Adaptive sampling strategy support for the unlined chromic acid pit, chemical waste landfill, Sandia National Laboratories, Albuquerque, New Mexico

Abstract

Adaptive sampling programs offer substantial savings in time and money when assessing hazardous waste sites. Key to some of these savings is the ability to adapt a sampling program to the real-time data generated by an adaptive sampling program. This paper presents a two-prong approach to supporting adaptive sampling programs: a specialized objectoriented database/geographical information system (SitePlanner TM) for data fusion, management, and display and combined Bayesian/geostatistical methods (PLUME) for contamination-extent estimation and sample location selection. This approach is applied in a retrospective study of a subsurface chromium plume at Sandia National Laboratories' chemical waste landfill. Retrospective analyses suggest the potential for characterization cost savings on the order of 60% through a reduction in the number of sampling programs, total number of soil boreholes, and number of samples analyzed from each borehole.

Extent of chromium contamination beneath the 60s pits in the Chemical Waste Landfill at Sandia National Laboratories, Albuquerque, New Mexico

“…Past decision support for the adaptive sampling program at the Mixed Waste Landfill Integrated Demonstration (MWLID) at Sandia National Laboratories, Albuquerque, New Mexico, has provided evidence that an adaptive sampling program approach t o site characterization can result in significant cost savings in three areas (Johnson 1993a). First, the cost per sample is reduced by emphasizing field screening results rather than relying on traditional off-site laboratory analyses.…”

“…The role of SitePlannerTM is qualitative support: data integration, management, and display. The role of PlumeTM is quantitative: fusing hard and soR information regarding contaminant location, estimating contaminant extent, measuring the uncertainty associated with those estimates, estimating the expected value t o be obtained from additional samples, and assisting in identifying the locations for those samples so that maximum benefit is realized (Johnson 1993a). In fiscal year 1992 and 1993, these two technologies were applied t o the unlined chromic acid pit in the Chemical Waste Landfill at Sandia National Laboratories, to assist in estimating the extent of subsurface chromium contamination and t o suggest locations for additional bores.…”

A Bayesian/Geostatistical Approach to the Design of Adaptive Sampling Programs

“…Adaptive sampling programs result in more cost effective characterizations by reducing the analytical costs per sample collected, by limiting the number of samples collected by strategically locating samples in response to field data, and finally by bringing characterization to closure in the course of one sampling program. Adaptive sampling programs can result in characterization cost savings on the order of 50% to 80% (Johnson, 1993).…”