Probabilistic risk assessment of water distribution system in Hyderabad, Pakistan reveals unacceptable health hazards and areas for rehabilitation.

Poor water quality exacerbates multidimensional poverty in developing nations. Often centralized treatment facilities generate acceptable water quality, but the water is contaminated during distribution. Methods to assess sources of contamination in water distribution systems are lacking. A case study of two methods, human risk assessment linked to water distribution system sampling was conducted in Hyderabad, Pakistan to determine areas requiring infrastructure rehabilitation. Water samples from source water (i.e., the Indus River), treatment plant effluent and from taps in the water distribution system were analyzed by atomic adsorption spectroscopy for metals and metalloids (As, Cd, Cr, Hg, and Pb) and water quality parameters (dissolved and suspended solids, pH, conductivity, and total organic carbon). Source water exceeded acceptable drinking water levels for As, Cd, total Cr, and Pb, while the treatment plant effluent concentrations were acceptable. Concentrations of all metals and metalloids, except Hg, increased in the water distribution system post-treatment, exceeding safe drinking limits in at least one location, suggesting contamination of the water during distribution. A deterministic and a probabilistic risk assessment were conducted to evaluate two scenarios: (1) unrestricted use of piped water for all household purposes, including as drinking water and (2) restricted use of the water for purposes other than drinking in the household, including only dermal and inhalation exposure pathways. The water was deemed unsafe for unrestricted use as the sole source of drinking water by both risk assessment methods. Yet when an alternative source of drinking water was assumed and the piped water was used only for bathing and dish washing, the probabilistic risk assessment revealed acceptable health risks to the population, while the overly conservative deterministic risk assessment suggested unacceptable risks. The combined methods of water sampling, risk assessment and correlation analysis suggested areas for rehabilitation of the water distribution system in Hyderabad, Pakistan and these methods can be adopted in other developing nations to target limited funds for infrastructure rehabilitation.