A Simulation Analysis of Thermal Effects on Evaporation From Soil

Abstract

The Richards equation, which expresses the conservation of water in an isothermal soil, has a more general form in a nonisothermal soil. In using the latter, it is necessary to know soil temperature, and modeling becomes considerably more complicated. A detailed, numerical simulation model quantifies the thermal effects for two hypothetical soils under two climates. During characteristic 4‐day climatic sequences, in a season of soil heating, diffusion of vapor due to thermally induced vapor concentration gradients suppresses evaporation. The suppression is greatest (5–15% in this set of experiments) under arid conditions. Under these conditions, such thermal vapor diffusion also distorts the usual diurnal pattern of evaporation. Evaporation is generally more sensitive to isothermal than to thermal vapor diffusion. Variations in time and depth of the soil temperature cause corresponding variations in the water transport coefficients. These, in turn, result in biases (2–5%) and diurnal distortions of evaporation rates. Liquid flow attributable to the dependence of matric potential on temperature accounts for about 1% of evaporation in our experiments. In simulations of 1 month duration for each combination of soil and climate, the joint neglect of all thermal effects mentioned above introduces an error of only about 1% in the average evaporation rate and does not distort its time distribution significantly.