Predicting the flow of water between the soil surface and the atmosphere is a critical issue which geotechnical engineers must resolve for many practical problems. The analysis of saturated-unsaturated groundwater flow problems requires the specification of the flux of water at the upper soil boundary. The flow of moisture at the ground surface is also important for many problems in soil mechanics. These include the analysis of volume change in expansive soils and the stability of slopes.

The transfer of water across the soil-atmosphere surface occurs as infiltration and evaporation. The mechanics of infiltration into the soil surface is well understood and has been widely addressed in the literature. Alternately, the

rocess of evaporation from the soil surface is poorly understood. Extreme difficulties frequently arise while predicting evaporation from unsaturated soil surfaces.

Empirical methods of estimating evaporation from unsaturated soil surfaces can be found in the literature:​ however, the suitability and accuracy of these methods can be questioned.

A theoretical approach for evaluating the rate of evaporation from an unsaturated soil surface is developed. The theory is based on the principles of Darcy's Law and Fick's Law to describe the flow of liquid water and water vapour in the saturated-unsaturated soil below the surface. Dalton's Law and a modified form of Penman's Method for evaporation are utilized to predict evaporation from the soil surface.

Drying tests were conducted using three distinct soil types of sand, silt and clay. The soil surfaces were found to evaporate at the same rate as free water surfaces when saturated. The rate of evaporation begins to decline once the soil surfaces become unsaturated and total suction exceed approximately 3000 kPa. The rate of evaporation is proportional to total suction and continues to decline as suction increases. This principle appears independent of soil type and universal for the three texturally distinct soils selected for testing. The rate of evaporation may be predicted on the basis of the water content of the soil and its moisture retention curve established using routine test procedures.

The proposed theory was used to simulate the results of a 42 day evaporation test for a column of fine, uniform, clean sand. Good agreement was generally found between the computed and measured values of evaporation rate, soil water content and soil temperature. Additional analyses were conducted using various values of the saturated coefficient of permeability and the pore-size distribution index. The computed evaporative fluxes were found to be very sensitive to the permeability of the soil. Varying the coefficient of molecular diffusion for water vapour was also found to influence the rate of evaporation.

The modified Penman expression was applied to an example evaporation problem for Saskatoon, Saskatchewan during a 10 day period in July. The evaporative fluxes were computed with the watertable positioned at several depths below the surface of the sand. Evaporative rates were found to vary widely between the full potential rate of 7.7 mm/day and 0.4 mm/day depending on the position of the water table. In general, the results showed that the rate of evaporation from a soil surface depends strongly on the groundwater conditions.