A Deterministic Algorithm for Determination of Optimal Water Quality Monitoring Stations

Water quality monitoring networks are usually designed according to statistical approaches and general criteria without a consistent or logical deterministic design strategy. In this research, a deterministic approach for allocating the most sensitive water quality monitoring stations was proposed. This approach was applied on the western part of the Al-Hammar Marsh. Two-dimensional hydrodynamic and water quality simulation models were used to estimate the distribution of total dissolved solids (TDS) within the marsh for all of the expected conditions. Subsequently, the spatial distribution of the variance of TDS was computed based on the results of these models and performed in a Geographic Information System (GIS) database layer. The standard acceptable TDS variation limits of ±5%, land-use map, land-cover map and other main selection criteria of the monitoring stations were set as constraints via GIS database layers. These layers were integrally applied to the variance layer to obtain the locations of the most sensitive monitoring stations. It was concluded that, the most representative monitoring network consists of 46 stations. This number can be reduced to 37 and 29 stations by increasing the acceptable TDS variation limits to ±10% and 15%, respectively. The developed approach can be used with limited data. Moreover, it can be applied to rivers, lakes or wetlands, considering all of the related constraints. In addition, the GIS database can be easily updated and analysed. These features are not available in other methods such as the Sanders method, multiple criteria decision making and dynamic programming approach.