A statistical method to determine pollutant sources

This investigation is directed to the purpose of enforcing water quality standards. The problem can be stated as follows: there is one monitoring station which at a given moment, measures the concentration of $r$ components, or pullutants; however, each component may have multiple sources within the drainage basin. It is assumed that the goal of the regulatory agency is to discover the source of illegal discharges by collecting an effluent sample while the violation is occurring. In order to accomplish the objective through the use of a stream monitoring station, it is necessary to make two predictions based on the current data: (1) When the next violation will occur. Time series analysis will be used to uncover the underlying process in the receiving water body and to predict the next illegal discharging event. (2) Which dischargers are contributing to the violation. The MSE method, based on the mean square error measure, can give us the function to be minimized. The mathematical program will help up to decide the sources of violation by minimizing the function. We find our methodology is significantly better than the conditional random sampling method and the completely random sampling method and can be a useful tool to the regulating agencies