| description abstract | In hydrologic modeling, the value of the model parameters is often influenced by the rainfall patterns of the calibration period. However, modeling studies typically only use a single calibration period, with parameters constrained by the associated hydrologic conditions. In this paper, we explore the use of multiple, hydrologically disparate calibration periods and investigate their influence on parameter selection. Specifically, we present a dual calibration approach designed to predict different aspects of the hydrograph: one model scenario is calibrated to replicate wet period conditions to support peak flow conditions, while the other experiment is calibrated to dry periods to support evaluating ecological flows. We use a surface–subsurface hydrologic model (SWAT+, enhanced with new groundwater module gwflow) applied to the semiarid, snowmelt-dominant, intensively irrigated Cache la Poudre River Basin (4,824 km2) in Colorado for the 2000–2015 period. The model is tested against monthly streamflow. We conduct a global sensitivity analysis to explore the seasonal sensitivity of streamflow, high flow, and low flow during wet and dry seasons and to evaluate the watershed-average hydrologic fluxes for the wet- and dry-calibration experiments, highlighting the key differences between the two calibration experiments. Sensitivity analysis shows seasonal shifts in the relation between parameters and simulated streamflow: during wet periods, runoff lag time and aquifer specific yield control, while dry periods highlight hydraulic conductivity and percolation. The model effectively simulates high flows during wet periods but has difficulties in dry conditions, underscoring the complexities of successfully designing and deploying a single mode and calibration strategy across varying hydrologic regimes. Low-flow simulations during dry calibration exhibit reduced precision; however, wet testing enhances performance, illustrating the model’s ability to accurately represent low flows in wetter settings. This research highlights the intricacies of simulating the full range of flow dynamics in complex watersheds. | |
