Economic evaluation of seawater desalination : a case study analysis of cost of water production from seawater desalination in Saudi Arabia

Abstract

As a result of the increasing scarcity of freshwater resources worldwide, many countries have resorted to the use of unconventional sources, of which seawater desalination is the most significant, for meeting the supply-demand gap. However, despite the recorded advances in desalination technologies of recent decades, desalination remains a very expensive operation and operators will be greatly assisted if reliable means of predicting the costs are available to aid effective decision making during planning of new plants or the operation of existing plants. To achieve this, it is important to fully understand the factors that contribute to desalination costs, which could then be used to develop appropriate models for predicting costs that can support budgeting and/or cost reductions decision making. Consequently, this project has investigated the development of such models for predicting monthly production costs using data from 16 operational plants in Saudi Arabia. Monthly and annual data spanning 2001 – 2010 were collected on total water production, type of desalination technique, sea water salinity, product water salinity, energy consumption, and total (capital and operational) unit cost of water production. Because of the way in which the data were archived, some of the variables only had the annual totals for some of the years, which made them unsuitable for the monthly scale adopted for the analyses. Consequently, disaggregation schemes based on several variants of the method of fragments widely used in hydrological studies were used to obtain monthly data from the annual data. Exploratory analysis showed that the monthly costs correlated most with the total water production, which then formed the lone independent variable for various tested regression model formulations. In general, an inverse regression model performed best during both calibration and validation. To enhance the usefulness of the predictive model for decision making, uncertainty limits of the predictions were constructed using a Monte Carlo simulation approach involving the seasonal, lag-1 autoregressive generation of equally likely realisations of the available historic records that have been transformed to remove the skewness. Extensive testing of the data generation technique showed that the assumed lag-1 auto-regressive dependence structure was adequate. This study thus provides for the first time a predictive model for costs of desalination in Saudi Arabia and its uncertainty range for effective budgeting and operational management. Although the models were developed using Saudi Arabia data, the fact that only one independent variable was used means that the replication of the methodology in other desalination-intensive countries can be readily carried out.