Cold region freeze-thaw cycles result in serious problems with winter frost heave and springtime snowmelt and associated flooding. Frost heave and freeze-thaw cycles result in damage to both paved and unpaved roads as far south as California, Texas, New Mexico and North Carolina. In northern tier states and Canada fall and spring frost heave problems can so weaken roads that seasonal load limits must be imposed that restrict fully loaded trucks and slow local commerce. Determining when load limits be imposed and when they may be lifted has always been a challenge. Since the late 1970’s several models have been developed for both frost heave and the frozen soil physics of the active layer. Most of these models have addressed either the mechanics of frost heave or the parameterization of thaw depth. Currently, however, while much progress has been made to incorporate frozen soil physics into active use in numerical weather prediction models the complexity of the frost heave models has resulted in no current use of a frost heave model in a numerical weather prediction scheme.

The authors present a parameterization whereby frost heave and frozen soil physics are both incorporated into the land-surface scheme of a numerical weather prediction model. The model output is a component of ongoing research into enhanced surface transportation decision support systems.