Spatial Knowledge-Based Decision Analysis System for Pond Site Assessment

Abstract

In the South Platte River Basin of Colorado, seasonally flooded wetland ponds are being developed for groundwater recharge and wildlife habitat. The selection and evaluation of pond sites is primarily done by local experts. This study presents an approach that combines the use of knowledge-based systems, multicriterion decision analysis techniques, and geographic information systems to develop a representation of the pond evaluation process. The multicriterion decision analysis technique is used to calculate a surrogate certainty factor; that is, a suitability factor, for processing rules within a forward-chaining knowledge-based system to determine the overall suitability of a specific site. Spreadsheet software was used to implement the rule processing, with site data and spatial analysis results provided through links with a geographic information system. The spreadsheets were designed to provide flexibility for users to change thresholds, rating scales, and relative importance weights. Scenario testing and testing of the system by local experts demonstrate its potential use for pond site evaluations.