Evaluating a parsimonious watershed model versus SWAT to estimate streamflow, soil loss and river contamination in two case studies in Tietê river basin, São Paulo, Brazil

Abstract

Study region: The Atibaia and Jacaré-Guaçu watersheds from the Tietê river basin, São Paulo, Brazil. Study focus: This study aims to compare estimates of flow, sediment yield and nutrients loads obtained from two distinct models with different structures and degree of complexity. The Generalized Watershed Loading Function (GWLF) and the Soil Water Assessment Tool (SWAT). We are particularly interested in understanding under which conditions the use of each model is to be recommended, namely when does the addition effort required to run the SWAT model leads to effective better results. As SWAT’s calibration procedure is cumbersome, the advantage of using a more detailed and distributed model fails to materialize when detailed data are not available or when monthly estimates are enough. GWLF model provides useful results with a reduced data gathering and calibration effort. New hydrological insights of the region: The joint calibration of both models to two watersheds offered a robust set of parameter values for prevalent conditions of Tiête river basin given the existing data set, although not all modelled variables are reproduced accurately. The performance of both models is adequate when estimating streamflow at a monthly time, but decreases when estimating daily flow, sediment yield, and nutrients loads. The poor monitoring of sediments and nutrients concentration hinders the ability to fully calibrate the model’s water quality component.

Keywords

• SWAT
• GWLF
• Parsimonious lumped hydrological model
• Distributed hydrological model