Impacts of model calibration on high-latitude land-surface processes: PILPS 2(e) calibration/validation experiments

In the PILPS 2(e) experiment, the Snow Atmosphere Soil Transfer (SAST) land-surface scheme developed from the Biosphere–Atmosphere Transfer Scheme (BATS) showed difficulty in accurately simulating the patterns and quantities of runoff resulting from heavy snowmelt in the high-latitude Torne–Kalix River basin (shared by Sweden and Finland). This difficulty exposes the model deficiency in runoff formations. After representing subsurface runoff and calibrating the parameters, the accuracy of hydrograph prediction improved substantially. However, even with the accurate precipitation and runoff, the predicted soil moisture and its variation were highly “model-dependent”. Knowledge obtained from the experiment is discussed.