Physical Processes Determining Water Loss from Soil

A field plot of uniform fine sandy loam free of plants was deeply irrigated with water containing 44 me./1. of chloride. The water and chloride content of the soil and the hydraulic head were measured as functions of depth and time for 59 days during summer. The water content, W, of surface layers of soil of thickness from 10 to 50 cm. were functions of time, T, that could be closely represented by equations of the form W = aT⁻ᵇ. This means that the rate of loss of soil water was inversely proportional to time after irrigation. Making use of this equation and its derivative with respect to time, it is possible to calculate the vertical water-flow velocity in the soil at various times and depths. Capillary conductivity values were calculated and agreed closely with published values obtained by steady state laboratory methods for soil from an adjacent area. The evaporation from the soil surface over the 59-day period of record was obtained from the water content of the 0- to 40-cm. soil layer by subtracting the downward drainage calculated from hydraulic gradient and capillary conductivity values. The evaporation thus obtained was 5.6 cm. The evaporation was also independently calculated from the change in water content and chloride content of the 0- to 10-cm. layer of soil. This method gave an evaporation figure of 6.0 cm. Early in the test period a marked decrease in the concentration of chloride in the soil solution developed in the 5- to 10-cm. depth interval. This depth interval apparently corresponds to a zone of condensation of water vapor that moves downward from the overlying and warmer surface layer of soil. However, after taking negative adsorption into account, it was concluded from the vertical distribution of chloride in the soil and the change of this distribution with time that the water transferred in the vapor phase below the 10-cm. depth in the profile was of negligible agricultural significance under the conditions of the test.