Assessing improvements in environmental flow estimation through the use of a coupled surface-groundwater model

Abstract

Environmental flow refers to the water required to maintain river ecosystems. Establishing methodologies to ensure these minimum flow rates is essential, particularly in highly anthropized regions, such as Spain. Hydrological models, including SWAT+, are used to simulate rivers in natural regime. However, accurately representing surface-groundwater interactions is crucial in groundwater-driven catchments, such as the Tagus River headwaters. Due to the standalone model limitations, we used the new coupled SWAT + gwflow model to appropriately represent surface-groundwater interactions in streamflow response. The calibration approaches of the (1) base SWAT+, and (2) SWAT + gwflow models focused on adjusting low flows, allowing for greater accuracy in daily flows to estimate environmental flows. In addition, (3) an experiment of monthly-to-daily disaggregation of the standalone model based on a daily flow pattern was conducted at two upstream stations. Our results indicated that the coupled model outperformed the standalone model across all statistical indicators, especially in low flows, where the Nash-Sutcliffe logarithm (LnNSE), an indicator of low flow performance, was inadequate for the standalone model because it underestimated of the simulated flows. This underestimation was also reflected in the environmental flow values, with variations ranging between −41% and −58% compared to the observed values. For the base model, a disaggregation experiment showed an improvement in environmental indicators, reducing their underestimation by half. Finally, the SWAT + gwflow model showed the best fit for environmental flows compared to observations, with average variations of ±10% in most cases, showing improvement in daily low flows due to a better representation of surface-groundwater interactions.