Cost-effective based Water Quality Monitoring Network design algorithm and  WebGIS Platform

Abstract

The Water Quality Monitoring Network (WQMN) design is most empirical in practices so far. Design parameters includes water quality indicators, monitoring sites, and sampling frequency, which are all linked to the cost. Among them, location is the most important, which directly affects the accuracy of data and the budget. Previously studies few considered the cost of monitoring stations in the optimization objectives, while means a design in practice need to meet the maximum budgets requirements.  This study main considers this kind of restriction for optimize the response effective of the network as if the pollution events comes. Therefore, the basic idea is straightforward:(1) minimizing the total cost of water quality monitoring stations and (2) minimizing the average detection time of the contamination events. The candidate sets of monitoring locations are selected by topology at first. The stations are represented by an adjacency matrix L of river network, wherein the element Lij indicates whether the i-th station is adjacent to the j-th station. The velocity of each river section between stations is represented by matrix V. The average detection time Tr of the network is calculated.                                             Where the distance of each station is represented by matrix D, in which the element Dij means the distance. N means the number of river sections.The total cost of stations C is calculated by the formula.                                                         The cost required to treat the contaminated water L is calculated by the following formula.                                         Where means the cost of sewage treatment per unit mass,  Qvij means river discharge. The optimization formula F is summarized by the formula.                                                               The F value of all potential observation stations is calculated, and the smallest F is the best site. We may also consider the risks distribution among each river reach, descripted by a matrix of R,  according to the local knowledge on pollution sources. Besides, it integrated a GIS-based module that can automatically identify the necessary of parameters, and calculating the optimal locations and number of monitoring sites. It does not rely on water quality monitoring records, nor on hydraulics. We take Maozhou River in Shenzhen, China, as an example to demonstrate the usability of the WQMN design tool and algorithm. A map of monitoring network is successfully produced with the number of WQMN stations reduced to 38. The platform for global application will be online soon for testing.