Calibration of a Freeze-Thaw Prediction Model for Spring Load Restriction Timing in Northern New England

A major problem with low-volume roads located in seasonal frost areas is their susceptibility to damage from trafficking during spring thaw. Therefore, seasonal load restriction (SLR) policies that limit the axle loads of heavy trucks during the spring thaw period have been implemented in many countries in an effort to minimize costly roadway damage. Several agencies have been addressing the question of when to place and remove SLRs and have expressed the need for a prediction model to aid them in the process of posting roads. Models are available which predict the depth of frost and thaw penetration based upon air freezing and thawing indices, requiring only air temperature data for input. Various forms of these models have been used by transportation agencies in the United States and Canada. When using any prediction model, a key element is model validation and calibration for local conditions. The purpose of the research described herein was to calibrate a freeze-thaw index model for use in SLR timing in northern New England. Atmospheric weather data and measured subsurface temperature data obtained from nine field test sites in New Hampshire over a period of three years were used in this analysis. Frost and thaw coefficients for the model were calibrated on a site-specific basis. Results suggest that frost-thaw patterns were reasonably estimated at most of the nine test sites using this model, although the model tended to be too conservative in estimating end-of-thaw dates, with estimated end-of-thaw dates falling after measured dates in many instances.