Designing an Optimal Monitoring Network of Water Quality Variable in the Semi-Enclosed Estuary

The semi-enclosed estuary with sea-dike has complicated hydrodynamic characteristics due to the effects of artificial freshwater discharge. Such inflow of external sources has a profound effect on coastal changes, so it is necessary to establish an appropriate monitoring network to detect them. In general, there are two types of monitoring approaches. One is on-site monitoring that is keeping the instruments and going directly to the site to collect data periodically. Another is real-time monitoring that is acquiring continuous data in real-time by installing the monitoring stations. In order to design an optimal monitoring network considering both of them, the present study focuses on optimizing the array of monitoring points in area with very high spatiotemporal variability due to the artificial sources. In order to determine the design variables for optimization, we first decomposed the principal components representing the total variance of the various monitoring properties using EOF and selected two representative properties that contribute to three main principal components. Afterward, the cosine angles using the three-dimensional eigenvectors for two representative monitoring properties were calculated and used as design variables. The optimal array of the monitoring points is graphically selected by using graphical optimization and evaluated by statistical metrics to determine whether the points are placed in a reasonable position with the minimum numbers. The optimal monitoring network presents a suitable array of on-site monitoring and recommends the installable areas of the real-time monitoring station that reflect the global and local characteristics of the semi-enclosed estuary.