The National Science Foundation (NSF) funded Instant Reference Sources, Inc. to develop a prototype expert system to aid scientists and engineers plan and execute more cost-effective environmental sampling and analysis. The expert system guides users through the complex choices involved in tailoring sampling and analysis plans to meet project specific needs while, at the same time, considering the many variables involved with the physical and chemical attributes of both the environmental sampling site and the target chemicals.
The software combines decision criteria based upon systematic planning (including EPA's Data Quality Objective (DQO) Process, the new Performance and Acceptance Criteria (PAC) Process, and EPA's new Triad Approach) with a user's specific project needs and the principles of Performance Based Measurement Systems (PBMS). Meta data available from a new National Environmental Methods Index (NEMI) is used to define specific project needs for data quality and also measurement system (i.e., analytical methods) requirements. Thus, the project was designed to help correct two current problems (misuse or no use of the DQO Process and PBMS) [1] while promoting the use of a new Internet tool (NEMI) to systematically determine the best methods to use for project specific needs.
Although EPA initiated the DQO Process over ten years ago it is still frequently misunderstood and used improperly or not at all. A recent EPA Science Advisory Board review found that adoption of the DQO Process by EPA's regions, program offices, states and tribes has been "sporadic to nonexistent." The report acknowledges that although the risk of making a wrong decision based on environmental data has always been present, the DQO Process requires that this risk be explicitly stated and quantified and this is "disquieting" to most project managers. [1]
The expert system ensures that the DQO Process is used properly and, at the same time, provides advice to help users understand that the degree of probability of error resulting from their choices is quantifiable and also easily changed by modifying their project requirements. There was no easy way of doing this prior to EMMA; people had to rely on lengthy guidance documents from multiple sources.
Many current protocols are based on "prescriptive" methods developed mostly in the 1980s and 1990s. Recent efforts to base environmental sampling and analysis on Performance Based Measurement Systems (PBMS) instead of only prescriptive methods have been hampered by a lack of understanding of the relationships between method performance, numbers of samples, data quality and decision information quality. [2] For example, with ambient water quality monitoring, many programs have used prescriptive methods without documenting associated method performance or data quality and this has resulted in unknown or poor quality data and uncertainty in the comparability of data collected across programs or organizations. [3]
The expert system provides users with advice that they may use to select the best methods from the NEMI database program based on project-specific requirements.
References
[1] U.S. EPA Science Advisory Board, Letter Report: EPA-SAB-EEC LTR-99-002 Review of the Agency-Wide Quality System. 1999 Science Advisory Board Report to EPA
[2] Lawrence H. Keith, Evaluation of Approaches to Improve the Quality and Cost-Effectiveness of Environmental Monitoring, American Chemical Society, Washington, DC, May 2000.
[3] Jerry Diamond, Andrew Eaton, Cliff Annis, Herb Brass, Larry Keith, Ann Strong, Dennis McChesney, and Merle Shockey. Towards a Definition of Performance-Based Laboratory Methods. August, 2001. Methods and Data Comparability Board. Available on the Internet at http://wi.water.usgs.gov/pmethods/PBMS/nwqmc.0102.pdf (this requires Adobe Reader to read). The primary web site is at http://wi.water.usgs.gov/pmethods/PBMS/PBMS.html.