Embedded Knowledge in Software: 2. Demonstration and Preliminary Evaluation

Authors

  • G. L. McClymont,

    1. Golder Associates Ltd., 7017 Farrell Road S.E., Calgary, Alberta, Canada T2H OT3.
    Search for more papers by this author
    • Gordon McClymont conducted this research as a graduate student in the Department of Geology, University of Alberta. He received a B.Sc. in Geology from the same institution in 1976. Prior to graduate studies, he was involved in research and consulting in western Canada. He is currently employed by Golder Associates Ltd. in Calgary, Alberta where he is conducting ground-water contamination studies and is involved in further applications of knowledge-based systems to environmental problems.

  • F. W. Schwartz

    1. Department of Geological Sciences, The Ohio State University, Columbus, Ohio 43210.
    Search for more papers by this author
    • Frank Schwartz was appointed in 1988 as the Ohio Eminent Scholar in Hydrogeology at The Ohio State University. He is known internationally for his work in ground-water modeling, field and theoretical aspects of contaminant hydrogeology, and ground-water geochemistry. In 1984, Schwartz was named a co-recipient of the 0. E. Meinzer Award presented by the Geological Society of America. He was also named as the John Birdsail Distinguished Lecturer for 1983-84 by the Geological Society of America. He recently completed work with Patrick Domenico on a textbook entitled Physical and Chemical Hydrogeology.


  • Discussion open until May 1, 1992.

Abstract

This paper demonstrates the operation of EXPAR, a knowledge-based system designed to assist in preparing a set of data for a contaminant transport model. The accompanying evaluation exercise provides preliminary indications of the usefulness of EXPAR. The exercise tested the ability of participants to simulate the distribution of three organic compounds at a hazardous waste site near Ottawa, Canada. A modest set of basic information was provided together with one of four sets of supplementary data. The embedded knowledge in the EXPAR system appeared to provide valuable assistance in modeling, and the system itself met our expectations in terms of the user-friendliness and robustness. Additional work will be required to remove the weaknesses found in the evaluation particularly with respect to biodégradation data and guidance in establishing flow directions. The results of the project have shown that knowledge-based systems show promise in solving problems associated with the implementation of a transport model.

Ancillary