SummaryThe work reported in this paper is part of the U.S. Department of Energy's Science and Technology Initiative to develop improved conceptual models of flow and transport in the vadose zone, particularly for the Hanford Site, Washington. The National Academy of Sciences has identified significant knowledge gaps in conceptual model development as one reason for the discovery of subsurface contamination in unexpected places (NAS 2000). Inadequate conceptualization limits include, not only the understanding of long-term fate and transport, but also the selection and design of remediation technologies. Current conceptual models are limited, partly because they do not account for the random heterogeneity that contributes to the extremes of very nonlinear flow behavior typical of the Hanford vadose zone. A major improvement in the conceptual model of the Hanford vadose zone includes a better understanding and description of soil anisotropy, a property that appears to control much of the subsurface flow and transport in layered sediments at the Hanford Site.This project used a combination of geophysical and soil-physics techniques to investigate the infiltration and redistribution of water and dilute tracers in a controlled field experiment at the Army loop Road clastic dike study area at the Hanford Site, near Richland, Washington. In the FY 2002 tests, surfacedeployed ground-penetrating radar was used to identify the discrete pattern of horizontal layering that comprises the coarse component of the heterogeneity along a 60-m transect. The transect was instrumented to allow water to be applied along its length from a line source. Local-scale water content, matric potential, and tracer concentrations were monitored as a function of spatial scale by multipurpose time-domain reflectometry probes and suction lysimeters (solution samplers). The resulting data were used to characterize fine-scale heterogeneity as well as correlation lengths of hydraulic and transport parameters. Tracer breakthrough data were used to determine longitudinal and transverse dispersivit ies and their scale dependence. Parameters were analyzed to identify a suitable averaging (upscaling) procedure for field-scale infiltration predictions. Distributions of water and solute were found to be spatially dependent and controlled by the fine-scale features present both at the clastic dike and near the horizontal sill that was found to emanate from the dike. A theory of anisotropy was developed and successfully tested based on observations at the clastic dike site. The results of this study will help to bridge the gap between local-scale transport observations and field-scale transport behavior. It will allow the validation of recently developed inverse procedures for predicting field-scale parameters and will improve our prediction capability for heterogeneous sediments at Hanford. The improved conceptualizations will permit DOE to make defensible corrective and remedial-action decisions at the Hanford Site.