Effectiveness of SAC-SMA model to simulate streamflow in the Colombian Andes: results for watersheds with different physical properties and high climate variability

Abstract

The Andean region of Colombia, characterized by hydrometeorological extreme events and watershed complex topography, give rise to the main rivers of the Colombia fluvial network, essential for agriculture, ecosystems, consumption, and hydropower generation. Therefore, promote hydrology research and direct policy making towards sustainable and integrated water resources management for the Andean region is needed. This study evaluates the spatially lumped Sacramento Soil Moisture Accounting (SAC-SMA) model for 12 Colombian watersheds located in the Andean region with different climate regimes and geomorphological features. The model inputs include rainfall and streamflow data from gauges stations and potential evapotranspiration and land cover satellite-​based data. The model calibration was performed using a Multistep Automatic Calibration Scheme coupled with the Shuffled Complex Evolution optimization algorithm. SAC-SMA Model performance was evaluated with Nash-Sutcliffe, Kling–Gupta efficiency, and Percent Bias. The model shows good performance, exhibiting NSE values > 0.5, KGE > 0.5 and Bias ±18% for the calibration period. Streamflow is underestimated for all watersheds in the calibration period and underestimated as well in the validation in 7/12 watersheds, however, the calibrated model was able to predict the observed outputs with reasonable accuracy. Correlation analysis show an impact of the watershed slope on the factors that control the different types of flows. The model results show the capabilities of SAC-SMA on representing the hydrograph recession and estimation of the baseflow in the study watersheds. This study provides information for land-use planning, agriculture, flood and drought risk management and water resources management in the Andean region.