SENSITIVITY OF SOIL FREEZING SIMULATED BY THE SHAW MODEL

Abstract

ABSTRACT Sensitivity analyses conducted on hydrologic models are useful for understanding the complicated interactions of hydrologic processes and for demonstrating the potential errors resulting from incorrect estimation of input parameters. The effects of typical changes (measurement errors or natural variability) in input parameters on soil freezing as simulated by the Simultaneous Heat and Water (SHAW) model were investigated. The SHAW model simulates the interrelated heat and water transfer within snow, residue, and soil using estimated or measured weather data. Simulated frost depth was found to be very sensitive to small changes in air temperature (1° C) and initial snow depth (10 cm). Simulated frost depth was also potentially sensitive to the specified soil temperature for the lower boundary depending on the proximity of frost depth to the depth of the simulated profile. Site, residue soil characteristics having the greatest impact on simulated frost were slope, thickness of the residue layer, thermal conductivity of soil particles, soil bulk density, and a surface roughness parameter. Soil hydraulic parameters had little effect on frost depth, but had a large impact on water movement toward the zone of freezing and ice content of the frozen soil.