A comparison of measured and simulated soil temperature using air temperature and soil surface energy balance as boundary conditions

Abstract

Numerical modelling of soil processes in the field can be based on soil surface temperature or soil surface heat flow as the thermal upper boundary condition. In this study the soil surface temperature and the heat flow were estimated from the air temperature and the soil surface energy balance, respectively, using a physically based numerical model. Simulated values of soil temperature, snow depth and frost depth for a lawn in northern Sweden were compared with measured data over a 22-year period. The air temperature boundary condition gave overestimates of soil temperature during summer and underestimates during autumn. These discrepancies were reduced to negligible levels when an additional resistance to soil heat flow, corresponding to a 4-cm-thick humus layer, was introduced. The heat flow boundary condition based on the soil surface energy balance underestimated soil temperature during autumn. Sensitivity tests indicated that this underestimation was related to the use of a constant crop extinction factor for net radiation. The results of both methods pointed to between-year differences in development of the grass cover. Simulated values of snow depth agreed with measured values. However, the predicted frost depth was an overestimate, probably because the thermal conductivity of snow was overestimated. The temperature boundary condition may be used for prediction purposes, whereas the heat flow boundary condition should be used for physically based studies of soil boundary processes.