Streamflow prediction based on the soil and water assessment tool in the Pajeú river basin, Brazilian semiarid

Abstract

It is considered highly important for developing strategies to conserve water resources and minimize environmental impacts, especially in places with low rainfall and high potential evapotranspiration, such as semiarid regions. In this context, the main objective of this study was to evaluate the ranking of the most sensitive parameters and the calibration and validation of the streamflow between 1976 and 2018 in the Pajeú river Basin using the SWAT hydrological model. The SUFI-2 algorithm, integrated into the SWAT-CUP software, was used for sensitivity analysis (SA), calibration (from 1976 to 1995), validation (from 1999 to 2018), and uncertainty analysis. Results showed that only seven parameters related to land use and land cover (initial SCS runoff curve number for moisture condition II – CN2), groundwater (baseflow alpha factor – ALPHA_BF, groundwater delay time – GW_DELAY, threshold depth of water in the shallow aquifer required for return flow to occur – GWQMN, and groundwater “revap” coefficient – GW_REVAP), and runoff (slope length for lateral subsurface flow – SLSOIL and lateral flow travel time – LAT_TTIME) were considered the most sensitive and were subsequently calibrated. Thus, CN2 was reduced by 28%, consequently decreasing surface runoff. ALPHA_BF showed a slow response to recharge, while GW_DELAY affected groundwater discharge by 47 days. GWQMN increased by approximately 2586.9 mm, improving streamflow simulation. GW_REVAP, with a value close to zero, exhibited a reduction trend in the groundwater table during the dry season and in the upward movement of water to the root zone. The basin slope class ≤8% increased SLSOIL to 29.41 m, which in turn increased LAT_TTIME by 99.47 days. SWAT also simulated maximum streamflows associated with extreme precipitation events in an acceptable manner. The Floresta gauging station achieved more satisfactory performance, as observed in the percentage of observations covered by the 95% prediction uncertainty – 95PPU (P-factor), the average thickness of the 95PPU band divided by the standard deviation of the measured data (R-factor), the coefficient of determination (R²), Nash-Sutcliffe efficiency (NSE), and percent bias (PBIAS), with respective values of 0.68, 1.05, 0.64, 0.63, and 6.5% for calibration. Therefore, investigating hydrological processes in this region allows for the development of water planning strategies, given the significant impacts on water resources and agricultural production, especially during drought events.