Application of Bayesian Belief Network to groundwater quality assessment

We describe the development of a prototype. Bayesian Belief Network (BBN) that models groundwater quality in the Sultanate of Oman, in particular. This model presents a unified approach to the analysis and interpretation of groundwater quality data in order to determine if qualitative or quantitative standards for groundwater quality have been exceeded. The approach is to use a graphic representation of a probabilistic distribution to represent the static and dynamic cause-and-effect relationships between groundwater quality constituents. Experts have been arguing that the current used techniques are not accurate means of measuring groundwater contamination. This is mainly because these techniques neglect the characteristics that are significant in understanding of pollution-generation processes from various sources. Furthermore, the data gathered from groundwater monitoring systems are uncertain, and the test methods used by environmental laboratories do not emphasize the accuracy.