H53H-01
Soil moisture: applications and needs in vadose zone hydrology
In this presentation, we address the state of the art in using soil moisture measurements to derive soil hydraulic properties, to quantify water and energy fluxes in the vadose zone, to retrieve spatial and temporal dynamics of soil moisture profiles, and to develop monitoring networks. We will discuss these issues at two different scales important in vadose zone hydrology: the field and the catchment scale. Analyzing the value of soil moisture measurements is motivated by our increasing ability to measure soil moisture due to the availability of novel non- invasive measurement techniques at the field and catchment scale, of remote sensing platforms and improved retrieval algorithms as well as of novel soil moisture network sensor technologies in providing high quality soil moisture data with high spatial and temporal resolution. We advocate that optimal use of soil moisture measurements will require further development of down- and upscaling algorithms to bridge the disparity in scales between hydrological measurements and mathematical models, to improve data assimilation techniques for retrieving the vertical and horizontal distribution of soil moisture including its temporal dynamics but also hydrological parameters driving the flow of water and to explore the potential in combining hydrogeophysical techniques with remote sensing measurement of soil moisture. With respect to the issue of upscaling we feel that stochastic upscaling theories developed in vadose zone research have not really been optimally used at scales larger than the field scale. This will be illustrated by applying stochastic theories in interpreting observed soil moisture fields. Applications of these theories might help in bridging the gap between model and measurement scale at larger scales.
H53H-02
What can vadose zone physics can tell us about hillslope flow processes?
Studies of hillslope-scale hydrologic processes often invoke simplifications that fragment time (discrete rainfall- runoff events), fragment space (division into saturated and unsaturated domains), or fragment process (separating vertical infiltration from lateral interflow). This study explores the hillslope flow behaviors that emerge from continuous simulations that solve the Richards' equation for variably saturated subsurface flow. Simulations use the finite element numerical model, HYDRUS-2D, which can represent continuous, dynamic subsurface flow processes. This model is used in an exploration of how soil characteristics, hillslope shape, and climate affect hillslope flow behavior. A series of simulations are created for two-dimensional (vertical, lateral/downslope) hillslope slices with variable shapes and water retention characteristics. Each simulation is forced with continuous hourly measurements of precipitation and evapotranspiration. Simulated hillslope outflow hydrographs and soil water content / pressure head patterns are saved at hourly time steps. Results show that some hillslope configurations and water retention behaviors result in continuous, gradual responses to rainfall, whereas other configurations exhibit pronounced threshold responses to rainfall. Shapes of soil moisture patterns and saturated zones in the simulations vary significantly with different hillslope characteristics and timing of rain events. Although hypothetical, the simulations show a breadth of possible hillslope flow behaviors and moisture patterns that can be produced with natural rainfall and a full solution of the Richards' equation. Including the dynamic, non-linear behavior of unsaturated zone flow in hillslope hydrology studies has strong potential to guide hypothesis development in future research on hillslope scale flow mechanisms.
H53H-03
Virtual Experiments to Explore Non-Linear Soil Moisture-Hydrology Interactions at the Hillslope Scale
The highly non-linear links between soil hydrology and runoff generation at the hillslope scale are poorly understood. As such, a framework for the general characterization of hillslopes is still lacking. While field studies have contributed to our understanding, such work is still of limited value because of the small number of places and events that have been characterized to date. This presentation explores how models can be used to identify characteristic forms of non-linearity on the hillslope scale (e.g. threshold behavior, hysteresis) and their controls on lateral flow generation. We present a number of virtual experiments with a 3D physically-based finite element model using the Richards equation to systematically investigate topographic and vadose zone controls on subsurface stormflow generation and the implications for the hillslope storage-discharge relationship. Topographic and stormflow data from an existing hillslope research site with an installed subsurface flow collection system were used to build and calibrate the model. The aim of the model calibration was to create a virtual model environment that reflects the general flow behavior of this hillslope, not to reproduce observations exactly. Experimental studies at this hillslope have demonstrated a highly non-linear subsurface flow response dependent on storm total precipitation (fill and spill behavior). The model domain was 28 m by 48 m with an average slope of 13 degrees, consisting of an irregular geometry based on GIS data for two layers that represent the topography of the soil surface and the bedrock surface, respectively. The parameterization of the soil and bedrock properties was based on field measurements of soil moisture and saturated hydraulic conductivity. After successful calibration, topography and vadose zone characteristics (i.e. slope angle, soil depth, surface and subsurface topography, antecedent moisture conditions) were systematically varied. Results are presented as a matrix for assessing the non-linear interplay between rainfall and soil moisture at the hillslope scale and for identifying the relative importance of the varied characteristics for the spatial and temporal evolution of subsurface moisture patterns.
H53H-04
Pathways of soil moisture controls on boundary layer dynamics
Soil moisture controls on precipitation are now receiving significant attention in climate systems because the memory of their variability is much slower than the memory of the fast atmospheric processes. We propose a new model that integrates soil water dynamics, plant hydraulics and stomatal responses to water availability to estimate root water uptake and available energy partitioning, as well as feedbacks to boundary layer dynamics (in terms of water vapor and heat input to the atmospheric system). Using a simplified homogenization technique, the model solves the intrinsically 3-D soil water movement equations by two 1-D coupled Richards' equations. The first resolves the radial water flow from bulk soil to soil-root interface to estimate root uptake (assuming the vertical gradients in moisture persist during the rapid lateral flow), and then it solves vertical water movement through the soil following the radial moisture adjustments. The coupling between these two equations is obtained by area averaging the soil moisture in the radial domain (i.e. homogenization) to calculate the vertical fluxes. For each vertical layer, the domain is discretized in axi-symmetrical grid with constant soil properties. This is deemed to be appropriate given the fact that the root uptake occurs on much shorter time scales closely following diurnal cycles, while the vertical water movement is more relevant to the inter-storm time scale. We show that this approach was able to explicitly simulate known features of root uptake such as diurnal hysteresis of canopy conductance, water redistribution by roots (hydraulic lift) and downward shift of root uptake during drying cycles. The model is then coupled with an atmospheric boundary layer (ABL) growth model thereby permitting us to explore low-dimensional elements of the interaction between soil moisture and ABL states commensurate with the lifting condensation level.
H53H-05
Simple Ecohydrological Models: Is Average Root Zone Soil Moisture an Adequate Driver in the Functions for Evaporation and Assimilation?
Dryland ecosystems are typically characterized by low annual precipitation, much of which is delivered in the form of small rainfall events that may only wet the top portion of the root zone. In these areas, evapotranspiration (ET) is limited by the availability of soil moisture rather than by atmospheric demand, i.e. ET << potential ET. Likewise, when optimal temperatures and soil nutrients are not limiting, the uptake of carbon by vegetation via photosynthesis, i.e. assimilation, is also limited by the availability of soil moisture. Though soil moisture is largely depth dependent, only average root zone soil moisture is used in typical simple models of ecohydrological processes. Here, we show that in semiarid grassland and shrubland, the surface soil layer is the primary source of water for ET, at least throughout the monsoon season. Conversely, we show that only large precipitation events (or series of small events) generate enough soil moisture below the influence of atmospheric demand to trigger carbon assimilation in these dryland ecosystems. From this we hypothesize that a realistic representation of ecohydrological processes in semiarid areas can not be made solely using average root zone soil moisture. In this study we utilize records of ET, assimilation, and soil moisture at several depths collected during 3 summer monsoons at the Sevilleta National Wildlife Refuge in central New Mexico using eddy covariance methods. Additionally we employ a simple bucket model of ecohydrological processes (e.g. Rodriguez-Iturbe et al. 1999, Daly et al. 2004) driven by average root zone soil moisture. We compare bucket model predictions of ET and assimilation to the actual data records. We show that (1) bucket model predictions of ET lack the dynamic temporal variability of actual observations, (2) declines in ET following peaks are significantly steeper in observed than in predicted times series of ET, and (3) peaks in bucket model predictions of assimilation occur before peaks in observed assimilation. Our results suggest that two or more layers of soil moisture are necessary for making more accurate predictions of ET and assimilation in semiarid ecosystems.
H53H-06
Separating Water Content Changes and Soil Texture Using Electromagnetic Induction Soil Imaging
The spatial distribution of soil texture is important for determining soil moisture storage and soil hydraulic transport properties. Electromagnetic induction (EMI) surveys of the soil apparent electrical conductivity (ECa) are being used to infer soil spatial heterogeneity at the field scale, due to their non-destructive nature, rapid response and ease of integration onto mobile platforms. The purpose of this study is to develop a procedure to non- invasively map field-scale soil texture patterns with minimal calibration and separate response due to water content change from static textural properties. Geo-referenced ECa measurements were taken using a DUALEM- 1S ground conductivity meter on multiple days with different field soil water contents on a 50 x 50 m field site at the Utah State University's Greenville Farm. Our results suggest that there are distinct zones with different textural properties and the lowest conductivity zone corresponds to observed gravelly area. Using temporal stability analysis these EMI maps reveal the spatial distribution of time-invariant subsurface properties and are informative for modeling and experimental design purposes in ecological, environmental and agricultural applications.
H53H-07
Soil Water and Shallow Groundwater Relations in an Agricultural Hillslope
Shallow water tables contribute to soil water variations under rolling topography, and soil properties contribute to shallow water table fluctutations. Preferential flow through large soil pores can cause a rise in the water table with little increase in soil water except near the soil surface. Lateral groundwater flow can cause a large rise in water table at toeslope and depressional landscape positions. As plants transpire, water can move up into the root zone from the water table and wet soil below the root zone. Roots can utilize water in the capillary fringe. The purpose of this study was to interface automated measurements of soil water content and water table depth for determining the importance of drainage and upward movement. In 2006 soil water and water table depth were monitored at three positions: shoulder, backslope, and toeslope. Neutron access tubes were manually monitored to 2.3 m depth, and automated soil moisture was measured using CS616 probes installed at 0.3, 0.5, 0.7, and 0.9 m depth. Water table depths were monitored manually and automated, but the automated measurements failed during the season at two sites. In 2007, similar measurements were made at one toeslope position, but the CS616 probes were installed at nine depths and better quality automated well depth equipment was used. The 2006 data revealed little landscape position effect on daytime soil water loss on a wetter date; however, on a dry day just before a rain, daytime water loss was greatest for the toeslope positon and least for the shoulder position. After a period of intense rain, a rapid and significant water table rise occurred at the toeslope position but little water table rise occurred at the other landscape positions. The rapid toeslope water table rise was likely caused by lateral groundwater flow whereas minor water table rise at the other positions was likely due to preferential flow since the soil had not wet up below 0.6 m. Use of automated equipment has improved our understanding of the relations of soil water to water table fluctuations in an agricultural field.
H53H-08
Augmentation of the Darcy-Buckingham-Richards Formulation to Account for Fast Response to Infiltration of Soil Moisture at Depth
The generally accepted quantitative model of unsaturated flow is Richards' equation, based on concepts put forward by Darcy and Buckingham. This continuum approach is normally implemented using a fine discretization and a representative elementary volume conceived on the scale of some fraction of a meter. Traditionally, it has typically been used where time scales of interest range from several hours for diurnal soil-water-plant response, to years for long-range contaminant transport. Its predictions depend strongly on quantitative knowledge of unsaturated hydraulic properties of soil and rock, and on initial and boundary conditions in a subsurface domain. A growing accumulation of observational evidence indicates that moisture in the unsaturated zone, even several meters below the land surface, may respond rapidly (minutes to hours) to larger-scale hydrologic conditions such as fluctuating rainfall intensity. Rapid responses have been documented in terms of water content, arrival of tracer, or water-table fluctuation. The Darcy-Buckingham-Richards (DBR) approach alone is often ill-suited for these phenomena because their short time scale and relatively large spatial scale contrast markedly with those of diffuse unsaturated flow occurring at the same time. The size of representative elementary volume needed to justify the continuum approach for diffuse flow may be much too small to also represent preferential flow. The model of Nimmo (2007) predicts travel times of the preferential component of flow using alternative concepts such as constant-rate progress through preferential paths, without regard to traditional unsaturated flow properties of the subsurface media. It formulates the problem with strong reliance on the dynamic character of the infiltration rather than particular properties of the soil and rock. Where preferential and diffuse flow are both significant, an interactive combination of this approach with the DBR approach improves prediction of net vertical unsaturated-zone fluxes in response to hydraulic inputs and the evolving distribution of soil moisture.