This article was produced in partnership with The Times-Picayune and The Advocate, which is a member of the ProPublica Local Reporting Network.
The data for our story and corresponding graphic comes from several sources. We provide details on each below.
The bulk of the analysis relies on data from the Risk-Screening Environmental Indicators (RSEI) model, which was developed by the U.S. Environmental Protection Agency. Manufacturing facilities with 10 or more employees in particular industries, which are in possession of chemical quantities above specific thresholds, are required to disclose information on their toxic emissions to the Toxics Release Inventory, a program administered by the EPA. The EPA releases this information online each year as required by the 1986 Emergency Planning and Community Right-to-Know Act. RSEI translates the TRI data, which is reported by weight, into values that reflect the relative risk to human health. These indicators allow regulators, companies and communities to assess risks and take action relative to a specific facility or waste stream.
Our analysis uses RSEI data from 2017, the most recent year available. To understand the estimated impacts of industrial activity in neighborhoods near facilities, we looked at data at the most granular unit available — 810-by-810-meter grid cells. By isolating chemical concentrations from individual facilities, we were able to identify the entities that pose the greatest potential health risks to communities along the lower Mississippi River.
In our interactive map, we visualized the estimated concentrations of cancer-causing chemicals generated from large industrial facilities. In order to compare those concentrations, we weighted them by a measure called the inhalation unit risk (IUR). A chemical’s IUR is an estimate of the increased cancer risk an individual incurs if, over the course of their lifetime, they are continually exposed to the chemical at a concentration of 1 microgram per cubic meter of air. IUR values are published in the EPA’s Integrated Risk Information System (IRIS). We refer to IUR-adjusted values throughout our article and graphic as “toxic levels of cancer-causing chemicals.” These values do not reflect non-cancer related toxicity.
Our tool lets readers see how cancer-weighted concentrations in their area compare with everywhere else in the following seven parishes on the lower Mississippi River: St. John, St. James, Ascension, Iberville, St. Charles, East Baton Rouge and West Baton Rouge. Our map uses a range of cancer risk similar to that of the EPA’s National Air Toxics Assessment (NATA), which we describe in more detail in a later section. On the lower end of our scale is a cancer risk of 1 in 100,000. That is, if 100,000 people are exposed to the same concentration of a pollutant continuously over a presumed lifespan of 70 years, one person would likely contract cancer from the exposure. The high end of our scale is 1 in 1,000. In the cases where we show cancer-weighted toxicity for individual chemicals and facilities, we scale the map between 1 in 1 million and 1 in 1,000. Although we scale our map by cancer risk, the values we show do not convey absolute risk to individuals residing in these areas. We are illustrating toxic levels of cancer-causing chemicals, as modeled by RSEI. On the advice of experts in this field, our tool does not publish risks at the grid cell level. The EPA says RSEI “does not perform a risk assessment, but is rather a screening-level tool to help identify situations of potential concern.”
The text version of our story also uses RSEI data to analyze Louisiana’s improvements in toxic emissions over time, compared with the rest of the country. For that analysis, we used historical RSEI data, aggregated at the level of census block groups (data for 810-by-810-meter blocks is unavailable for years prior to 2015 in the most recent version of RSEI). The historical analysis included all chemicals, not just carcinogenic ones. We analyzed facility-level aggregated RSEI data from 1988 to 2017, every year that it was available.
In addition to determining the locations of greatest concern, ProPublica hired RSEI expert Michael Petroni to estimate increases in toxic levels of cancer-causing chemicals from new industrial plants and plant expansions that have either received approval or are currently pending approval from the Louisiana Department of Environmental Quality (DEQ). He modeled data for new industrial projects using the RSEI methodology. The data sources for that analysis are detailed in the following two sections.
For our investigation, we needed to determine where chemical companies are planning to build new and expanded plants along the lower Mississippi River. We filed several public records requests with the DEQ to obtain permit files of industrial projects that are pending approval or have already been approved by the agency.
We asked for all approved and pending air permits dating back to Jan. 1, 2015, in the seven parishes listed above.
Once we received the permit files, we excluded any projects that are not expansions or new facilities (many of the files we received were modifications of existing permits). Additionally, we only included projects that the EPA deems to be “major sources” of toxic air pollution. Such projects are required by federal law to receive Title V air permits before they can begin operating. A column in the data specifies each project’s permit type. We filtered for initial Title V air permits, excluding other types of projects.
EDMS and ERIC Online
Once we had a list of approved and pending major-source industrial projects between Baton Rouge and New Orleans, we had to determine each project’s proposed emissions. Inputs to the RSEI model include chemical name, pounds of the chemical released and other information specific to the source of the release, such as stack height and exit velocity. Stack height refers to the height at which pollutants are released to the air, through pipes, vents or other enclosed air streams. Exit velocity measures the speed at which pollutants escape into the air.
For industrial projects that have already received approval to begin operating, we pulled information on their allowable chemical releases from the Emissions Reporting and Inventory Center (ERIC), an online database administered by the DEQ. We also looked at proposed emissions information from Formosa’s two pending facilities in the DEQ’s Electronic Document Management System (EDMS) because its St. James complex is one of the largest and most expensive industrial projects in the state’s history, and the company was recently sued for polluting waterways in neighboring Texas. Formosa has said that the conditions of the settlement it reached with the state demonstrate its commitment to manufacturing its products “in a safe and environmentally friendly manner.”
The National Air Toxics Assessment (NATA) is a screening tool administered by the EPA to help state and local agencies identify pollutants, emission sources and locations of potential risks to public health. Unlike RSEI, which publishes relative toxicity information, NATA assigns estimated cancer risk scores to census tracts across the country. While our analysis did not use NATA data, we checked our results against it to ensure our results matched the locations NATA flagged for high cancer risk.
Spatial data in our interactive map came from several sources. Louisiana keeps a shapefile of parish boundaries on its website. The shapefile with 810-by-810-meter grid cells is downloadable from the U.S. EPA’s RSEI site. The shapes of land parcels in St. James Parish owned by industrial companies were provided by Justin Kray, a data analyst with the Louisiana Bucket Brigade. Kray pulled information on existing industry in St. James from the Parish Assessor’s Office and 2017 aerial photography from the U.S. Department of Agriculture. Data on recent sales and transfers came from the St. James Parish Clerk of Court.
Our graphic includes information on demographics in St. James Parish. These estimates were taken from the U.S. Census Bureau’s 2013-2017 American Community Survey 5-Year Estimates.