Towards a better understanding of catchment hydrology in Brazil

Abstract

The catchment hydrology in Brazil is still poorly explored. The hydrological behavior, similarities, and functioning of Brazilian catchments are unknown and must be investigated. It is important to better understand the hydrological processes and responses to changes in climate to increase the ability to deal with hydrological extremes. The main objective of this doctoral thesis is to improve the understanding of catchment hydrology in Brazil. The GCMs/RCMs evaluation showed a good agreement (bias up to 10%) of downscaled annual simulations over the Amazon and Cerrado and large biases (reaching 40%) in the Pampa. I showed that HadGEM2-ES can represent long-term means for large areas and Eta RCM improves MIROC5 simulations. I also presented the CABra, which is a multi-source large-sample dataset including long-term data for 735 Brazilian catchments in eight attribute classes. The results of the satellite rainfall products evaluation showed that they performed better than ERA5 in estimating precipitation against ground observations. Streamflow and hydrologic signatures are better modeled with SM2RAIN-ASCAT and GPM+SM2RAIN. The catchment classification showed the existence of six groups of similar catchments: “non-seasonal”, “dry”, “rainforest”, “savannah”, “extremely-dry”, and “extremely-wet”. The streamflow into groups is mainly driven by the aridity index. The results found in this doctoral thesis provide benchmark material to benefit catchment hydrology investigations, exploration of new hypotheses and thereby advance our understanding of catchments’ behavior.