Soil moisture and discharge modeling in a representative watershed in northeastern Brazil using SWAT

Abstract

Abstract Soil moisture and discharge are interrelated variables which play a major role on water availability, environmental conservation and ecohydrological functioning of a watershed. In order to investigate their spatio-temporal dynamics, hydrological models are valuable tools. However, to minimize uncertainties of the modeled estimates, good quality distributed observational data sets are required for realist predictions. The aim of this study is to apply the Soil and Water Assessment Tool (SWAT) in a representative watershed of the northeastern Brazil, and to evaluate the effects of using both discharge and soil moisture datasets on the model uncertainties and predictions. The study area is located in the Mundau River Basin (4090.39\xa0km2). For streamflow, the values of the Nash–Sutcliffe model efficiency (NS) among gauge stations ranged from 0.71 to 0.92 in the calibration phase for the annual time step, and between 0.55 and 0.78 in the monthly time step. In the validation phase, NS values ranged from 0.53 to 0.76 for the annual time step, and between 0.62 and 0.72 for the monthly time step. Calibration and validation against daily soil moisture data resulted in a NS values of 0.53 and 0.52, and PBIAS of 0.4% and −1.1%, respectively. These results represent “satisfactory” and “very good” performances for discharge and soil moisture. Although there is still some amount of uncertainty, the use of complementary data, such as soil moisture, to calibrate and validate the SWAT model is useful, especially when discharge data are scarce, as for some watersheds in the semiarid zone.