Comparative Evaluation of the Performance of SWAT, SWAT+, and APEX Models in Simulating Edge of Field Hydrological Processes

Hydrologic and water quality models are often used in assessing the response of environmental processes to human activities and climatic change. However, these models differ in terms of their complexity, requirements, underlying equations, and assumptions, and as such their performance in simulating landscape processes varies. Consequently, a key question that has to be addressed is to select the most suitable model that gives results closest to reality for an intended purpose. In this study, the performance of the basin-wide older version of SWAT is compared with that of the small watershed model APEX to assess the performance of both models at a field scale level. The new restructured version of SWAT (SWAT+) is compared with the older version to determine whether the new changes incorporated in SWAT+ have improved model performance, particularly for small watersheds. The three models were used to simulate the edge of field processes for a 6.6 ha plot located at the USDA-Agricultural Research Station near Riesel, Texas, and to predict water yield, soil, and mineral phosphorous loss from the micro watershed. Results showed that all the uncalibrated models over-predict soil and phosphorous loss in a micro watershed. Uncalibrated SWAT and SWAT+ models simulated water yield satisfactory albeit with low-performance metrics. The calibrated versions simulated water yield with indices close to optimal values. PBIAS as a performance assessment metric was determined to be overly sensitive and prone to numerical errors. SWAT+ will be helpful in the understanding of hydrological and water quality processes at micro watersheds considering that it addresses structural flaws associated with the older version, and the manually calibrated version matches the performance of both APEX and SWAT, despite the latter two undergoing rigorous automatic calibration.