Implementing to PSD program - ACS Publications - American

fare are not at risk? By the courts' interpretation, the. Clean Air Act requires an air quality program aimed not only at improving air quality where ...
0 downloads 0 Views 7MB Size
Implementing the €%Dprog6m An NAS studyfinds that the prevention of significant deterioration program is basically sound but that flexibility in administering this provision must be assumed by the states Myron F. Uman National Research Council Washington, D.C. 20418

article the findings, conclusions, and recommendations of the academy's study are summarized.

John T.Middleton Private Consultant Washington, D.C. 20008

The PSD committee As is customary, the academy, through its National Research Council, assigned the study to a committee organized expressly for the purpose, in this instance under the auspices of the NRC's Environmental Studies Board. Committee members were chosen for their technical expertise in the physics and chemistry of the atmosphere; air quality modeling; visibility; effects of air pollution on plants; air pollution control technology; environmental law, economics, and decision making; political science; and industrial management (see box). Academy committees are not ordinarily asked to make recommendations about what public policy should

Clean air is a valuable resource. In most cases, industry and others exploit that resource just as they'consume other natural resources. Few would argue that degradation of air quality sufficient to threaten public health and welfare should not be prevented. But should we manage air quality where it is good, where public health and welfare are not at risk? By the courts' interpretation, the Clean Air Act requires an air quality program aimed not only at improving air quality where the air threatens public health and welfare but also at preventing serious degradation of air quality where the air is relatively clean ( I ) . Congress affirmed this concept in the 1971 amendments to the act by enacting Part C of Title 1, which deals with prevention of significant deterioration (PSD) of existing clean air resources (2). But the PSD provisions have been controversial, both in the principle they represent and in their implementation. In section 323(f) of the act as amended in 1977, Congress also specified that the National Academy of Sciences should conduct a study of the implementation of the PSD provisions. A similar study was required of the National Commission on Air Quality. Both studies were presented toCongressonMarch2,1981. Inthis 1000 Environmental Science 8 Technology

NAS committee for PSD $tu John 1.Middleton, chairman Myron F. Uman, NRC staff officer Thomas D. Crockw, University Wyomilig J. ClDavifia, The collservation

Foundation

'

Charles H. Kruger, Jr., Stanford Uni-

versity John D. Ludwig, EPA (retired) John H. Seweld, Califanii Instituteof

Technology Lelan F. Sillin, Jr.. Northeast Utilities Service Co. Mi@ Soper, University of Michigan John Trljonk, Santa Fe Research WP.

be; rather, they are asked to analyze scientific and technical aspects of policy options in the hope that the analyses will shed light on issues of concern in debates about policy. So it was with this committee. It was not asked to judge whether the purposes of the PSD provisions were good or bad; it was asked to examine their implementation in light of the objectives of the act and to make recommendations for improving the provisions and their implementation so that the objectives might be more readily achieved. We did this by studying the legislative history of the act; examining the provisions; assessing the state of scientific knowledge and the capabilities and limitations of analytical techniques; interviewing more than 80 people from industry, government, and academia; and conducting an extensive case study. The PSD provisions The PSD provisions, which apply in regions of the country where the air is cleaner than the minimum acceptable levels of air quality (defined by the National Ambient Air Quality Standards), create a decision-making process for managing the air resource. The objective is to strike a balance between industrial development and protection of values associated with clean air. The committee concluded that the PSD provisions provide a basically sound means of attaining this objective, but that they have generally been interpreted too narrowly by EPA, state regulatory agencies, and industry. As a result, the program has been implemented with a rigidity that iscontrary

0013-936X/81/0915-1000$01.25/0 @ 1981 American Chemical Society

to congressional intent and that will Class 111 parts of three counties in defeat the act's purpose. southwestern Illinois. Congress understood that there is For each of the three classes of land, neither adequate scientific knowledge the act specifies "maximum allowable nor a national consensus about the increases," or increments, over a relative worth of existing good air baseline concentration for sulfur quality tosociety toestablish uniform dioxide and total suspended particunational standards. The worth of lates (TSPs) (Table I ) . Most misinmaintaining good air quality where it terprerations of the PSD prouisions exists-relative to other social con- concern the increments. We often cerns, such as economic growth, en- heard the increments described as ergy development, or employmentsupplementary standards applicable in varies, depending on environmental, attainment areas (as the primary and economic, and social conditions. secondary standards apply everyCongress therefore established in the where). The committee concluded that PSD provisions a process for deter- this interpretation is incorrect, if only mining the geographically variable because the states may change the invalue of good air quality, a process that crements applicable to most parcels of it intended the states to implement. land (by reclassifying) subject only to Explicit in the provisions are the the administrative procedures detailed tools to be used in the decision-making in the act. As we shall see, there are process. Land in clean air areas is di- other reasons for rejecting the intervided into three classes according to pretation of increments as standards, the relative importance of maintaining including the compelling one that acexisting air quality. In certain areas, cepting it would defeat the purposeof such as national parks and wilderness striking a balance locally between enareas, existing air quality was deemed vironmental and industrial values. by Congress to be very valuable. These At the heart of the PSD process is areas are designated Class 1; in them, the preconstruction review of major a very small degradation of air quality new stationary sources or modificais considered significant. Figure 1 tions of existing stationary sources. shows the federal mandatory Class I The act defines a major new source as areas. one of a class of 28 types of industrial All other lands in clean air areas facilities that has the potential for were initially designated Class 11, in emitting 100 tons or more per year of which a greater degradation of air a regulated pollutant, or any type of quality is deemed acceptable. State facility that might emit 250 tons or and tribal authorities may redesignate more per year. The potential to emit most lands as being in any of the three has been interpreted by the courts (3) classes. The Northern Cheyenne In- to mean taking any pollution control dian Reservation in Montana has been equipment intoaccount. Thedefinition designated Class I by the tribe and a of a major modification is contained in proposal is pending to redesignate as regulations and depends on the pollu-

tants emitted. The effects of the projected emissions of a proposed facility on air quality are assessed using air quality models. In addition, part of the preconstruction review is the determination of the facility's emission limits. Each major source must use the best available control technology (BACT), which, according to the act, must be determined on a case-by-case basis by taking into account energy, environmental, and economic impacts, as well as other m t s . The emission limitations of BACT must, however, be at least as stringent as those of the New-Source Performance Standards (NSPS) for that type of facility, if such standards exist. Another important tool in the PSD process, the effects-related test, is employed when Class I areas are involved. A PSD permit may be granted or denied on the basis of a demonstration of the effects of the proposed emissions on air-quality-related values, one of which is visibility ( 4 ) . If air quality values are not adversely affected, a permit can be issued, the status of the increments notwithstanding. An application for a waiver of Class I increments based on an effects test at Olympic National Park is pending. It is apparent that since the Class I increments do not establish a final basis for approving or disapproving a permit application ( 5 ) . they cannot accurately be regarded as supplementary standards. Reliance on the states Such reliance to make decisions about the relative importance of exVolume 15. Number 9. September 1981

1001

Maximum allowable increases (increments) in concentrations over the baseline in PSD areas and national ambient air quality standards for SO2 and TSP (jtglms) POIIY1.rd

Partlcler Annual geometric mean

awl

IncrCWII

5

19

CIanUl

37

NAAQS Fllmary -aq

75

2 M o w maximum' Sulfur dloxldn Annual arithmetic mean 2431our maximum' 3-How maximum*

For example, the Class I1 and 111 increments should be general measures of air quality. Based on current knowledge about the statistical behavior of the atmosphere and about air quality modeling, we concluded that this objective can be met with a single increment each for SO2 and TSP in Class I1 and Class 111 areas without changing the effect of the current set of Class 11 and Class 111 increments. That is, instead of Class I 1 and 111 increments for the annual average and for three-hour and 24-hour maximum concentrations, an annual-or seasonal-average would suffice. The committee also found a need to improve the set of Class I increments. Ambient concentrations of fine particles are more closely related to adverse effects of pollution on the more commonly perceived air-quality-related values than are concentrations of SO2 and TSP (8). We therefore concluded that Class I increments for concentrations of fine particles are needed to improve the relationship between preconstruction review and the effects-related test, especially when visibility is of interest. There are circumstances, however, in which S020r TSP is important for specific effects, so we did not recommend deleting any of the existing Class I increments. To enhance the possibility of using the effects-related test, it would be helpful if the responsible federal land managers were to publish descriptions of the air-quality-related values of importance in the respective Class I areas and to give guidance on what effects-related tests might include or how they might be performed. Much scientific work needs to be done in this area.

I isting clean air resources requires a program that is flexible enough to take account of local needs and values. Quite understandably, ambient standards to protect public health and welfare are uniform across the country. But where the air is clean relative to the ambient standards, local conditions may influence the social value of preventing degradation of existing good air quality. If local communities are to make the required judgments, the system must be adaptable to various local conditions. The required state implementation plan, land classification, BACT determination, increment allocation, and effects-related tests all were intended to provide flexibility to the states. To date, however, only six states have implementation plans approved by EPA, while 13 administer all or parts of EPA's regulations (6). In the other states, EPA enforces the federal regulations. N o state has yet reclassified its lands according to its own interests. As a general rule, except where increments have been threatened, BACT has been equated with NSPS (7). A first-come, first-served allocation system is being used everywhere. And, mostly because access to Class I increments is still available almost everywhere, there has been only one attempt to apply an effects-related test. Nonetheless, we concluded that as states become familiar with PSD and appreciate its utility as a resource management tool, the elements of flexibility are more likely to be employed. The committee regarded flexibility as a desirable administrative characteristic for this type of regulatory program: but predictability isalsodesirable in regulation. Industrial firms would usually rather face predictable systems, for example, so that they know what emission limitations they

must meet. Predictability simplifies business decisions, whereas flexibility is more likely to result in accommodation to local conditions. Whether accommodation means more or less stringent protection of existing good air quality or more or fewer opportunities for industrial development depends on the outcomes of the political processes that underlie PSD decision making.

The increment system The commitee concluded that the increment system is as good as any other for achieving the objectives of the act. Both permitting agencies and industry, however, have treated the increments as absolute constraints on degradation of air quality, tantamount to supplementary ambient standards for attainment areas. But the act provides an effects-related test for relief from the Class I increments and reclassification for relief from the Class I 1 increments, if relief is deemed socially and politically desirable. The N A S report discusses the relationships between the SO2 and TSP increments and air-quality-related values, including visibility. The committee concluded that the increments are not substitutes for the effects-related test for Class I areas. Thus, the Class 1 and I1 increments should be viewed only as signals in the process-the former to prompt consideration of actual effects, the latter to trigger the process by which states may make explicit decisions about how clean they want their air to be. Only if reclassifcation proves to be undesirable or impracticable will the Class 11 increments serve as absolute limits to deterioration of air quality. In light of the purposes of the PSD part of the act, the committee concluded that the set of increments shown in Table 1 could be improved.

Modeling

There is no way to estimate the effect on air quality that a proposed new source will have without using air quality models. While the committee found the basic modeling requirements in the statute to be sound, we concluded that models currently being used in the PSD process are inappropriate for many of the applications intended. For example, conservative screening models are being used to determine emission limitations: models designed for short-range studies are being used in situations in which long-range transport is the critical feature; and steady-state models are being used to describe dynamic conditions. The committee said that the use of conservative models to screen permit applications is not inappropriate as a means for reducing administrative Volume 15, Number 9, September 1981 1009

' ...

. - .

-?.

'.

New Mexiropueblos. In Son Juan Ciinl Rosin, a Class I1 a r m

burdens of the preconstruction review process; but when such models are used to determine the availability of increments or to set emission limitations, predictions of consumption of increments will be unrealistic and emission limitations may be more stringent than necessary. The committee also found that the short-term increments (three-hour and 24-hour maximums that may be exceeded only once per year) pose difficult problems for modeling. The short-term maxima are rare events; the results of modeling for them are inherently uncertain. Different models give different results owing to differences in treatment of the details of the meteorology. As general rules, modeling for average values (rather than extremes) gives results that are less deuendent on the model used. and the loiger the averaging period. the closer the agreement among different models. The committee concluded that the problems dswiated with uncertainties in modeling for short-term increments (for examJe, arguments about which model is most appropriate) could be alleviated by changing the criterion for compliance from one based on extreme conditions (one violation per year) to one with a firmer base in statistics. The World Health Organization has suggested, for example, that a number be selected that could be exceeded during 2% of the intervals over a year (9). Other possibilities are 90 or 95% compliance (Le., 36 or 18 days of violation, respectively, of a 24-hour number). Using any of these statistical measures of compliance need not lead to a change in the level of protection afforded by the existing system since 1004 EnvironmentalScience & Tedrmlw

knowledge of the statistical behavior of the atmosphere will suggest what new number (that may be exceeded more frequently) corresponds to the same level of air quality as thecurrent number (that may be exceeded only once). Such a change in the measure of compliance with PSD increments could have far-reaching effects on all air quality programs, since the arguments also apply to modeling shortterm ambient standards. Because of the limits of our study, we were unable to explore all the ramifications. Further study is needed to determine the advantages and disadvantages of alternative measures of compliance, and we strongly recommended that such study be undertaken. Emission limitations One of the more important tools in the PSD program by which states can manage their clean air resources is the determination of BACT. If, as we concluded. PSD was intended to be a Drocess for manaeine the use of a iimited resource, I't makes sense for both the environment and the economy to allow for the flexibility that the BACT determination provides. The value of the air resource varies from place to place, implying that the costs of using that resource should also depend on location. Consider an example in which the availability of mineral or energy resources makes it likely that there will be more intense pressures for development in the future; it makes economic sense to require that each new facility use the minimum amount of the air resource feasible in order to reserve some of the air resource for use by others and thereby to increase the

total opportunity for industrial development and economic growth in the area. Most PSD permits issued to date define BACT as the technology necessary lo meet an applicable NSPS. the least stringent limitation that could be applied under the act. Exceptions have occurred mainly when projected .emissions threatened to cause an increment to be exceeded. If we assume that PSD is intended as a resource management tool, we might be disappointed in this record, because it implies that the resource is beingconsumed at a maximum rate. Decisions, a priori, to substitute NSPS for BACT will result in consumption of the resource at the maximum rate and in minimization of the aggregate opportunity for development. Individual cases must be looked at todetermine whether morestringent limitations are practical. A factor that contributes to the existing record is that most PSD permits have been issued by EPA regional offices in the absence of EPA-approved state implementation plans for PSD. EPA appears to be reluctant to impose more than the minimum limitation and, in some cases, has established BACT for classes of sources rather than doing case-by-case evaluations. There may be circumstances in which time-dependent emission limitations could be incorporated into BACT without harming the environment and with considerable benefit to industrial development. According to the act, BACT limitations must be at least as stringent as applicable NSPS limitations, and NSPS limitations must be based on continuous emission reductions. It is not clear whether BACT limitations may, to the extent that they are more stringent than NSPS limitations, include limitations imposed as a function of meteorological conditions. Intermittent There are circumstances in which intermittent controls might be appropriate for PSD purposes, particularly in the West. These circumstances and the conditions under which using intermittent controls would be consistent with the purpose of the PSD part of the act are described in detail in the NAS report. Briefly, we suggest that intermittent controls may be attractive where emissions from a proposed facility already using the best available continuous control technology still threaten to cause increments to be exceeded in a distant Class I area under infrequent meteorological conditions.

A critical factor in deciding on intermittent controls in each case is the acceptable level of confidence with which specific meteorological conditions can be predicted. It should be realized that use of intermittent controls would increase the overall regional emissions burden, even though they provide protection from effects of plume blight. On the other hand, if a facility could not be built in the most desirable location (for example, at the mouth of a mine) becauseof a Class I increment, but could be built and operated elsewhere in the same airshed with higher costs (due, for example, to having to ship the coal), it might make sense to use additional intermittent controls at the more desirable location.

a substantial number of sources is required. PSD areas are frequently characterized by few sources, some of them with the potential of consuming substantial portions of the available increments. Thus it is not obvious that the desirable properties of competition inherent in economic allocation schemes would consistently emerge in PSD areas.

Conclusion A number of the committee's conclusions are a t odds with those of others who are concerned about the PSD part of the act. The differences, we believe, can be attributed todifferences in understanding of the intent of the PSD part of the act, which we concluded is to establish decision-making processes by which states can decide on Regional haze & acid precipitation the relative importance of maintaining The committee found that the cur- clean air and of competing values. The implementation of PSD has, in rent PSD program is of limited usefulness for addressing regional air- general, not been consistent with the quality problems caused by pollutants intent of the act. Very few states have transported long distances from many approved PSD implementation plans. sources, such as large-scale hazes or Our informal survey indicated that a acid precipitation. Large-scale mul- good many states have been reluctant tistate air quality problems are not yet to prepare implementation plans beamenable to modeling, which cannot cause of a history of frequent legislaseparate out influences of individual tive and judicially mandated changes sources hundred of kilometers away. in the program. For the PSD program to work, the Furthermore, the act contains no effective administrative procedures for states must adopt their own programs; dealing with problems that span many only in this way can they be assured that decisions on the appropriate balstates. PSD also focuses mainly on new ance between maintaining existing sources, whereas for the next one or clean air resources and encouraging two decades the principal sources of industrial development will take into SO?, which contributes to these prob- account local concerns and values. lems through its transformation Another reason that PSD does not products, will be poorly controlled appear to have been implemented in a manner consistent with its intended existing sources rather than well-controlled new sources. An alternative goal is that very few cases have arisen approach to regional air-quality in which increments have been conproblems is establishment of limita- sumed; thus, both applicants and pertions on total regional emissions. mitting agencies are faced with the question of what to d o next. It may be Alternatives that only after the cases now pending have been analyzed will a full appreAlternatives to the current regulatory system for implementing PSD ciation of the PSD provisions be obhave been proposed, but the committee tained. found that none appears to offer clear Clearly, even with major changes in advantages in practical applications. the PSD provisions, the time will come Among the alternatives most often when the political processes set forth discussed are schemes employing in the act will result in decisions that foreclose further industrial developeconomic incentives rather than adment in specific locations. The air reministrative regulations. Emission taxes and marketable emission permits source is, after all, finite. The PSD are two prominent examples (10). program provides a means for alloStudies of the advantages and appli- cating that resource since there is no cability of these types of systems have practicable economic market for doing been conducted primarily for situa- so. I n this light, the current PSD protions that apply in nonattainment gram appears to perform two overridareas and therefore may be of limited ing functions. One is to ensure that the use in deciding the applicability of states determine how much of the reeconomic incentives to PSD. source should be allocated. The second For most incentive systems to work, is to provide means, through decisions

on siting and emission limitations, for making possible the greatest amount of industrial development for a given degradation of air quality. References (I)Sierra Club et al. v. Ruckelshaus [344 F. Supp. 253 (D.D.C.)] was affirmed on a peal

[Env. Rep.Cas. 1815(D.C.Cir. 1972)rand again affirmed by an equally divided Supreme Court 1412 U S . 541 (1973)l. (2) Sections IM) through 169A of the Clean Air Act(42 USC 7470). enacted in the Clean Air Act Amendments of 1977 (P.L. 95-95, 91 Stat. 685). (3) Alabama Power v. Castle. 606 F.2d 1068 (D.C.Cir. 1979). (4) Visibility is the onlv air-qualitv-related value explicitly drscribid m the aci(Scction 169Al. Federal land managers haw an affirmatibe responribilil) to proleit airqual. it)-rcldied baIuc\ ISwtion I 6 5 id)(2)(B,I Scc45FdXrgirr 15016,Aprd IO. 1918.for a listing and dbcripuon of awyualil)-rclated ,d"eS

( 5 ) Mwkic. E. (bnx. RPC. 1977. June 8 . 91hY (61 S l s l a with approved PSD implementation plans as of Nabember 1980 arc Mainc. Vcrmont.Cmrgia.Tcnncnct.. honh Dakuta,and H p m i n g Statrr uilhaulhurit) tucnfarccall or parts of the lcdrral PSD regulations arc Alabama. Florida. Kentuck). Wow4ppi. North Carolina. South Carulina. Illinots. Indiana. Michigan. Mmncwt3. Ohio. WIE. cunvn. and hebraska (71 "Stud) of PSI) and OlTsct Pcrmiu": Dams and Moore S c ~ l t l e .IYXO. (n) "Protecting vl\lbliltY. ~n EPA ~ c p o nIO CongreK. C S I:PA. Research Trisnglc Park. h C.. 1979, "Sulfur Oxide". Corn. mitleenn Sulfur Oxides. Udtional Acadcmy of Scicnccr Washington. D.C.. 197R: " K k I r a l s An Envirunmcnlal A-menl". Panel un \ilr.at~s. \ationdl .\cadmy ofScicnco: Washington. r)(: , 191X (9) " 4 1 r Qualit) Crwrra for Urban Air PolIuldnt~".Uu. 5Ob. Horld Hwlth Organimlion: (icncba. 1972. ( I O ) Russell. C.S. fnuiron. Sci Technot. 1981,

IS. 24-ZR

(II) "On Prcucnuon of Significant Deterioration 01 Air Qualit)"; Uational Acadcm) P m r : Waihinpton. D.C.. 17 50 pcr copy

Myron F. Unnn (left) is senior staflofficer and associate executive secretary of the Environmental Studies Board of the N a tional Academy of Sciences, National Academy of Engineering. He has been with N A S since 1973. H e received his Ph.D. in electrical engineering from Princeton University in I968. John T. Middleton (right) is a private consultant who was thefirst administrator of NAPCA (National Air Pollufion Control Administration), a predecessor organization of EPA. H e consultsfor the United Nations and has served as a U . N . represenlafive in Malaysia. Yugoslavia, the Philippines, and Thailand. Volume 15. Number 9. September 1981

IO05