Characterization of a Perchlorate Contaminated Site

Conventional methods of subsurface assessment for remediation or monitoring purposes often involve field sampling and laboratory analyses of soil and water samples for specific contaminants species. Even though these procedures are well established and produce reliable results, they have a number of disadvantages. Among others, they are not measured in real time, and they are sometimes destructive because excavations are needed to obtain soil samples. Furthermore, the sampling and testing processes can be quite laborious and expensive. Various investigations have been carried out to develop alternative, nondestructive methods for such routine measurements. The application of artificial neural networks (ANN) in environmental site characterization has proved to be an effective modeling method for the prediction of migration paths of environmental contaminants. However, the uses of ANN modeling for the migration of explosives-related contaminants (in particular perchlorate) in water and soil, have not been widely reported in the literature. For this reason, this study will explore the potential use of neural network modeling for predicting the amount and distribution of perchlorate at military installations.