Entropy-assisted approach to determine priorities in water quality monitoring process

Effective determination of water quality and water pollution assessment is crucial and challenging processes. Evaluating water quality in rivers, researchers have referred to various statistical, probabilistic and stochastic methods to obtain efficient information from the monitoring network. As data are greatly random, the information content can be obtained by utilizing various methods including but not limited to the "entropy." Monitoring is a difficult process due to high measurement costs, while it is also difficult to optimize the network in terms of time, space, and especially the variable to be monitored. In the presented study, it is aimed to create an effective approach to be used in optimizing the monitoring network by determining the "prior" variables by entropy that measures the uncertainty by using all the data without time difference. The presented study proposes an alternative method to define the water quality variables that should be monitored much more frequently. Study is exemplified for demonstrating its potential use in a case study level, Grand River in Canada, by assessing water quality data obtained from 15 water quality monitoring stations. Results showed that BOD, Cl, and NO₂-N among examined 8 different variables are as the "prior" variables should be monitored. It is being proven that the prior variable that should be monitored for optimization of the network can be easily determined with the information obtained from the data statistically evaluated with entropy, and it can be stated as an effective method for managers to use in the decision-making process.