| Summary: | Process-based calibration of a hydrological model is an important step to ensuring model fidelity, or how ‘faithfully’ the model reproduces reality, which is even more meaningful for the catchments in northern latitudes subjected to the complexity of cold regions processes. The effectiveness of process-based calibration is examined using the Hydrological Predictions of the Environment (HYPE) model implemented for the Nelson Churchill River Basin (NCRB) using multi-objective optimization to both streamflow and soil moisture observations. The calibration process is guided by time-variant sensitivity analysis using flow signatures, which was influential in detecting highly seasonal parameters that previously went undetected by conventional methods. The model calibration is further improved by vertical discretization of the default three soil layers in HYPE to seven soil layers, which improved soil thermodynamic processes and, ultimately the simulation of soil moisture and evapotranspiration over longer-term periods. Spatial evaluation of soil moisture suggested the seven-layer discretization better represents surface soil moisture storage, which is essential for long-term water balance, agricultural water management, and climate change studies. Finally, given the importance of model fidelity for long-term simulation, climate change impact assessment on permafrost degradation was examined using the discretized HYPE model. Results showed a reduction in permafrost coverage up to 82% by the end of the mid-future period under the RCP 8.5 scenario within the NCRB. The novelty of this work includes utilizing multi-objective optimization to improve process representation of soil moisture and evaporation across a large domain hydrologic model. This study also underscores the importance of long-term water balance projection at the continental scale, which is valuable for large-scale planning and implementation of sustainable development principles and guidelines for decision-making. May 2023
|
|---|