Office of Advocacy

Date: December 8, 1998

To: Jere Glover, Chief Counsel

From: Kevin Bromberg

Subject: SBA Recommendation on Draft TRI PBT Rule

Introduction

EPA proposes to lower the reporting thresholds on 24 persistent bioaccumulative toxics (PBT), in order to capture potentially significant releases to the environment that are not reported under current thresholds used in the Toxics Release Inventory (TRI) program. EPA has postulated that PBTs, in extremely small quantities, can have a significant effect on human health and the environment. However, consistent with EPA's past policy, the agency has provided no analysis of the releases it expects to be reported, nor any risk assessment about which PBT releases it expects may have a significant effect on human health or the environment. Thus, Advocacy has been required to complete our own analyses to determine whether the reporting facilities will be reporting releases of sufficient quantity to warrant reporting.

This rule will have a potentially costly impact on many of the same facilities that will be subject to the new reporting requirements under the April 1997 industry expansion rule (reports due in June 1999), specifically chemical and petroleum wholesalers. In the absence of any evidence of significant releases, EPA proposes to require 15,000 facilities, of which over 5,000 facilities are owned by small businesses, to report releases of mercury, cobalt, PCBs, PACs and other chemicals. The largest group of reporters are the petroleum bulk plants (1,229, of which 875 are small). In addition to the current requirement to report insignificant air emissions, these new PBT releases are very unlikely to exceed even 0.1% of nationwide (or local) emissions of these chemicals (see December 7^th analysis "PBT Emissions of Petroleum Bulk Plants are Small"). In light of these small business impacts, and impacts on other small business industries, SBA believes this proposal deserves very significant scrutiny. Furthermore, the Administration should honor its April 1997 commitment to achieve TRI burden reductions. This proposal to add 1.9 million paperwork burden hours is not consistent with that commitment.

As two examples, EPA estimates first year costs of $32 million for petroleum bulk plants and $5 million for coal mines, while we would estimate PBT emissions, except for PACs, are negligible. With reports at approximately $5600/report, the cost of reporting mercury and other PBTs will exceed $100 million/ton of PBT reported for many of these facilities (using 0.01 pounds as the maximum release for mercury, cobalt, vanadium and PCBs). Why not require reports of the facilities that EPA has already chosen to regulate under section 112(c)(6) of the Clean Air Act Amendments? Why are many of these reporting facilities not on this list? Shouldn't EPA, at least, discuss this dichotomy in the preamble?

Regulatory Flexibility Act

It is our contention that this rule should not be certified as having no significant economic impact on a substantial number of small entities for two reasons. First, by itself, the PBT rule will have a significant impact. The impacts on the manufacturing sector and the petroleum sector may well trigger the criterion in this rule alone. Second, when combined with the impacts resulting from the TRI requirements finalized in April 1997, the impacts will be even more burdensome. Today, the new reporting industries have not yet reported. These rules could have been promulgated together. The effect of this rule is to add additional reports to the first reporting rule.

Advocacy has always believed that closely related rules should be considered together, irrespective of the fact that the agency has separated them. Advocacy has claimed since 1986 (indeed the first TRI rules) that all the right-to-know rules should be viewed together (sections 311, 312 and 313 at that time). A rule should not be certified simply because EPA chose to promulgate these rules in two parts, rather than one.

As such, we believe the impacts on SIC 5169 and 5171 of the industry expansion rule triggered significant impacts (substantial loss of profits). This rule adds to these costs. In view of the serious impacts on these industries from the first rule, and the new PBT rule, and the significant public policy questions, this rule warrants the SBREFA process. Therefore, a certification of no significant economic impact is not appropriate.

Additionally, over 4300 of the 5600 new small business reporters are in the manufacturing sector. EPA's economic impact analysis averaged over the entire manufacturing sector, making it impossible to separate the economic effects on individual three-digit or four-digit SIC code industries, or even two-digit SIC code industries, where the impacts are quite variable. The number of reports for these facilities are estimated to vary between one and 17 reports per facility, a substantial reporting burden on facilities facing the maximum number of reports.

Discussion

Several options are available to reduce these impacts, while still addressing right-to-know objectives. First, reporting of two of the chemicals, dioxins and mercury, should be deferred, pending additional analysis. Alternatively, the mercury reporting threshold can be raised from 10 to 100 pounds, reducing the number of reporting sources. The dioxin threshold should be raised from 0.1 gram to 1.0 grams. In addition, the Form A alternative threshold (using a "reportable amount" under 500 pounds) should be used to further eliminate sources that are otherwise subject to the new PBT thresholds (see attached data for cobalt and mercury for several industrial categories).

A. Analysis of Inventory Data Collected by EPA

EPA has expended substantial sums to complete inventories of sources of mercury and dioxin in recent years. An eight-volume report to Congress on mercury was completed in December 1998, which includes a national map of emission sources. A dioxin report was completed in January 1998, and a facility inventory is scheduled for completion in January 1999. Earlier this year, the agency compiled an emissions inventory of HAPS under section 112(c)(6) of the CAAA, which accounts for 90% of the emissions for six of the PBTs in this rule.

In light of these facts, SBA is cautiously evaluating the "value-added" of these new reporting requirements. EPA has failed to provide an analysis of the existing data regarding availability, completeness, or accuracy. It has provided no information indicating whether there are natural sources which would overwhelm the contribution of local or regional sources (which is the case for mercury). The agency has provided no information regarding whether the inaccuracy of emission estimates for larger sources would overwhelm the emissions from individual or collective smaller sources. It has provided no information about whether many sources would be subjected to reporting with releases that would be an insignificant fraction of the reporting threshold. Finally, EPA has no information about whether any given threshold would constitute what percentage of reporting of the universe of sources and presents no risk information about these sources.

EPA, as part of a broad and intensive international mercury control strategy, is still compiling emissions data, including a costly testing requirement imposed on utilities to better characterize utility mercury emissions. What will individual facilities reporting under the new rule add to the knowledge base established by the 1997 inventory? What information will be added to the inventory beyond what has already been reported in the voluminous 1997 Report to Congress (which is not mentioned in the preamble)? Is this level of reporting consistent with the mercury strategy and the 112(c) air toxics strategy to regulate industrial sources accounting for 90% of the releases? What are the potential sizes of releases? What is the significance of the natural component to emissions? What are the paths of health impact for mercury and its compounds? What is the significance of small or large sources to the neighborhood, region, or the nation? What significant anthropogenic sources, such as residential oil boilers, will be omitted in the TRI reporting? Despite the overwhelming amount of data, mostly compiled by other EPA offices, this work has not been used to consider how reporting might be structured to capture significant releases.

B. Control of PBT Emissions and Reduction of Risks

Our evaluation of the inventory data leads to several conclusions. First, the PBT rule is unlikely to add to community right-to-know unless the data added reflects significant emissions, in comparison to the total amount of such emissions. For example, it makes good policy sense to require reporting for sources that EPA or the states may wish to consider as candidates for emission reductions. As part of the North American Commission on Environmental Cooperation mercury plan, the US has targeted reductions in mercury emissions. Major reductions are already underway for incinerators and medical waste combustors (approximately 30% of total national anthropogenic emissions). Additionally, the latest estimates are that about 20-30% of the total emissions for mercury is the result of recycling of emissions from natural sources. Also, because a large component of the mercury health risk is derived from regional or distant transport, small local sources do not play a significant role in the health risks.

C. PBT Standards

Lowering the PBT standard is not required by the statute. In this light, EPA should be careful in extending this authority. EPA should take care that these PBTs contribute to risk somewhere, and that knowledge of these PBT releases could cause the community or the facility to consider taking action to reduce risks. Reporting simply for the sake of reporting is not appropriate. Congress intended some benefits to accrue to the community from the reporting burdens placed on industry. While EPA may not be required to calculate risks quantitatively to support this rulemaking legally, it should be using semi-quantitative methods to ensure that emissions that may bear risks are reported, and those that do not, are not reported.

Despite a request by our office, the agency has provided no specific rebuttal as to why alternate definitions of PBTs by the Canadian government or the RCRA office differed from the draft TRI definition, or why specifically those PBT lists include or exclude PBTs listed in this draft rule. The RCRA office has already produced approximately 100 pages of comments and analysis on the extensive ranking scheme it used to produce its PBT scheme, yet there is no comparable discussion from the TRI office. (Although it uses additional ranking criteria for RCRA purposes, the PBT-related criteria are directly relevant here.)

Regulatory Alternatives

A. Establish Thresholds that are Technically and Environmentally Efficient

For dioxin and mercury, EPA needs to evaluate existing inventories, and determine how they should be supplemented, if at all, by TRI reporting. If all significant sources are already identified, and the degree of uncertainty is greater than the size of the sources that would be reporting, this would imply that the source "cutoff" could be set at the level of "significant sources." In the case of mercury, estimated at 316,000 pounds/year, with an additional 20-30% natural component, no significant sources exist that emit under 50 pounds/year. Many sources, particularly utilities, are subject to emission factor uncertainties of the order of 30%.

Using the dioxin inventory, an appropriate cutoff could be set. The reporting threshold has been proposed for 0.1 gram, but nationwide emissions is of the order of hundreds of thousands of grams. What proportion of reporting would be captured by a 1 or 10 gram threshold? Many small business sources are expected to report dioxin. How would these small businesses be expected to estimate dioxin? Estimation is a significant problem even for large utilities. If many of the midsize or largest sources have no detectable emissions, what effect will this have on the inventory?

API reported in its 1997 TRI dioxin listing comments that five refineries in the San Francisco Bay area contributed 0.05 grams of dioxin air emissions, whereas stormwater runnoff accounted for about 98 percent of the dioxin emissions into the air. In a 1996 report, the Bay Air Quality Management Authority stated "[t]he currently approved health risk assessments completed for facilities in the Bay Area under AB-2588 indicate that dioxin emissions are not contributing to significant health risks."

Since reporting in toxicity equivalents (TEQ) is the most common reporting convention at EPA (e.g., EPA 112(c)(6) air toxics inventory) and elsewhere, and since it carries risk information that is not obtainable from a report of the amount of grams discharged, shouldn't TEQ be the preferred form of TRI reporting? We have not reviewed any EPA reports of dioxin emissions that is reported in grams; all report releases in the form of TEQ.

For PACs, EPA has included a 3 1/2 page 1990 inventory of emissions from about 130 industrial categories. Total emissions are estimated to exceed 400,000 tons/year. "These are the byproduct of human and natural activities that are produced or emitted during thermal processes such as the incomplete combustion of organic compounds, pyrolysis, or the processing of fossil fuels, bitumins or nonfossil fuels." Sources include forest fires, volcanoes, internal combustion engines, cigarette smoke, open burning and other forms of combustion. Appendix J to Economic Analysis, page J-2. Why would EPA require reporting of non-combustion plants with a throughput of 10 pounds are year, when combustion is the source of emissions? EPA estimates that over 4,000 reports will be required annually for PACs. The PAC threshold should be set, using the actual data provided in EPA's detailed inventory, compiled under section 112(c)(6). What sources of PACs does EPA propose to regulate, and how does this compare to the proposed reporting facilities?

B. Defer Reporting of Mercury and Dioxins

A better approach would be to defer reporting of dioxins and mercury until the agency first performs an evaluation of the issues outlined above. Thus, reporting could be tied to existing information, existing regulatory strategies, and risk evaluations. We believe that "right-to-know" information should be examined for its costs and benefits, like any other regulatory requirement with costs and benefits. This is simply good public policy, consistent with the Regulatory Flexibility Act and EO 12866.

C. Decrease Reporting Frequency

As a requirement not imposed by statute, EPA could require these "extra-statutory" requirements only every three or five years. Because PBTs are trace elements in processes that are integral to industrial manufacturing, PBT emissions are unlikely to change significantly from year to year.

D. Delete Vanadium and Cobalt from PBT List

As we understand it, the EPA references for vanadium and cobalt do not support the EPA finding of bioaccumulation. Why should vanadium be listed, when the reference cited by EPA declares that the high bioaccumulation values relied on by EPA are an anomaly among the species? Why should EPA list cobalt and vanadium as bioaccumulative when there is no evidence that these metals accumulate anywhere where it is causing harm (unlike mercury, or PCBs, for example). These metals do not appear on any of the other established international or agency PBT lists. If this is only the initial list of PBTs to be developed by EPA, why should these highly questionable chemicals be included now, particularly where these two chemicals account for about one quarter of the total reporting costs?

E. Employ Form A to Eliminate Small Releases with Lower "Reportable Amount"

To eliminate small sources and achieve burden reduction, the Form A was used to eliminate sources of total wastes below a certain threshold. The same philsophy is applicable to PBT releases, although the threshold would need adjustment to reflect the increased concern about smaller quantities of releases. Analagously to the current Form A, the PBT Form A should be used to eliminate sources of emissions below a given release "goal" of the emissions inventory. Therefore, if the goal is to require reporting for all sources in excess of 50 pounds of emissions, the alternate reporting threshold can be set at 100,000 pounds, with the Form A "reportable amount" threshold at 50 pounds. This revision alone could eliminate the major portion of the small business burdens. It is a major issue in this rulemaking. At a minimum, why should the Form A not be used at 10 pounds, the agency minimum throughput for reporting?

In addition, in order to correct a problem discussed in the draft preamble, we propose that the Form A "reportable amount" include all wastes reported, including remedial actions. We agree with the agency that remedial actions should be reported on the Form R, if the amount exceeds the reportable amount. This oversight should be corrected in the Form A, as it exists today, and for the PBT Form A.

We would propose that EPA develop a two tier system for the Form A in this rulemaking. First, it should modify the current reportable amount from 500 to 5,000 pounds for non-PBT chemicals, in order to provide some paperwork savings, in accordance with the Administration promise. This captures under the Form R, even under EPA's conservative accounting, 99.9% of the data, while saving about an additional 150,000 paperwork hours. (By comparison, this rule accounts for a 1,900,000 annual paperwork increase.) We propose a 50 pound Form A for the PBT rule. Thus, the two tiers are: 50 pounds for PBTs, and 5,000 pounds for non-PBTs.

For example, consider the following mercury data derived from Appendix F of the EPA EA. The following table provides a comparison of the source releases to the throughput amounts (process, manufacture or use) which are used to trigger the reporting requirement.

MERCURY RELEASES
SOURCES	Amount Reported In 1994-1995 (lbs/year)
Combustion Sources	275,400
Utility Boilers (99% coal)	103,600
Municipal Waste Combust	59,200
Ind. Boilers (75% coal)	56,800
Medical Waste Incinerators	32,000
Hazardous Waste Combustors	14,000
Manufacturing Sources	31,200
Miscellaneous Sources	2,800
POINT SOURCES	309,400
AREA SOURCES	6,800
MOBILE SOURCES (not estimated)	--
TOTAL	316,000
MERCURY THROUGHPUTS
		Amounts Expected To Be Reported (lbs/year)
Metal Mining (trace amounts in ore)		131,000
Coal Mining (trace amounts in coal)		433,850
Paper & Allied Products (impurities in wood and chem)		3,800
Chemicals & Allied Products (5% as trace)		414,000
Petroleum & related Indus. (trace amounts in pet. Crude)		11,360,000
Primary Metals Smelting & Refining (trace in ore)		192,000
Secondary Metals S&R (recovered from scrap)		984,000
Electric and Electronic Equip (in bulbs etc.)		175,000
Instrument Related products (part of prod.)		164,000
Electric Power (not calculated -- trace in fossil fuels)		---------
Petroleum Bulk stations & Bulk Terminals (trace in oil)		2,219,600
TOTAL in NON-combustion situations		16,300,000
TOTAL for facilities burning coal and oil (SICs 20-39)		54,000
TOTAL		16,354,000

For non-combustion facilities, the throughput of 16 million pounds vastly overwhelms the total emissions, which is under 50,000 pounds. Thus, emissions are a minute fraction of the throughput for these hundreds or thousands of facilities. For the great majority, if not almost all, the emissions would be well below the ten pound throughput. The Form A would be used to eliminate these very small sources from the full Form R. In constrast, for the over 1,000 oil and coal combustion sources, the throughput of 54,000 pounds is likely comparable to the emissions of these sources (data not available in Appendix F), and the Form A would only be needed for the smallest facilities.

In 1995, three refineries reported mercury emissions to TRI. One refinery, Koch Refining, reported 32 pounds, and two others reported 0 pounds (presumably less than 0.5 pounds). Since these facilities processed over 25,000 pounds, the other refineries (Appendix F estimates throughputs in excess of 1,000 pounds for all but one of 173 refineries) will be reporting a fraction of these releases. If releases under 0.1 pounds are reported as zero (new draft PBT protocol), many refineries and many other new PBT reporters can be expected to report zero emissions. Perhaps, as a simple solution, non-combustion sources should be exempt from PBT reporting where combustion processes are the source of almost all emissions (e.g. mercury).

Alternatively, sources that do not perform chemical reactions or combustion could be exempt from reporting, in a way analogous to the alloy reporting and the dioxin "manufacture only" exemption. EPA’s explanation for "manufacture only" is "These [chemicals] are ubiquitous in the environment…, facilities that process raw materials would be required to report simply because the raw material contains background levels of these chemicals. … This will mean that only those [chemicals]

That are manufactured at the facility, including those coincidentally manufactured, will be the subject of reporting under section 313." Preamble, p. 47 This rationale is also applicable to those facilities that do not perform chemical reactions or combustion for the relevant chemical. Some of the PBTs are trace elements that are ubiqitous in the environment (e.g. cobalt, mercury, PCBs).

Based on the bulk plant report cited earlier, bulk plant emissions (except for PACs) are expected to be under 0.01 pounds, and will, therefore be reported as zero releases. These chemicals include PCBs, cobalt, vanadium and mercury. These all involve non-combustion. EPA estimates about 3,850 reports from this sector alone for these particular chemicals.

Our review of a substantial portion of manufacturing plants in SIC 20-29 (specifically, SIC 20, 22, 25,28 and 2911) reporting air emissions of cobalt compounds suggests that manufacturing facilities that do not currently report cobalt will be reporting zero releases. Current TRI reporters of cobalt compounds from 1995 and 1996 report zero releases in 70% of the reported emission values for air. Similar results are found for cobalt reporting. Current reporters report releases with throughputs exceeding 25,000 pounds, and with mixtures in excess of 0.1% cobalt compounds. The new reporters will be reporting releases with throughputs of up to a factor of 2500 lower, with no limit on the concentration of the mixture. Many of these new reporters will be smaller facilities in the same industries that are already reporting. Since these facilities are most often using the same industrial processes, new reporters can be expected to report releases that are only a fraction of their reporting counterparts. Thus, given the substantial decrease in throughputs, and the elimination of the de minimis threshold, the 70% figure can be expected to increase. (The new protocol specifies a report of zero, if the release is below 0.1 pounds; current protocol specifies a zero report for releases under 0.5 pounds)

Current cobalt compound reporters have average emissions of 100-200 pounds. Seventy percent of the 1996 emissions, or 46,000 pounds, were contributed by only 15 emission points (all but two are at DuPont facilities). Conversely, 352 of 367 emission points contributed a total of 11,000 pounds, or about 30 pounds each. The new emissions will be dramatically lower than the currently reported emissions. Thus, we expect that the top 15 emission points (46,000 pounds) will exceed the total of all the new reporters by a factor of more than ten to hundred in the same industries. Industries that are not currently reporting at all would be expected to have even lower releases.

According to table A-9 to the EPA Economic Analysis, only about 10,000 of the 32,000 reports are expected to involve combustion. Many, if not most, of these 22,000 reports are likely to involve zero releases. A separate analysis of petroleum bulk plants reveals several thousand reports of zero releases.

Many, if not most of the 22,000 reports will involve zero releases, and certainly could benefit from use of the Form A. In addition, the Form A would be useful to facilities releasing below the "target goal" of releases. We do not believe there is practical utility to completing the full Form R for releases of, for example, PCBs from petroleum bulk plants. These facilities will report zero emissions. There are no Form R activities to report. Also, to the extent that the Form A "reportable amount" is set high enough, the amount of calculations needed to determine whether the facility qualifies for the Form A could be reduced. For example, based on our calculations for the four chemicals identified above, no petroleum bulk stations need perform any calculations for these chemicals to file the Form A, even if the Form A "reportable amount" were one pound, because these releases are so low under any circumstances for all bulk plants. Thus, the costs of filing 3,850 Form A’s would be almost negligible. Similarly, the use of the Form A for other "zero" or "near-zero" reporters would result in substantial savings, at no cost to the right-to-know.

Attachments: Cobalt and Mercury Throughput/Emissions