EPA Region 9 Air Quality NITROGEN DIOXIDE LOCATION and SEVERITY of NAAQS Violations* 1984-85 NEVADA North-South Profile ARIZONA Based on 1984-85 monitoring data 8 8 3 ♬ 8 8 8 8 Annual Mean (ug/m3) NAAQS = 100 ug/m3 annual mean concentration. * Higher of 1984 or 1985 annual mean concentration. At the Subcommittee's request, the General Accounting Office The report is particularly timely because we understand EPA is actively considering these matters for decision. It raises some concerns about the basis for some of those decisions. request your views and comments. We also request that you consider GAO's findings in your decisionmaking. EPA's Decisions for Regulating Benzene Emissions on the We 1. At the onset, let me express concern about the cost Honorable Lee M. Thomas Page 2 Please explain why this process is so costly, timeconsuming, and resource intensive and what EPA is doing to reduce these costs, etc., significantly for other substances. Are there alternative controls that are as effective? To what extent are these costs driven by the requirements of sections 112 and 307 of the Act, taking into consideration the "lengthy process" described in your December 19, 1985 letter to the Subcommittee? What are the expenditures (and over what period) for each of the other 20 substances considered under section 112 to date by your agency? What future expenditures are required for substances now listed and those that you intend to list? What is the impact of Gramm-Rudman on section 112? It 2. The GAO report confirms what Mr. Ruckelshaus told the Subcommittee at our 1983 hearing that he had "shifted EPA's hazardous air pollutant standard-setting process away from a technology-based approach in favor of a risk-based approach." states that this June 1983 EPA "decision" is "consistent with a 1983 National Academy of Sciences report ... that recommended risk assessment and risk management be used by regulatory agencies dealing with scientific uncertainties." Under this EPA policy, "risk to the public is the overriding factor used in the decision," although the policy "requires information on control technologies, their effectiveness, and costs." The GAO report (pages 24 - 29) discusses EPA's handling of health data concerning benzene which gives the Subcommittee concern. The report indicates that EPA "planned to reevaluate all recent data on benzene health risks" and issued a final report last month. Science Advisory Board (SAB) is to review the "EPA analysis of the new benzene health data." Was this reevaluation done by EPA employees or by contract? What is its status and the schedule for SAB review and report? What will be the next action by EPA and when? What is the relationship of this analysis to the petitions filed administratively and in court? What is the status of those proceedings? The 3. GAO states that "emission and population data used in its benzene decision were not always accurate" and modeling of benzene health and exposure data was based on several assumptions that add to the uncertainty of the benzene risk assessment numbers. Based on its review of EPA files and site visits of plants affected by proposed standards, GAO shows that data were "not current or were based on assumptions" never verified. On prior occasions, we, like GAO, noted that EPA has ample authority under section 114 of the Act to get needed data and verification. EPA never disagreed. But there is little evidence that this authority is often used in regard to standard setting. Costs are cited for not doing so. But there is also a cost in data that are relied upon not being accurate. What are the so-called cost factors? When is section 114 used? What, if any, reasons exist as to why it is not useful in listing substances or setting standards under section 112? The GAO says: Honorable Lee M. Thomas Page 3 Topographical maps are another resource available EPA should examine other ways, like this one, to ensure that its data are accurate. Please explain why such maps have only recently come into use and when EPA will use them. 4. According to EPA, "it is not possible to directly link actual human cancers with ambient air exposure to chemicals, such as benzene," thus EPA "relies on mathematical modeling techniques to estimate human health risks." EPA used a simplified model which has "limitations." Both the simplified and complex versions "assume that (1) people are exposed for 70 years, (2) the population is immobile, (3) sensitive populations are not represented, (4) the terrain is flat, (5) the emissions are constant, and (6) exposure occurs at computer-generated 'population centroids' rather than precise locations." GAO states: According to an engineer in EPA's Pollutant Assessment Branch, EPA can improve its model by accounting for potentially sensitive groups, population, mobility, and terrain differences. EPA had a contractor evaluate the feasibility of making these improvements. EPA officials told us in September 1985 that they had decided to improve the model but had not yet determined the extent to which the improvements would be made. They expected to make a final decision on this issue in late 1985. Please explain when, how, and to what extent EPA will improve this model. What will be the cost and how long will it take? What limitations will remain? Is the model used for all 112 substances? When was the model developed and at what cost? Were these limitations recognized then? 5. GAO states that EPA's Pollutant Assessment Branch plans to develop an operating manual this fiscal year. GAO's recommendation is: |