhibited determination coefficients (R²) exceeding 0.89, reaching up to 0.99. This high degree of correlation highlights the robustness and effectiveness inherent in the adopted methodology, emphasizing the association of GEE platform data with the Scilab programming environment.