Filtering of Cyclical Infiltration Forcings in the Vadose Zone

• Main Author: Dickinson, Jesse
• Other Authors: Ferré, Paul A.
• Language: en_US
• Published: The University of Arizona. 2016
• Online Access: http://hdl.handle.net/10150/623171; http://arizona.openrepository.com/arizona/handle/10150/623171

Infiltration and downward percolation of water in the vadose zone are important processes that define the availability of water resources in many areas of the world. Flow in the vadose zone can vary spatially and temporally because of the complex exchange of water and energy between the land surface and atmosphere. Precipitation and infiltration forcings at the surface are filtered in the vadose zone in terms of the lag time between the forcing at the land surface and a response at any depth, and the damping of the magnitude of flux variability with depth. Climate projections call for changes in both the timing and magnitude of precipitation and land surface forcings, which increases the importance of understanding how the vadose zone filters these forcings to predict the impacts of climate variability and change on groundwater resources. This dissertation research presents a theoretical framework for assessing how cyclical variations in one-dimensional, vertical flow are filtered in the vadose zone. The filtering properties are described using analytical and numerical solutions. The analytical solution linearizes Richards equation by representing the diffusive properties of the soil as constant through time. The numerical solution uses the full Richards equation. Three implications for filtering in the vadose zone using a linearized and full Richards equation are investigated in three modeling experiments. In the first experiment, the analytical solution is used to identify subregions of aquifers where infiltration variations are sufficiently damped so that recharge can be approximated to be steady through time. The linearized solution overestimates the diffusive properties of soils, thus the amount of damping and the area of subregions of steady recharge are both under predicted. In the second experiment, the linearized analytical solution is superimposed vertically to represent the lag time and damping in layered soils. The superposed linearized solutions do not represent transitions of soil-water properties that occur between real soil layers. As a result, the filtering can be over or under predicted because of systematic errors in the estimated water capacity in the analytical solution. The filtering in homogenous and layered soils (first and second experiments) is more accurate when the water content and diffusivity variations are small, and when soil layers are relatively thick compared to the depth over which the damping occurs. In the third configuration, a numerical solution which solves the full Richards equation is used to evaluate how multiple asynchronous infiltration cycles interfere constructively and destructively in homogeneous soil. A new cyclical variation in infiltration is generated within the vadose zone through the nonlinear interaction of cycles with similar frequencies. The emergent cycle may result in prolonged periods of both enhanced and decreased recharge.