The ‘force‐restore’ approach is commonly used in order to calculate the surface temperature in atmospheric models. A critical point in this method is how to calculate the deep soil temperature which appears in the restore term of the ‘force‐restore’ equation. If the prognostic equation for calculating the deep soil temperature is used, some errors in surface temperature calculation and consequently in partitioning the surface energy and land surface water can be introduced. Usually, these errors should appear as a result of incorrect parameterization of surface energy terms in the prognostic equation based on ‘force‐restore’ approach.
In this paper, the sensitivity of the ‘force‐restore’ model for surface temperature to the: (a) changes of soil heat flux; (b) variations of deep soil temperature and (c) changes in soil water evaporation is examined. In addition, the impact of the deep soil temperature variations on partitioning the surface energy and land surface water is discussed. Finally, a new procedure for calculating the deep soil temperature based, on climatological data of soil temperature and its exponential attenuation in the deep soil layers is suggested. All numerical experiments with the LAPS land surface scheme were performed using two data sets, obtained from the micrometeorological measurements over a bare soil at Rimski Šančevi (Yugoslavia), RS, and Caumont (France), HAPEX. Copyright © 1999 Royal Meteorological Society