What Happened
BlueConduit released a national mapping tool at LeadOutMap.org that uses Artificial Intelligence (AI)-powered predictive models to estimate the percent of homes likely to have a lead service line (LSL) within a utility.1 LSLs are lead pipes that connect the water main under the street to its internal (in-home) plumbing system.
Users can enter any address to get the estimate. They can also zoom out to see the area served by their utility and how their utility compares to others in the region and state. Areas marked in grey indicate no data was yet available on which BlueConduit could develop an estimate, or there is no utility serving the area.
The data is based primarily on EPA’s 7th Drinking Water Infrastructure Needs Survey and Assessment (DWINSA) and consists of information reported through states to EPA from 93% of 708 large utilities, 25% of about 9,000 medium utilities; and 1.7% of about 40,000 small utilities. It also pulls in data from several state agencies (e.g., California, Indiana, Michigan, Wisconsin) that had already compiled information about the count of lead service lines from some of its utilities.
Because those data sources include significant numbers of service lines with unknown material, LeadOutMap fills the gaps using BlueConduit’s machine learning technology to create estimates based on wide-ranging national and local data, including water system reports, census and real estate data, historical records, and other sources.
Why it Matters
Americans commonly use maps as a tool to help plan a trip, check the weather, find a home, or learn more about a community, a neighborhood, or even individual homes. Maps provide them with familiar means to interactively evaluate information, make decisions, and take action.
When it comes to exposure to lead, we have seen federal agencies make promising efforts to map the risks at the census tract level:
- HUD has a Deteriorated Paint Index interactive map based on percent of occupied housing built before 1980 with large areas of deteriorated paint.
- CDC has a Lead Exposure Risk Index (LERI) static map based on the same housing measure as HUD but couples it with environmental, sociodemographic, and geographic factors with the variables weighted on NHANES biomonitoring date from 2005-16.
- EPA has an EJ Screen that includes “Lead Paint EJ Index” interactive map based on percent housing units built before 1960 that is combined with demographic index based on percent of households with families that have low incomes or are people of color.
- EPA, HUD, and CDC released a study earlier this year evaluating Lead Exposure Hotspots that included various static maps. See our February blog on the study.
States are getting involved, too. In March, Unleaded Kids published a blog describing similar mapping efforts by California, Indiana, New Jersey, and Ohio.
Blue Conduit is the first organization to have a national interactive map for LSLs. The company’s predictive models using AI tools were first developed by University of Michigan researchers to assist Flint’s LSL replacement efforts. BlueConduit works directly with water systems across the country to support LSL inventory and replacement, in which they provide address-by-address predictions within water systems. The Rockefeller Foundation and Google.org supported the development of the LeadOutMap.
Our Take
We applaud BlueConduit as well as EPA, HUD, and the states for developing interactive mapping tools and encourage CDC to do the same with its LERI. The tools are an important means to help professionals better understand and communicate the risks of lead exposure in their community.
However, the tools may be frustrating to public users who seek detailed information at the address level and do not think in terms of census tracts or utility service areas. We recognize that this is primarily due to the fragmented and incomplete data on which they rely. In general, we lack sufficient address-specific information on the presence of LSLs, lead-based paint, or other sources of lead risks. AI tools such as those by BlueConduit can help fill in the gaps and better convey the information, but more data is needed.
We also are concerned that tools presenting the risk of single sources of lead exposure, such as LSLs or lead paint, may mislead the public because they lack context, especially when framed as showing conditions that are “lead-free” or “lead-safe.” Even with LSLs removed, lead may still be in drinking water from other sources, such as solder, brass, bronze, and galvanized coatings. And there is no safe level of exposure to lead.
We are hopeful that the situation will change starting this October when all utilities must make their service line inventories publicly accessible pursuant to 40 C.F.R. § 141.84 of EPA’s revised Lead and Copper Rule (LCR) signed in December 2020. The inventory must include the specific location of each service line and whether it is made of lead, 2) is non-lead, 3) is galvanized requiring replacement; or 4) its status is unknown.
For the 1,000 utilities serving more than 50,000 people, the inventory must be available online. If a utility has only non-lead service lines, its publicly accessible inventory may consist of a statement with a general description of the sources it used to make that affirmative determination.
We are encouraged to learn that some states such as Indiana, Ohio, and New Jersey are planning to integrate this address-specific information into interactive maps. They will be much more useful to people who do not have the patience to check with each utility. We also encourage BlueConduit to integrate the address-level information into its LeadOutMap as it becomes available from utilities, and we hope it will expand its tool to serve as a platform to track progress toward replacing all LSLs. In addition, we encourage federal agencies to collaborate on efforts like CDC’s LERI and their joint Lead Exposure Hotspots that will provide context for the risks posed by LSLs. It would be unfortunate if the public looks at a map of LSLs without understanding the risks from lead in paint, soil, and dust.
- By utility, we are referring to Community Water Systems under the Safe Drinking Water Act. ↩︎