Evaluating sampling locations in river water quality monitoring networks: application of dynamic factor analysis and discrete entropy theory

In this paper, a methodology is proposed for evaluating sampling locations in an existing river water quality monitoring network. The dynamic factor analysis is utilized to extract the independent dynamic factors from time series of water quality variables. Then, the entropy theory is applied to the independent dynamic factors to construct transinformation–distance (T–D) curves. The computation time in the case of using dynamic factors is significantly less than when the raw data is used because the number of independent dynamic factors is usually much less than the number of monitored water quality variables. In this paper, it is also shown that by clustering the study area to some homogenous zones and developing T–D curves for each zone, the accuracy of the results is significantly increased. To evaluate the applicability and efficiency of the proposed methodology, it is applied to the Karoon River which is the most important river system in Iran.