An Expert System for Decisionmaking in the Use of Desalination for Augmenting Water Supplies

(This project was the subject of Andrea Traviglia's masters research, click here for an abstract of her paper in the Journal of Water Resources Planning and Management or here for the final report to the Bureau of Reclamation)

Desalination has traditionally been considered to be one of the most expensive means of augmenting water supply. Over time, however, improving technology has lowered desalination costs while water scarcity has driven up the expense associated with many alternative means of water resource development. Given these trends, an important but largely unanswered question arises: where does desalination “fit” within the current cost hierarchy of water supply options? Brackish or saline sources are often dismissed as an option based on the cost gap between desalination and conventional treatment, a comparison that neglects many other important factors. The true costs of any water supply option must be evaluated based on the costs of both acquisition and treatment, but also subject to constraints on maintaining regulatory compliance and a high level of reliability.

The primary objective of this project is to develop an expert system that will allow communities to compare the costs of augmenting water supply through desalination with those of alternative supplies. Doing this will require a framework that allows for a more comprehensive evaluation of the acquisition and treatment costs of each option within specified regulatory and reliability constraints. The system will consist of a cost model and supporting guidance material, with the cost model designed as an integrated complement to USBR’s existing Water Treatment Estimation Routine (WATER). Guidance material will provide a step-by-step decisionmaking framework, which will include graphical and numerical decision tools for evaluating a range of water supply alternatives. The cost model (in Microsoft Excel format) will allow for multiple levels of investigation and will be based on inputs of data that are readily available to public water systems through regulatory records, USGS databases, and standard utility planning documents. Screening level evaluations involving minimal site specific data will be facilitated through use of default model inputs based on typical values drawn from the literature and similar projects. Communities that have undertaken a more thorough exploration of alternatives will have an ability to easily customize model inputs through use of site specific data. The model will include routines for evaluating the costs of achieving regulatory compliance and reliability goals for individual supply options using water quality and water availability data. These will be integrated with cost estimation algorithms for a range of alternative sources (e.g., wellfields, dams/reservoirs, pipelines, aqueducts, aquifer storage and recovery (ASR)) based on selected technologies, source yield, and infrastructure capacity.