At the point scale, we apply Darcy's law to field and lab measurements of hydraulic properties to estimate the deep water flux or recharge rate. This method uses an accurately measured unsaturated hydraulic conductivity (K) of a core sample, knowledge of the driving force for water flow, and field water content at the point from which the sample originated. At the regional scale, water balances can be made indirectly, for example, by using GIS databases that contain information on soil type and vegetation in combination with meteorological measurements.

• In the Middle Rio Grande Basin (MRGB) of New Mexico, our project is developing interpretive and integrative techniques to apply to point recharge measurements with depth and over topographically diverse locations, to estimate basin-scale aquifer recharge. On the basis of hydrogeomorphic characteristics and 16 Darcian point estimates of local recharge rates, we divided the 640-square-km region encompassing Abo Arroyo into four subregions and integrated to estimate the total recharge for the arroyo and adjacent watershed. The results indicate recharge rates in the middle of the range of estimates by other techniques. A major further goal is to permit the use of properties that are easier to measure than unsaturated K in order to explore the spatial variation of recharge rates over regions of diverse terrain. Ultimately this should permit basin-scale recharge determinations from information that can be obtained relatively easily.

• In the Mojave Basin of California, we have made Darcian flux measurements at the point scale and applied them at the profile scale. We determined hydraulic properties from a few good measurements and a great deal of qualitative borehole logging data and used these in evaluating profile-scale (to 50 m depth) unsaturated flow, including both deep percolation and lateral flow. We developed soil-property models of the layered profile at locations where lithologic information was available from boreholes. At a typical location within Oro Grande wash, the model has 42 discrete layers over 32 m of depth. Using these profile property models, we numerically solved Darcy's law in one and two dimensions to obtain values of matric pressure and water flux as a function of position. The simulated results support the validity of the critical assumptions of our computed downward flux values to within a factor of two and suggest that horizontal spreading of the water as it moves downward may account for the measured variations in vertical flux density. We evaluated the sensitivity of the results to various influences, showing the approximating assumptions to have little or negligible effect. The greatest sensitivity was found with layer thickness, a factor-of-ten decrease in which causes a decline in confidence from 96 to 80%. A single minimally conductive layer can by itself account for a high degree of spreading and decrease of flux density consistent with observations. The high confidence associated with point measurements suggest that this modified approach is valid for strongly layered deposits such as those of the Mojave Desert, and that water moves laterally away from the streambed as it moves downward through contrasting layers.

• In a succeeding study in the Mojave Basin, we have investigated the relationship between pore-size and particle-size distributions for 10 core samples from two different washes, to interpret the effect of depositional environment. The interrelated properties in this include large-core water-retention curves measured by our modified controlled-volume technique, and measured particle-size distributions. We found that textural factors, such as the range of particle sizes and mean particle size, rather than structural effects, such as grain arrangement and stratification, are the main controls on water retention for these media. This is encouraging for the possible use of particle size measurements and scaling theory to predict unsaturated hydraulic properties in this and similar areas.

• We have modeled unsaturated flow in the surficial sediments at the INEEL. We have assessed the importance of stratification, soil-water hysteresis, choice of grid sizes and node spacings, and other factors. Besides improving understanding of unsaturated flow in the top few meters of the subsurface, this work permits faster resolution of the numerical modeling problems that may arise in simulating variably saturated flow in the deep sedimentary interbeds. These efforts are directed toward the simulation of field tests. The immediate purpose is to let us evaluate field-experimental decisions such as borehole placement, instrument selection and placement, and mode of water injection or removal. Our plans for field testing by adaptation (for deep boreholes) of constant-head well permeameter and instantaneous-profile methods are complicated by the existence of perched water zones within and above the interbeds at the location of drilling. Thus we are formulating and evaluating new possibilities that would work in initially saturated as well as initially unsaturated conditions.