Evaluaton of population risk due to a contaminated multi-input water distribution network

The city of Milan (Italy) lies over an aquifer, which supplies drinking water to about 1,700,000 inhabitants. Groundwater is polluted by improperly disposed of chlorinated hydrocarbons in the 15 to 150 ppb range. The distribution network is fed by wells and pumping plants scattered all over the city area and yield water containing different contaminant species in different amounts, which also vary in time. Our aim has been to evaluate the amounts of pollutants which reach the end user and eventually to estimate the risk they entail. A number of sampling points was established soon after pollution was first detected (1977). [Veramente, fu scoperta nel 1975.] Concentration data from said points allowed the "zoning" of the city: areas were looked for where time-averaged (1 year) concentration values fell within a given range. Zones were defined by a simple statistical preprocessing of data followed by deterministic interpolation. Zone contour sensitivity with respect to data from additional sampling points has been studied. The second problem, i.e. risk assessment, is now being considered. A database containing a detailed city map and the number of inhabitants living in each house block has been created. We can therefore compute: a) how many people drink water containing given levels of given contaminants; b) how high is the toxicological risk in a given city area. These results will be applied to water distribution management and to aquifer reclamation. Moreover, they will be fed back into the data collection procedure, in order to optimise thee location of sampling points and the time intervals between samples.