Thermal gradients and water transfer in unsaturated soil.

Abstract

An investigation into the flux of soil water under the influence of a thermal gradient was conducted in closed soil systems in the laboratory. A thermal gradient was imposed across the sandy loam soil columns and measured continuously for the duration of the experiment. The movement of soil water was monitored periodically using a gamna ray attenuation device until the columns reached an apparent steady state condition with no net flow. Imposition of boundary conditions enabled delineation and evaluation of the system parameters. Values of net water flux in soil columns were analyzed using the Taylor-Cary irreversible thermodynamic and the Philip-de Vries theory of water movement. Application of the Taylor-Cary equation to describe the flow reveals that for initial time periods the flow is slightly overestimated, however, this small difference may be within the realm of experimental error. The flow predicted by the Taylor-Cary equation for the succeeding time periods greatly exceeded the measured flow rates. The Philip-de Vries theory of soil-water movement predicted greater net water movement than was observed in soil columns with a temperature gradient of 2.67 ° C/cm and with an average soil water content of 10.5 to 11.5 cm³/cm³. The following conclusions were reached after analyzing the data for the sandy loam soil material. Water content and temperature influence the diffusion transfer coefficient, β*, in a closed soil system. Thus the transfer coefficient is not an independent entity. Hysteresis is present in the wetter part of the system, although the magnitude of hysteresis involved is unknown. The β* coefficient can be evaluated only in regions where hysteresis is not present. The transfer of soil water was greater in a leached soil with no air gap than in an unleached soil with no air gap. There was supporting evidence that liquid water continuity did not exist throughout the column since there was no appreciable solute transport. The observed change in soil water content distribution at 18.0 cm³/cm³ to the imposed temperature gradient was not significant for the sandy loam soil material. The observed soil water flux increased in response to the imposed temperature gradient as soil water content decreased from 18.0 to 10.5 cm³/cm³. No analytical procedure is presently available that will describe thermally induced flow under all conditions. The approach of Philip and de Vries requires that the physical properties of the soil must be known accurately so that correct estimates can be made of the individual diffusivities. This approach has been found to predict the flow with some success on relatively dry soils, however, it possibly would be in error when estimating thermally induced flow in regions where liquid continuity exists and up through to a saturated system. The Taylor and Cary equation may be adequate for describing water flow to predict trends or obtain comparative values, however, much additional work needs to be done before it will adequately describe the flow for transient conditions or steady state conditions where nonuniform water content distributions are present.