RECEPTOR MODELS TO STUDY GROUNDWATER CONTAMINATION

This paper describes the use of receptor modeling to accomplish three tasks: estimating the number of sources making significant contributions to groundwater contamination at a receptor site, revealing the chemical signatures of effluent from each pollutant source, and apportioning responsibility for contamination at a receptor site among the sources. Physical and chemical processes cannot be ignored in groundwater modeling as they usually are in air pollution receptor modeling. If the contaminants undergo differential chemical reactivity and retardation in the soil column, estimates of the number of pollution sources will be inflated by the number of important physical chemical processes. This problem is addressed by coupling pattern recognition and factor analysis methods. Chemical reactivity and retardation cause changes in apparent source signatures with time and contribute to error in source apportionment with linear receptor models. The receptor modeling process reveals the nature and rates of physical chemical processes occurring in the soil column, allowing correction of apparent source signatures. Model error can be reduced by use of non-linear receptor models employing corrected source signatures.