An object-oriented approach to site characterization decision support

Effective decision support for site characterization is key to determining the nature and extent of contamination and the associated human and environmental risks. Site characterization data, however, present particular problems to technical analysts and decision-makers. Such data are four dimensional, incorporating temporal and spatial components. Their sheer volume can be daunting -- sites with hundreds of monitoring wells and thousands of samples sent for laboratory analyses are not uncommon. Data are derived from a variety of sources including laboratory analyses, non-intrusive geophysical surveys, historical information, bore logs, in-field estimates of key physical parameters such as aquifer transmissivity, soil moisture content, depth-to-water table, etc. Ultimately, decisions have to be made based on data that are always incomplete, often confusing, inaccurate, or inappropriate, and occasionally wrong. In response to this challenge, two approaches to environmental decision support have arisen, Data Quality Objectives (DQOS) and the Observational Approach (OA). DQOs establish criteria for data collection by clearly defining the decisions that need to be made, the uncertainty that can be tolerated, and the type and amount of data that needs to be collected to satisfy the uncertainty requirements. In practice, DQOs are typically based on statistical measures. The OA accepts the fact that the process of characterizing and remediating contaminated sites is always uncertain. Decision-making with the OA is based on what is known about a site, with contingencies developed for potential future deviations from the original assumptions about contamination nature, extent, and risks posed