Theodore Roosevelt Administrative History

Degradation of the air quality over the park is the most significant threat to park resources.

One does not often find such a sweeping declaration in the park management literature, but there it is, in a prominent place in the draft of Theodore Roosevelt's Natural Resources Management Plan. Nor are any punches pulled when it comes to identifying the cause: "The air quality within the park is being threatened by increasing development outside."

A more straightforward assessment of the predicament could not be given. Declining air quality, says the plan, is the park's "number one resource management problem" because air "is the most important, yet the most difficult to manage" natural resource. [1]

As we have seen, air pollution has decidedly affected the aesthetic quality of the park. This chapter will discuss its effects on ambient (i.e., overall) air quality, whose degradation is a physical threat to resources ranging from vegetation to historic structures. The many and subtle consequences of air pollution are what make it so hard to mitigate and why it rightly heads the list of Theodore Roosevelt's priorities.

Monitoring

The most remarkable aspect of the air quality problem is how suddenly it arose. Before the mid-1970s the park's air was "uncompromised," substantially the same as it was during the life of Roosevelt. When the Clean Air Act of 1970 created nationwide ambient air quality standards, it had no immediate effect on the park—all the standards were already being easily met. [2]

The park began monitoring air quality just at the start of North Dakota's latest oil boom. In 1974 a total suspended particulates (TSP) collector was installed in the mixpit maintenance yard of the South Unit, near Peaceful Valley Ranch. A private company working under an Environmental Protection Agency contract ran the unit for the first year; it has since been operated by the state. Monitoring did not really hit stride, though, until the height of the boom in 1979; indeed, the staff believes that any data gathered before then did not register the influence of all the new gas and oil wells. [3] By 1982 air quality monitoring was one of the park's major resource management activities. Aside from sampling particulates, Service personnel now analyze the acidity of atmospheric deposition and take teleradiometer readings. NPS also cooperates with the State Department of Health in its operation of monitoring equipment in the park (including measuring ozone levels in the North Unit). Also, as a result of a temporary cooperative venture between NPS and the Bureau of Land Management, Forest Service, and Fish and Wildlife Service, the state of North Dakota now keeps track of hydrogen sulfide (H₂S) levels using a permanent station located in the Lone Butte Oil Field east of the North Unit. Finally, under the auspices of the Theodore Roosevelt Nature and History Association, the Park Service sponsored a study of lichens as indicators of changes in air quality. They are the plant species most sensitive to sulfur dioxide and should, therefore, be the first to exhibit adverse effects. [4]

The Service does not yet have enough information about recent changes to be able to predict future trends in air quality. Each monitoring project is currently at a different stage, producing a like effect in terms of the staff's knowledge of each pollutant.

Since they were the first to be studied, more is known about trends in particulates than the others. Even so, the park now has only ten years of data from the South Unit and five from the North Unit, since a TSP unit was not installed there until 1979. In that same year the University of California at Davis received a contract from the EPA to put in a fine particulate sampler at the South Unit mixpit and two high volume samplers (one each in the North and South units) to obtain a breakdown of TSP. These aerosol samplers are part of the Western Fine Particle Monitoring Network. Particles trapped in them are laboratory-tested for suspended sulfate, nitrite, and fine sulfur. [5]

During the boom years of 1979 to 1981 the park's overall TSP level stayed below legal standards (refer to Appendix E) but rose each year. Still, individual pollutants, such as suspended sulfates and nitrites, exceeded or very nearly approached their state-set limits. [6] Scientists studying these findings for the State Department of Health theorized that the cause might have been abnormally dry weather, but "industrial emissions . . . associated with the 1,500 megawatt expansion of coal-fired electrical generating capacity in west-central North Dakota from 1979 to 1981 also may have contributed to these recent concentrations." [7] Yet at the same time, sulfur dioxide, nitrogen dioxide, and nitric oxide levels "were generally less than the detection limits of the analysis instruments." [8]

If the information on TSP can be said to indicate no general trend in ambient air quality, the same also seems to be true of what is known about changes in the acidity of air over the park. Theodore Roosevelt's acid deposition monitoring equipment, situated in the North Unit, is part of the National Atmospheric Deposition Program, which is designed to chart changes in the acidity of air over all the United States by means of a rigorous, systematic sampling and analysis procedure. (The first annual summary of NADP data—for the year 1982—only became available in 1985.) All NADP sites operate in the same way. Wet- and dry-fall are monitored using separate catchments, a recorder of discrete instances of precipitation, conductivity and pH meters, and a weight scale. Deposition is collected at regular intervals and sent to a central laboratory for analysis. Results are returned to the park and to NPS's Water Resources Field Support Laboratory so that baseline "normal" levels of acidity can be determined. [9]

No such baseline has yet been postulated, but preliminary indications from other sources have recently become available. In 1982, Clifford M. Wetmore, a scientist with the University of Minnesota, was engaged to do a lichen flora of all three units of the park and to make some tentative judgments on how the Park Service could use natural sources to monitor air quality. Wetmore, who had done a similar study for Voyageurs National Park, completed his research at Theodore Roosevelt in 1983. The next year, permanent photographic plots of the rarest lichen species were established to monitor their status annually, and a year later permanent plots of aspens were made to keep track of the park's ozone levels, since these trees are sensitive to ozone. Also in 1985, the United States Geological Survey completed a baseline study of toxic heavy metals in the park's soil and vegetation. In all three of these projects encouraging results were returned: Wetmore reported that in the course of his work he noticed little deterioration in the physical condition of the lichens, even though he found many species which are known to be particularly sensitive to acid deposition; no foliar injury to the aspen trees was found in 1985, indicating low ozone levels; and baseline amounts for toxic heavy metals were all within normal ranges, with no elevated sulfur levels. [10]

Acid deposition and TSP are but two of NPS's concerns at Theodore Roosevelt. In the absence of a "normal" baseline it cannot be said with certainty how much of a problem they have been or will be. However, an impartial observer would concede that overall air quality in the park has worsened over the past decade. The question for policy makers, though, is not merely how much of a decline has occurred, but how much of a decline is acceptable. For this a legal baseline has been established: the Class I Prevention of Significant Deterioration standards of 1977.

Prevention of Significant Deterioration: PSD

The Clean Air Act Amendments of 1977 introduced a system to classify the nation's lands which was designed to help maintain the air quality associated with different locales. This Prevention of Significant Deterioration program divided up those areas of the United States already meeting both the national secondary ambient standards (standards set at a level thought adequate to prevent detrimental effects to property, the environment, or other components of the public welfare) and the stricter primary ambient standards (which are meant to protect public health) into three groups: classes I, II, and III. Each class represents a different level of acceptable deterioration of ambient air quality. Class I areas have the strictest PSD standards; in other words, the quality of the air over these areas is allowed to deteriorate the least of the three.

It must be understood that the PSD classes are not, strictly speaking, indicative of an area's current air quality. The classification is merely intended to assist in preventing deterioration of existing air quality, what ever its current state might be. Improvement of existing air quality is addressed by individual state implementation plans and not by PSD. However, because of the nature of the classification system it is true that many Class I areas already have good ambient air quality as compared with most of the country. This is certainly true of Theodore Roosevelt, which, like many other sizable NPS units outside of urban areas, was designated Class I. [11]

PSD standards are important because they provide, at least in theory, a means by which the Park Service can protect air quality (and values related to air quality) in units of the System. To help regional offices and individual parks in pursuit of this goal, NPS set up an Air and Water Quality Division in Washington, D. C. The division is involved not only in PSD matters, but in every aspect of the subject as it pertains to national parks. [12]

Each Class I area has a "base level" of air quality and an "increment." The base level is variously interpreted as the ambient quality of the air at the time of the first PSD application, or as the quality prevailing under perfectly "natural" conditions. [13] Obviously the two are radically different and the interpretation greatly affects how much overall pollution will be allowed in a Class I area. The increment is simpler and written right into the law: it is the amount of allowable deterioration to the base level air quality. Increments can be exhausted by new sources of pollution. For example, if a new power plant opens near a Class I area and puts out sulfur dioxide in quantities equal to the increment, the PSD increment for sulfur dioxide is said to be "consumed" and no other sources emitting sulfur dioxide would be allowed to locate near the Class I area unless they obtained a special certification from the federal government (for which see below).

Two things should be noted. First, the Environmental Protection Agency has defined increments only for sulfur dioxide and particulates. Second, the PSD regulations apply only to "new sources" of pollution—that is, to emissions from facilities for which construction started after January 6, 1975.

The federal government does not always handle the entire PSD process. Some states have been given charge of implementing the law, but only if they have developed programs which are both acceptable to the Environmental Protection Agency and are at least as stringent as the federal PSD regulations. North Dakota is such a state, with the Department of Health responsible for enforcing PSD regulations. It performs the computer modeling estimation of the amount of new pollution, estimates which decide whether the increment has been consumed. It also evaluates privately-done analyses. [14]

Variances

In the jargon of the Clean Air Act, the Secretary of the Interior is referred to by the generic term "federal land manager." According to the law, the federal land manager (whose responsibilities in the Theodore Roosevelt case were assumed by Interior's assistant secretary for fish, wildlife, and parks) "has an affirmative responsibility to protect the air quality related values . . . of a Class I area and to consider . . . whether a proposed major emitting facility will have an adverse impact on such values." [15] As this statement implies, there is a way for a PSD increment to be legally exceeded. If the company proposing a new facility near a Class I area can demonstrate to the satisfaction of the federal land manager that the pollution produced in excess of the amount allowed by the increment will have no unacceptable adverse impact on air quality related values in the park, a "notice of certification" will be issued to the state. If the state accepts this certification of no unacceptable adverse impact, it may in turn issue a "variance," a waiver of the Class I PSD regulations so that the company can go ahead and build the facility. Even should the federal land manager find that a new facility would produce an unacceptably adverse effect, it is possible for the company to obtain a variance by convincing the state governor—or, if the federal land manager still objects, the president—to overrule. If the decision reaches the chief executive, the president's decision is final and cannot be reviewed in court. [16]

Variances are a major loophole in the PSD regulations. It is possible for the Service's recommendations on new polluting sources to be overridden on any of three levels: by the Department of the Interior, the state, or the chief executive. Theodore Roosevelt has been involved in controversy over variances from the time it was named a Class I area. Much of the park's struggle to maintain its ambient air quality turns on this crucial process.

The variance controversy arose quickly because the Class I PSD increment for sulfur dioxide was consumed quickly. In 1977 five pending new-source energy facilities were evaluated for their effect on sulfur dioxide levels in Theodore Roosevelt. The increment for sulfur dioxide, as measured in 24-hour concentrations, is five micrograms per cubic meter. Therefore, the combined daily output of sulfur dioxide from these five facilities could not exceed five micrograms per cubic meter. Each was assigned a place in the increment in the order in which its permit application was received. The UPA/CPA coal conversion plant was the first to submit its application to the state, so its contribution of 1.4 micrograms per cubic meter found a place in the increment. The results for all five facilities are summarized in Table 6.1.

The entire sulfur dioxide increment was immediately consumed by these new sources. [17] This did not sit well with energy developers wanting to process the coal, oil, and natural gas of the state, for unless the existing point sources relocated or cut back their emissions so others could find a place in the increment, or variances were granted, no new facilities could be built. The former being most unlikely, developers devoted their efforts to obtaining variances.

The consumption of the increment soon had the attention of the state's journalists. Any proposed energy plants which were to be located within an imaginary triangle linking the North and South units and the Coal Creek plant near Underwood (Figure 6.1) were considered subject to the PSD new source increment. Determined by an approximation of the path pollutants would take if the wind blew steadily from Coal Creek to the park, the triangle was considered part of Theodore Roosevelt's airshed. It was dubbed the "Smokestack Triangle" by journalists, and became to the general public a readily understandable, if somewhat inaccurate, symbol of the PSD variance process. [18]

Being in the middle of a situation portrayed as a battle between economic development and the environment was not a pleasing prospect to the park staff. By 1979 they were noting with dismay the number of editorials appearing in the region's small newspapers urging a "balance" between the use of mineral resources and preservation of air quality. "There will be a confrontation sometime soon about the Class I air over the park," wrote Superintendent Harvey Wickware the next year. "We will deepen our 'bad guy' image because of our mandated position [to protect Theodore Roosevelt's ambient air quality]. There will be hard times for us in North Dakota in this regard." [19]

He was only too right. By the end of 1980, eight construction applications for the Smokestack Triangle had been received. The applicants all contended that the computer models the State Department of Health had used in 1977 to predict sulfur dioxide emissions from the original permittees were outdated and should be replaced by their own, newer models. The state's response was encouraging: "If the newer long-range models can provide a more realistic prediction of changes in air quality and can show that increment is available beyond the 1977 prediction, then some energy development will be allowed." [20] Of the eight projects announced for the Smokestack Triangle, six went all the way through the variance process.

Models and meteorology

The models referred to are computer simulations of pollutant dispersion; PSD decision-making depends on being able to predict where airborne pollutants will go, and in what concentrations, after they leave a given smokestack. Models are necessary because the numerical increments are too small to be reliably measured in the field. But modeling is imprecise because of the assumptions necessitated by complicated atmospheric processes. That fact contributed to the controversy in North Dakota, for the way the computer analysis was performed—what data were chosen, which formulas were used, what assumptions were made—profoundly influenced the results. Models were developed for each of the six facilities, and each was somewhat blighted by technical inadequacies.

In 1980, soon after their applications were filed, the six companies submitted their own models to the State Department of Health. [21] The results were difficult to judge because of the variety of models used. The Department reviewed them, held a public hearing to assess the adequacy of the proposed techniques, and after a thorough survey chose from the applicants's models the ones it thought best suited for making uniform estimates. The Environmental Protection Agency, while not approving for general use the state's choice of models, found its basic approach sound and legally defensible, and mostly concurred in its findings of fact. [22]

The Department of Health's re-evaluation of its 1977 methodology originally was to have used a generalized screening model to select "critical scenarios," with these then being run through a more refined model. However, the Department revised the screening model so much (partly in response to comments made at the public hearing) that it never got to the refined-model stage. Park Service officials attended the assessment hearing and met with the Department to review the estimates of air quality produced by the revised screening model. For the most part NPS also could not fault the state's modeling approach. [23]

Most computer models are based on Gaussian plume dispersion theory, which combines engineering and climatic data to produce results. The importance of accurate weather statistics to the efficacy of these models cannot be overstated. [24] The Department of Health was criticized for using only one year of meteorological data gathered in 1964 from a station at Bismarck, well over a hundred miles from the park. The state justified the chosen data base by asserting that the weather for Bismarck and the badlands was generally the same; they also pointed out that 1964 was the last year in which hour-by-hour readings were taken. [25] But one analyst argued that the state's models could be run using every third hour of surface readings, thus enabling the use of a broader data base—the years 1971 to 1975—to better ascertain the frequency of occurrence of meteorological conditions. [26]

Another problem with the computer models was their applicability. Through 1979 estimates of pollution effects on Theodore Roosevelt were prepared using Gaussian steady-state models. The EPA recommends their use only to a radius of fifty kilometers from the emitting source, but the park was farther away than that from some of the six proposed facilities. The Department of Health knew this when it did its original analyses, but had no choice: the Clean Air Act Amendments of 1977 required that the amount of consumption of the PSD increments be determined as soon as possible, and at the time no larger-scale models had been adequately tested. [27] It was an unusual instance of legislation outdistancing theory.

By 1980 so-called "mesoscale" (medium- to long-range) models had been developed, so the companies had a point when they demanded that the analyses be redone. Still, the new mesoscale models did not account for ruggedness of terrain any better than the steady-state ones. Computer models assume that the land between the polluting source and the receptor is perfectly flat. Introducing variations in terrain alters both the predicted dispersion and concentration of the pollutants. One would think that, of all places, North Dakota would fit the models' assumption best. But the terrain of Theodore Roosevelt National Park is certainly not flat, and this fact could not be compensated for in the mesoscale modeling. [28]

Once the data was inserted into the models, it could be analyzed in any number of ways, with the results subject to varying interpretations and reasonable differences of opinion. At one point the Environmental Defense Fund charged the Park Service, whose Air and Water Quality Division did the technical review of the computer models, with making a mathematical error in the equations used to predict deterioration of visibility. Using higher coefficients, the Environmental Defense Fund maintained that if the variances were granted to all the applicants, the new sources of pollution would reduce visibility seventeen to thirty-four days a year rather than the six predicted by NPS. [29] Other discrepancies and questionable computations can be found. [30]

Finally, the expense of doing the computer simulations prevented them from being as complete as they perhaps should have been. For example, the Department of Health did not always perform a comprehensive examination of particulate levels because of "the relatively low TSP emission rates of the sources involved, and the somewhat less restrictive Class I increments for TSP. Further," it added, "the computer costs of modeling would substantially increase with a detailed TSP analysis." [31] Instead, it modeled a worst-case TSP episode for the North Unit, and found it to be well within the increment for 24-hour concentrations. The Department did not analyze average annual TSP concentrations. [32]

The technical review and waiver

The Department of Health's mesoscale modeling indicated that, if all six plants were built, sulfur dioxide concentrations higher than the 3-hour increment in the North and South units and much higher than the 24-hour increment in all three units would occur. So in 1982 the state informed five of the companies that they would have to apply to the federal land manager for a variance. The sixth, Phillips Petroleum, did not need a variance because its predicted emissions were below the "significance level" as subsequently set forth in state government regulations implementing the Clean Air Act Amendments of 1977. But in the technical review of the variance applications and the state's modeling procedures, NPS's Air and Water Quality Division (acting as proxy for the federal land manager) went ahead and evaluated all six of the projects. [33]

The technical review took four parts: A BACT (best available control technology) analysis; an evaluation of the state's s air quality modeling; a study of the new source's effects on values associated with air quality; and a summary of findings.

Each company had to demonstrate that it would use the best available combinations of equipment to control emissions. One of the applicants, Warren Petroleum, had to redo its initial BACT because the Air and Water Quality Division found alternative sets of equipment that would bring better sulfur recovery. [34]

The evaluation of the Department of Health's findings on air quality focused on three individual pollutants and visibility. NPS concurred with the state and found that, if anything, it had overestimated sulfur dioxide concentrations. The Service also agreed with the state's decision not to estimate average annual concentrations of TSP because "it is obvious from the results of the 24-hour screening analysis that the annual average Class I increment for particulate matter (5 micrograms per cubic meter) will not be exceeded with the introduction of the new sources. Also, concentration estimates for Elkhorn Ranch were not made because of the extremely low estimates for the other units and because there is no reason (based on the SO₂ analysis) to expect higher particulate matter concentrations at Elkhorn Ranch." Even though expected concentrations of fluoride, the third pollutant considered in the technical review, were below the government's significance level and so not subject to PSD regulations, the Air and Water Quality Division made an estimate since fluoride is so toxic to vegetation. [35]

Each applicant met the Environmental Protection Agency's criteria for plume perceptibility; that is, considered separately, none of the sources were expected to cause a smoke plume visible from Theodore Roosevelt. Still, NPS decided that a study of the cumulative effect of these new sources on visibility should be made. The division estimated the reduction in visibility with the help of the teleradiometer and fine particulate data gathered from the South and North units. They predicted that permitting the new sources would cause less than a two percent reduction in the park's standard visual range. [36]

The Air and Water Quality Division also evaluated the effects of the proposed projects on "air quality related values"—species of flora and fauna, cultural resources, soil types, and surface waters sensitive to air pollution. Five out of the six applications contained no discussion of these effects and the sixth's was inadequate.

Of the park's flora, lichens were the species thought most at risk. In the North Unit, sulfur dioxide had already reached concentrations which elsewhere had been detrimental to lichens. In early June 1982, NPS personnel made a field trip to the park to look for symptoms of injury to sensitive vegetation. No sulfur dioxide damage to lichens was found, nor was ozone damage seen in five indicator species: ponderosa pine, clematis, globemallow, alfalfa, and milkweed. They were examined because the presence of sulfur dioxide lowers the ozone concentration threshold for injuring some plants.

The Air and Water Quality Division's fluoride concentration estimates were not high enough to ascribe an adverse effect to vegetation or insects. But the technical review did note that fluoride cannot be metabolized and accumulates until toxic levels are reached; therefore, perennial plants and herbivores who feed on them would be at risk from long periods of exposure to low levels of it. The division's review could not predict what effect increased TSP levels might have on birds at Theodore Roosevelt, nor could it do more than remark on the possibility that fish and amphibians might be hurt by increased acid deposition. [37]

The technical review admitted to "considerable uncertainty" in its determination of the effects of the new polluting sources, enumerating a long list of cautions. [38] After all was said and done, though, the Air and Water Quality Division strongly implied that even if all the permits were denied, concentrations of the three pollutants were already high enough to produce some ill effects in the park's ecosystem, although none had been observed. "In fact," the reviewers wrote, "the five applicants contribute relatively small percentages of the maximum predicted concentrations" and "none of the probable effects are expected to occur on a scale large enough to irreversibly change ecosystems functions and structures or to affect visitor use and enjoyment." With this finding of no unacceptable adverse impact, the division, acting for the federal land manager, recommended that the state approve all the applications for PSD variances. [39]

The NPS technical review did not go uncriticized. The National Parks and Conservation Association thought the visibility and vegetation analyses were not good enough. [40] During the public comment period, many substantial criticisms (fifty pages in all) were made by the Dacotah Chapter of the Sierra Club, the Dakota Resource Council, the Technical Information Project, and the Environmental Defense Fund. Some of them were effectively rebutted by the Service; others, less so. [41]

Many of these criticisms were specifically prompted by the preliminary finding of no unacceptable adverse impact, which had been announced on July 12, 1982. In late August the public comment period was reopened briefly upon the receipt of regional air quality statistics showing a sharp increase in wintertime sulfur dioxide levels over the first six months of 1982. The Air and Water Quality Division subsequently issued a supplementary technical review (with an expanded discussion of visibility); in it the Service found no reason to change its preliminary finding. [42] On September 20, a final determination of no unacceptable adverse impact was published in the Federal Register. The way was cleared for the state to grant variances to the projects. [43]

The real effects of Theodore Roosevelt's PSD experience will probably not be felt for a few years. "At the current time the State and the Service are on the same wavelength on this issue," noted an in-park report in late 1982. "With the Service certifying no adverse impact from the six major sources most of the weight of the decision is now on the State Health Department. The day of reckoning is not far off, however. The first and subsequent denial of such certification on the part of the 'Federal Land Manager' will bring great weight to bear upon the Service by state officials and the leadership of the major industries involved. Major litigation is all but a foregone conclusion." [44]

Coal leasing in the Fort Union Region

The "day of reckoning" may come if and when the Bureau of Land Management completes the long-delayed leasing of federally-owned coal in the Fort Union Coal Region, an administrative appellation encompassing eastern Montana and western North Dakota. BLM estimates that about 1.6 billion tons of accessible coal underlie the Region. Originally, BLM planned to offer twenty-four tracts for lease (Figure 6.2).

While most of the Fort Union coal would be shipped out for burning, some would probably remain to fuel existing and proposed power plants in the Northern Plains. [45] This presents a problem to NPS because some of these facilities would be west of Theodore Roosevelt, with the prevailing winds bringing their pollutants directly to the park. All the new sources involved in the variance process discussed above are east of Theodore Roosevelt—still within its airshed, but likely to produce only episodic degradation of the park's ambient air quality. Hence, the federal land manager's certification of no unacceptable adverse impact was, in their cases, a relatively straight forward decision. Granting variances to new sources west of the park would be quite another matter.

BLM's own Fort Union regional environmental impact statement predicted that, no matter which leasing alternative was approved, TSP concentrations would go beyond state and federal standards, the sulfur dioxide increment for Theodore Roosevelt would be considerably exceeded, and visibility within the park would decline. BLM anticipated that companies interested in leasing would voluntarily comply with the Class I PSD regulations to gain access to the federal tracts. [46]

In early 1983, as the Fort Union leasing date neared (it has since been delayed), environmentalists charged that the Park Service, given its recent recommendation of a no unacceptable adverse impact certification, was preparing to make Theodore Roosevelt a "sacrifice area" by not opposing the BLM leasing program. This the Service denied, and reported to the newspapers its official concern about the BLM plan. [47] NPS was particularly worried about three tracts close to the park. Perhaps responding to that concern, in June 1983 the secretary of the interior withdrew two of them (North Wibaux-Beach and Zenith) from the lease sale and made leasing of the third (South Wibaux-Beach) subject to special stipulations requiring the lessor to protect Theodore Roosevelt's air quality. [48]

The most difficult resource to manage

Air quality management is the most complicated, and the most important, resource management question facing the park. The air over Theodore Roosevelt is still relatively clean but how much it will be allowed to deteriorate is the essence of the controversy.

As we have seen, the PSD regulations, which are meant to limit deterioration, are imperfect. Variances are not the only loophole. The park staff suspects that most of the hydrogen sulfide sour gas coming into the park comes from aggregate oil well flaring, from individual wells which are not yet considered "point sources" and therefore not subject to PSD regulations. They also put out considerable amounts of other pollutants: a study of 235 wells near the park showed that from July 1981 through June 1982 they emitted about 20,000 tons of sulfur dioxide. [49] Yet these data only suggest another complication: the variance process can actually work to the betterment of the park's ambient air quality. Three of the six applicants—Phillips, Warren, and Amoco—wanted to build treatment plants to process natural gas which otherwise would be flared or vented into the atmosphere from individual wells. Even though gas treatment plants themselves produce pollution, their construction could cause an overall reduction in hydrogen sulfide and sulfur dioxide emissions from non-point sources around Theodore Roosevelt as more and more wells hook up with pipelines connected to the treatment plants. [50]

The oil boom of the 1970s and early 1980s is now over, and as the rush to develop new fields slackens there are indications that some citizens in North Dakota are beginning to share the park staff's concern about ambient air quality. [51]

Public support for the park's position is of course essential if NPS is to keep air quality degradation to a minimum. Gaining that support requires demonstrating to the public how quickly the problem has arisen and how severe it can just as quickly become. After all, as recently as 1976 the park's own superintendent could rather nonchalantly report that, while state-sponsored air quality research "is not directly related to the park, some information, valuable to the park, will be derived." [52]

Then again, the park's air was considerably cleaner in 1976 than it is today.