Coupled heat and water flow dynamics in dry soils : application to a multilayer waste cover

Unsaturated flow plays an important role in numerous environmental phenomena. It is complex in arid regions, where liquid water fluxes are small and vapor fluxes become relevant, so that heat, water and solute mass transport are needed to understand evaporation. This thesis aims at gaining insight evaporation and vapor flow mechanisms and the relevance of matric potential, temperature and osmotic gradients. These issues are especially relevant for soil salinization, whose mechanisms are poorly understood despite their global impact. We studied them in open soil evaporation column experiments. We found that a water separation process occurs in the soil. Above a very narrow evaporation front, the soil contains high salt concentrations due the solutes transported by the upward liquid flux and concentrated by evaporation. Below, concentrations are lower than the initial ones, because vapor flows downwards below the evaporation front driven by temperature gradients. Condensation of this downward vapor flux dilutes upflowing water, improving soil conditions and providing an area where root plants could live. We modeled nonisothermal multiphase flow and reactive transport during the experiments to quantify the actual processes and to understand the nature of the downward vapor flux. Modeling required modifying the retention curve to represent oven dry conditions. Our model supports the traditional division of soil, by an evaporation front, into a dry region and a moist region. This view may suffice for evaluating evaporation rates and water mass balances, but not for assessing salt processes, which require acknowledging that not only liquid water flow but also vapor diffusion occur below the evaporation front. Unsaturated flow concepts control the design of covers to isolate solid waste. Their goal is to protect the waste from infiltration during long periods of time by promoting surface runoff and lateral drainage and by hindering biointrusion. Two cover designs were built, both consisting of an evapotranspiration layer, a biointrusion barrier and an infiltration barrier. We analyzed their performance for two years (2009 and 2010, which was very wet) by a thorough monitoring system. We conclude that the upper portion of one cover did work as planned, but the lower infiltration barrier did not, which suggests some design improvements: increase the retention capacity of the sand layers, include filter layers and facilitate lateral drainage by increasing the slope