But EPA says approach pays few dividends for RMP
For several years, EPA’s Office of Inspector General (OIG) has been trying to persuade several Agency programmatic offices to conduct cross-checks wherein facility information collected by one program can indicate if the facility is potentially not complying with requirements in another program. Generally, the OIG believes this approach can be useful in alerting facilities about their obligations or provide the information the EPA needs to take enforcement actions against noncompliance. The OIG first looked at the benefits of cross-checking for facilities that fail to file required reports—called nonfilers—in 2009, and more recently issued a follow-up report that it sent in draft form to two Agency offices. One office agreed with OIG’s draft findings and said it would undertake the recommended actions; a second office disagreed with virtually all the findings and recommendations. As a result, the OIG modified the findings and recommendations in its final report.
The OIG looked specifically at three EPA programs: the Emergency Planning and Community Right-to-Know Act’s (EPCRA) Toxics Release Inventory (TRI); the Clean Air Act’s (CAA) Risk Management Program (RMP); and permitting and discharge monitoring reports (DMRs) required under the Clean Water Act’s (CWA) National Pollutant Discharge Elimination System (NPDES). The OIG found that by using RMP reports, the EPA found very few facilities that did not file required TRI reports. Similarly, after looking at TRI reports, the Agency found very few facilities that did not file required RMP reports. Regarding NPDES permitting, the OIG found that using DMR information in TRI reports was not effective in identifying facilities that should have NPDES permits but do not.
The Agency’s Office of Chemical Safety and Pollution Prevention (OCSPP), which oversees the TRI program, agreed with OIG’s findings and said it would take several recommended actions to improve the way TRI, RMP, and DMR data are used to identify TRI nonfilers. EPA’s Office of Land and Emergency Management (OLEM), which oversees the RMP, told the OIG that using TRI data has been found to “be a very inefficient and ineffective method of identifying RMP non-filers.” The OLEM did agree with an OIG recommendation that a modification to the TRI reporting software can be effective in notifying RMP nonfilers of their potential RMP reporting requirements.
While it does not seem that the changes being undertaken as a result of OIG’s report will significantly alter management of the TRI program and RMPs, facilities that have decided not to report under either of these programs should be aware that the EPA will soon have more tools to identify who they are.
The OIG investigation concentrated on reports required by three federal programs.
- TRI. Under Section 313 of EPCRA, facilities must file TRI reports with the EPA if the following conditions are in place:
- The facility is in an industry sector that is required to file.
- The facility has 10 or more full-time equivalent employees.
- The facility manufactures or processes more than 25,000 pounds (lb) or otherwise uses more than 10,000 lb of a listed chemical. Persistent, bioaccumulative, and toxic chemicals have lower reporting thresholds.
TRI reports filed with the EPA include the quantity of each chemical released to air, water, and land. Information reported also includes the maximum amount of the chemical on-site at the facility during the reporting year.
- RMP. Pursuant to CAA Section 112(r) and RMP regulations (40 CFR Part 68), owners or operators of facilities holding more than a threshold quantity of a regulated substance in a process must file an RMP document that includes a hazard assessment detailing the potential effects of an accidental release, an accident history of the last 5 years, and an evaluation of worst-case and alternative accidental releases. RMPs must be revised and resubmitted to the EPA every 5 years.
- NPDES. Under the CWA, any discharger to waters of the United States must obtain a NPDES permit from the EPA or a state authorized to implement the NPDES program. A NPDES permit generally specifies an acceptable level of a pollutant or pollutant parameters that may be discharged into a receiving water. The TRI database includes wastewater discharge data from DMRs. The EPA has developed its TRI and the DMR Comparison Dashboard, a tool that allows users to compare DMR data search results against TRI data search results and vice versa.
OIG’s 2009 report
OIG’s 2009 report recommended that the EPA conduct programmatic cross-checking related to airborne chemical releases. Specifically, the OIG said the EPA should incorporate TRI and other effective methodologies into guidance for EPA regions to use to identify potential RMP nonfilers. All those recommendations have been implemented, reports the OIG. The current report indicates that the OIG believes that more can be done.
For example, during its 16-month investigation for the current report, the OIG identified hundreds of potential RMP nonfilers based on analysis of 2014 TRI data. EPA efforts, conducted during the course of the investigation, also found hundreds of potential RMP nonfilers. The EPA distributed lists of potential nonfilers to EPA regions for follow-up compliance and enforcement actions. To date, results indicate few actual nonfilers (only 4 out of 563).
Similarly, to evaluate the utility of using RMP data to identify TRI nonfilers, the OIG analyzed data for toxic chemicals regulated by both the RMP and TRI program. The OIG focused on facilities categorized in the chemical manufacturing sector. About 8 percent of the facilities looked at were found to be potential TRI nonfilers. After a review of 2011 through 2015 data based on RMP data, the EPA found only 4 percent to be actual nonfilers.
“While review of TRI and RMP data identifies many potential non-filers, very few of these have been confirmed as actual noncompliant facilities,” the OIG concluded. “Further, these follow-up efforts have been characterized by EPA staff as resource-intensive.”
In comments on OIG’s draft report, the OCSPP and EPA’s Office of Enforcement and Compliance Assurance (OECA) noted that they deferred all Agency responses about finding RMP nonfilers to the OLEM. But regarding TRI nonfilers, the OCSPP agreed to develop a mechanism to annually identify potential TRI nonfilers by using an automated comparison of RMP and TRI data. The OCSPP informed the OIG that its recent comparative analysis of RMP reports with facilities that filed TRI reports to identify facilities that may be noncompliant with the TRI reporting requirements produced a finding of 4 percent noncompliance with the TRI reporting requirements by those facilities that did not file TRI reports for the 2011, 2012, 2013, 2014, or 2015 reporting years.
“Nonetheless, OCSPP will assess the feasibility and practical utility of developing an automated process of using RMP information to identify facilities that may be non-compliant with the TRI reporting requirements by September 30, 2018,” the OCSPP told the OIG.
In its comments on the draft, the OLEM questioned the benefits of analysis of TRI data to identify actual—in contrast to potential—nonfilers of RMP reports. The OLEM pointed out that during its 2009 investigation, the OIG identified 39 potential RMP nonfilers. Based on this finding, the OIG recommended that the EPA strengthen its controls to identify RMP nonfilers, and the OLEM subsequently implemented a policy to search for RMP nonfilers every 3 years. However, the OLEM eventually determined that none of the 39 potential nonfilers were subject to the RMP rule.
“This fact was raised with the OIG auditors on several occasions throughout their evaluation, including during the opening and closing meetings on July 26, 2016, February 9, 2017, and May 23, 2017; however, it does not seem to have been incorporated in this evaluation and report,” commented the OLEM. “It performs a surface-level comparison of the TRI and RMP databases to develop a list of ‘… Hundreds of Potential RMP Non-Filers With Millions of Pounds of Potentially Unreported Chemicals.’ However, the actual results are virtually the same as in 2009. Very few of OIG’s ‘potential’ RMP non-filers have been found to be actual RMP non-filers.”
According to the OLEM, the reasons that comparing TRI and RMP filings is a relatively ineffective method of discovering RMP nonfilers include the following:
- Different minimum concentration criteria. For some substances that are common to both the RMP and TRI chemical lists, both rules do not cover substance mixtures or solutions where the substance is present below 1 percent concentration (under TRI, this is referred to as the “de minimis concentration”). However, for other substances, the de minimis concentration under TRI is different from the minimum reportable concentration under the RMP rule. For example, for 24 substances on both the RMP and TRI lists, the TRI de minimis concentration is 0.1 percent, but the RMP minimum concentration is 1 percent. Additionally, the RMP rule specifies much higher minimum concentration cutoffs than the TRI for several substances, including aqueous ammonia (20 percent), hydrochloric acid (37 percent), hydrofluoric acid (50 percent), and nitric acid (80 percent). Due to these differences, many facilities that report more than RMP threshold quantities of one or more substances under TRI are not subject to the RMP rule if they hold the substances below the applicable RMP concentration cutoff.
- Quantity on-site vs. quantity in a process. To be covered under the RMP rule, a facility must have a process containing more than a threshold quantity of a regulated substance. Process means “any activity involving a regulated substance including any use, storage, manufacturing, handling, or on-site movement of such substances, or combination of these activities.” Under this definition, if a facility has several separate areas for storing, manufacturing, or using chemicals, but no single area contains more than a threshold quantity of a regulated substance, the facility is not subject to the RMP rule. The TRI rule, on the other hand, requires facilities to report the “maximum amount of the toxic chemical on-site at any time during the calendar year.” Therefore, the TRI database contains some facilities that hold threshold quantities of RMP substances on a sitewide basis but not in a single process. These facilities are not subject to the RMP rule.
- Bulk storage quantities. Some facilities erroneously report very large “maximum amount on-site” quantities of RMP-covered substances on their TRI submission, but the substances are not actually held on-site above RMP threshold quantities at any one time. For example, the EPA has identified numerous examples of facilities reporting combustion by-products, such as hydrofluoric and hydrochloric acid, above RMP threshold quantities. Only when the EPA contacts or visits the facility does it become apparent that these substances are not actually being held on-site above an RMP threshold.
- Flammable mixtures. Some toxic chemicals covered under TRI are listed as flammable substances under the RMP. Where these substances are present as part of flammable mixtures (for example, at a petroleum refinery), the entire mixture can be reported in the RMP database under the name flammable mixture. While a comparison of the TRI and RMP databases for one of these substances may appear to identify an RMP nonfiler, in some cases, these facilities have, in fact, submitted an RMP for a flammable mixture containing the named flammable substance.
The OLEM stated that determining whether a potential nonfiler is an actual nonfiler is a labor-intensive process that involves preparing legal documents and information requests and may also require a site visit. Given that very few potential nonfilers are actual nonfilers, the effort seems counterproductive and detracts from resources needed to inspect high-risk RMP facilities to ensure they are operating safely and protecting the surrounding local community from the consequences of accidental chemical releases, said the OLEM.
OIG’s report is here.
William C. Schillaci