There are many locations throughout the world where subsurface contamination impacts the natural environment, with potentially serious consequences for the quality of our water and for human heath. At U.S. Department of Energy (DOE) sites alone there are estimated to be “about 6.4 billion cubic meters of contaminated soil, groundwater and other environmental media” (DOE Environmental Management Science Program announcement 02–03). One of the initial steps in dealing with a contaminated site is that referred to as site characterization. During site characterization, measurements are made that allow for the development of an accurate model of the physical, chemical, biological, and hydrogeological properties of the subsurface. Such a model is required to design an appropriate plan for remediation of a contaminated site and can also be used, and continually updated, for short-term or long-term monitoring of the site. Site characterization can involve locating and identifying a known or suspected contaminant, and can also involve determining the properties of the subsurface controlling the fate and transport of the contaminant. The challenging problem we face, at many sites, is identifying an approach to site characterization that provides the required information about the subsurface while minimizing the risks associated with contacting the contaminated region. The most common approach to site characterization involves drilling and directly sampling the near surface (top ∼100 m) of the earth. Boreholes are used to extract samples for laboratory analyses, for borehole logging, and to conduct on-site testing. Examples of on-site testing include water sampling for the direct detection of contaminants and pump tests, slugs tests, and tracer tests to estimate hydraulic conductivity. While the borehole-based methods can provide direct and accurate measurement of subsurface properties of interest, they are limited in terms of the spatial density of the sampling (dictated by the number of boreholes) and the volume of …