Articles | Volume 8, issue 4
Hydrol. Earth Syst. Sci., 8, 706–716, 2004

Special issue: Assessing nitrogen dynamics in European ecosystems: integrating...

Hydrol. Earth Syst. Sci., 8, 706–716, 2004

  31 Aug 2004

31 Aug 2004

A simple model for predicting soil temperature in snow-covered and seasonally frozen soil: model description and testing

K. Rankinen1, T. Karvonen2, and D. Butterfield3 K. Rankinen et al.
  • 1Finnish Environment Institute, P.O. Box 140, FIN-00251 Helsinki, Finland
  • 2Water Resources Engineering, Helsinki University of Technology, P.O. Box 5200, FIN-02015 HUT, Finland
  • 3Aquatic Environments Research Centre, Department of Geography, University of Reading, Reading, RG6 6AB, UK
  • E-mail for corresponding author:

Abstract. Microbial processes in soil are moisture, nutrient and temperature dependent and, consequently, accurate calculation of soil temperature is important for modelling nitrogen processes. Microbial activity in soil occurs even at sub-zero temperatures so that, in northern latitudes, a method to calculate soil temperature under snow cover and in frozen soils is required. This paper describes a new and simple model to calculate daily values for soil temperature at various depths in both frozen and unfrozen soils. The model requires four parameters: average soil thermal conductivity, specific heat capacity of soil, specific heat capacity due to freezing and thawing and an empirical snow parameter. Precipitation, air temperature and snow depth (measured or calculated) are needed as input variables. The proposed model was applied to five sites in different parts of Finland representing different climates and soil types. Observed soil temperatures at depths of 20 and 50 cm (September 1981–August 1990) were used for model calibration. The calibrated model was then tested using observed soil temperatures from September 1990 to August 2001. R2-values of the calibration period varied between 0.87 and 0.96 at a depth of 20 cm and between 0.78 and 0.97 at 50 cm. R2-values of the testing period were between 0.87 and 0.94 at a depth of 20cm, and between 0.80 and 0.98 at 50cm. Thus, despite the simplifications made, the model was able to simulate soil temperature at these study sites. This simple model simulates soil temperature well in the uppermost soil layers where most of the nitrogen processes occur. The small number of parameters required means that the model is suitable for addition to catchment scale models.

Keywords: soil temperature, snow model