Predicting forest soil temperatures from monthly air temperature and precipitation records

Abstract

A process-oriented forest soil temperature model, FORSTEM, is presented. FORSTEM considers vertical heat conduction as well as freezing and thawing, and it lumps the effects of forest canopies on soil surface temperature with the surface heat transfer coefficient. It runs in conjunction with the forest hydrologic model, FORHYM. FORSTEM and FORHYM input is limited to (i) air temperature; (ii) precipitation and its snow fraction; and (iii) descriptive site information (latitude, elevation, slope, aspect, forest coverage, and soil layer thickness and texture). FORSTEM uses generalized parameters derived from existing empirical information. The model was applied to 10 different cover type–site conditions, including lawns, deciduous forests, and coniferous forests before and after clear-cutting in Ontario, Quebec, New Brunswick, and Colorado. The only model parameter we calibrated for different sites was the effective ground/air conductance ratio. The ratio was found to be a function of incoming solar radiation and vegetative area index. Differences between monthly simulations and field measurements fell within ± 1.5 °C for at least about three-quarters of the data cases at individual sites. Major exceptions occurred when temperature measurements showed no damping down the soil profile or with soils containing large air gaps between coarse rock fragments.