This report presents the results of an investigation into the potential use of expert system technology as an effective tool for drought forecasting and management. Historical data derived from the Kentucky River Basin was used to test the resulting decision support system. This study has demonstrated that expert system technology can serve as an effective platform for use in assisting the decision maker in both characterizing the nature of an existing drought and in selecting and implementing the required management policy.
The success of any decision making process will of course be dependent upon the quality of the data upon which those decisions are made. In the current study, the critical data were the forecasted streamflow and the forecasted system demand for the coming week. Several different model structures were investigated for use in forecasting both streamflow and system demand. While a reasonable level of accuracy was obtained for the demand forecasts, only limited success was obtained in forecasting future weekly average streamflows. As a result, a probabilistic model structure was developed for use with streamflow forecasts that allowed more interaction with the decision maker. Although the resulting model evaluations failed to produce what may be considered satisfactory results, the developed model structure does provide a flexibility that may lead to improved performance by augmentation with additional rule based heuristics. Although not fully explored, several potential heuristic structures are identified. Additional model improvement can be expected by further refinement of the underlying deterministic streamflow forecast model structure.
Digital Object Identifier (DOI)
The work upon which this report is based was supported in part by funds provided by the U.S. Department of the Interior, Washington D.C. as authorized by the Water Resources Act of 1990 (Public Law 101-397.
The contents of this report were developed under a grant from the Department of the Interior, U.S. Geological Survey.
Ormsbee, Lindell E. and Jain, Ashu, "Development of a Decision Support System for Drought Characterization and Management: Application to Lexington, Kentucky" (1992). KWRRI Research Reports. 26.