Epsilon Open Archive

Abstract

Modelling of ecosystem processes often requires soil temperature as a driving variable. Since soil temperature measurements are seldom available for regional applications, they have to be estimated from standard meteorological data. The objective of this paper is to present a general, simple empirical approach for estimating daily depth profiles of soil temperature from air temperature and a surface cover index (LAI; leaf area index) mainly focusing on agricultural soils in cold temperate regions. Air and soil temperature data measured daily or every fifth day at one to six different depths was acquired from all meteorological stations in Sweden where such records are available. The stations cover latitudes from 55.65 to 68.42 N and mean annual air temperatures from +8.6 to -0.6 oC. The time series spanned between two and ten years. The soils at the stations cover a wide range of soil textures, including two organic soils. We calibrated the model first for each station and then for all stations together and the general parameterization only slightly decreased the goodness of fit. This general model then was applied to two treatments in a field experiment: bare soil and a winter rape crop. The parameters governing the influence of LAI on heat fluxes were optimized using this experiment. Finally, the model was validated using soil temperature data from two barley treatments differing in LAI taken from another field experiment. In general, the model predicted daily soil temperature profiles well. For all soils and depths at the meteorological stations, 95% of the simulated daily soil temperatures differed by less than 2.8 oC from measurements. The corresponding differences were somewhat higher for the validation data set (3.9 oC), but bias was still low. The model explained 95% of the variation in the validation data. Since no site-specific adjustments were made in the validation simulations, we conclude that the application of the general model presented here will result in good estimates of soil temperatures under cold temperate conditions. The very limited input requirements (only air temperature and LAI) that are easily obtainable from weather stations and from satellites make this model suitable for spatial applications at catchment or regional scales.