United States v. Louisiana-Pacific Corp. , 682 F. Supp. 1141 ( 1988 )

682 F.Supp. 1141 (1988)

UNITED STATES of America, Plaintiff,
v.
LOUISIANA-PACIFIC CORPORATION, a Delaware corporation, Defendant.

Civ. A. No. 86-A-1880.

United States District Court, D. Colorado.

March 22, 1988.

Daniel S. Maus, Asst. U.S. Atty., Teresa N. Lukas, Asst. Regional Counsel, U.S.E. P.A., Denver, Colo., Cyrus S. Picken, Jr., Environmental Enforcement Section, Land and Natural Resources Div., U.S. Dept. of Justice, Washington, D.C., for U.S.

William A. Hillhouse, II, Zach C. Miller, Felicity Hannay, Davis, Graham & Stubbs, Denver, Colo., for Louisiana-Pacific Corp.

FINDINGS OF FACT AND CONCLUSIONS OF LAW

ARRAJ, District Judge.

This is a civil enforcement action brought by the United States of America, as plaintiff, on behalf of the U.S. Environmental Protection Agency ("EPA") for violations of the Clean Air Act, 42 U.S.C. § 7401, et seq., and the regulations promulgated thereunder concerning the prevention of significant deterioration ("PSD")^[1] of air quality by the defendant, Louisiana-Pacific Corporation ("LPC"). Plaintiff seeks to enjoin defendant from further alleged violations of the PSD regulations, which are set forth at 40 C.F.R. § 52.21. Additionally, plaintiff seeks the assessment of civil penalties against LPC for alleged violations of these regulations.

The case was tried to the court on January 19 through 22, and January 25 and 26, 1988. Written closing arguments were submitted by the parties, and oral closing argument was heard on February 17, 1988. Having heard the testimony and arguments, and having reviewed the voluminous transcripts and exhibits, I find that the matter is ripe for disposition. The following shall constitute the court's findings of fact and conclusions of law in conformance with Fed.R.Civ.P. 52(a).

I. BACKGROUND: THE PSD PROGRAM

The Clean Air Act establishes minimum air quality standards to be achieved in all regions of the country. In 1977, Congress amended the Act to establish a program for the "prevention of significant deterioration ("PSD") of air quality." The PSD statutes and regulations are designed to protect areas of the country where the air is relatively clean. The goal of the program is to prevent the air quality in areas where it exceeds the statutory minimum from degenerating to that level.

To achieve this result, areas of the country where the air is cleaner than required by the National Ambient Air Quality Standards are identified by the states and designated as "attainment areas." 42 U.S.C. §§ 7407, 7471 (1983). The attainment areas are further divided into three classes: Class I for areas that have very clean air (such as national parks) where little or no deterioration is permitted; Class II for areas where moderate deterioration of air quality may occur; and Class III for areas where more economic growth and resulting air quality deterioration is allowed. Id. §§ 7472, 7474. The thrust of the PSD program is that new "major emitting facilities" may not be constructed within these areas before certain permits have been obtained. Id. § 7475. The permits, in turn, allow the new facility to contribute to air pollution only up to specified incremental amounts. Id. § 7473(b). Of central importance to this case is the fact that LPC's Kremmling and Olathe facilities are located within attainment areas.

The Clean Air Act provides that "[n]o major emitting facility ... may be constructed in any [attainment area] unless a permit has been issued for such proposed facility in accordance with this part setting forth emission limitations for such facility...." 42 U.S.C. § 7475(a)(1) (emphasis added). The Act further provides that the term "major emitting facility" includes any source with the potential to emit 250 tons per year (TPY) or more of any air pollutant. Id. § 7479(1).

The PSD regulations go into more detail and establish the rule that no "major stationary source" or "major modification" of a major stationary source "shall begin actual construction without a permit" which states that the source or modification will meet the emission requirements set forth in the regulations. 40 C.F.R. § 52.21(i) (1983). The term "major stationary source" is defined to include any facility which emits, or has the potential to emit, 250 TPY of any air pollutant. Id. § 52.21(b)(1)(i)(b). A "major modification" is defined as any physical change or change in operation that would result in a significant increase in the emission of any one of several pollutants. Id. §§ 52.21(b)(2)(i), 52.21(b)(23). With regard to the pollutants that are relevant in the present case, a net emissions increase of 100 TPY of carbon monoxide (CO) or 40 TPY of volatile organic compounds (VOCs) would be significant, and thereby constitute a major modification. Id.

Permits may be issued only to sources that satisfy two principal requirements. First, the source must demonstrate that emissions from the construction or operation of the facility will not violate any applicable emission standard of the act. 42 U.S.C. § 7475(a)(3). Second, the proposed source must be subject to the best available pollution control technology. Id. § 7475(a)(4). To facilitate its review, the EPA requires that new sources submit air monitoring information necessary to determine the impact on air quality of the proposed source. 40 C.F.R. § 52.21(m). Generally, such monitoring must be gathered one year in advance of submission of the PSD application. The EPA then has up to one year to review and grant or deny the application. 42 U.S.C. § 7475(c). As a result, it may take up to two years before the source is allowed to commence actual construction of the new facility.

Where the EPA determines that the provisions of the Clean Air Act and its implementing regulations have not been complied with, it may issue a notice of violation ("NOV") to the alleged offender. 42 U.S. C. § 7413(a)(1). If the alleged violation continues for more than 30 days after the issuance of the NOV, the EPA is then empowered to bring a civil enforcement action. Id. § 7413(b)(2). If a violation is established, the Act authorizes the court to issue a temporary or permanent injunction, or to assess a civil penalty of up to $25,000 per day of violation, or both. Id.

II. FINDINGS OF FACT

Defendant LPC came to Colorado in 1983, with the encouragement of the state government, to establish the industry of waferwood manufacturing.^[2] Since that time, LPC has built two waferwood plants in Colorado, the first in Kremmling, and the other near the town of Olathe. The air pollution emissions from these two plants, and the failure by LPC to obtain PSD permits from the EPA, form the basis of the present litigation.

A. "Waferwood"

In order to fully appreciate the issues before the court in this case, it is necessary to have some familiarity with the process by which LPC's Kremmling and Olathe facilities turn aspen and pine logs into "waferboard." First, when the logs are ready to be processed, they are cut by a saw into lengths of about eight feet. Once cut, the logs are moved into pools of heated water, called "hot ponds," to condition the bark for removal.^[3] From the hot ponds, the logs go to the "debarker" which, not surprisingly, is a machine that removes the bark. After the bark is removed, the logs move on to the "slasher," which cuts the logs into three-foot pieces, and then to the "waferizer," which chops these pieces into one-and-a-half to three-inch chips, or "wafers." The wafers then go to storage bins.

From the storage bins, the wafers go to the "wafer dryer," which is a machine that combusts wood and sawdust to produce a heated "exhaust gas." The hot exhaust gas is brought into direct contact with the wood chips and thereby dries them. The chips are blown by the exhaust gas into a cyclone which, using principles of centrifugal force, separates the dried wood chips from the exhaust gas. The dried wafers then move on to a "screening" process where they are separated into two different sizes and stored.

Once the chips have been screened, they move from the storage bins to a "blender," where they are mixed with adhesives and waxes for the forming process. The chips are then laid on a mat, with larger chips on the top and bottom and smaller chips in between. The material on the mat is split by a "cross-cut saw" into sections measuring eight feet by sixteen feet. These sections are then loaded into the "press," which heats and compresses the material into "waferboard." From the press, the sections of waferboard are trimmed and cut into sheets measuring four feet by eight feet by the "trim saw." These fourby-eight sheets of waferboard are the final product.

The process just described creates air emissions in a number of ways. First, wet bark and sawdust from the slasher and debarker are combusted in a device known as a "Konus" thermal oil heater to generate much of the heat required by the plant. The main purpose of the Konus is to provide heat to the presses by means of a hot oil system, which is similar to a boiler system. The heat from the Konus is used to heat oil which, in turn, transfers that heat to the presses. A secondary purpose of the Konus is to supply heat to the hot ponds. Finally, heat from the Konus is also used to heat the building itself in the wintertime. The emissions generated by the Konus include carbon monoxide ("CO") and volatile organic compounds ("VOCs"), as well as particulates, from the complete and incomplete combustion of the wet bark and wood that is used as fuel for the device.

Particulate emissions from the Konus combustion process are removed from the exhaust gas in two ways. First, the gas is blown into a "cyclone," which is a cylindrical device that causes the exhaust to rotate around in it. As a result of the rotation, solid material in the gas stream is thrown to the side of the device and is collected. Second, the gas exiting the cyclone is blown into a "baghouse." A baghouse is a pollution control device that operates in much the same way as a household vacuum cleaner. It consists of several fabric bags through which the exhaust is blown. The fabric catches particulate matter as the gas passes through.

In addition to the Konus, the wafer dryer process creates a second source of air emissions. As with the Konus, the combustion process again creates CO, VOCs, and particulate emissions. Additionally, when the wood chips are heated and dried in this fashion, natural resins are released from the wood.

As noted above, exhaust gas from the combustion of wood and sawdust is blown, along with the wood wafers being dried, to a primary cyclone where the wafers are separated from the gas. The exhaust gas continues on from the primary cyclone to a number of smaller cyclones operating at a higher velocity which remove more particulate matter from the gas stream. Under the original design, the gas exiting the smaller cyclones was vented directly to a stack. Subsequently, however, LPC added an additional pollution control device, known as an "electrified filter bed" ("EFB"), to remove more particulates from the exhaust.

The presses give rise to a third source of emissions. VOCs result at this point as the heat and pressure from this process release more of the natural resins from the wood. These emissions are exhausted through the "press vents." Finally, the various saws make up a fourth source of emissions, since they generate sawdust which must be controlled.

B. The State Permits

LPC applied to the Colorado Air Pollution Control Division (APCD) in June of 1983 to obtain air emission permits for the Kremmling plant. The application requested permits for four emission sources: the Konus hot oil heater, the wafer dryer, the crosscut saw, and the trim saw. In October of 1983, LPC submitted a similar application for the Olathe plant. LPC then commenced on-site construction at Kremmling and Olathe in July and November of 1983, respectively.

In January of 1984, the Colorado APCD issued four air emission permits for the four emission sources at Kremmling referenced in LPC's application. These permits contained restrictions on the amount of fuel that could be combusted and on the amount of waferboard that could be produced by each source. The wafer dryer permit restricted that source to 20,000 tons per year of wood fuel and 93,000 tons per year of production. The permit for the Konus limited the annual fuel input for that device to 19,000 tons of bark and wood. Finally, the two permits for the saws limited production to 49,950 four-by-eight foot sheets of waferboard per year.

In February of 1984, the APCD received comments from the Colorado State Council of Carpenters to the effect that the public notices issued for the Kremmling and Olathe facilities failed to contain any information concerning formaldehyde emissions. As a result, the APCD requested information from LPC concerning the possibility that formaldehyde was being emitted from the press vents. LPC responded to this request on March 8, 1984, by supplying the APCD with the data from one of four previous press vent tests it had conducted at its waferboard plant in Hayward, Wisconsin. These four tests were conducted in September of 1981, May of 1983, July of 1983, and the early part of 1984. LPC sent the APCD the preliminary results of the 1984 test as soon as they were available. While these test results were the most recent and current, they also showed the lowest emission rates.^[4]

In addition to supplying this test data, LPC invited the APCD officer who had made the inquiry, Mr. Abe Vasquez, to observe another test of formaldehyde emissions from the press vents at the Hayward, Wisconsin plant. Vasquez accepted, and the test was conducted in May of 1984. LPC subsequently applied for a permit for the Kremmling press vents in October of 1984, and such a permit was issued by the APCD in April of 1985. This permit limited waferboard production to a maximum of 49,950 tons per year and 160 tons per day.

In September of 1984, the APCD issued five air emission permits for the Olathe plant. Four of these five permits were for the four emission points referenced in LPC's application, and the fifth was issued for the Olathe press vents. These permits contained combustion and production limitations similar to those issued for the Kremmling plant. Specifically, the wafer dryer was restricted to 20,000 tons per year of wood fuel and 80,127 tons per year of production, the Konus was restricted to 19,000 tons per year of bark and wood fuel, and the crosscut and trim saws, as well as the press vents, were limited to 49,950 tons of production annually and 160 tons of production daily. Revised permits for the Olathe Konus and the Olathe dryer were issued in May of 1985.

The APCD informed LPC by letter in June of 1985 of its intention to revoke the wafer dryer permits for both Kremmling and Olathe on the ground that LPC had violated certain conditions of the permits relating to opacity. A hearing on this matter was held before the Air Quality Control Commission on September 5, 1985, and by written order (dated September 23, 1985, nunc pro tunc September 5, 1985) the Commission ruled that the Kremmling dryer permit would be revoked effective October 15, 1985, and that the Olathe dryer permit would be revoked effective November 15, 1985. The order further provided, however, that LPC could continue to operate the plants if it obtained new dryer permits by these dates. The purpose of the order was to give LPC some additional time to install electronic filter beds ("EFBs") to further control emissions from the dryers. LPC did install EFBs in the fall of 1985, and opacity tests were subsequently performed which indicated compliance. As a result, replacement permits for the dryers were issued in October and November of 1985. These permits contained various restrictions on emissions and output, the amounts of which were determined "based on" 8000 hours per year of operation.

The APCD again in early 1986 informed LPC of its intention to revoke the same wafer dryer permits, as well as the permit for the Konus heater at Olathe. As with the 1985 revocations, however, LPC appealed this action to the Air Quality Control Commission, and the revocation decisions were stayed pending a hearing before the Commission.

Subsequently, LPC and the Commission entered into a settlement agreement to resolve the problem. The settlement set forth a number of improvements and modifications for the air pollution control system, and provided that the decision to revoke would be withdrawn if LPC made all of the specified improvements and modifications. After a hearing was held on December 18, 1986, the Commission issued its order, dated January 6, 1987, finding that LPC had "complied in all respects with the terms and conditions of the Settlement Agreement," and ordering that the "suspended decisions" revoking the permits in question were vacated in all respects.

The most restrictive limitation^[5] contained in the state emission permits issued for Kremmling and Olathe limited annual production at both facilities to 49,950 tons of waferboard per year.^[6] Taking into account the weight of a sheet of waferboard that measures three-eighths of an inch in thickness, undisputed expert testimony established that the mathematical equivalent of 49,950 tons is roughly 90 million square feet on a three-eighths inch basis. While LPC kept production within this amount in 1985 and prior years, production exceeded this permit limitation in 1986 and 1987. Specifically, production in square feet at Olathe and Kremmling amounted to 105 million and 106 million in 1986, and 124 million and 94 million (through November) in 1987, respectively.

Desiring to increase production at Kremmling and Olathe beyond the limits on production contained in the original permits, LPC applied to the APCD for new permits allowing increased production. Revised permits limiting production to 78,216 tons per year were issued for all five of the emission sources at Kremmling in July of 1987. Revised permits for the Olathe plant had not been issued as of the time of trial.

C. The PSD Permits

It is undisputed that the LPC had not submitted PSD permit applications for either of its Colorado waferboard plants to the EPA prior to initiating construction and operation of these facilities. At the time of trial, LPC had submitted PSD permit applications, but actual PSD permits for Kremmling and Olathe had not been issued.

In September of 1983, Mr. Steven Frey of the United States EPA was driving to an inspection when he stopped to visit the Kremmling construction site. Frey stopped because he noticed a large amount of smoke being emitted from a "wigwam burner" at the site. Frey visited the Kremmling operation a second time in December of 1984 because he was aware that the APCD had been conducting frequent inspections of the facility. Frey informed LPC at or around the time of this second visit that the wigwam burner probably constituted a "major stationary source" of air emissions as that term is defined in the PSD regulations. As a result, the new waferboard plant could be considered a "major modification" of the wigwam burner, and could therefore be in violation of the PSD program.

A "wigwam burner" is a tepee-shaped incinerator used to burn wood waste from a sawmill. Such a wigwam burner and a sawmill were already in existence at the Kremmling plant site when the property was purchased by LPC in 1982. A permit which allowed emissions of 500 TPY of CO from the wigwam burner was transferred to LPC in August of 1983. As a result of Frey's warning, LPC quickly closed operation of the wigwam burner and, by June 4, 1985, it had completely dismantled and removed that facility.

In December of 1984, Robert Jorgenson of the Colorado APCD sent a letter to LPC requesting that air emission tests (or "stack tests") be performed at the Kremmling and Olathe plants. The division required test data for a number of pollutants, including CO and VOCs. LPC accepted bids from a number of companies specializing in this kind of testing and recommended by the APCD. After reviewing the bids, LPC selected Interpoll, Inc. to conduct the tests, and scheduled them for March of 1985.

Alex Slivinsky was hired by LPC in January of 1985 and given direct responsibility for the stack testing to be done in March of 1985. Interestingly, he had no previous experience in air emissions testing. Similarly, Jorgenson, who had a background in wildlife biology and public administration when he was hired by the APCD in 1984, had never observed an emissions test for CO prior to the March, 1985 tests at Kremmling and Olathe. Slivinsky and Jorgenson worked together to prepare the protocol^[7] for the March, 1985 emissions test.

Jorgenson and Slivinsky ran into some confusion in preparing the protocol for the Konus heater test. Although Jorgenson had no previous experience with the design of the Konus and did not review the specifications for the device, he did learn from an informational brochure that the Konus could generate a maximum heat output of 28 million BTU. As a result, in preparing the protocol, and in administering the test at Olathe,^[8] Jorgenson insisted that the Konus be operated to provide this maximum heat output.

An undisputed fact of critical importance, established by the testimony of numerous expert and lay witnesses, is that the Konus is designed to match heat output with heat demand. As noted above, the sources which demand heat from the Konus include the press (hot oil system), the hot ponds, and the building itself. A thermostat within the Konus works to operate an automatic fuel feed system. When heat demand exceeds heat output, fuel will automatically be added. When heat output and demand are approximately equal, or output exceeds demand, the system will automatically stop supplying fuel. Aditionally, if the fire gets too hot, a second system will automatically turn off the fans which supply the air for the combustion, and the fire will smolder. The purpose behind these automatic systems is to achieve maximum combustion and heat output with the smallest amount of fuel.

The emissions test for the Konus heater at Olathe was performed on March 12, 1985.^[9] Although he tried, Slivinsky was never able to generate the maximum heat output called for in the protocol for a number of reasons. First, the fuel was fed not automatically, but rather at a pre-calculated rate. By estimating the amount of BTUs that a fixed amount of fuel would generate, Jorgenson and Slivinsky had hoped to be able to create 28 million BTUs by pouring in a pre-calculated amount of fuel. Unfortunately, the fuel created a greater amount of heat than had been estimated. Second, even though the hot ponds, the press, and the building had been allowed to cool the night before the test, and even though the building heat was turned up to maximum and hot ponds were heated to a temperature forty percent higher than normal operations, these sources did not generate a large enough heat demand. These two facts, combined with the fact that the Konus will not generate more heat than required, worked together to create a cycle of problems.

As too much fuel was fed in, and because the heat demand was too low, the system would overheat and the fans would shut down. With the air supply cut off, the fire would "smolder" rather than "burn."^[10] Once the smoldering caused the unit to cool down, more fuel would be added to what was already too much, smothering what little fire there was.^[11] When the fire got to burning again, the cycle would repeat itself. The ultimate result of this operation was that fuel was fed into the Konus in "lumps," rather than continuously, and that the fuel primarily "smoldered," rather than "burned."

The Konus heater at the Kremmling facility was tested the following week on March 19, 1985. As a result of the problems experienced at Olathe, Slivinsky arranged with Jorgenson to operate the Konus differently. Specifically, although Slivinsky still pre-calculated the amount of fuel to be burned, he calculated a lower fuel-feed setting. The result was that the amount of heat created more closely matched the heat demand, and the Konus therefore operated continuously, and at a relatively stable rate, throughout the test. Using significantly less fuel, the device actually generated more heat than at Olathe, and the plant as a whole was able to operate (that is, produce waferboard) for a greater percentage of the testing time. It is important to note that the representatives of the EPA and the APCD who testified at trial did not consider any of the Kremmling test results to be incorrect or misleading.

The test results processed by Interpoll and returned to LPC indicated that CO emissions were three times greater at Olathe than they were at Kremmling. This discrepancy is due to the fact that CO is a product of incomplete combustion. Since there was so much more incomplete combustion associated with the Olathe test, it naturally follows that the CO emissions there would be greater.

Steven Frey of the EPA reviewed the March, 1985 stack test results and used them to calculate the potential to emit various pollutants from the two plants. Using this data, he concluded that the Olathe facility had the potential to emit more than 250 TPY of CO, and therefore constituted a "major stationary source" of air emissions (as that term is defined in the PSD regulations). Similarly, Frey calculated that the Kremmling facility had the potential to emit more than 100 TPY of VOCs, and therefore qualified as a "major modification" of the wigwam burner. Accordingly, the EPA issued two Notices of Violation ("NOVs") to this effect on June 5, 1985.

Frey's original calculations did not take into account any of the restrictions on operation contained in the state permits. Rather, his original figures are based on the assumption that the Kremmling and Olathe plants could operate at an unrestricted 8760 hours per year. Accordingly, he combined this figure and the March emission data from Olathe to calculate that the Olathe plant had the potential to emit 437.9 TPY of CO. Similarly, he used the March data from Kremmling and EPA Method 25 to conclude that the Kremmling plant had the potential to emit 265.0 TPY of VOCs. These calculations formed the basis for the issuance of the June, 1985 NOVs.

After comparing the results of the March stack tests at Kremmling and Olathe, and considering Slivinsky's report on the different methods of operation at each facility, LPC concluded that the test data for the Olathe Konus was inaccurate because the unit was not operated as designed. LPC contacted the EPA and the APCD to explain this conclusion. It informed both agencies of its decision to retest the Olathe Konus in June, and invited both agencies to attend. Jorgenson accepted the invitation and attended for the APCD. Frey responded that the maximum capacity of the Konus could not be tested in the relatively warm month of June. As a result, he stated that the June test results would have no effect on his conclusion and that he would not be in attendance.

LPC did retest emissions from the Konus at Olathe in June of 1985. The fuel feed was operated in the automatic mode, and, as with the test at Kremmling, heat output was matched with heat demand. Predictably, the emission rate for CO was drastically lower than the March test at Olathe and similar to the emission rate measured at Kremmling.

On July 10, 1985, representatives of LPC and the EPA held a conference to discuss the NOVs that were issued the previous month. At this conference, Frey explained the reasoning behind the EPA's position that the plants were in violation of the PSD regulatory scheme. In response, Slivinsky explained why LPC felt that no violation had occurred. With respect to Olathe, Slivinsky explained that the March stack tests were unreliable because the plant would never actually be operated so badly that the Konus fire would go out. Addressing the EPA's concern that maximum heat demand could not be tested in June, Slivinsky offered to retest the Konus the following winter. With respect to Kremmling, Slivinsky informed the EPA that the wigwam burner, the alleged major stationary source, had been dismantled.

At this conference, Frey was informed by LPC that the restrictions in the state permits effectively limited the plants to 8000 hours of operation per year.^[12] Applying this limitation to the data from the March stack tests at Olathe, he calculated that the Olathe plant had the potential to emit 399 TPY of CO. Similarly, the Kremmling data, when applied to this limitation, indicated that the Kremmling facility had the potential to emit 242.1 TPY of VOCs measured in accordance with EPA Method 25.

Upon learning that the wigwam burner had been dismantled before the NOVs were ever issued, the position of the EPA gradually became that the Kremmling facility constituted a major source in its own right. At this point, unconvinced that Method 25 was the appropriate method for measuring VOCs in the PSD context,^[13] Frey recalculated the potential to emit VOCs at Kremmling using a new and unpublished methodology that he conceived and that he felt was preferable. The basic difference between the two methods is that under Method 25, VOCs are expressed as carbon, but under Frey's method, VOCs are expressed as formaldehyde. Since the molecular weight of formaldehyde is greater than the atomic weight of carbon, Frey's method results in a greater VOC emission rate than Method 25. Using his new method, Frey calculated the potential to emit VOCs at Kremmling to be 293.5 TPY for 8760 hours of operation and 265.3 TPY for 8000 hours of operation.

Although nothing transpired at the July conference to change LPC's position that its Kremmling and Olathe facilities were not subject to the PSD program, it decided after this meeting to apply for PSD permits anyway. This decision represented both an attempt to satisfy the EPA and a realization that a significant expansion of these operations in the future might really trigger the PSD program. Before any such applications were ever submitted, the EPA issued an administrative order to LPC on September 27, 1985. The order directed LPC to submit a PSD permit application for its Olathe waferboard facility within 60 days of the effective date of the order. The order stated that it would become effective 15 days after its issuance. However, in a display of the efficiency for which the public sector is so famous, the order was neither signed nor dated when it was issued.

One of the components of a complete PSD application is air "monitoring" data.^[14] Since this requirement can be waived by the administrator,^[15] LPC requested such a waiver from the EPA on November 7, 1985. Although only the Olathe plant was subject to the administrative order, LPC asked the EPA to consider a waiver for both Kremmling and Olathe because the plants were so similar and because it was preparing to submit applications for both plants. EPA responded to LPC's request in the negative on December 3, 1985, but the response only addressed the Olathe plant. As a result, Slivinsky continued to wait for a response which addressed the Kremmling plant. When it appeared that such a response would not be forthcoming, he submitted PSD applications for both facilities on January 15, 1986.

At the time these applications were submitted, a state implementation plan ("SIP") for Colorado had not yet been approved by the EPA.^[16] Accordingly, the EPA was responsible for the administration of the PSD program in Colorado, and any application for a PSD permit submitted during this period should have been submitted to the EPA. Nonetheless, under the terms of an "interim agreement" between the EPA and the Colorado APCD, the substantive review of the application was performed by the APCD. Thus, when a PSD permit was submitted to the EPA, it was shortly forwarded to the APCD for review.

Aware of this procedure, Slivinsky submitted the PSD permit applications, contrary to the directions in the administrative order, directly to Jim Geier of the APCD. Slivinsky left a message with the APCD that Geier should contact him if the latter had any questions or if there were any problems with what was submitted. Shortly after receiving the application, Geier conferred with Frey over the fact that the PSD applications had been submitted. Neither Frey nor Geier made any attempt, either by cover letter or phone call, to inform LPC that the applications had been submitted to the wrong agency. LPC was informed of the problem by way of a letter from EPA's regional counsel, on March 25, 1986.

LPC hired Mr. Charles Bray in February of 1986 as a consultant to assist LPC in the PSD permitting process for the Kremmling and Olathe facilities. Bray reviewed the data from the stack tests that had been conducted in March and June of 1985 and used these test results to calculate the Kremmling and Olathe plants' potential to emit various pollutants. In contrast to Frey's conclusions, however, Bray's calculations indicated that the Olathe facility did not have the potential to emit 250 TPY of CO, and that the Kremmling facility did not have the potential to emit 250 TPY of VOCs. In short, Bray's calculations indicated that neither of LPC's Colorado facilities was a major stationary source of air emissions subject to the PSD program.

The different conclusions reached by Frey and Bray can be explained by the fact that Bray's calculations differ from Frey's in a couple of important respects. First, with regard to Olathe, Bray used the CO emission data from the June, 1985 test rather than the March, 1985 test. Bray believed it would be inappropriate to use the March results because the Konus heater was operated at that test in a manner contrary to its design. Second, with regard to both facilities, Bray concluded that the most restrictive permit limitation was the annual limit on production of 49,950 TPY which is contained in the original saw and drier permits. Frey (it will be recalled) used a limit of 8000 hours per year of operation. Third, Bray used Method 25 (rather than Frey's new method) to calculate VOC emissions.

Applying the permit limitation on annual tons of production, Bray concluded that the potential of the Kremmling plant to emit VOCs was 193.7 TPY under Method 25. Although he believed that Method 25 was the proper methodology to employ in calculating the weight of VOC emissions, he also calculated the potential to emit VOCs at Kremmling to be 216 TPY using Frey's new and unpublished methodology. Using the test results of the June, 1985 stack test, and applying the permit limitation on tons of production, Bray calculated that the potential to emit CO at the Olathe plant was 196 TPY. He noted that if he had used the results of the March stack test at Kremmling (instead of the data from the June test at Olathe) that the potential of the Olathe plant to emit CO would have been lower still (by about ten percent).

After reviewing LPC's original PSD permit applications, the EPA noted a number of deficiencies. In response to the agency's complaint that the applications did not contain a "complete" monitoring plan, Bray submitted revised monitoring plans for both plants in June of 1986. In an effort to address the other deficiencies, LPC submitted revised PSD applications to the EPA in July and August of 1986 for the Olathe and Kremmling facilities, respectively. In September of 1986, EPA informed LPC that the revised monitoring plan was also deficient, and, in October of 1986, EPA informed LPC of a number of problems with the second set of PSD permit applications. Yet another monitoring plan was submitted by LPC in April of 1987, and a third set of PSD applications (which EPA has since found to be complete) were received by EPA in July of 1987. PSD permits for the two facilities had not been issued as of the time of trial.

D. Procedural Posture of the Case

The United States filed its complaint in this case on September 12, 1986. The complaint contained two claims for relief. The first claim alleged that the Kremmling facility constituted a "major modification" of the pre-existing wigwam burner, and the second alleged that the Olathe plant itself was a "major stationary source." These claims charged that the plants were in violation of the PSD program because they were constructed and were being operated in the absence of PSD permits.

On February 3, 1987, the EPA issued yet another NOV to LPC alleging this time that the Kremmling plant constituted a "major stationary source" in its own right. The United States then moved to amend its complaint to add a first claim for relief in the alternative based on the violation alleged in the 1987 NOV. The government also sought to add a third claim for relief based on LPC's failure to comply with the administrative order issued in September of 1985. This motion to amend was granted. The first claim for relief was dismissed by Memorandum Opinion and Order of this court dated October 30, 1987, and the third claim for relief was dismissed on defendant's motion at trial made at the close of plaintiff's case-in-chief.

As a result of these rulings, only the first claim for relief in the alternative and the second claim for relief remain for resolution. The narrow questions they present are whether the Olathe plant had the potential to emit 250 TPY of CO, and whether the Kremmling plant had the potential to emit 250 TPY of VOCs. While these issues might at first appear to present questions of fact, their resolution actually turns on the legal construction of the term "potential to emit."

III. CONCLUSIONS OF LAW

A. The Thirty Day Notice Provision of 42 U.S.C. § 7413

42 U.S.C. § 7413(a)(1) provides as follows:

Whenever, on the basis of any information available to him, the Administrator finds that any person is in violation of any requirement of an applicable state implementation plan, the Administrator shall notify the person in violation of the plan ... of such finding. If such violation extends beyond the 30th day after the date of the Administrator's notification, the Administrator ... may bring a civil action in accordance with subsection (b) of this section.

42 U.S.C. § 7413(a)(1) (1983) (emphasis added). Subsection (b), in turn, empowers the EPA to bring a civil enforcement action for an injunction, or civil penalty, or both, whenever the owner of a major stationary source "violates any requirement of an applicable implementation plan ... more than 30 days after having been notified by the Administrator under subsection (a)(1) of this section of a finding that such person is violating such requirement." Id. § 7413(b)(2) (emphasis added).

These provisions make it clear that, in enacting the PSD program, Congress envisioned a system where, before the EPA has jurisdiction to bring a civil enforcement action, (1) the source which is allegedly in violation must be notified by the EPA of the violation, and (2) the source must disregard the warning and persist in the alleged violation for 30 days. The EPA is empowered to bring such a civil suit only on the basis of the specific violation alleged in the NOV and only where that specific violation has continued for 30 days. United States v. Louisiana-Pacific Corp., 682 F.Supp. 1122, 1128 (D.Colo.1987) (hereinafter Memorandum Opinion). As a result, not every violation of the PSD provisions is actionable, but only those where the alleged offender is notified of the violation and persists in the violation for 30 days thereafter. Id. at 1129.

A primary legal question raised in this case and which must necessarily be resolved at the outset is the proper construction of the 30 day period referred to in 42 U.S.C. § 7413. Defendant contends that this provision should be given the narrowest possible construction. It argues that in considering whether the 30 day requirement is met, the court must look only to the 30 day period immediately following the issuance of the NOV. It urges that any other events transpiring after this period are irrelevant. Thus, if the facilities in question became major stationary sources (the specific violation alleged in the NOVs at issue) 31 days after the NOVs issued, and this violation continued thereafter, LPC would contend that such a violation is not actionable because it began more than 30 days after the notice was issued. If the EPA wished to bring an action on this violation, the argument goes, then it would have to issue a second NOV alleging the same violation and wait another 30 days.

Applied to the facts of the present case, LPC urges that because the NOV for the Olathe plant (which alleged that the facility had the potential to emit 250 TPY of CO and was therefore a major stationary source) was issued on June 5, 1985, this court should only consider whether this facility had the potential to emit 250 TPY of CO between June 5, 1985, and July 5, 1985. Similarly, since the NOV for the Kremmling facility (which alleged that the facility was a major stationary source because it had the potential to emit 250 TPY of VOCs) issued on February 3, 1985, LPC would have the court narrow its inquiry to whether the Kremmling plant had the potential to emit 250 TPY of VOCs between February 3, 1985, and March 5, 1985.

Plaintiff argues for a broader construction of the 30 day requirement. It asserts that this jurisdictional prerequisite exists solely for the purpose of giving the source fair warning of the problem and a reasonable period of time to clean up its act. Thus, in contrast to the position taken by LPC, the government urges that this jurisdictional requirement has been met if the source commits the specific violation alleged in the NOV anytime after the 30 day grace period has run.

I conclude, again,^[17] that the latter construction now being urged by the government is indeed the correct one. The Clean Air Act taken as a whole, and a plain reading of its provisions, both clearly indicate that, in enacting the notice requirement at issue, Congress' intention was to give an alleged source a brief period of time within which to evaluate its options before the substantial penalties available under the act could become a possibility.^[18] It did not intend to create a jurisdictional technicality that could be abused to prevent even the most reckless and chronic polluter from being brought to trial.

Where a source is truly in violation, the PSD program is designed to allow and encourage the source to correct the problem. To further this goal, the provision being considered should be construed in such a way as to create an incentive for the source to permanently correct the problem, not merely to correct it for 30 days. To achieve this permanent correction, the EPA's power to enforce the violation alleged in the NOV must be ongoing rather than extending merely for 30 days.

Were this court to accept the construction being urged by LPC, it would create a loophole in the enforcement scheme large enough to swallow the entire PSD program. Under such a construction, an irresponsible source could chronically and even intentionally avoid the PSD program by temporarily correcting the violation alleged near the end of the 30 day period. After that period had passed, the source could return to business as usual and continue to operate in violation until the next NOV was issued. In light of the fact that one NOV is sufficient to put a source on notice, I fail to see what possible purpose could be served by forcing the EPA to continually issue identical NOVs to the same offender.

In sum, the jurisdictional requirement of 42 U.S.C. § 7413 has been met if the source commits the specific violation alleged in the NOV anytime after the 30 day grace period has run. Therefore, in the case at hand, if the EPA can show that the Olathe facility had the potential to emit 250 TPY of CO anytime after July 5, 1985, it has shown a violation of the PSD program actionable under 42 U.S.C. § 7413. Similarly, if it can prove that the Kremmling plant had the potential to emit 250 TPY of VOCs anytime after March 5, 1987, it has made out an actionable violation. This ruling does nothing to increase the exposure to liability of a source that, upon receiving notice of a violation, does what is necessary to meet its responsibilities to society by pursuing a policy of permanently complying with the law. Rather, the practical effect of this holding extends only to sources who would take advantage of a perceived technicality in the law and whose long term strategy and policy is to continue to violate the Clean Air Act even after having been warned.

B. "Potential to Emit"^[19]

The PSD regulations define the term "potential to emit" as follows:

"Potential to emit" means the maximum capacity of a stationary source to emit a pollutant under its physical and operational design. Any physical or operational limitation on the capacity of the source to emit a pollutant, including air pollution control equipment and restrictions on hours of operation or on the type or amount of material combusted, stored, or processed, shall be treated as part of its design if the limitation or the effect it would have on emissions is federally enforceable."

40 C.F.R. § 52.21(b)(4). In order to resolve the seemingly narrow issues of the potential to emit VOCs and CO at Kremmling and Olathe, it is necessary to grapple with some perplexing (and as yet unanswered)^[20] legal questions raised by this definition and the unique facts of this case. First, what is meant by the "maximum capacity" of a source to emit a pollutant under its "physical and operational design"? Second, are the operational limitations contained in the state permits "federally enforceable"? Third, if they are, which of the several permit restrictions should be used in the calculation? Fourth, should such operational limitations be included in the calculation of a source's potential to emit even where such restrictions are routinely and knowingly violated? The court will now address each of these issues in turn.

1. "Maximum Design Capacity"

LPC argues that the results obtained from the March, 1985 test of the Konus heater should not be used in the calculation of potential to emit. LPC urges the court that it would be inappropriate to use such data because the concept of potential to emit clearly contemplates the unit being operated as designed, and that the Olathe Konus was operated contrary to its design at the test in March of 1985. The government responds that the March, 1985 Olathe data is acceptable because the term "potential to emit" really means the maximum emissions that a source can possibly generate, regardless of whether it is being operated as designed. The government argues that, even though the operation of the Konus at this test may have been incorrect, it was still possible to operate the unit in this way, and that this data is therefore useful for determining the maximum emissions the source can generate. For the several reasons that follow, I find the government's position on this issue untenable, and hold that the concept of potential to emit refers to the maximum emissions a source can generate when being operated within the constraints of its design.

The PSD regulations themselves define the potential to emit as the maximum capacity of a source to emit pollutants under its physical and operational design. 40 C.F.R. § 52.21(b)(4). The plain reading of this language indicates that test data must meet two requirements before it may properly be used in the calculation of a source's potential to emit. First, the unit being tested must be operated during the test in the manner in which it is designed to be operated. Second, within that constraint, the unit must be operated at maximum capacity, or "full throttle," throughout the test.

Any analysis of the definition of "potential to emit" must include a reference to the case of Alabama Power Co. v. Costle, 636 F.2d 323 (D.C.Cir.1979) because the current definition above was promulgated in response to the D.C. Circuit's holding in that case. The reasoning in the Alabama Power opinion indicates that the government's construction should not be accepted. At the time this case was before the D.C. Circuit, the EPA by regulation defined potential to emit as referring to the projected emissions of a source when operating at full capacity, with the projection increased by hypothesizing the absence of air pollution control equipment designed into the source. Id. at 363. The court rejected such an interpretation, and remanded the regulations to the EPA with instructions to the agency to include the effect of in-place control equipment in defining potential to emit. Id. at 355. Holding that potential to emit refers to a facility's "design capacity," the court reasoned that since air pollution control equipment was part of the overall design of the source, it must be considered in the calculation of potential to emit. Id. at 353.

The broad holding of Alabama Power is that potential to emit does not refer to the maximum emissions that can be generated by a source hypothesizing the worst conceivable operation. Rather, the concept contemplates the maximum emissions that can be generated while operating the source as it is intended to be operated and as it is normally operated. Of course, it is possible that a source could be operated without the control equipment designed into it or that a Konus heater could be operated so badly that the fire would go out. Yet, Alabama Power stands for the proposition that hypothesizing the worst possible emissions from the worst possible operation is the wrong way to calculate potential to emit.

Additionally, it serves no legitimate purpose to test the emissions from a source when that source is being operated in a way it would never be operated in actual practice. Such data is valueless unless EPA's purpose is to require every source in attainment areas to be subject to the PSD program. It is clear, however, that this was not Congress' intention, since it expressly exempted small sources.

The government makes much of the fact that it is theoretically possible to operate the Konus in the manner that was done at the March, 1985 test at Olathe, and that it was even possible to operate the plant (produce waferboard) when the Konus was being misused in this way. While this statement may be correct, this argument fails to meet the court's concern that any emission data gathered during such operation would be valueless. For example, it makes as much sense to add so much fuel to the Konus that the fire goes out as it does to fuel the unit (which is designed to accept wet bark and sawdust) with coal. Certainly it might be possible to do both, and the unit might even generate sufficient heat to produce waferboard. Yet, either course of action would be contrary to the unit's design, and neither would yield any useful emissions data.

In the present case, there can be no doubt that the Konus at Olathe was operated during the March, 1985 emissions test in a manner contrary to its design. First, it is uncontroverted that the Konus is designed to match heat output with heat demand, whether the unit is operated in the automatic or semi-automatic mode, and that this was not done at the test. Second, the Konus is designed to generate heat by way of complete combustion, but the fire primarily smoldered, rather than burned, during the test in question.

Moreover, the manner in which the Konus was operated during the March, 1985 test at Olathe would never occur during normal operations. First, the function of the Konus is to generate heat. The testimony was uncontroverted (and common sense would also indicate) that, in light of this purpose, the Konus would never be operated so badly that the fire would actually be smothered. Second, the Konus is designed to be fuel efficient, generating the greatest amount of heat or power from the least amount of fuel. Since resorting to outside sources for fuel would be an expense to the business, the realities of a competitive marketplace suggest that LPC would act to conserve its internal fuel supply by operating the unit fuel-efficiently as it is designed. Third, and perhaps most important, the fuel-feed setting was pre-calculated to provide an amount of fuel that would generate 28 million BTU. Although the unit was often run on semi-automatic, this kind of fuel feed setting would never occur in actual practice because (even allowing the plant to cool for a full winter night, and heating the hot ponds to temperatures forty percent above normal) the Olathe facility will simply never generate that great a heat demand.

In sum, the results of the March, 1985 test of the Konus heater at Olathe cannot be used to properly calculate the potential of that source to emit CO because during that test the device was operated in a manner contrary to its design and in a manner that would never occur in normal operations. The government's only evidence that the potential to emit CO at Olathe exceeded 250 TPY consisted of Frey's calculations, all of which were based on data from the March, 1985 test at Olathe. Since (for the reasons expressed above) this evidence is unreliable, and in light of the fact that the CO results from the Kremmling test were unchallenged by the government and were so radically different from the Olathe CO data, I find the government's evidence on this matter unpersuasive.^[21] Accordingly, since plaintiff has failed to carry its burden of proof, the second claim for relief will be dismissed.

2. "Federally Enforceable" Restrictions

A crucial aspect of LPC's defense in the present case is its assertion that the operational limitations contained in the state emission permits must be considered in calculating the potential of the Kremmling plant to emit VOCs. With regard to such restrictions, the PSD regulations provide that any operational limitation to which a source is subject, including "restrictions on hours of operation or on the type or amount of material combusted, stored, or processed," should be taken into account in determining the source's potential to emit, but only if the limitation or the effect it would have on emissions is "federally enforceable." 40 C.F.R. § 52.21(b)(4). In the present case, since the permit limitation upon which LPC chiefly relies (an annual limitation on the amount of waferboard which may be produced) is clearly a restriction on the amount of material processed, it should indeed be included in the calculation of potential to emit if it is "federally enforceable."

The PSD regulations provide that the term "federally enforceable" refers to all limitations and conditions which are enforceable by the EPA. 40 C.F.R. § 52.21(b)(17). The term is broadly defined to include any requirement or limitation contained in or created pursuant to any SIP, whether it be a SIP to enforce the national ambient standards or a SIP to enforce the PSD program. Additionally, the term embraces any requirements or limitations imposed to enforce new source performance standards or created pursuant to a new source review process. Id.

Caselaw confirms the proposition that restrictions on emissions imposed by a state in or pursuant to its SIP are federally enforceable. In the leading case of Union Electric Co. v. EPA, 515 F.2d 206, 211 (8th Cir.1975), aff'd, 427 U.S. 246, 96 S.Ct. 2518, 49 L.Ed.2d 474 (1976), reh'g denied, 429 U.S. 873, 97 S.Ct. 189, 50 L.Ed.2d 154 (1976), the court held that the requirements of an EPA-approved SIP "have the force and effect of federal law and may be enforced by the [EPA] in federal courts." Accord Friends of the Earth v. Carey, 535 F.2d 165, 171 n. 6 (2d Cir.1976), cert. denied, 434 U.S. 902, 98 S.Ct. 296, 54 L.Ed.2d 198 (1977). Even state-adopted emission limitations which are more stringent than necessary to meet the federal ambient air standards are federally enforceable. Friends of the Earth v. Potomac Electric Power Co., 419 F.Supp. 528, 533 (D.D.C. 1976).

The state permits at issue in this case were issued under the terms of Colorado's air quality regulation No. 3, 5 C.C.R. § 1001-5. This regulation was part of Colorado's approved SIP for the enforcement of the National Ambient Standards. Thus, since the restrictions in question were established pursuant to a SIP, they are federally enforceable by definition.

3. Which Restrictions to Apply

Restrictions contained in state permits which limit specific types and amounts of actual emissions ("blanket" restrictions on emissions) are not properly considered in the determination of a source's potential to emit. Memorandum Opinion at 1132. However, federally enforceable permit conditions which restrict hours of operation or amounts of material combusted or produced are properly included in the calculation. Id. Within the latter category, however, where the permits at issue contain a number of different restrictions, a question arises as to the proper restriction to use in the calculation. The expert testimony on this issue was uncontroverted that the "most restrictive" of the several permit limitations is the one that should be employed in determining the potential to emit.^[22] I find that I agree with that proposition, and so hold.

In this particular case, however, such a ruling does not dispose of the issue, since the experts in this case were in disagreement over which permit limitation should be considered the most restrictive. Frey's calculations, it will be recalled, were based on an annual limit on operations of 8000 hours. In contrast, Bray employed the annual limit on production, contained in the original saw and press permits, of 49,950 tons.

To state the issue a bit more precisely, there was never any question about which limitation was the more restrictive of the two. All other factors and variables being equal (that is, if the parties had otherwise used the same methodology and test data), the limitation on annual tons of production would always yield a lower figure for potential to emit than the limitation on annual hours of operation. Thus, in that sense at least, the restriction utilized by Bray was clearly the more restrictive. Rather, the controversy on this issue stemmed from the government's contention that the restriction on tons of production was not an "effective" limit on operations and should not have been used at all.

Essentially, the government's position was that this restriction did not really limit production to 49,950 TPY because it applied only to finished production. Since some of the waferboard produced is removed during the trimming process, the government argued that more than 49,950 tons could actually be produced under this limitation. For example, if the LPC produced 49,950 tons of finished product, and in the process removed 1000 tons of waferboard as trim, the government would contend that 50,950 tons had actually been produced.

LPC's response to this concern was that Bray took the trimming process into account in making his calculations. In computing the "emission factor" upon which his results were based, Bray took the amount of total emissions generated during the test and divided by the total weight of finished product to come up with a figure of emissions per ton of finished production. Of central importance is the fact that the emission factor was based on production after the trimming process. Bray then multiplied the emission factor by the annual limit of 49,950 tons to determine the annual potential to emit.

After a thorough examination of the calculations submitted by the experts in this case, I find that the annual limitation on tons of production, properly employed, is indeed as effective a restriction on operations as any of the others contained in the permits. I further find that this restriction was properly utilized by Bray. Since the emission factor he computed stated the emissions generated per ton of finished product, the emissions generated in producing the waferboard that was ultimately trimmed were included in the potential to emit figure. Moreover, if it is valid to assume that the emissions generated during a four-hour test are representative of and can be used to compute the emissions generated throughout the year, it is just as valid to assume that the amount of trim removed during such a test is representative of the trim removed throughout the year. Accordingly, since the annual limitation of tons of production is the most restrictive permit limitation, and since it is as effective a limitation on operations as any of the other restrictions contained in the permits, I find that it was the proper limitation to employ for purposes of determining potential to emit in the present case.

4. The Proper Effect of Permit Limitations That Are Willfully and Regularly Violated

Federally enforceable restrictions on operations that are contained in state permits are properly considered in determining potential to emit. 40 C.F.R. § 52.21(b)(4). Where a number of such restrictions exist, the "most restrictive" of the several provisions is the one that should be employed. In the case at hand, the annual limitation on tons of production is both federally enforceable and the most restrictive. Nonetheless, the government argues that this limitation should not be considered in this case.

The government argues generally that a source which knowingly and routinely violates the conditions of a permit should not get the benefit of those conditions in the computation of the source's potential to emit. Thus, since LPC regularly and knowingly violated the restriction on annual tons of production, the government urges that this restriction should not be considered in the present case. For the reasons which follow, I agree with the government on this point, and rule that conditions contained within state emission permits are not to be considered in the determination of a source's potential to emit, notwithstanding 40 C.F.R. § 52.21(b)(4), where such conditions are knowingly and regularly violated.

First, as already noted, the definition of potential to emit at issue here was promulgated in response to the D.C. Circuit's holding in Alabama Power, 636 F.2d at 323. In that case, it will be recalled, the court ruled that the effect of pollution control equipment designed into a source must be considered in calculating the source's potential to emit. Id. at 355. While that rule of law is a good one, it is clear from the opinion that this holding is based upon the assumption that the control equipment in question will be used. Id. at 353-55. See also (prior opinion in same case) Alabama Power Co. v. Costle, 606 F.2d 1068, 1076 (D.C.Cir.1979) ("The ``potential to emit' of any stationary source must be calculated on the assumption that air pollution control equipment incorporated into the design of the facility will function to control emissions in the manner reasonably anticipated when the calculation is made."). As a result, I am unconvinced that the D.C. Circuit would extend this protection to a source where the control equipment was never used, inoperable, or disconnected.

The EPA went beyond the narrow holding of the Alabama Power case when it drafted the new definition of potential to emit to encompass not only "air pollution control equipment," but also federally enforceable "restrictions on hours of operation or on the type or amount of material combusted, stored, or processed." 40 C.F. R. § 52.21(b)(4). In the same way that the court's holding in Alabama Power assumes that the control equipment will be used, however, I believe that the latter part of this definition contemplates that emission limitations appearing within state permits will be complied with. Thus, as I am unconvinced that the Alabama Power court would extend the protections offered by its opinion to sources which fail to utilize their pollution control equipment, I am similarly unwilling to extend the rule that federally enforceable permit limitations are a component of potential to emit to a case where such limitations are repeatedly ignored or violated.

Second, to hold that permit limitations which are repeatedly violated should nonetheless be considered in determining potential to emit would give better treatment to sources which knowingly violate such conditions than the treatment currently afforded sources which comply with the law. For example, consider a source which has a potential to emit pollutants of less than 250 TPY solely by virtue of operational limitations contained within state permits issued to it. When faced with the need to expand operations, such a source can choose to either 1) apply for new permits with less restrictive limitations and comply with the old permits until the new ones are issued, or 2) violate the conditions contained within its current permits. Should it choose to obey the law and follow the former course of action, and should the relaxation of its permit limitations cause its potential to emit to exceed 250 TPY, it will become subject to the PSD program as soon as the new permits are issued. This is because the regulations currently provide that when a particular source becomes a major source solely by virtue of the relaxation of a federally enforceable limitation on operations, the source shall at that time become subject to the permit requirements of the PSD program. See 40 C.F.R. § 52.21(r)(4).

In the present case, it is established that LPC knowingly violated the annual restriction on tons of production contained in the state air emission permits at both Kremmling and Olathe. As a result, this limitation (upon which Bray's calculations were based) may not be employed in determining potential to emit in this case. Therefore, my conclusion as to the potential to emit VOCs at Kremmling is based upon unrestricted operations.

In addition to the calculations based upon unrestricted operations, Frey also calculated the potential to emit VOCs at Kremmling employing an annual limitation on operations of 8000 hours per year. I have not considered these calculations in reaching my conclusion for a number of reasons. First, it does not appear that any of the Kremmling permits really do limit operations to 8000 hours per year. The only permit issued for Kremmling even containing a reference to 8000 hours of operation is the drier permit dated November 20, 1985, but the terms of that permit merely state that some of the specific restrictions that are set out in that permit were determined "based on" 8000 hours of operation per year.^[23] Second, even if this permit did limit operations to 8000 hours of operation per year, such that it were necessary for me to decide the question, I would hold, for the reasons expressed above, that a regular and willful violation of one permit limitation (such as the annual restriction on tons of production) should eliminate consideration of any other permit limitations (such as the annual restriction on hours of operation) which would otherwise apply to the source.

Third, even if the rulings above are found to be too harsh, the ultimate conclusion regarding the potential to emit at Kremmling should still be based upon unrestricted operations, since both the permit containing the 49,950 ton limitation and the permit containing the 8000 hour reference were superceded in July of 1987. The new permits issued for Kremmling do not contain the 8000 hour reference, and raise the restriction on annual production to 78,216 tons. Of course, if a violation were to be based upon this fact rather than upon the legal rulings in this opinion, the date of the violation would be July 20, 1987, rather than the end of November, 1986.

Without considering any restrictions on operations, Frey calculated the potential of the Kremmling plant to emit VOCs to be 265 TPY under EPA Method 25 and 293.5 TPY using his own unpublished methodology. Under either approach, the Kremmling plant obviously qualifies as a major stationary source.^[24] Under the reasoning I have employed, the plant would have become a major source around November of 1986, which is when LPC first violated the limitation on production upon which it had been relying. Accordingly, I conclude that the violation alleged in the February 3, 1987 NOV (that the Kremmling plant was a major stationary source without a PSD permit) not only existed on that date, but persisted for more than 30 days thereafter. Therefore, I find in favor of the plaintiff on its First Claim for Relief in the Alternative.

C. Penalty

Where the EPA files a civil enforcement action and successfully establishes that a violation of the PSD regulatory scheme existed for more than 30 days following the issuance of an appropriate NOV, the court is empowered to assess a civil penalty of up to $25,000 per day of violation. 42 U.S.C. § 7413(b)(2). Generally, "[d]etermination of the amount of [a civil penalty] is committed to the informed discretion of the district judge." United States v. Ancorp Nat'l Services, Inc., 516 F.2d 198, 202 (2d Cir.1975). However, the penalty provision at issue expressly provides that

In determining the amount of any civil penalty to be assessed under this subsection, the courts shall take into consideration (in addition to other factors) the size of the business, the economic impact of the penalty on the business, and the seriousness of the violation.

42 U.S.C. § 7413(b).

All three of the factors enumerated in 42 U.S.C. § 7413 are important and should be considered. United States v. Chevron U.S.A., Inc., 639 F.Supp. 770, 779 (W.D.Tex.1985). Contra United States v. General Motors Corp., 403 F.Supp. 1151, 1164 (D.Conn.1975).^[25] However, there is nothing to indicate that all three factors are equally important or deserve equal weight. As a result, a nominal fine may be imposed upon even the largest enterprise in the appropriate circumstances. General Motors, 403 F.Supp. at 1164. For purposes of computing the appropriate fine, the penalty period begins when the source first commits the violation, and not later when the NOV is issued. United States v. SCM Corp., 667 F.Supp. 1110 (D.Md.1987). Delay on the part of the government in bringing the enforcement action should neither increase nor decrease the penalty amount. Id. at 1128.

There is little precedent providing guidance on how to assess the "seriousness" of the violations at issue. One recorded case in which a fine was imposed for a violation of the PSD program is United States v. Chevron, 639 F.Supp. at 770. In that case, the oil company knowingly allowed treatment of hydrogen sulfide to cease for a period of 17 months at its El Paso refinery. This action greatly increased emissions of sulphur dioxide, a harmful chemical and principal cause of "acid rain." Id. at 772. The PSD rules were violated because the cessation of treatment constituted a "major modification" for which the company had failed to obtain a PSD permit. Due to the fact that Chevron had numerous opportunities to treat and control these emissions and "choose not to do so for purely economic reasons," the company was fined $1000 per day for 522 days of violation. Id. at 779.

In contrast, the General Motors case dealt with a violation of the Clean Water Act. United States v. General Motors, 403 F.Supp. at 1151. In that case, vandals had entered an abandoned manufacturing facility that General Motors was trying to sell. Once inside, they opened the valves on the plant's oil storage tanks, causing oil to spill onto the ground and drain into a nearby creek which fed into the Pequabuck River. When General Motors acquired knowledge of the spill, it promptly notified the appropriate state and federal authorities, and directed a thorough clean-up operation which prevented all but about 25 of the 6-8000 gallons spilled from reaching the river. Id. at 1153. In light of these efforts, and the fact that the spill had been caused by third parties, a violation was found, but the court assessed a fine of only one dollar. Id. at 1165.

1. Mitigating Factors

In the present case, a number of factors going to the "seriousness" of the violation mitigate against the imposition of a heavy penalty. First, in LPC's defense, it should be noted that the PSD provisions create a most unusual and perplexing regulatory framework. These provisions prohibit the construction of a major stationary source until after a PSD permit is not only applied for, but actually received. Yet, one of the very propositions illustrated by this case is that it is impossible to know with certainty whether a source will qualify as a "major" source until after it is constructed and emission tests are performed.

As a result, the PSD framework makes no provision for a source which constructs in the good faith belief that it is not subject to the program, only to find out after operations are commenced that it is a major source.^[26] In such a situation, the most a source can do (other than cease operations) is apply for PSD permits, and this was promptly done by LPC upon receipt of the NOVs.^[27]

Second, the only purpose to be served in requiring a new source to submit a PSD permit application—the only real purpose of the PSD permitting program—is to ensure that the new source contains the best available control technology ("BACT"). I am aware that the determination of what controls constitute BACT for a particular source is an agency determination to be made by the EPA, and not by this court. However, the testimony of numerous experts at trial did establish the fact that the pollution control equipment "pioneered" by LPC^[28], and which was installed at Kremmling and Olathe at considerable expense, was the most effective control equipment for the particular application at issue that technology could provide. While this court cannot and does not hold that this equipment was BACT, I can and do hold that, in light of the ultimate purpose of the PSD program, these actions taken by LPC mitigate against the imposition of a heavy penalty.

Third, there is no evidence that the emissions from Kremmling and Olathe caused environmental damage in the sense that air quality standards were violated. In addition to the installation of BACT, the other requirement of the PSD permitting process is for the owner to demonstrate that operation of the source will not cause emissions in the area to exceed the National Ambient Air Quality Standards ("NAAQS") or any "increments" established for particular pollutants. The government conceded that no "increments" have been set for the pollutants at issue in this case, and that therefore a source need only stay within the NAAQS. Additionally, the evidence was undisputed that the existing ambient air quality, with the plants in operation, is far better than the NAAQS require for the pollutants at issue.

Fourth, I am unconvinced that LPC reaped any economic benefit from its delayed compliance with the PSD program. The benefits of delayed compliance are properly computed by attempting to quantify the savings a source obtains by installing the control equipment required by the PSD program not when it is legally required, but rather at some later point in time. The benefit consists of both the deferral of capital investment in the equipment and the complete avoidance of the expenses of operation and maintenance which would have been incurred if the equipment were in place. The economists proffered by both sides agreed that the benefit should be computed by determining the cost of the equipment as of the data of noncompliance, and then bringing that value forward to the date of compliance using an appropriate discount rate. The maintenance and operational expenses also create savings, and this cash flow must be discounted as well.

The date of noncompliance is the date that the control equipment that would have been required by the BACT analysis should have been paid for and installed. This, of course, must be a date when the source is in violation of the PSD program and when the equipment was technically available. The date of compliance is the date when the equipment is paid for, installed, and operational.^[29] The economists that testified reached different conclusions because they employed different discount rates and were given different dates (by the parties) as the date of noncompliance. All of them used the date that the EFBs were installed and operational as the date of compliance.

In the present case, there was no economic benefit from delayed compliance for two reasons. First, the Kremmling and Olathe plants were the first plants of their kind in the country to install EFBs to control emissions. Since the control equipment required by the PSD program was installed as soon as it became commercially available,^[30] it cannot be said that LPC delayed in installing this equipment. Second, and perhaps more important, is the fact that the EFBs were installed, and the required modifications were complete, by the end of 1986. As established in Part III B above, however, the violation at Kremmling did not occur until about the end of November, 1986, since it was at this time that the permit restriction on annual tons of production was first violated. Thus, since compliance (in economic terms) occurred at the same time the PSD program was first implicated, there cannot be said to have been any delayed compliance or resulting economic benefit.

I note for the record that the government proposed an alternative methodology for computing the proper penalty in a case such as this. The approach is to assess as a fine a percentage of the profits generated by the source for the period that it was in violation. This approach is rejected because it seems to this court to be so arbitrary and simplistic as to not really qualify as a "methodology" at all. If this method were used, two companies of exactly the same size could commit exactly the same violation, yet two drastically different fines would be imposed if one company were profitable and the other were not. Moreover, if the percentage is based solely on the magnitude of the violation as suggested, this approach leaves no room to consider the culpability of the offender. Thus, where a large emission or spill occurs, the method leaves no room to assess a nominal fine against a profitable defendant, as was properly done in the General Motors case.

2. Aggravating Factors

Notwithstanding the several factors above which mitigate against the imposition of a heavy penalty, I conclude that some penalty must be assessed nonetheless on the unique facts of this case. Initially, I note that LPC did knowingly violate the restriction on annual production contained within its state emission permits. Moreover, it was this willful act that caused the defendant to be in violation of the PSD program. In this sense, therefore, the violation in this case (however serious) was the result of a deliberate and willful act, and cannot be characterized as an accidental or inadvertent transgression.^[31]

In determining whether a source is subject to the PSD program, the EPA, in good faith, takes into account state-imposed restrictions on operations. However, the definition of the term "potential to emit"—and therefore the PSD program as a whole—is based on the assumption that a source subject to such restrictions will make a good faith effort to comply. Were this court to assess a nominal penalty only in this case, it would give sanction to a willful disregard of the PSD regulatory framework, and encourage other sources in the future to disregard other lawful restrictions on operations whenever convenient to do so.

As I have already noted, the regulatory framework at issue may be unusually difficult to comply with because it requires a source to guess what its emissions will be prior to construction and the commencement of operations. Nonetheless, there must be no question that the burden of guessing correctly remains with the source, and that a mistake in this process can indeed result in a penalty. Otherwise, future sources that are unsure of whether they will qualify as a major source will have no incentive to apply for PSD permits which, undisputably, is a burden. Rather, they will build first and wait for the issuance of an NOV before initiating the permit application process.

Finally, failure to assess a penalty might wrongly give some indication that the PSD provisions were somehow complied with in this case. LPC urges that by submitting PSD applications and installing state-of-the-art pollution control equipment, it complied "in substance" with the PSD program all along. Whatever effect these actions may have on the "seriousness" of the violation, they do not, in and of themselves, constitute compliance with the PSD regulatory framework. Although a source which has done these things has probably done all that the PSD program requires it to do, to hold that this constitutes compliance would be to entirely obliterate the EPA's role in the process. Rather, the requirements of the program have been met only upon receipt of PSD permits (not submittal of applications) after agency review and determination of BACT. As a result, the PSD framework still remains to be complied with in this case.

The violation at the Kremmling plant began around November of 1986 and continues to the present time. Since more than 30 days have passed since the NOV alleging this violation was issued on February 3, 1987, this court may impose a fine of up to $25,000 per day of violation. On the basis of the several considerations discussed above, I find that a fine of $65,000.00 is the proper penalty to impose in this case.

D. Injunction

The purpose of an injunction is to prevent future violations. United States v. SCM Corp., 667 F.Supp. at 1128; United States v. W.T. Grant Co., 345 U.S. 629, 633, 73 S.Ct. 894, 898, 97 L.Ed. 1303 (1953). As a result, before an injunction may properly issue, the court must find that there exists some cognizable danger of recurrent violation. The moving party bears the burden of satisfying the court that such danger exists and that injunctive relief is necessary. Id.

Rule 65 of the Federal Rules of Civil Procedure requires that "[e]very order granting an injunction ... shall be specific in terms [and] shall describe in reasonable detail ... the act or acts sought to be restrained ...". Fed.R.Civ.P. 65(d). One purpose of these requirements is to avoid the possible founding of contempt citations on an order that is too broad or vague. Schmidt v. Lessard, 414 U.S. 473, 94 S.Ct. 713, 38 L.Ed.2d 661 (1974); Calvin Klein Cosmetics Corp. v. Parfums de Couer, Ltd., 824 F.2d 665, 669 (8th Cir.1987). Thus, broad language in an injunction that essentially requires a party to obey the law in the future is improper because it is basic to the intent of Rule 65(d) that those against whom an injunction is issued should receive fair and precisely drawn notice of what the injunction actually prohibits. Schmidt v. Lessard, 414 U.S. at 476, 94 S.Ct. at 715; Calvin Klein, 824 F.2d at 669.

In the present case, LPC has submitted PSD permit applications that the EPA has found to be complete, and all indications are that the control equipment already installed will be found to constitute BACT. As a result, the government has failed to establish that there presently exists some danger of recurrent violation. Moreover, the type of injunction requested by the government—that this court enjoin LPC from further violations of the Clean Air Act and the Colorado SIP—would merely require LPC to "obey the law." As such, it would fail to meet the specificity requirements of Fed.R.Civ.P. 65(d). Accordingly, the government's prayer for an injunction will be denied.

CONCLUSION

Based upon the above and foregoing,

IT IS HEREBY ORDERED, ADJUDGED, AND DECREED that Plaintiff United States of America's Second Claim for Relief is DISMISSED with prejudice;

IT IS FURTHER ORDERED, ADJUDGED, AND DECREED that the DISMISSAL with prejudice previously entered in this case of Plaintiff United States of America's First Claim for Relief is hereby CONFIRMED;

IT IS FURTHER ORDERED, ADJUDGED, AND DECREED that the DISMISSAL with prejudice previously entered in this case of Plaintiff United States of America's Third Claim for Relief is hereby CONFIRMED;

IT IS FURTHER ORDERED, ADJUDGED, AND DECREED that the court finds in favor of Plaintiff United States of America and against Defendant Louisiana-Pacific Corporation on Plaintiff's First Claim for Relief in the Alternative; therefore

IT IS FURTHER ORDERED that a civil penalty is hereby assessed against Defendant Louisiana-Pacific Corporation in the amount of $65,000.00. The Clerk is ordered to enter final judgment in this amount in favor of the plaintiff and against the defendant;

IT IS FURTHER ORDERED that Plaintiff United States of America's prayer for injunctive relief be, and the same hereby is, DENIED.

Costs shall be assessed to the defendant upon plaintiff's filing of a bill of costs as provided by law.

NOTES

[1] The PSD Program, added to the Clean Air Act by Congress in 1977, is designed to protect areas where the air is relatively clean. It requires that a special permit be obtained before a "major stationary source" of air pollution, or a "major modification" of a major stationary source, may be constructed in such an area.

[2] Waferwood is a plywood substitute product made of resinated wood chips, or "wafers," which are compressed into boards.

[3] Additionally, the hot ponds perform the function of thawing out any logs which may, in the wintertime, be frozen.

[4] The 1984 tests showed formaldehyde emissions from the press vents of 9.14 lbs/hour. In contrast, the tests from May and June of 1983 indicate emissions of 19.05 and 31.92 lbs/hour, respectively.

[5] The concept and term "most restrictive permit limitation" recognizes the fact that a permit limitation, while it may be issued in reference to a particular piece of equipment in the process flow, is effectively a limit on the whole facility. For example, in a waferboard plant possessing a single waferizer and a single press, if the waferizer was limited to 200,000 tons of production per year, and further down the line the press was limited to 100,000 tons per year, the latter limitation would obviously be the more restrictive of the two. Moreover, it would effectively limit production for the entire facility (including the waferizer) to 100,000 tons per year.

[6] This permit limitation was contained in the wafer dryer permits for both Kremmling and Olathe, as well as the permits for the Olathe cross-cut and trim saws. I must admit some confusion over the fact that the permits for the Kremmling cross-cut and trim saws limit production to 49,950 four-by-eight foot sheets of waferboard annually. Assuming that one four-by-eight foot sheet of waferboard weighs less than a ton, this later restriction on sheets of production would clearly seem to be more restrictive than the former limit on tons of production. However, since neither plaintiff nor defendant argued that this latter limitation was the most restrictive, I will ignore this discrepancy as well.

[7] A "protocol" is a written plan or program which specifies how the emissions testing is to be conducted.

[8] As a representative of the APCD, Jorgenson was present to observe the testing at Kremmling and Olathe.

[9] Various emissions tests were performed at Olathe on March 12, 13, and 14, 1985.

[10] Roughly translated from layman's terms into more precise terms, "burning" would correspond to "complete combustion," and "smoldering" would correspond to what the experts referred to as "incomplete combustion."

[11] This method of operation was so unusual that at one point the Konus fire actually went out completely for 15 to 20 minutes because the large amount of fuel added (consisting of wet bark and sawdust) smothered it. One expert compared operation of the Konus to burning a small pile of wet leaves in the backyard. Operating the Konus as it is designed would be like adding wet leaves to the fire a few at a time. In contrast, the operation at the March, 1985 test at Olathe would be akin to putting out the fire by throwing a full bushel of wet leaves onto the pile all at once.

[12] Interestingly, none of the many permits issued for the Kremmling and Olathe facilities, by their terms, expressly limit operations to 8000 hours per year. This figure does not even appear at all in 15 of the 19 permits that were ultimately issued, including the original ten permits and the five permits issued for Kremmling in 1987. Four of the permits—the Olathe Konus and dryer permits dated May 28, 1985, the Olathe dryer permit dated October 21, 1985, and the Kremmling dryer permit dated November 20, 1985—do contain a reference to 8000 hours of operation. However, these actually state only that various other specific restrictions on emissions that are expressly contained in those permits were determined "based on" 8000 hours of operation per year.

[13] Method 25 is a method for VOC emission testing and analysis promulgated by the EPA and published at 40 C.F.R. § 60 App. A. It was originally developed in the context of new source performance standards, but the regulations state that all of the methods contained in Appendix A have potential applicability in other contexts. The government's position is that a methodology arising in the context of new source performance standards "is not necessarily applicable to sources subject to the prevention of significant deterioration requirements." In enacting the PSD program in 1976, Congress ordered the EPA to promulgate regulations giving specific guidance for a number of pollutants, including VOCs. 42 U.S.C. § 7476(a) & (c). The government does not dispute the fact that the EPA has never complied with this directive, and that the deadline set by Congress passed several years ago.

[14] The PSD regulations generally require that the air quality of the area in which the new emission source is to be located is to be monitored over a period of at least one year. See 40 C.F.R. 52.21(m).

[15] See 40 C.F.R. 52.21(i)(8).

[16] Colorado's SIP for its PSD program was approved by EPA in September of 1986.

[17] See Memorandum Opinion at 1131 (wherein the approach now being urged by the government was applied by this court without comment at a time before the issue had specifically been raised).

[18] The specific language chosen by Congress expressly contemplates the effect of an NOV extending beyond the 30 days immediately following its issuance. Eg. 42 U.S.C. § 7413(a)(1) ("If such violation extends beyond the 30th day ..."); 42 U.S.C. § 7413(b)(2) (EPA has jurisdiction to sue whenever the owner of a source commits a violation "more than 30 days after having been notified ...").

[19] A thorough analysis of the term "potential to emit", including a history of its definition and construction, is set forth in the Memorandum Opinion at 1131-34.

[20] There is precious little prior authority dealing, even in a general way, with the proper construction of the term "potential to emit." Moreover, with regard to the narrow and unique issues enumerated and discussed in this litigation, the parties have been unable to supply any helpful citation in their briefs, and the court has similarly been unable to locate any caselaw bearing directly on these points. Thus, since the issues raised in this case appear to present novel questions of law, the court must address them without the benefit of any precedent with which to guide the analysis.

[21] Additionally, I note that there is no evidence in the record (presented by either side) to indicate that the CO results from either of the other two tests (March, 1985 at Kremmling or June, 1985 at Olathe) would yield a (potential to emit CO at Olathe of 250 TPY).

[22] For an explanation of the concept behind the term "most restrictive permit limitation," see supra note 5.

[23] See supra note 12.

[24] Accordingly, I need not reach the issue of whether Method 25 or Frey's methodology is the proper approach for calculating the potential to emit VOCs.

[25] Since LPC is one of the largest businesses in the United States, it urges this court to rule that the first two factors enumerated in 42 U.S.C. § 7413 should not be considered, and in support thereof cites the case of United States v. General Motors, 403 F.Supp. at 1151. While I agree with the district judge in General Motors that the seriousness of the violation may well be the most important factor of the three, I am not prepared to say that the other two factors are irrelevant. First, the General Motors court was interpreting a different provision than the one at issue in this case and, while similar, it is not identical. Second, to ignore two of the three factors expressly listed in the statute would be contrary to both common sense and the clear instructions of the Congress. Third, I believe that the General Motors court was not inclined to consider the first two factors because the defendant was an enormous enterprise and the court had concluded that a nominal fine was appropriate under the unique circumstances of that case.

[26] Where the owner of a proposed source does not believe that the PSD program is applicable, there is every incentive not to submit a PSD application, since the permitting program may legally take two to three years and, in practice, can take an infinitely long time.

[27] In response to the government's contention that these applications did not contain "complete" monitoring information, it defies logic to criticize a source in this context (already constructed, and application required immediately) for failure to include a year's worth of pre-construction monitoring information in its PSD application.

[28] Use of EFBs to control emissions at Kremmling and Olathe represented the first successful commercial application of that technology in the waferwood industry.

[29] This is the date of "compliance"—regardless of whether PSD permits have been issued—because the expenditure is tied to this date, and it is the avoidance of this expenditure that is being studied.

[30] In stating that EFBs are the control equipment that would be required by the PSD permitting process, I do not mean to make any ruling that such equipment is BACT. As already noted, BACT is an agency determination. Nonetheless, the evidence at trial was overwhelming that the state-of-the-art equipment installed at Kremmling and Olathe would constitute BACT when that determination is ultimately made. Moreover, I note that the government's own economist used the date the EFBs were installed as the date of compliance in making her calculations. Thus, while I do not rule that the EFBs constitute BACT, I have, for purposes of computing the penalty in this case, no reason to believe that they do not.

[31] However, I do not wish to characterize LPC's actions as a knowing or willful violation of the PSD program. Prior to the issuance of this opinion, at least, a knowing violation of the conditions contained within a state-issued air emission permit was not necessarily the equivalent of a knowing violation of the PSD program.