Simulação hidrológica na bacia hidrográfica do Rio Araguaia

Abstract

Hydrological modeling is an important tool of water management. There are different techniques for hydrological modeling whose ultimate goal may be flow prediction. In addition, global climate change can influence water production and quality around the world. The thesis is divided into two parts, in the first part: literature review on the subject of hydrological modeling with emphasis on semi-conceptual hydrological models and empirical hydrological models by Artificial Neural Networks (ANN). The second part consists of two articles, and the first aimed to calibrate and validate two models (SWAT and ANN) for sub-basins of the Tocantins-Araguaia watershed. The models were calibrated and validated and met the quality criteria based on the Nash-Sutcliffe Efficiency Ratio values. For the Cristalino River basin, RNA modeling presented better performance than conceptual modeling. In the other basins, the semi-conceptual and empirical models presented similar performances. The second article aimed to evaluate the possible hydrological impacts of global climate change resulting from projections based on Representative Concentration Pathways (RCPs) 4.5 and 8.5 for Tocantins-Araguaia watershed sub-basins. The monthly and annual average flows, the reference flows Q50, Q90 and Q95 were evaluated. It is concluded that climate models indicate the possibility of rainfall reductions in both CPRs being more critical in CPR 8.5. The Eta-CanESM model projects more severe reductions, while the Eta-HadGEM2-ES model is the most conservative. The Cristalino River basin will be most impacted in climate change scenarios.