SOUZA NETO, P. F.; http://lattes.cnpq.br/4939680141716056; SOUZA NETO, Pedro Fernandes de.
Abstract:
The São Francisco River basin (BHSF) is the main basin of the Brazilian Northeast water
system due to its capillarity in the water systems of this region, contributing to ensure the water
availability of several states in the Northeast. Simulating the rainfall and runoff patterns of this
basin along its four physiographic regions, spatial division along the main river, is substantially
strategic for the management of water resources in the States that benefit from the use of water
in this basin. Thus, the objective of this study was to analyze the hydroclimatology of the São
Francisco River basin using regional climate modeling. In this sense, the RegCM climate model
was used to simulate precipitation, temperature, potential evapotranspiration and runoff in the
São Francisco River basin for a period of ten years (2000 to 2010). The model was integrated
using ERA-interim reanalysis data as initial and lateral boundary conditions. Sensitivity tests
of the cumulus parameterization in the ocean and on the continent were carried out due to the
close relationship between precipitation and runoff, and the uncertainty in choosing this
parameterization in the simulations. In these tests, the most adequate parameterization in the
representation of precipitation patterns in the basin was evaluated. This evaluation was
conducted from an objective analysis using functions that quantify the dexterity of the simulated
rain in relation to the observed rain from three different sources. The uncertainty in the
definition of the grid domain used in the simulations motivated the performance of a sensitivity
test to evaluate this aspect, configured as follows: the first test, with a larger domain, and the
second with a smaller domain, whose limits have a configuration significantly away from and
close to the basin, respectively. In general, among the evaluated parameterizations, Emanuel's,
both in the ocean and on the continent, was the one that best represented the precipitation in the
basin, and with the use of a greater grid domain. As for hydroclimatic modeling, the model
satisfactorily simulated seasonal and annual rainfall patterns, evapotranspiration and basin
runoff. The Submédio and Alto São Francisco sub-basins presented the best results in the
simulation of rainfall and potential evapotranspiration. However, although the simulations
represent the hydroclimatic patterns of the basin, rainfall and potential evapotranspiration were
underestimated, which implied an overestimation of runoff in the basin. The continuity of
hydroclimatic modeling studies with RegCM at BHSF requires an analysis and probable
improvements in the surface modulus of the model in order to adjust the simulated
evapotranspiration, and therefore, adjust the precipitation and runoff in the basin.