Three existing soil models (SOIL, SHAW and Benoit's) capable of simulating soil temperatures and soil freezing were evaluated for their ability to predict soil frost depth and timing in northern and midwestern North America. The evaluation was carried out by form of comparison of field data, which came from two sites, one near Delta Junction, AK and the other near Morris, MN to simulated data. The Delta Junction site was characterized by a cryaquept with a grass cover, while the Morris site was a haploboroll with corn stubble. All three models were found to predict frost depth with reasonable accuracy, when the simulated snow depth closely matched the actual snow depth. When models did not accurately simulate snow depth, their predictions were often far from the measured data. The results illustrate the importance of snow cover in controlling soil frost as well as the difficulty in accurately simulating snow cover and soil frost.

Ian Kennedy
USDA-ARS
North-Central Soil Conservation Research Laboratory 
N. Iowa Ave.
Morris, MN 56267
   
Phone: 320-589-3411
ikennedy@mail.mrsars.usda.gov