Soil-moisture deficit simulations with models of varying complexity for forest and grassland sites in Sweden and the U.K.

Using soil moisture observations from a forest area in the U.K. and from forest and grassland sites in southern Sweden, a range of simple, daily soil moisture deficit models were tested. Each model consists of potential evaporation, interception losses, and a routine by which transpiration was calculated from the potential via a soil water regulating function. In the case of the Swedish sites, snow melt and temperature effects on transpiration rates were also included.Despite the large range of vegetation and soil types, the same conbinations of models generally gave the best fit to the observations. Also, in most cases, the most successful model used the simple subroutines.At each site, it was found essential to optimize the transpiration regulating function. The most successful model was a linear formulation using the concept of available soil water. A good relationship was found between the model-derived available water and field-capacity values, and the measurements of soil-water potentials. This suggests that these models could be calibrated using simple measurements on the soils.Finally, the analysis illustrates some important physical processes, notably the negative feedback between transpiration and high evaporative demand, the limitation of transpiration by low temperatures, and the possibility of drainage of rainwater through unsaturated soils.