Assessing Lead in Drinking Water

Elevated lead levels in Flint, Michigan have raised awareness in the United States of the continued presence of lead in drinking water. Lead is a neurotoxin and harmful to human health at any level. Negative health effects include neurological issues and behavioral issues, and effects are exacerbated in children and pregnant women. The presence of lead is common in older industrial cities, which saw large expansion during the early 19th century when lead was a preferred material in drinking water systems. Lead was most commonly used in service line pipes (the pipe connecting a house to main line in the street), indoor piping and plumbing fixtures, making this a unique public health question involving public and private plumbing systems interacting with water treated by public utilities.

Lead levels in drinking water are regulated through the Lead and Copper Rule (LCR). The LCR mandates operation of corrosion control systems, which adjust select water chemistry parameters to reduce lead release from plumbing. LCR sampling is conducted at a limited number of customer homes that are known to contain lead plumbing and requires that the 90th percentile of samples be less than the action level of 15 ppb. The LCR’s approach is operationally focused. Further investigation is required to determine the public health implications and overall water quality in a large drinking water system.

Pittsburgh Water and Sewer Authority (PWSA) has been actively managing lead through its corrosion control program for many years. However, recent results (2016 and 2017) of mandatory sampling indicated elevated levels in more than 10% of the test locations. As with many water utilities, PWSA has insufficient information on areas of the system that are at higher risk due to presence of lead piping or the use of lead-containing fixtures in homes. Compliance testing locations may be insufficient to identify potential hot spots; however, PWSA’s extensive customer-initiated testing program provides much more data across a wider selection of city neighborhoods. To prioritize efforts to reduce lead levels in drinking water and to assess the value of multiple potential interventions, novel sampling and analysis techniques are required.

In this project, a statistical model predicting regions at risk for elevated drinking water lead levels is paired with a decision-making optimization model to identify the best combination of lead reductions strategies across a distribution system to meet both cost and multiple water quality objectives (e.g. overall lead reduction versus prioritization of lead reduction for sensitive populations).