Measuring water sources pollution using intervention analysis in time series and log-normal model (a case study on Tigris River)

This article presents two different techniques, to analyse the effects of water pollution sources on water quality in the presence of dependent noise structure. In this article we measure the pollution level in the water of Tigris River after passing through the Baghdad metropolitan area, where many sewages and industries dump their refuse into the river. The data analysis methodology depends on the presence of unusual observations and the validity of the underlying assumptions. First we developed an adaptation to the lognormal model, assuming that the measurements taken across the river are correlated. Secondly we assumed the data present a time series with intervention. The Box and Taio method is used to analyse the data. Results showed significant level of pollution in the river after passing through the specified region. Copyright (C) 2002 John Wiley Sons, Ltd.