John E. Wilson

An Environmental Sampling Model for Combining Judgment and Randomly Placed Samples

In the event of the release of a lethal agent (such as anthrax) inside a building, law enforcement and public health responders take samples to identify and characterize the contamination. Sample locations may be rapidly chosen based on available incident details and professional judgment. To achieve greater confidence of whether or not a room or zone was contaminated, or to certify that detectable contamination is not present after decontamination, we consider a Bayesian model for combining the information gained from both judgment and randomly placed samples. We investigate the sensitivity of the model to the parameter inputs and make recommendations for its practical use.

Accounting for False Negatives in Hotspot Detection

Hotspot sampling designs are used in environmental sampling to identify the location of one (or more) contiguous regions of elevated contamination. These regions are known as hotspots. The problem of how to calculate the probability of detecting an elliptical hotspot using a rectangular or triangular grid of sampling points was addressed by Singer and Wickman in 1969. This approach presumed that any sample which coincided with a hotspot would detect the hotspot without error. However, for many sampling methodologies, there is a chance that the hotspot will not be detected even though it has been sampled directly--a false negative. We present a mathematical solution and a numerical algorithm which account for false negatives when calculating the probability of detecting hotspots that are circular in shape.

Acceptance sampling using judgmental and randomly selected samples

We present a Bayesian model for acceptance sampling where the population consists of two groups, each with different levels of risk of containing unacceptable items. Expert opinion, or judgment, may be required to distinguish between the high and low-risk groups. Hence, high-risk items are likely to be identifed (and sampled) using expert judgment, while the remaining low-risk items are sampled randomly. We focus on the situation where all observed samples must be acceptable. Consequently, the objective of the statistical inference is to quantify the probability that a large percentage of the unsampled items in the population are also acceptable. We demonstrate that traditional (frequentist) acceptance sampling and simpler Bayesian formulations of the problem are essentially special cases of the proposed model. We explore the properties of the model in detail, and discuss the conditions necessary to ensure that required samples sizes are non-decreasing function of the population size. The method is applicable to a variety of acceptance sampling problems, and, in particular, to environmental sampling where the objective is to demonstrate the safety of reoccupying a remediated facility that has been contaminated with a lethal agent.

Version 2.0 Visual Sample Plan (VSP): Models and Code Verification

Visual Sample Plan (VSP) is an easy-to-use visual and graphic software tool being developed by the Pacific Northwest National Laboratory (PNNL) to select the right number and location of environmental samples so that the results of statistical tests performed to provide input to environmental decisions have the required confidence and performance. It is a significant help in implementing the Data Quality Objectives (DQO) planning process that was developed by the U. S. Environmental Protection Agency. Gilbert et al. (2001) documented the quality assurance (QA) procedures that were conducted to assure that Version 0.91 of VSP was operating correctly. Subsequently, Version 0.91 was renamed Version 1.0 and placed on the internet at http://dqo.pnl.gov/vsp . Since that time VSP has been enlarged and improved and is now available as Version 2.0. The current document is an expansion of Gilbert et al (2001) to include the QA procedures and testing that were conducted to assure the validity and accuracy of the new features added to Version 1.0 to obtain Version 2.0.

Statistical Methods and Tools for Uxo Characterization (SERDP Final Technical Report)

The Strategic Environmental Research and Development Program (SERDP) issued a statement of need for FY01 titled Statistical Sampling for Unexploded Ordnance (UXO) Site Characterization that solicited proposals to develop statistically valid sampling protocols for cost-effective, practical, and reliable investigation of sites contaminated with UXO; protocols that could be validated through subsequent field demonstrations. The SERDP goal was the development of a sampling strategy for which a fraction of the site is initially surveyed by geophysical detectors to confidently identify clean areas and subsections (target areas, TAs) that had elevated densities of anomalous geophysical detector readings that could indicate the presence of UXO. More detailed surveys could then be conducted to search the identified TAs for UXO. SERDP funded three projects: those proposed by the Pacific Northwest National Laboratory (PNNL) (SERDP Project No. UXO 1199), Sandia National Laboratory (SNL), and Oak Ridge National Laboratory (ORNL). The projects were closely coordinated to minimize duplication of effort and facilitate use of shared algorithms where feasible. This final report for PNNL Project 1199 describes the methods developed by PNNL to address SERDPs statement-of-need for the development of statistically-based geophysical survey methods for sites where 100% surveys are unattainable or cost prohibitive.

Identifying Minefields and Verifying Clearance: Adapting Statistical Methods for UXO Target Detection

It may not be feasible to completely survey large tracts of land suspected of containing minefields. It is desirable to develop a characterization protocol that will confidently identify minefields within these large land tracts if they exist. Naturally, surveying areas of greatest concern and most likely locations would be necessary but will not provide the needed confidence that an unknown minefield had not eluded detection. Once minefields are detected, methods are needed to bound the area that will require detailed mine detection surveys. The US Department of Defense Strategic Environmental Research and Development Program (SERDP) is sponsoring the development of statistical survey methods and tools for detecting potential UXO targets. These methods may be directly applicable to demining efforts. Statistical methods are employed to determine the optimal geophysical survey transect spacing to have confidence of detecting target areas of a critical size, shape, and anomaly density. Other methods under development determine the proportion of a land area that must be surveyed to confidently conclude that there are no UXO present. Adaptive sampling schemes are also being developed as an approach for bounding the target areas. These methods and tools will be presented and the status of relevant research in this area will be discussed.