Integrating Daily CO2 Concentrations in SWAT-VSA to Examine Climate Change Impacts on Hydrology in a Karst Watershed

Abstract. Highlights We used SWAT-VSA to assess the effects of climate change with rising CO2 on the water balance of a karst basin. For future climate, SWAT-VSA with rising CO2 yielded 7.1% less ET and 6.3% more runoff than standard SWAT-VSA. Rising CO2 also affected variable source areas, with greater ET declines and runoff increases in the wettest soils. Findings suggest CO2 effects on water balance should be included in future climate change studies with SWAT-VSA. Characterizing the effects of climate change on hydrology is important to watershed management. In this study, we used SWAT-VSA to examine the effects of climate change and increasing atmospheric CO2 (CO2) on the water balance of Spring Creek watershed, a mixed land-use karst basin in the Upper Chesapeake Bay watershed. First, we modified the stomatal conductance and leaf area index (LAI) routines of SWAT-VSA‘s Penman-Monteith evapotranspiration (ET) procedure and enabled the model to accept daily CO2 data. Using downscaled climate projections from nine global climate models (GCMs), we then compared water balance estimations from baseline SWAT-VSA against two modified versions of SWAT-VSA. One SWAT-VSA version integrated daily CO2 levels (SWAT-VSA_CO2), while another version added flexible stomatal conductance and LAI routines (SWAT-VSA_CO2+Plant) to the dynamic CO2 capacity. Under current climate (1985–2015), the three SWAT-VSA models produced generally similar water balance estimations, with 51% of precipitation lost to ET, and the remainder converted to runoff (10%), lateral flow (9%), and percolate (30%). For future climate (2020–2065), water balance simulations diverged between baseline SWAT-VSA and the two modified SWAT-VSA models with CO2. Notably, variable stomatal conductance and leaf area index (LAI) routines produced no detectable effects beyond that of CO2. For the 2020–2065 period, baseline SWAT-VSA projected ET increases of 0.7 mm yr-1, while SWAT-VSA models with CO2 suggested annual ET could decline by approximately -0.4 mm yr-1 over the same period. As a result, the two CO2-based SWAT-VSA models predicted streamflow increases of almost 1.6 mm yr-1 over the 2020–2065 period, which were roughly double the streamflow increases projected by baseline SWAT-VSA. In general, SWAT-VSA models with CO2 effects produced 22.4% more streamflow in 2045–2065 than the SWAT-VSA model without CO2. Results also showed that adding daily CO2 to SWAT-VSA reduced ET in wetter parts of the Spring Creek watershed, leading to greater runoff losses from variable source areas compared to baseline SWAT-VSA. Findings from the study highlight the importance of considering increasing atmospheric CO2 concentrations in water balance simulations with SWAT-VSA in order to gain a fuller appreciation of the hydrologic uncertainties with climate change.