A methodology to prioritize spatio-temporal monitoring of drinking water quality considering population vulnerability

The optimization of drinking water monitoring becomes increasingly complex with the size of a water distribution system. Municipal water managers have to combine their experiences with different types of information (historical water quality variability, infrastructure, water residence times, sociodemographic profiles, etc.) that are available in different forms (qualitative, quantitative, geographical, etc.) to be able to select the monitoring locations for regulatory compliance and routine water quality management and control. Therefore, the integration of such information requires to select suitable variables and use the appropriate data mining and aggregation methods. This work aims to develop a methodology that helps optimize drinking water quality monitoring programs by considering the different components of population vulnerability that vary both spatially and temporally. This project was conducted in a distribution system that supplies approximately 510 000 citizens. Due to the high seasonal climatic variations and the size of the network, there are also considerable spatial and temporal variations in water quality throughout the year. An index representing the spatio-temporal population vulnerability (combination of population exposure, sensitivity and adaptation capacity) to the degradation of drinking water quality was developed by selecting the relevant parameters and aggregation methods. The population vulnerability index was calculated by aggregating spatio-temporal water quality data (representing microbiological and chemical risks) and distribution network characteristics (number of leakages, pipes type and age). This information was then compared with sociodemographic data related to population sensitivity (percentage of children and the elderly, and the number of health care centers) and the population's adaptive capacity (social and material deprivation). A fuzzy synthetic evaluation method is used for parameter aggregation and to calculate the different indexes. By considering variable locations and periods of time that may better represent the population vulnerability, the results of this project are useful for drinking water managers to optimize their drinking water monitoring strategies.