Site of the 30 meter eddy covariance micrometeorological tower. Approximately 6% slope to the NW and SE over a 1km distance.

In semiarid regions, the partitioning of precipitation into surface runoff, infiltration and potential recharge is highly variable in space and time. Understanding the spatial and temporal variability of these processes at a range of scales improves our ability to quantify and manage the available water resources. This research group is interested in understanding and quantifying the relationships among soils, their hydraulic properties, and hydrologic processes (rainfall, infiltration, and runoff) on upland areas and within ephemeral channels. Developing relationships between easily measured soil properties and vegetation characteristics and hydrologic processes is essential for up scaling from point, plot, and sub-watershed to basin scale.

Individual projects within this subarea are:

• Quantification of recharge within ephemeral channels using ERT
• Characterization of soil hydraulic properties
• Estimating infiltration and runoff response of semi-arid rangelands

Quantification of recharge within ephemeral channels using ERT
T. Ferré (UA-HWR), A. Furman (UA-HWR), A. Hinnell (UA-HWR), J. Simunek (UCR)

The electrical resistance tomography (ERT) method is currently used to qualitatively assess the two- or three-dimensional distribution of water and solutes in the subsurface. Our efforts have been aimed at 1) improving the practice of ERT to allow for monitoring transient hydrologic processes; 2) improving the inversion of ERT measurements to allow for quantitative subsurface imaging; and 3) providing guidelines for subsurface hydrologists to determine whether ERT is appropriate for their specific investigations.

The focus of recent efforts has been application of the analytical element method to ERT. Typically, several tens of ERT arrays are measured to construct a single subsurface image. While this is acceptable for defining static properties such as geologic structure, it limits the application of ERT to the monitoring of transient hydrologic processes. As a critical first step to adapting ERT to monitoring transient hydrologic processes, this research group has used the analytic element analysis approach to identify the optimal set of ERT arrays for use in conducting an ERT survey . The remainder of the summer 2002 will be focused on the continued development of coupled hydrologic and geophysical numerical models in conjunction with Simunek at the USSL.

The analytic element method is well suited to modeling the flow of electricity through a heterogeneous subsurface. Our initial approach has been to conduct a perturbation analysis to determine the spatial distribution of sensitivity of an ERT array throughout the subsurface. This method allows one to define the change in electrical potential throughout the subsurface due to the addition of a body that has an electrical conductivity that is different from the background. Based on these results, the response of any four-electrode ERT array can be predicted. Because the analytic element is so rapid, all possible arrays for each subsurface location can be investigated. Then, the sample area of each array can be described and the array can be uniquely identified with the greatest sensitivity to each point in the subsurface. A surprising result is that the most sensitive (and therefore preferred) array type is neither of the commonly applied arrays. That is, we have demonstrated that the current practice of ERT measurement is not optimal and we have developed guidelines to improve ERT measurements. Further efforts will focus on the optimization of ERT arrays for specific hydrologic applications. The goal of the research is to optimize the use of ERT to monitor transient subsurface hydrologic processes.

We have just begun to see the benefits of our fundamental research into the ERT method. Over the next two years, analysis will be advanced through the development of coupled hydrologic and geophysical models. In parallel, the improved ERT method will be applied to monitor infiltration and recharge beneath ephemeral streams and, in conjunction with other SAHRA researchers (Conklin and Grimm: TA2) and to monitor solute transport through streambed sediments.
(return to top)

Characterization of soil hydraulic properties
R. van Genuchten (USDA-ARS), B. Mohanty (TA&M), F. Leij (UCR), M. Schaap (UCR), P. Shouse (USDA-ARS), J. Simunek(UCR), J. Zhu (TA&M)

Soil hydraulic properties are the key constitutive relationships for water flow in the vadose zone. These properties quantify how porous media retain and transmit water while they also greatly affect the relative amounts of subsurface and surface flow. Measurement of unsaturated hydraulic properties is notoriously cumbersome and impractical for the simulation of subsurface flow and transport at the catchment scale. The main objective of our work in Riverside is to develop a pragmatic approach to estimate hydraulic properties for the San Pedro basin and to quantify the uncertainty of our estimates. To meet this objective, we have been measuring soil hydraulic and other basic soil properties, developing neural network models to predict hydraulic properties from soil and topographical attributes with pedotransfer functions (PTFs), and investigating the up-scaling of hydraulic properties from the sample scale. Our research will be applicable to several other SAHRA activities. Among them are the spatial distribution of hydraulic properties to elucidate patterns of runoff flow and erosion and the simulation of hill-slope hydrology in the San Pedro Basin. Furthermore we are involved in the inverse modeling and calibration of ERT surveys to conveniently estimate water content and solute concentration in the vadose zone.

Approximately 120 soil samples are being analyzed for their unsaturated soil hydraulic properties at the Salinity Laboratory. Samples were taken from several sites in Arizona, especially from the San Pedro basin. Water retention was measured using Tempe cells, and hydraulic conductivity was estimated using numerical inversion of the multi-step outflow data. Twenty-two soil cores are measured in each experimental run, which may last up to three months. Additional retention data are obtained at suctions of 1000, 3000, 8000, 15000 cm using pressure plate extractors. Soil texture and related soil physical properties have also been measured. Some 50 samples have been analyzed thus far.

Topography affects soil formation and some correlation has been found between topographical attributes and infiltration, surface water content and retention points. The effect of topography on hydraulic properties may be particularly pronounced for hillslopes. In collaboration with the investigators from the University of Naples in Italy, we therefore studied how well PTFs perform that use both basic soil properties and topographical attributes for a hillslope in Basilicata, Italy.

We continued the work on averaged hydraulic properties and water fluxes for a heterogeneous soil with an emphasis towards implementation and application. Pixels in meso- and regional-scale Soil-Vegetation-Atmosphere Transfer (SVAT) schemes of hydro-climatic models for semi-arid regions may represent an area between several hundred m2 to km2. Guidelines were developed how to best average hydraulic parameters for different textures and correlations.

Plans

Hydraulic properties will be required to model subsurface flow and transport for the Rio Grande and San Pedro basins and such properties may also be useful indicators for soil quality. In year 4 we will finish the measurement of soil and hydraulic data on the core samples and we also hope to garner some results from our involvement with ERT. We will then implement our findings in the SAHRA context using the general framework of predicting hydraulic properties over different scales that we developed in the first three years. Specifically, we will:

1. Retool the Rosetta and UNSODA software for the San Pedro basin and, possibly, other areas
2. Cooperate with other SAHRA researchers to utilize the data and to develop meaningful conclusions and guidelines for a range of scales and applications.

It is conceivable that the predictions with the newly calibrated PTFs will lead to systematic differences between estimated and observed hydraulic properties. PTFs for soils from the Southwestern U.S. will undoubtedly differ from those developed on more general databases that contain many soils from regions with humid climates and moderate temperatures. Several strategies may be pursued to improve estimates, including the use of Bayesian statistics. A central theme of our predictive methods is that we want to quantify the uncertainty associated with the estimation, including the distribution of predicted hydraulic parameters. Up-scaling can be achieved by running the Bayesian models on NRCS texture maps, developed from soil surveys, of the Rio Grande and San Pedro basins.
(return to top)

Estimating infiltration and runoff response of semi-arid rangelands
D. Goodrich (USDA-ARS/UA), A. Warrick (UA-SWES), G. Paige (USDA-ARS), J. Stone (USDA-ARS)

This research objective is to understand and quantify the relationships among soils, their hydraulic properties, and hydrologic processes (rainfall, infiltration, and runoff) in semi-arid upland areas. Developing relationships between easily measured soil properties, vegetation characteristics, and hydrologic processes is essential for up-scaling from point, plot and sub-watershed to the basin scale. Field research is being conducted on key ecological sites (soil - vegetation complexes) within the upper San Pedro Basin to quantify the relationships among rainfall intensity, ecological site characteristics and the resulting hydrologic processes, infiltration and runoff. Rainfall simulator experiments are being conducted to quantify these relationships at the plot and hillslope scales.

The full range of infiltration rates for four ecological sites has been determined. These rates are being related to easily measured soil and vegetation characteristics. The relationship has been developed and appears to be robust; however, it is still being validated. The field experiments and analysis will continue through the summer on a range of ecological sites. The results from the rainfall simulator experiments are being compared with results from soil hydraulic property research being conducted under subtask 1.2.2. Soil cores were taken within the ecological sites where the rainfall simulator experiments are being conducted. The results from the simulator experiments will be compared with the results from the soil cores and the pedotransfer functions. These will be used to improve our ability to scale up from point (soil cores) to plot and hillslope (rainfall simulator) and eventually to the watershed scale. This research is being conducted on uplands within the upper San Pedro River Basin. The results from this research will be important for linking the hydrologic processes on the uplands with the riparian areas in the basin-scale modeling efforts of SAHRA.

The recent research has had two distinct focal points:

• Infiltration theory and scaling
• Improving methods and methodologies for measuring infiltration and runoff on rangelands

Significant improvements have been made to the rainfall simulator. Both the mechanics for the simulator in general and the programming of the timing for the lower rainfall intensities (25 - 40 mm/h) have been significantly improved. The improved timing for the lower rainfall intensities has made it possible to measure infiltration and runoff processes right at the threshold intensity where runoff begins on most of the ecological sites within the San Pedro Basin.

Rainfall intensity-infiltration relationships for the four sites are being refined and validated using measurement results from the rainfall simulator experiments on ecological sites and hydrologic simulation models. In addition, extensive work has been conducted in collaboration with scientists at the USSL to develop pedo-transfer functions for the ecological sites on which we are conducting rainfall simulator experiments.

The results from the rainfall simulator experiments show distinct relationships between rainfall intensity and infiltration and runoff processes for the specific ecological sites on which we have conducted our research. These relationships are defined by the vegetation and cover characteristics of the site as well as the soil. The results indicate that the relationships are unique for a given ecological site and that they can be quantified. The implications of the results to date are that a range of infiltration parameters can be defined for a given ecological site. These results, in conjunction with the research being conducted in TA1.2.2 should significantly improve our ability to model rainfall, infiltration, and runoff processes on uplands within the upper San Pedro Basin.

Plans

Field experiments are planned at five more ecological sites this summer. The validation of the intensity-infiltration relationships will be extended to the watershed scale on sites located within the Walnut Gulch Experimental Watershed. A preliminary analysis of the hydrologic properties of the soils from the ecological sites was conducted by USSL from soil texture data. These will be validated with the 50+ soil cores that we collected from 9 ecological sites this spring. USSL scientists are currently conducting the analysis of the soil cores. These results will be compared with the plot scale results from the rainfall simulator experiments being conducted this summer. The resulting relationship(s) will be used to develop methods to scale up from point scale measurements and properties to plot and site scale rainfall runoff processes.
(return to top)