Calibration of an energy balance model to simulate wintertime soil temperature, soil frost depth, and snow depth for a 14 year period in a highland area of Iran

A physically-based heat and mass transfer model, CoupModel, is calibrated to simulate wintertime soil temperature, soil frost depth, and snow depth for a 14-year period in a highland area of Iran. A Monte Carlo based approach is used for calibration process based on subjective performance criteria. Sensitivity and uncertainty analyses of the model were performed by selecting 30 parameters and the model was run using 22,000 samples taken from the uncertainty range of the parameters. By using the Nash-Sutcliffe Index to evaluate the performance of the model and applying a cutoff threshold for the performance to snow depth and soil temperature, 161 behavioral simulations were recognized and considered as the accepted ensemble to represent the field conditions. Sensitivity analysis of the model revealed some parameters associated with soil evaporation, soil hydraulic properties, and snow modeling as sensitive and highly important parameters. Uncertainty analysis of the model for wintertime soil temperatures showed a reasonable agreement between simulations and observations in most cases. However, a systematic error occurred at some periods because of high uncertainty of the actual snow density and details of snow melting. Uncertainties were also due to the simplified model assumptions regarding snow thermal properties and temperature within snow cover. The snow depth at the accumulation and melting stages were described well by the model in most cases. (C) 2015 Elsevier B.V. All rights reserved.