Optimization of water quality monitoring networks on rivers

The design of water quality monitoring networks and the improvement of existing networks is an ongoing problem to be resolved today. In particular, research is still underway on the grounds that observations are not continuous due to certain reasons, the frequency of measurement is not sufficient and the problem of micro location is not solved in the selection of observation locations. Moreover, a great majority of the presented methods focus only on the information value carried by the presented data, but do not take into account the elements of the stations themselves, such as station availability, frequency of observations, continuity, technological infrastructure or evaluate very little. Methods that assess the efficiency of newly established monitoring networks with low data, or operational performance of established ones serving specified purposes, have not been fully developed yet. Furthermore, most of the methods in the literature are relying on weight coefficients based on biased decisions of designers and the decisions given for improvement of the network are therefore based on subjective judgements. The aim of this thesis study is to develop a method for the assessment of network performance regarding the monitoring objectives specified, and to create a methological basis for the re-design if the it is deemed to be necessary. In addition, the "Reference Point Method" developed in the study can be used to rank the performance of the existing stations for evaluation with the asset of not having the obligation of assignment of weighting coefficients for network's monitoring objectives. Based on the water quality data observed between 1985-2000 in Gediz Basin, the performance of the existing network is evaluated, the quality parameters are divided into different groups according to the types of pollution, and the relations of the parameters with each other and relations with the other groups are investigated. The data collection period of each station, the frequency, the total number of observations for each quality variable, and the station variables of the quality variables are considered. Observation data, meta data and station specific characteristics are calculated for each station separately. With these calculations, the purpose of the observation network was investigated and it is determined which station characteristics serve these purposes. Reference Point Method is used to rank station performances in the network according to these specific purposes and results are evaluated with the comparison of other results obtained in previous studies.