1

Dakota Resource Council

Rocky Mountain Office of Environmental Defense

Sierra Club

July 12, 2007

Terry L. O’Clair, P.E., Director Division of Air Quality North Dakota Department of Health 918 East Divide Avenue, 2nd Floor Bismarck, North Dakota 58501

RE: Comments on North Dakota Department of Health’s Air Quality

Effects Analysis and Proposed Permit to Construct for the Gascoyne 500

Power Plant Dear Mr. O’Clair: The Dakota Resource Council (DRC), the Rocky Mountain Office of Environmental Defense (ED), and Sierra Club respectfully submit the following comments on the North Dakota Department of Health’s May 2007 Air Quality Effects Analysis (NDDH Permit Analysis) and Permit Application of Westmoreland Power to authorize construction of the Gascoyne 500 Generation Station (Gascoyne 500) and the Gascoyne Mine by Westmoreland Power, Inc. (Westmoreland). I. NDDH DID NOT ADDRESS THE PUBLIC NOTICE REQUIREMENTS OF

THE PSD REGULATIONS The North Dakota Department of Health (NDDH) has failed to meet proper public notice requirements for the proposed permit to construct for Gascoyne 500. Section 165(a)(2) requires that, in order for a PSD permit to be issued, “the proposed permit has been subject to a review in accordance with [section 165 of the Clean Air Act]. . .and a public hearing has been held with opportunity for interested persons. . .including representatives of the Administrator to appear and submit written or oral presentations on the air quality impact of such source, alternatives thereto, control technology requirements, and other appropriate considerations.” In EPA’s implementing regulations for PSD SIPs, it is stated that the public notice for a proposed permit must provide “the degree of increment consumption that is expected from the source.” 40 C.F.R. §51.166(q)(2)(iii). North Dakota has a similar requirement in Section 33-15-15-01.2(g)(2)(c) of the North Dakota Administrative Code. The EPA’s Environmental Appeals Board has interpreted these provisions as meaning that the public notice for a PSD permit must include the degree of

2

increment consumption that is expected in all of the locations impacted by the proposed source. IN THE MATTER OF HADSON POWER 14- BUENA VISTA, PSD Appeal Nos. 92-3, 92-4, 92-5, 4 E.A.D. 258, 272-3 (EAB 1992). In particular, the EAB noted “Different potential commentors may have an interest in different areas to be impacted and would want, and would reasonably be entitled to, available data on increment consumption at the area of their particular concern.” Id. at 273. With respect to Class I area increment consumption, NDDH’s public notice for Gascoyne 500 only listed the percentage of Class I increment consumption in South Unit of Theodore Roosevelt National Park. No information on the amount of increment consumption expected at the Lostwood Wilderness area was provided in the public notice, and no information on the cumulative amount of increment consumption at Class I areas was provided in the notice even though such cumulative increment consumption information was provided in the notice for the Class II increment. Given the significant issues associated with existing SO2 increment violations in Class I areas in North Dakota, it is imperative that the public be given proper notice of increment impacts at all North Dakota Class I areas. In addition, as discussed further below in our comments, the Gascoyne 175 MW facility and mine and the Gascoyne 500 facility and mine should all be reviewed as one source under the PSD program and, as such, NDDH should have provided the cumulative impacts on increment considering the entire Gascoyne facility. The imperative to provide public notice of increment consumption at specific Class I areas flows directly from the core statutory purposes of the PSD program. Section 160(2) of the Clean Air Act plainly provides that a central statutory purpose of the PSD program is “to preserve, protect, and enhance the air quality in national parks, national wilderness areas, national monuments, national seashores, and other areas of special national, scenic, or historic value.” Congress also instructed that the PSD program is intended “to assure that any decision to permit increased air pollution in any area to which this section applies is made only after careful evaluation of all the consequences of such a decision and after adequate procedural opportunities for informed public participation in the decisionmaking process.” CAA Sec. 160(5). Adequate notice is a necessary predicate to informed public participation in the PSD permit process. Thus, NDDH failed to adequately inform the public of the degree of increment consumption expected by Gascoyne 500, and the Gascoyne facility as a whole (i.e., the 175 MW unit and the 500 MW plant), in all areas to be impacted by the proposed facility and, accordingly, NDDH must re-issue its public notice to comply with its public participation requirements.1

1 As discussed later in these comments, NDDH also failed to develop an adequate analysis of impacts on soils and vegetation prior to issuing the proposed permit and did not make a meaningful soils and vegetation analysis available prior to convening a public hearing as required by the North Dakota’s PSD regulations. NDDH must also remedy this procedural flaw in the Gascoyne 500 permit.

3

II. THE PERMIT SHOULD BE DENIED BECAUSE OF THE IMPACTS OF THE

GASCOYNE FACILITY’S GREENHOUSE GAS EMISSIONS Neither Westmoreland or NDDH addressed the carbon dioxide (CO2) or other greenhouse gases to be emitted from plant. Yet, the Gascoyne facility will be a significant emitter of greenhouse gas pollutants. Those emissions will contribute significantly to global warming and its adverse impacts on the health, welfare, economy, and environment of the State of North Dakota and the planet as a whole. In February 2007, the Intergovernmental Panel on Climate Change (“IPCC”) released a summary of the contribution of Working Group I to its Fourth Assessment Report. The Summary, a copy of which is attached as Attachment 1 to this letter, concludes, among other things:

● The global atmospheric concentration of carbon dioxide has increased from a pre-industrial value of about 280 parts per million (ppm) to 379 ppm in 2005;

● The atmospheric concentration of carbon dioxide in 2005 exceeds by far

the natural range over the last 650,000 years;

● The primary source of the increased atmospheric concentration of carbon dioxide since the pre-industrial period results from fossil fuel use;

● There is at least a 9 out of 10 chance that the global average net effect of

human activities since 1750 has been one of warming;

● Warming of the climate system is unequivocal, as is now evident from observations of increases in global average air and ocean temperatures, widespread melting of snow and ice, and rising global average sea level;

● At continental, regional and ocean basin scales, numerous long term

changes have been observed. These include changes in arctic temperatures and ice, widespread changes in precipitation amounts, ocean salinity, wind patterns and aspects of extreme weather including droughts, heavy precipitation, heat waves and the intensity of tropical cyclones;

● There is greater than a 90% likelihood that most of the observed increases

in global average temperatures since the mid-20th century are due to the observed increases in anthropogenic greenhouse gas emissions;

● For the next two decades, warming of about 0.2 Degrees Celsius per

decade is projected for a range of emission scenarios;

4

● There is greater than a 90% likelihood that hot extremes, heat waves and heavy precipitation events will continue to become more frequent; and

● Anthropogenic warming and sea level rise would continue for centuries

due to the time scales associated with climate processes and feedbacks, even if greenhouse gas concentrations were to be stabilized.

In April 2007, the IPCC released a Summary of the Contribution of Working Group II to its Fourth Assessment Report. The Summary, a copy of which is attached as Attachment 2, concludes, among other things:

● Temperature increases have had effects on agriculture and forestry management at Northern Hemisphere higher latitudes;

● Drought-affected areas will likely increase in extent. Heavy precipitation

events which are very likely to increase in frequency, will augment flood risk; and

● In North America, major challenges are projected for crops that are near

the warm end of their suitable range or depend on highly utilized water resources.

On or about May 4, 2007, the IPCC released a Summary of the contribution of Working Group III to its Fourth Assessment Report. The summary, a copy of which is attached hereto as Attachment 3, concludes, among other things:

● Global greenhouse gas (GHG) emissions have grown since preindustrial times, with an increase of 70% between 1970 and 2004;

● The largest growth in global GHG emissions between 1970 and 2004 has

come from the energy supply sector (an increase of 145%);

● With current global climate change mitigation policies and related sustainable development practices, global GHG emissions will continue to grow over the next few decades;

● There is substantial economic potential for the mitigation of global GHG

emissions over the coming decades, that could offset the projected growth of global emissions or reduce emissions below current levels;

● There are mitigation opportunities with net negative costs, in other words,

for which the benefits such as reduced energy costs and reduced emissions of pollutants equal or exceed their costs to society, excluding the benefits of avoided climate change;

5

● Fuel switching from coal to gas, renewable heat and power (hydropower, solar, wind, geothermal and bioenergy), and early applications of carbon capture and storage (e.g., storage of removed carbon dioxide from natural gas) are key mitigation technologies and practices currently commercially available;

● Near-term health co-benefits from reduced air pollution as a result of

actions to reduce GHG emissions can be substantial and may offset a substantial fraction of mitigation costs;

● It is often more cost-effective to invest in end-use energy efficiency

improvement than in increasing energy supply to satisfy demand for energy services. Efficiency improvement has a positive effect on energy security, local and regional air pollution abatement and employment;

● Renewable energy generally has a positive effect on energy security,

employment and on air quality; and

● In order to stabilize the concentrations of GHGs in the atmosphere, emissions would need to peak and decline thereafter.

NDDH should consider the entirety of the Fourth Assessment Report in making its decision on the permit for Gascoyne 500 facility. We request that NDDH make the entire Fourth Assessment Report part of the administrative record for Gascoyne 5002. Due to the severe impacts of the proposed facility’s greenhouse gas emissions on the health, welfare, economy, and environment of the State of North Dakota, the nation, and the planet as a whole as described in the IPCC report, NDDH should deny the proposed permit. The proposed permit contains no provisions whatsoever designed to eliminate or minimize carbon dioxide emissions. III. NDDH FAILED TO ADDRESS BACT REQUIREMENTS FOR THE

CARBON DIOXIDE AND OTHER GREENHOUSE GAS EMISSIONS TO BE

EMITTED FROM GASCOYNE 500 NDDH did not evaluate best available control technology (BACT) for CO2 or other greenhouse gases to be emitted from Gascoyne 500, nor did NDDH consider in the BACT analysis the environmental impacts of the greenhouse gases to be emitted from the proposed circulating fluidized bed (CFB) boiler technology for producing electricity at the Gascoyne facility as compared to other methods for producing electricity from coal. It is clear following the Supreme Court’s April 2, 2007 decision in Massachusetts v. EPA, ___ U.S. ___, 127 S. Ct. 1438 (2007), a copy of which is attached as Attachment 4 to this letter, that NDDH must conduct a best available control technology (BACT) analysis for carbon dioxide and other greenhouse gas pollutants, and set BACT emission

2 The IPCC report is available at http://www.ipcc.ch/activity/ar.htm.

6

limitations for these greenhouse gas pollutants in any permit that it issues for the Gascoyne source. The proposed permit for the Gascoyne facility is significantly flawed in the absence of such a review and BACT emission limits for CO2 and other greenhouse gases. The federal Clean Air Act and North Dakota regulations prohibit the construction of a new major stationary source of air pollutants except in accordance with a prevention of significant deterioration construction permit issued by NDDH. Clean Air Act § 165(a), 42 U.S.C. § 7475(a); 40 C.F.R. §52.21(a)(2)(iii) incorporated by reference into North Dakota Administrative Code (NDAC) §33-15-15-01.1. NDDH must conduct a BACT analysis and include in the construction permit BACT emission limitations “for each pollutant subject to regulation under [the Clean Air Act]” for which emissions exceed specified significance levels. Clean Air Act, §§ 165(a), 169, 42 U.S.C. §§ 7475(a), 7479; NDAC § 33-15-15-01.2 (adopting by reference 40 C.F.R. §§ 52.21(b)(1), (b)(2), (b)(12), (b)(50), (j)(2)). In NDAC §33-15-15-01.1, NDDH adopted, largely verbatim, the Environmental Protection Agency’s (“EPA”) Prevention of Significant Deterioration regulations set forth at 40 C.F.R. § 52.21.3 The regulations provide that “[a] new major stationary source shall apply best available control technology for each regulated NSR pollutant that it would have the potential to emit in significant amounts.” 40 C.F.R. § 52.21(j)(1) (emphasis added). Section 52.21(b)(50) defines “regulated NSR pollutant” as including “any pollutant . . . subject to regulation under the Act.”4 Specifically, the regulation provides: Regulated NSR pollutant, for purposes of this section, means the following:

(i) Any pollutant for which a national ambient air quality standard has been promulgated and any constituents or precursors for such pollutants identified by the Administrator (e.g., volatile organic compounds are precursors for ozone);

(ii) Any pollutant that is subject to any standard promulgated under Section

111 of the Act;

(iii) Any Class I or Class II substance subject to a standard promulgated under or established by title VI of the Act; or

3 NDAC §33-15-15-01.2 adopts by reference each of the subsections of 40 C.F.R. §52.21 cited herein verbatim, with the caveat that the term “administrator” as used in Section 52.21 means the NDDH except for those authorities which cannot be delegated to NDDH by EPA in which case Administrator means the Administrator of EPA. 4 References to the “Act” in this case pertain to the federal Clean Air Act, 42 U.S.C. 7401 et seq. NDAC §33-15-01-04.1 states that references to the term “Act” in the Article 33-15 of the NDAC means the North Dakota Century Code, Chapter 23-25, except “when the context indicates otherwise.” In this term “regulated NSR pollutant,” references to the Act clearly are referring to the federal Clean Air Act, since there is no section 111 or 112 or title VI in Chapter 23-25 of the North Dakota Century Code. Indeed, the context of the federal PSD regulations of 40 C.F.R. §52.21, which North Dakota has essentially incorporated by reference at NDAC §33-15-15-1.2, provides that any references to “the Act” are referring to the federal Clean Air Act.

7

(iv) Any pollutant that otherwise is subject to regulation under the Act; except

that any or all hazardous air pollutants either listed in section 112 of the Act or added to the list pursuant to section 112(b)(2) of the Act, which have not been delisted pursuant to section 112(b)(3) of the Act, are not regulated NSR pollutants unless the listed hazardous air pollutant is also regulated as a constituent or precursor of a general pollutant listed under section 108 of the Act.

40 C.F.R. § 52.21(b)(50) [emphasis added]. Section 52.21(b)(12), which defines BACT, also makes clear that BACT requirements apply to all air pollutants subject to regulation under the Clean Air Act. The regulation states:

Best available control technology means an emissions limitation (including a visible emission standard) based on the maximum degree of reduction for each pollutant subject to regulation under Act which would be emitted from any proposed major stationary source or major modification which the Administrator, on a case-by-case basis, taking into account energy, environmental, and economic impacts and other costs, determines is achievable for such source or modification through application of production processes or available methods, systems, and techniques, including fuel cleaning or treatment or innovative fuel combustion techniques for control of such pollutant.

40 C.F.R. § 52.21(b)(12); see also 42 U.S.C. 7479(3). [Emphasis added.] The BACT analysis that NDDH must conduct for each pollutant subject to regulation under the Clean Air Act must include a case-specific review of relevant energy, environmental and economic considerations that is informed by detailed information submitted by the applicant. See 42 U.S.C. § 7479(3); 40 C.F.R. 52.21(b)(12), (n). Based on its BACT analysis, NDDH must set emission limitations in its permit. See 42 U.S.C. § 7479(3) (BACT means “an emission limitation”); 40 C.F.R. 52.21(b)(12)(same). It is undisputed that the Gascoyne source is subject to BACT requirements for a number of air pollutants for which emissions will exceed specified significance levels. See NDDH Permit Analysis at 19, Part II (7 pollutants are subject to BACT review because they exceed PSD significance levels). Nor is there any dispute that the source has the potential to emit CO2 emissions in excess of any applicable BACT significance threshold.5 See Information Request Response to Request No. 3 from Westmoreland Power, Inc., to IWLA, Fresh Energy, Wow, and Minnesota Center for Environmental Advocacy (MCEA), dated April 12, 2007, included as Attachment 5 to this letter (acknowledging that the potential to emit of the proposed lignite coal-fired generating station is over 5 million tons of carbon dioxide without carbon capture). The proposed

5 Section 52.21(b)(23)(i), 40 C.F.R., does not set forth a significance level for carbon dioxide. Therefore,

pursuant to 40 C.F.R. § 52.21(b)(23)(ii), any emissions of carbon dioxide are significant.

8

permit is subject to BACT requirements for carbon dioxide because carbon dioxide is an “air pollutant” subject to regulation under the Clean Air Act. Section 302(g) of the Clean Air Act defines “air pollutant” expansively to include “any physical, chemical, biological, radioactive . . . substance or matter which is emitted into or otherwise enters into the ambient air.” 42 U.S.C. § 7602(g)(emphasis added). As the Supreme Court recently found, CO2 and other greenhouse gases are clearly considered “air pollutants” under the Clean Air Act. See April 2, 2007 decision in Massachusetts v. EPA, ___ U.S. ___, 127 S. Ct. 1438 (2007), included as Attachment 4 to this letter. Specifically, the Court found:

The Clean Air Act’s sweeping definition of “air pollutant” includes “any

air pollution agent or combination of such agents, including any physical, chemical . . . substance or matter which is emitted into or otherwise enters the ambient air . . . .” §7602(g) (emphasis added). On its face, the definition embraces all airborne compounds of whatever stripe, and underscores that intent through the repeated use of the word “any”. . . Carbon dioxide, methane, nitrous oxide, and hydrofluorocarbons are without a doubt “physical [and] chemical . . . substance[s] which [are] emitted into . . . the ambient air.” The statute is unambiguous.

127 U.S. 1438 (2007), at 1460. Thus, the Court in Massachusetts v. EPA dispensed with any uncertainty whether carbon dioxide is an “air pollutant” under the Clean Air Act.6 The plain meaning of Section 165(a)(4) of the Clean Air Act’s mandate that BACT applies to “each pollutant subject to regulation under [the Clean Air Act]” extends not only to air pollutants for which the Act itself or EPA or the States by regulation have imposed requirements, but also to air pollutants for which EPA and the States possess but have not exercised authority to impose such requirements. Section 202 of the Act requires standards applicable to emissions of “any air pollutant” from motor vehicles, and Section 111, which requires standards of performance for emissions of “air pollutants” from new stationary sources. 42 U.S.C. §§ 7411, 7521. Regulation under Sections 202 and 111 is required where air pollution “may reasonably be anticipated to endanger public health or welfare.” 42 U.S.C. § 7411(b)(1)(A); 42 U.S.C. § 7521(a)(1). Moreover, carbon dioxide is not just “subject” to regulation under the Act, but is currently regulated. Section 821 of the Clean Air Act Amendments of 1990 required EPA to promulgate, within 18 months after enactment of the Amendments, regulations to require certain sources, including coal-fired electric generating stations, to monitor carbon dioxide emissions and report monitoring data to EPA. 42 U.S.C. § 7651k note. In 1993, EPA promulgated such regulations, which are set forth at 40 C.F.R. Part 75. The regulations generally require monitoring of carbon dioxide emissions through

6 EPA’s then general counsel, Jonathan Z. Cannon, opined in 1998 that carbon dioxide is within the Clean

Air Act’s definition of “air pollutant” and that EPA has the authority to regulate carbon dioxide. More recently, however, EPA has advanced an interpretation that is contrary to the plain language of Section 302(g) and the Massachusetts v. EPA opinion.

9

installation, certification, operation and maintenance of a continuous emission monitoring system or an alternative method (40 C.F.R. §§ 75.1(b), 75.10(a)(3)); preparation and maintenance of a monitoring plan (40 C.F.R. § 75.33); maintenance of certain records (40 C.F.R. § 75.57); and reporting of certain information to EPA, including electronic quarterly reports of carbon dioxide emissions data (40 C.F.R. §§ 75.60 – 64). Section 75.5, 40 C.F.R., prohibits operation of an affected source in the absence of compliance with the substantive requirements of Part 75, and provides that a violation of any requirement of Part 75 is a violation of the Clean Air Act.7

If there were any doubt that carbon dioxide is subject to regulation under the Clean Air Act following Massachusetts v. EPA, 127 S. Ct. at 1459-63, the President issued a May 14, 2007 Executive Order, reconfirming that EPA can regulate greenhouse gases, including carbon dioxide, from motor vehicles, nonroad vehicles and nonroad engines under the Clean Air Act. The Executive Order directs EPA to coordinate with other federal agencies in undertaking such regulatory action.

A challenge to EPA’s failure to establish emission limits for carbon dioxide emissions from power plants under Section 111 of the Clean Air Act is pending before the United States Court of Appeals for the District of Columbia Circuit. State of New York, et al. v. EPA, No. 06-1322. EPA refused to establish such emission limits solely on the ground that EPA lacked the authority to regulate carbon dioxide under the Clean Air Act. Based on Massachusetts v. EPA, petitioners, on May 2, 2007, asked the Court of Appeals to vacate EPA’s determination that it lacks authority to regulate carbon dioxide emissions under Section 111, and to remand the matter to EPA for further proceedings consistent with the Massachusetts v. EPA decision.

EPA and the State’s regulations cited above echo the mandate of Section 165(a)(4) of the Clean Air Act that BACT applies not only to pollutants for which regulatory requirements have been imposed, but also to air pollutants for which EPA and the States possess but have not exercised authority to impose regulatory requirements.8 The regulations provide that BACT applies not only to air pollutants for which there are national ambient air quality standards under Section 109 of the Act, standards of performance for new sources under Section 111 of the Act, or standards under or established by Title VI of the Act (relating to acid deposition control), but also to “[a]ny pollutant that is otherwise subject to regulation under the Act.” 40 C.F.R. § 52.21(b)(50). Carbon dioxide is an air pollutant subject to regulation under the Clean Air Act for which NDDH must comply with BACT requirements. Thus, for all of the above reasons, NDDH cannot issue a permit for Gascoyne 500 until it evaluates BACT for CO2 and proposes a BACT emission limit for CO2 at Gascoyne 500.9

7 North Dakota has adopted the carbon dioxide monitoring requirements of 40 C.F.R. Part 75. NDAC §

33-15-21-09. 8 Indeed, EPA and North Dakota lack the authority to promulgate regulations diluting the mandate of

Section 165(a)(4) of the Clean Air Act. 9 In its April 12, 2007 Information Request Responses to IWLA, Fresh Energy, Wow, and Minnesota Center for Environmental Advocacy (MCEA) (Response to Request No. 3 in Attachment 5 to this letter), Westmoreland also concedes that carbon dioxide is a regulated pollutant subject to BACT. Specifically, it

10

Not only must NDDH determine BACT for CO2 at Gascoyne 500, but NDDH must also evaluate BACT and propose BACT emission limits for nitrous oxide (N2O) and methane (CH4) to be emitted from Gascoyne. As discussed above, the Supreme Court found that the Clean Air Act definition of “air pollutant” clearly includes both of these pollutants in addition to CO2. 127 S. Ct. 1460 (2007). Thus, as pollutants subject to regulation under the Clean Air Act, NDDH must determine BACT for these pollutants as well as CO2. The National Coal Council identifies fluidized bed combustion, the coal combustion technology that Westmoreland proposes to use at Gascoyne, as an especially large source of the greenhouse gas N2O:

“N2O has a GWP (Global Warming Potential) 296 times that of CO2. Because of its long lifetime (about 120 years) it can reach the upper atmosphere, depleting the concentration of stratospheric ozone, an important filter of UV radiation. N2O is emitted from fluidized bed coal combustion; global emissions from FBC units are 0.2 Mt/year, representing approximately 2% of total known sources. N2O emissions from PC units are much lower. Typical N2O emissions from FBC units are in the range of 40-70 ppm (at 3% O2). This is significant because at 60 ppm, the N2O emission from the FBC is equivalent to 1.8% CO2, an increase of about 15% in CO2 emissions for an FBC boiler. Several techniques have been proposed to control N2O emissions from FBC boilers, but additional research is necessary to develop economically and commercially attractive systems."10

Based on information in the permit application and default AP-42 emission factors, the Gascoyne 500 facility has a potential to emit 5,771 tons of nitrous oxide each year.11 The nitrous oxide that would be released from Gascoyne 500 is equivalent, in Global Warming Potential, to an additional 1,708,216 million tons per year of carbon dioxide. Methane will also be emitted from the Gascoyne 500 facility. According to the EPA, conditions that promote N2O formation will also favor methane formation.12

failed to identify the potential to emit of nitrous oxide and methane to be emitted from Gascoyne 500, stating that the pollutants were “[n]ot a regulated pollutant and no BACT emission rate is required.” Westmoreland did not include a similar claim with respect to carbon dioxide emissions, and instead provided the potential CO2 emissions expected from the Gascoyne lignite coal-fired generating station without carbon capture. 10 “Coal-Related Greenhouse Gas Management Issues,” National Coal Council, May 2003, at page 7. 11 Emissions of CO2 and N2O were calculated based on default AP-42 emission factors for lignite combustion in atmospheric fluidized bed boilers, on the design coal carbon content of the Gascoyne coal (from Table 2-1 of the May 14, 2004 Gascoyne 175 MW permit application, which was used because carbon content of the coal was not provided in the Gascoyne 500 permit application), and on the maximum coal feed rate for Gascoyne 500 of 527 tons/hour (from Appendix B-1 of the Gascoyne 500 permit application) assuming continual operation throughout the year. 12 See AP-42 at 1.7-3 (9/98), available at http://www.epa.gov/ttn/chief/ap42/index.html. AP-42 does not provide specific emission factors for methane emissions from lignite-fueled power plants.

11

As with CO2, because no significance level is identified for N2O or methane in the definition of “significant” in the PSD regulations, any increase in emissions of these pollutants must be considered significant. See 40 C.F.R. §52.21(b)(23)(ii) which is incorporated by reference into North Dakota’s regulations at NDAC §33-15-15-01.2. Thus, the Gascoyne 500 power plant will be significant for N2O and methane emissions and, accordingly, BACT must be met for these pollutants as well as CO2. See 40 C.F.R. §52.21(j)(2) which is incorporated by reference into North Dakota’s regulations at NDAC §33-15-15-01.2. Not surprisingly, several western states have recently in fact adopted carbon dioxide emission limitations for new coal-fired power plants. California and Washington have both adopted carbon dioxide emission limitations of 1100 pounds per MW-hr. And Montana recently adopted a minimum sequestration mandate, providing that new coal plants must capture and sequester a minimum of 50% of the carbon dioxide reduced.

12

The table below summarizes these carbon dioxide emission standards and limits adopted by other western states. These provide manifest evidence of BACT-level emission limits. Table 1: Western State Carbon Dioxide Emission Limitations (as of July 2007)

STATE LAW

STANDARD APPLICABILITY EFFECTIVE DATE

State of Montana, HB 0025, signed into law by Gov. Schweitzer on May 14, 2007

Mandate for the facility to capture and sequester a minimum of 50% of the carbon dioxide produced.

Applies to new electric generating units “primarily fueled by coal.”

January 1, 2007

State of Washington, SB 6001, signed into law by Gov. Gregoire on May 3, 2007

The lower of 1100 pounds of greenhouse gases per megawatt-hour or the average available GHG emission output of new combined cycle natural gas thermal electric generation turbines commercially available and offered for sale.

Triggered upon long-term financial commitments: (1) new ownership interest or upgrade to baseline power plant, or (2) new/renewed contract with a term or five years or more.

Standard takes effect on July 1, 2008

State of California, SB 1368, signed into law by Governor Schwarzenegger on Sept. 29, 2006

Greenhouse gas emissions performance standard shall be established by administrative agency at a rate that is no higher than the rate of emissions of greenhouse gases for combined-cycle natural gas baseload generation; CPUC recently established 1100 pounds of CO2 per MW-hour as the operative standard

Applies to long-term contracts for baseload power of five years or longer

CPUC rules for IOUs take effect February 1, 2007

13

Even if greenhouse gases were not pollutants subject to regulation under the Clean Air Act, Westmoreland would be required to consider emissions of greenhouse gases in its BACT analysis for Gascoyne. Specifically, the federal Environmental Appeals Board (EAB) has interpreted the definition of BACT as requiring consideration of unregulated pollutants in setting emission limits and other terms of a permit, since a BACT determination is to take into account environmental impacts.13 A published article entitled Considering Alternatives: The Case for Limiting CO2 Emissions from New

Power Plants through New Source Review by then EPA Assistant General Counsel Gregory B. Foote discusses the regulatory background to support consideration of CO2 impacts when permitting a new source and, in particular, a new coal-fired power plant.14 This article indicates that it is entirely appropriate to consider CO2 emissions and other pollutants that impact climate change when evaluating environmental impacts under the new source review permit program, and the paper also provides suggested approaches for evaluating technologies in terms of climate changing emissions. The proposed permit for the Gascoyne 500 does not contain a BACT emissions limitation for CO2, N2O, or methane. NDDH has not conducted a BACT analysis for CO2, N2O or methane. NDDH has made no effort to identify or evaluate available “production processes or available methods, systems and techniques” for control of these pollutants. See 40 C.F.R. § 52.21(b)(12). NDDH has failed to do so, despite the fact that Westmoreland has indicated the ability to design Gascoyne 500 with a carbon capture option that could consist of oxy-fuel combination, ammonia scrubbing, amine scrubbing, and membrane separation. (See Attachment 5, Response to Request No. 3). NDDH and Westmoreland must conduct BACT review and analyses for CO2 and the other greenhouse gases to be emitted by Gascoyne 500, including providing a review of the reductions in those emissions that might be achieved by through use of lower emitting technologies. For example, Westmoreland could use a more efficient electricity production process such as integrated gasification combined cycle (IGCC). This letter and attachments provide detailed information (further below in this letter) to show that IGCC facilities are more efficient in producing electricity. Not only are IGCC facilities more efficient in producing electricity from coal than a circulating fluidized bed boiler, the CO2 in the synthesized gas can be captured and sequestered at a fraction of the cost of post-combustion carbon capture and sequestration at other coal-fired power plants. Such IGCC facilities can also meet lower emission rates of other PSD pollutants, including mercury, than conventional coal-fired power plants have been permitted at today, as well as very low levels of SO2 and NOx. There are also more efficient options for producing electricity from coal from conventional coal-fired power plants, including supercritical fluidized bed boilers. This technology is also discussed in this letter, and it would result in lower CO2 emissions by producing electricity from coal in a more efficient manner than the currently proposed CFB boiler for Gascoyne 500. Westmoreland must also consider the use of or blending

13 See In Re North County Resource Recovery Associates, 2 E.A.D. 229, 230 (Adm’r 1986), 1986 EPA App. LEXIS 14. 14 Considering Alternatives: The Case for Limiting CO2 Emissions from New Power Plants through New Source Review, Gregory B. Foote, 34 ELR 10642, 7-2004, Attachment 6.

14

with other fuels to lower CO2 emissions. One example would be blending coal with biomass. Another option would be use of or blending with K-fuels, which is a fuel that has been treated by a pre-combustion process that improves the quality of the coal, increasing the heat content of the coal and thus resulting in less CO2 emissions as well as lower emissions of SO2, NOx and mercury.15 After evaluating more efficient technologies and options to prevent or reduce the amount of CO2 and other greenhouse gas emissions, Westmoreland must also evaluate as BACT for Gascoyne 500 add-on technologies to capture and sequester the greenhouse gas emissions. The U.S. Department of Energy is the primary federal agency working on research and development of CO2 capture and sequestration technologies, and thus information on carbon capture and sequestration technologies is available on the U.S. DOE website.16 The IPCC issued a report in 2005 that discussed the main options currently available to capture CO2 from fossil fuel-fired power plants:17

• Post-combustion capture using a liquid solvent to capture the CO2

• Oxy-fuel combustion capture, which uses oxygen instead of air for combustion to produce a flue gas that is mainly water vapor and high concentrations of CO2

• Pre-combustion capture, which would be used at power plants using IGCC.

According to the IPCC, commercial CO2 capture systems can reduce CO2 emissions by 80-90% per kilowatt-hour.18 CO2 capture systems are available today and have been applied to several small power plants19. NDDH must require Westmoreland to evaluate the available CO2 capture systems and to evaluate such CO2 control systems at the proposed CFB facility. Westmoreland has clearly been evaluating these technologies. See Attachment 5. NDDH must require that Westmoreland provide a BACT analysis of these technologies. Further, NDDH must also require Westmoreland to evaluate the more efficient approaches to producing electricity from coal, including an IGCC plant and a supercritical fluidized bed boiler, with CO2 capture systems. All of these options must then be evaluated in top-down BACT analyses for CO2, nitrous oxides, and methane. In addition to violating the mandate to comply with BACT requirements for carbon dioxide because carbon dioxide is a “pollutant subject to regulation” under the Clean Air Act, NDDH has unreasonably, arbitrarily, capriciously and unlawfully failed to eliminate or limit carbon dioxide and other greenhouse gas emissions under other applicable provisions of the North Dakota regulations. NDAC §33-15-01-15.1 provides that no person shall permit or cause air pollution. “Air pollution” is defined in NDAC §33-15-01-04.3 as “the presence in the outdoor atmosphere of one or more air contaminants in such quantities and duration as is or may be injurious to human health, welfare, or

15 See Fact Sheets on K-Fuel at http://www.evgenergy.com/about.shtml. 16 See http://www.fossil.energy.gov/programs/sequestration/capture/. 17 2005 IPCC Special Report on Carbon dioxide Capture and Storage, Technical Summary, at 25. See also Chapter 3 of this report. (Both the Technical Summary and Chapter 3 are included as Attachment 7; entire document is available at http://arch.rivm.nl/env/int/ipcc/pages_media/SRCCS-final/IPCCSpecialReportonCarbondioxideCaptureandStorage.htm). 18 Id. at 107 (Chapter 3). 19 Id.

15

property or animal or plant life, or which unreasonably interferes with the enjoyment of life or property.” “Air contaminant” is defined in NDAC §33-15-01-04.2 in a very broad manner, similar to the definition of “air pollutant” in the Clean Air Act addressed in Massachusetts v. EPA.20 The Massachusetts v. EPA decision makes clear that NDDH may rely on §33-15-01-15.1 to eliminate or limit carbon dioxide and other greenhouse gas emissions from the Gascoyne facility. In light of the serious adverse impacts of carbon dioxide and other greenhouse gas emissions on human health and welfare, property, and the environment, NDDH cannot lawfully refuse to exercise its authority under NDAC §33-15-01-15.1 to eliminate or limit carbon dioxide and other greenhouse gas emissions in taking action on the proposed Gascoyne 500 permit. See Discussion in Section II, above. Indeed, the Supreme Court in Massachusetts v. EPA, even without the benefit of the most recent IPCC Reports, noted that the “[t]he harms associated with climate change are serious and well recognized.” 127 S. Ct. at 1455. The Supreme Court also acknowledged “the enormity of the potential consequences associated with man-made climate change.” Id. at 1458. NDDH’s failure to conduct a BACT analysis for carbon dioxide, nitrous oxides, and methane and to establish emission limitations for these greenhouse gas pollutants must be rectified before NDDH may lawfully issue a permit for the Gascoyne 500 facility. It appears that Westmoreland has not provided NDDH relevant information as part of its permit application sufficient to allow NDDH to conduct the required analysis. NDDH cannot propose issuance of the PSD permit for Gascoyne until this significant issue is addressed. IV. FEDERAL AND STATE CLEAN AIR LAWS REQUIRE WESTMORELAND

TO CONSIDER THE APPLICATION OF PRODUCTION PROCESSES AND

AVAILABLE METHODS, SYSTEMS AND TECHNIQUES TO LOWER

AIRBORNE CONTAMINANTS

IGCC is an available, demonstrated advanced coal combustion technology with significant emission reduction benefits including fewer emissions of criteria and hazardous air pollutants, the opportunity for capturing greenhouse gases, such as CO2, that cause global warming, and a general increase in efficiency over other coal burning technologies. However, the permit application for Gascoyne 500 does not evaluate advanced coal technology and instead focused solely on CFB coal combustion. NDDH also did not evaluate IGCC technology and gave no indication as to why IGCC was not evaluated. NDDH has a duty as part of the core of the BACT determination process to provide a reasoned justification for rejecting an available control technology.

20 NDAC §33-15-01-04.2 defines “air contaminant” as “any solid, liquid, gas, or odorous substance, or any

combination thereof.”

16

North Dakota and Federal Law Require a Thorough Evaluation of IGCC as Part of the BACT Analysis. Section 165(a)(4) of the Clean Air Act (CAA) provides that “no major emitting facility on which construction is commenced after August 7, 1977, may be constructed in any area to which this part applies unless…the facility is subject to the best available control technology for each pollutant subject to regulation under this chapter emitted from, or which results from, such facility.”21 The requirement for conducting a BACT analysis is codified at 40 CFR § 52.21(j), which has been incorporated by reference into North Dakota’s regulations at NDAC §33-15-15-01.2. Federal and State law further requires that “the owner or operator of a proposed source. . . shall submit all information necessary to perform any analysis to make any determination” required including “information necessary to determine that best available control technology would be applied.”22 BACT is then defined under North Dakota law as follows:

an emissions limitation (including a visible emissions standard) based on the maximum degree of reduction for each pollutant subject to regulation under the [Clean Air] Act which would be emitted from any proposed major stationary source or major modification which the Administrator, on a case-by-case basis, taking into account energy, environmental, and economic impacts and other costs, determines is achievable for such source or modification through application or production processes or available methods, systems, and techniques, including fuel cleaning or treatment or innovative fuel combustion techniques for control of such pollutant.23

This definition includes coal gasification. The legislative history of the amendment adding the term “innovative fuel combustion techniques” to the Clean Air Act’s definition of “BACT” is clear. Coal gasification must be considered. The relevant passage of the debate is excerpted below:

Mr. HUDDLESTON. Mr. President, the proposed provisions for application of best available control technology to all new major emission sources, although having the admirable intent of achieving consistently clean air through the required use of best controls, if not properly interpreted may deter the use of some of the most effective pollution controls. The definition in the committee bill of best available control technology indicates a consideration for various control strategies by including the phrase “through application of production processes and available methods systems, and techniques, including fuel cleaning or treatment.” And I believe it is likely that

21 42 U.S.C. §7475(a)(4). 22 40 C.F.R. §52.21(n)(iii), incorporated by reference into North Dakota’s regulations at NDAC §33-15-15-01.2. 23 40 C.F.R. §52.21(b)(12), emphasis added, incorporated by reference into North Dakota’s rules at NDAC §33-15-15-01.2. See also 42 U.S.C. §7479(3).

17

the concept of BACT is intended to include such technologies as low Btu gasification and fluidized bed combustion. But, this intention is not explicitly spelled out, and I am concerned that without clarification, the possibility of misinterpretation would remain. It is the purpose of this amendment to leave no doubt that in determining best available control technology, all actions taken by the fuel user are to be taken into account--be they the purchasing or production of fuels which may have been cleaned or up-graded through chemical treatment, gasification, or liquefaction; use of combustion systems such as fluidized bed combustion which specifically reduce emissions and/or the post-combustion treatment of emissions with cleanup equipment like stack scrubbers. The purpose, as I say, is just to be more explicit, to make sure there is no chance of misinterpretation. Mr. President, I believe again that this amendment has been checked by the managers of the bill and that they are inclined to support it. Mr. MUSKIE. Mr. President, I have also discussed this amendment with the distinguished Senator from Kentucky. I think it has been worked out in a form I can accept. I am happy to do so. I am willing to yield back the remainder of my time.24

EPA and federal courts have consistently interpreted the BACT provisions found in the CAA and the agency’s regulations as embodying certain core criteria that require the permit applicant either to implement the most effective available means for minimizing air pollution or justify its selection of less effective means on grounds consistent with the purposes of the Act. In Citizens for Clean Air v. EPA,25 the Ninth Circuit held that “initially the burden rests with the PSD applicant to identify the best available control.” As stated in long-standing EPA guidance, “[r]egardless of the specific methodology used for determining BACT, be it ‘top-down,’ ‘bottom-up,’ or otherwise, the same core criteria apply to any BACT analysis: the applicant must consider all available alternatives, and [either select the most stringent of them or] demonstrate why the most stringent should not be adopted.”26 Accordingly, the PSD permit applicant not only must identify all available technologies, including the most stringent, but it must also provide adequate justification for dismissing any available technologies. Consistent with these core criteria, the EPA’s New Source Review (NSR) Workshop Manual establishes that, as the first step in the “top-down” BACT analysis, the applicant must consider all “available” control options:

The first step in a "top-down" analysis is to identify, for the emissions unit in question (the term "emissions unit" should be read to mean emissions unit, process or activity), all "available" control options. Available control options are those air pollution control technologies or techniques with a practical potential for application to the emissions unit and the regulated

24 95th Congress, 1st Session (Part 1 of 2) June 10, 1977 Clean Air Act Amendments of 1977 A&P 123 Cong. Record S9421. 25 959 F.2d 839, 845 (9th Cir. 1992) 26 Memorandum from John Calcagni, Director of EPA Air Quality Management Division, to EPA Regional Air Directors (June 13, 1989), at 4 (emphasis added).

18

pollutant under evaluation. Air pollution control technologies and techniques include the application of production process or available methods, systems, and techniques, including fuel cleaning or treatment or innovative fuel combustion techniques for control of the affected pollutant. This includes technologies employed outside of the United States. As discussed later, in some circumstances inherently lower-polluting processes are appropriate for consideration as available control alternatives.27

EPA further explains that potential control options can be categorized in three ways:

• Inherently lower emitting processes/techniques

• Add-on controls

• Combinations of inherently lower emitting processes/techniques and add-on controls.28

With respect to inherently lower emitting processes, EPA explains that “[l]ower-polluting processes should be considered based on demonstrations made on the basis of manufacturing identical or similar products from identical or similar raw materials or fuels.”29 “The term ‘available’ is used…to refer to whether the technology ‘can be obtained by the applicant through commercial channels or is otherwise available within the common sense meaning of the term.’”30 In keeping with the stringent nature of the BACT requirement, EPA has repeatedly emphasized that “available”

is used in the broadest sense under the first step and refers to control options with a “practical potential for application to the emissions unit” under evaluation. . . . The goal of this step is to develop a comprehensive list of control options.31

EPA adjudicatory decisions also examine the core requirements for the BACT determination process. “Under the top-down methodology, applicants must apply the best available control technology unless they can demonstrate that the technology is technically or economically infeasible. The top-down approach places the burden of

27

NSR Manual, at p. B.5 (emphasis added). A complete copy of the NSR Manual is included as

Attachment 43 to this letter. 28 Id. at B.10. 29 Id. 30 In re: Maui Electric Company, PSD Appeal No. 98-2 (EAB September 10, 1998), at 29-30 (quoting NSR Manual at B.17). 31 In re: Knauf Fiber Glass, PSD Appeal Nos. 98-3 – 98-20 (EAB February 4, 1999), at 12-13 (quoting NSR Manual at B.5) (emphasis added by EAB); see also In re: Steel Dynamics, Inc., PSD Appeal Nos. 99-4 and 99-5 (EAB June 22, 2000), at 29 n.24 (citing Knauf with approval); NSR Manual at B.10 (“The objective in step 1 is to identify all control options with potential application to the source and pollutant under evaluation.”); id. at B.6 (emphasizing that a proper Step 1 list is “comprehensive”).

19

proof on the applicant to justify why the proposed source is unable to apply the best technology available.”32 Whatever analytical process is utilized for determining BACT, these core criteria – the requirement to consider all available technologies, including the most stringent, and to provide adequate justification in the administrative record for dismissing any of the technologies based on relevant statutory factors – must be satisfied. Thus, to conduct a BACT analysis consistent with the requirements of state and federal law, Westmoreland must thoroughly evaluate all available control measures. IGCC is commercially available today. North Dakota and federal law therefore require that this technology be thoroughly evaluated as part of the BACT analysis. Any Arguments that IGCC Does Not Need to Be Considered Because It Would Be “Redefining the Source” are Flawed and Must Fail In its air quality construction permit application, Westmoreland referred to a December 13, 2005 letter from Stephen D. Page of EPA regarding whether IGCC had to be evaluated as part of the BACT review for a proposed coal-fired power plant in which EPA stated that it would not require evaluation of IGCC in the BACT analysis because it would redefine the basic design of the source. Gascoyne 500 permit application at 5-1. But this document may not be given any effect that fixes rights or obligations. A number of environmental groups filed a petition for judicial review in the U.S. Court of Appeals in Washington, D.C. challenging the lawfulness of this December 13, 2005 letter. The parties reached a Settlement Agreement providing that “EPA agrees and stipulates that the December 13, 2005 document is not final agency action and creates no rights, duties, obligations nor any other legally binding effects on EPA, the states, tribes, any regulated entity or any person.” See NRDC, et al. v. EPA, No. 06-1059 (D.C. Cir.); see also 71 Fed. Reg. 61,771 (Oct. 19, 2006). EPA has argued in other contexts that the concept of “redefining the source” may relieve it of certain obligations under the PSD program.33 In particular, in the Prairie State case before the EAB, EPA argued as a matter of statutory interpretation that the Clean Air Act did not contemplate that permitting authorities would require a permit applicant to consider building a source other than the one it had proposed. In that case, the issue involved whether a proposed Illinois coal-fired power plant, that was being planned in conjunction with a new coal mine, needed to consider (as an element of its BACT

32 In re: Spokane Regional Waste-to-Energy Applicant, PSD Appeal No. 88-12 (EPA June 9, 1989), at 9 (internal quotation marks omitted) (emphasis in original); see also In re: Inter-Power of New York, Inc. PSD Appeal Nos. 92-8 and 92-9 (EAB March 16, 1994) (“Under the ‘top-down’ approach, permit applicants must apply the most stringent control alternative, unless the applicant can demonstrate that the alternative is not technically or economically achievable.”); In the Matter of Pennsauken County, New Jersey Resource Recovery Facility, PSD Appeal No. 88-8 (EAB November 10, 1988) (“Thus, the ‘top-down’ approach shifts the burden of proof to the applicant to justify why the proposed source is unable to apply the best technology available.”) 33 See In re Prairie State Generating Co., PSD Appeal 05-05, 13 E.A.D. __ (Sept. 24, 2006).

20

analysis) using coal that was lower in sulfur than the coal that the co-located mine would produce. EPA argued (as did Illinois EPA) that requiring the source to use coal other than that from the co-located mine would constitute an impermissible redefinition of the source. Ultimately, in a very narrow ruling, the Board in the Prairie State case held that the use of coal from the co-located mine was so integral to the very purpose and intent of the project that requiring the permit applicant to consider using some other source of coal instead would defeat the purpose of the original permit application. Accordingly, the Board ruled that the Illinois EPA did not “clearly err when it determined that consideration of low-sulfur coal, because it necessarily involves a fuel source other than the co-located mine, would require Prairie State to redefine the fundamental purpose or basic design of its proposed Facility, and that, therefore, low-sulfur coal could appropriately be rejected from further BACT analysis at step 1 of the top-down review method.” Prairie State at 36-37.

Even assuming that the Board’s decision in Prairie State was consistent with the Clean Air Act, that decision clearly demonstrates that Westmoreland’s failure to consider innovative combustion technologies as process options for controlling emission from the Gascoyne 500 plant is fundamentally flawed. First, the EAB’s ruling recognized that the default assumption under the Clean Air Act’s PSD provisions is that the use of potentially cleaner fuels (such as low-sulfur coal) will normally be a required part of the BACT analysis.34 Only where some unique element of the facility’s basic purpose made the particular BACT-related consideration fundamentally incompatible with the permit application, did the EAB recognize that further analysis of that BACT-related consideration might be unnecessary.35

In the end, even the Board’s decision in Prairie State reflects an understanding that the concept of redefining the source must be subordinate to the primary objectives of the BACT analysis. That is, the specific requirements inherent in the definition of BACT will define the obligations of permit applicants and permitting authorities, unless some specific fundamental conflict exists. Moreover, while the Board concluded that the permit issuer should look “in the first instance” at “how the permit applicant, in proposing the facility, defines the goals, objectives, purposes, or basic design for the

34 Prairie State at 22 (“Petitioners correctly observe that . . . consideration of ‘clean fuels’ must be a part of the BACT analysis. Specifically, . . . the Agency must consider both the cleanliness of the fuel and the use of add-on pollution control devices.”). Indeed, numerous other PSD permits have identified the use of clean fuel (including low sulfur coal) as BACT for new major sources. See, e.g. In re AES Puerto Rico 8 E.A.D. 324 (EAB 1999); In re Encogen Cogeneration, 8 E.A.D. 244 (EAB 1999); In re Hawaiian

Commercial & Sugar Co.y, PSD Appeal No. 92-1 at 5, n.7 (EAB, July 20, 1992). 35 In Prairie State the Board concluded that the mine and the coal-fired power plant were proposed together as a single source under the PSD provisions, and the mine was intended to supply the entirety of the power plant’s fuel throughout the plant’s entire operating life. Therefore, the EAB concluded, the plant and the mine were integral parts of a single proposal and the use of coal from another source would undermine the purpose of that proposal. If the mine were capable of supplying less than the full fuel needs of the power plant over its entire life cycle, for example, the Board’s analysis would likely have been different; the Board’s decision suggests that in such a case the consideration of low-sulfur supplemental fuel would have been required.

21

proposed facility,” the permit applicant cannot manipulate the definition of the facility as a mechanism to avoid appropriate BACT analysis. Prairie State at 29-30. In evaluating the permit, the permit issuer must “discern which design elements are inherent to [the] purpose [of the facility], articulated for reasons independent of air quality permitting, and which design elements may be changed to achieve pollutant emissions reductions without disrupting the applicant’s basic business purpose for the proposed facility.” Id. at 30.

Significantly, the Board specifically recognized that cost savings are not a valid purpose for a particular facility design; similarly, “the business objective of avoiding risk associated with new, innovative or transferable control technologies is not treated as a basic design element.” Prairie State at 30 n.23. Rather, cost and risk considerations are appropriately addressed during the later steps of the top-down BACT analysis. Westmoreland’s position on this issue is out of sync with both the Clean Air Act itself and with the EAB’s treatment of the concept of “redefining the source,” as well as North Dakota law. First, as discussed above, the Clean Air Act and North Dakota’s PSD regulations specifically calls for consideration of “the application of production processes and available methods, systems, and techniques, including fuel cleaning, clean fuels, or treatment or innovative fuel combustion techniques for control of each pollutant.” CAA § 169(3); NDAC §33-15-15-01.2. This language, on its face, requires as a part of the BACT analysis the consideration of innovative technologies like IGCC that make the generation power from coal significantly cleaner.36 Further, the two early decisions by the EPA Administrator that introduce the “redefining the source” policy, identify a policy that is much more limited than that put forth by Westmoreland. In In re Pennsauken County, New Jersey, Resource Recovery Facility the petitioner asked the EPA Administrator to deny a PSD permit to a municipal waste combustor and, instead, require the county to dispose of its waste by co-firing it with coal in existing power plants. See PSD Appeal No. 88-8 at 10 (Adm’r, Nov. 10, 1988). In effect, the petitioner wanted the EPA to order the applicant to engage in a different type of activity: electricity generation, rather than waste disposal. The Administrator rejected this option because the petitioner’s argument was based on his objection to a waste combustor generally, not to the conditions in the permit. Thus, the Administrator held, the petitioner was asking EPA to “redefine the source” from a waste combustor to a power plant.37 The Administrator subsequently reaffirmed the Pennsauken County

36 As discussed above, the legislative history of the Clean Air Act is equally as clear that the definition of BACT contemplates consideration of technologies like IGCC. 37 “Petitioner Filipczak’s fundamental objections to the Pennsauken permit are not with the control technology, but rather, with the municipal waste combustor itself. He urges rejection of the combustor in favor of co-firing a mixture of 20 percent refuse derived fuel and 80 percent coal at existing power plants. These objections are beyond the scope of this proceeding and therefore are not reviewable under 40 C.F.R. 124.19, which restricts review to “conditions” in the permit. Permit conditions are imposed for the purpose of ensuring that the proposed source of pollutant emissions-- here, a municipal waste combustor-- uses emission control systems that represent BACT, thereby reducing the emissions to the maximum degree possible. These control systems, as stated in the definition of BACT, may require application of “production processes and available methods, systems, and techniques, including fuel cleaning as treatment or innovative fuel combustion techniques” to control the emissions. The permit conditions that define these

22

decision and explained that “source,” within the newly created “redefining the source” policy, refers to a source category.38

After clarifying the “redefining the source” policy as only preventing a change in the “fundamental purpose,” i.e., the source category, the Administrator further explained that the “redefining the source” policy did not allow the permitting agency to blindly accept the source design proposed by the applicant. Hibbing, 2 EAD at 842-843. In Hibbing, the permit applicant wanted to burn petroleum coke at its taconite plant, but EPA required the applicant to consider burning natural gas – a lower polluting process and cleaner fuel – as part of a BACT determination. Id. The Administrator specifically rejected the idea that requiring consideration of cleaner fuel constitutes “redefining the source” because the fundamental purpose, or source category, remains the same.39

In other words, from its inception, prior to the 1990 Manual, the “redefining the source” policy has merely stood for the concept that EPA will not require an applicant to abandon its intended purpose for some other industrial venture. To the extent EPA’s subsequently-issued draft NSR Workshop Manual is inconsistent with prior Administrator interpretations in Pennsauken and Hibbing, which constitute the agency’s official position, the draft Manual is not entitled to any deference.40

systems are imposed on the source as the applicant has defined it… [T]he source itself is not a condition of the permit.” Pennsauken County at 10-11 (emphasis added). 38 “In Pennsauken, the petitioner was urging EPA to reject the proposed source (a municipal waste combustor) in favor of using existing power plants to co-fire a mixture of 20 percent refuse derived fuel and 80 percent coal. In other words, the petitioner was seeking to substitute power plants (having as a

fundamental purpose the generation of electricity) for a municipal waste combustor (having as a

fundamental purpose the disposal of municipal waste).” In re Hibbing Taconite Company, 2 E.A.D. 838, 843 at n. 12 (Adm’r 1989) (parentheticals original, emphasis added). 39 [O]ne argument that could be made is that the Region, by requiring the burning of natural gas to be an alternative to be considered in the BACT analysis [for a petroleum coke-fired plant], is seeking to "redefine the source." Traditionally, EPA has not required a PSD applicant to redefine the fundamental scope of its project… [The redefining the source] argument has no merit in this case. EPA regulations define major stationary sources by their product or purpose (e.g., "steel mill," "municipal

incinerator," "taconite ore processing plant," etc.), not by fuel choice. Here, Hibbing will continue to

manufacture the same product (i.e., taconite pellets) regardless of whether it burns natural gas or

petroleum coke… The record here indicates that there are other taconite plants that burn natural gas, or a combination of natural gas and other fuels. Thus, it is reasonable for Hibbing to consider natural gas as an alternative in its BACT analysis. Id. (parentheticals original, emphasis added). 40 In addition to simply being wrong, the NSR Manual’s application of the “redefining the source” policy is due no deference because it conflicts with the agency’s prior interpretations. See Pauley v. Beth-Energy

Mines, 501 U.S. 680, 698 (1991) (no deference to agency interpretations that are inconsistent with previously held view); see also Malcomb v. Island Creek Coal Co., 15 F.3d 364, 369 (4th Cir. 1994) (deference is not due to an agency interpretation of its own rules that is inconsistent); Brotherhood of Locomotive Engineers v. Atchison, Topeka Santa Fe R.R. Co., 116 S.Ct. 595, 133 L.Ed.2d 535 (1996). Other Supreme Court precedent confirms that “Chevron deference” is not due to an agency’s interpretations of the statutes that it operates under when such interpretations are the product of informal processes such as adoption of manual provisions rather than formal processes such as notice and comment rulemaking. See United States v. Mead Corp., 120 S.Ct. 2164 (2001), Skidmore v. Swift & Co., 323 U.S. 134 1944).

23

Because the Clean Air Act specifically calls for consideration of production processes and innovative fuel combustion techniques as means for reducing emissions from industrial sources regulated under the PSD program, even the Board’s analysis in Prairie

State would require evaluation of IGCC as part of the BACT analysis, unless there were

a specific, objectively discernable reason why doing so would be fundamentally at odds

with the primary objective of the project, based on appropriate considerations not related

to cost or the avoidance of risk.41 For Gascoyne 500, neither Westmoreland or NDDH have articulated such rationale. As discussed above, this position is simply untenable as a matter of statutory interpretation. Moreover, it also runs counter to the EAB’s favorable consideration of Illinois EPA’s requirement for permit applicants to consider IGCC. In Prairie State, the Petitioners argued that the scope of EPA’s “redefining the source” policy lacked any “principled standards,” and would therefore allow permit applicants to define-away basic elements of the BACT analysis. See Prairie State at 33. The EAB rejected this argument, but in doing so relied specifically on Illinois EPA’s policy of requiring consideration of IGCC to demonstrate why the policy was not fatally overbroad.42 Id. 33-37. The Board noted that Illinois EPA “required Prairie State to submit a detailed analysis of [IGCC] as a method for controlling emissions from the proposed Facility.” Prairie State at 35.43 The Board explained, “IGCC is not simply an add-on emission control technology, but instead would have required a completely redesigned ‘power block.’ . . . [Illinois EPA’s] demand that Prairie State provide a detailed analysis of IGCC, which [Illinois EPA] noted has the promise to achieve greater [emissions] reductions, demonstrates that [Illinois EPA’s] application of the policy against redefining the design of the source through application of BACT did not treat “very few” design changes as consistent with the proposed Facility’s basic design. . . . To the contrary, [Illinois EPA’s] consideration of IGCC demonstrates that [it] gave due regard to Prairie State’s objective in submitting a permit application for the proposed Facility, namely development of an electric power generating plant that would be co-located and co-permitted with a 30-year supply of fuel, and then explored every potential add-on technology and potentially lower-emitting production processes or methods consistent with that basic design to determine the maximum emissions reductions achievable for the Facility.” Id at 35-36.44

41 “The assertion, and finding, that the design is for reasons independent of air quality permitting must be reasonable and supported by the record.” Prairie State at 34 n.29. For Gascoyne, however, Westmoreland has failed to even make an evidence-based finding that IGCC is incompatible with the purpose of the project. 42 If the EAB affirmed Illinois EPA’s authority to require consideration of IGCC, such consideration must be within the permitting authority’s discretion under the statutory definition of BACT, and therefore cannot be a fundamental “redefinition” of the source that is impermissible under the Clean Air Act. 43 The Board references a letter from Donald Sutton, Illinois EPA to Diana Tickner, Prairie State (March 29, 2003), that letter is incorporated by reference here. 44 In its analysis, the Board specifically recognized that EPA guidance requires consideration of process-related technology advances like IGCC. Prairie State at 33 (“The NSR Manual also states with respect to production processes, that where ‘a given production process or emission unit can be made to be inherently less polluting’ ‘the ability of design considerations to make the process inherently less polluting must be

considered as a control alternative for the source.’”). The Board went on to explain that “viewing the proposed facility’s basic design as something that generally should not be redefined through BACT review does not prevent the permit issuer from taking a ‘hard look’ at whether the proposed facility may be

24

In contrast, for the Gascoyne facility, NDDH and Westmoreland have completely abrogated their BACT-related responsibilities when it comes to identifying “every potential add-on technology and potentially lower-emitting production processes or

methods consistent with that basic design to determine the maximum emissions reductions achievable.” While the Board ultimately concluded in Prairie State that IGCC was not required at the facility, that determination resulted from the Board’s conclusion that IGCC was essentially equivalent to the proposed boiler technology in terms of its potential emission control effectiveness. See Prairie State at 47. That conclusion was the unfortunate result of a poor record. As discussed at length below, it is very clear that IGCC is capable of achieving a level of emissions performance for virtually every regulated PSD pollutant that is significantly better than the performance of a PC boiler.45 Moreover, as discussed already, IGCC plants have a multitude of collateral environmental benefits: they provide opportunity for higher reductions in hazardous air pollutants like mercury, they produce less solid waste, they use less water, and they both emit less CO2 and provide the ability to capture CO2 emissions for permanent storage to help address global warming at lower costs than at a conventional coal-fired power plant. Accordingly, the Board’s justification for rejecting IGCC in Prairie State was based on solely on the facts of that case, which as will be shown below are simply inapplicable to the Gascoyne 500 plant.46

improved to reduce its pollutant emissions.” Id at 33-34. By “hard look” it is clear that the Board means a real, substantive BACT examination that explains in detail the technological, engineering, process, and/or design factors that make a particular emission control option incompatible with the projects objectives. See

Prairies State at 34 (citing Knauf, 8 E.A.D. 121, 127 (EAB 1999)). The Board explained that a permit issuer’s failure to take a sufficiently hard look at the design issues has “the potential to circumvent the purpose of BACT, which is to promote use of the best control technologies as widely as possible.” Prairie

State at 34 (quoting Knauf, 8 E.A.D. at 140). Significantly, the EAB gave short shrift to EPA’s essentially meaningless “alternatives analysis” which would have relegated consideration of any process, technique or alternative approach to pollution control to an analysis separate and apart from the BACT determination. Westmoreland’s treatment of IGCC in the Gascoyne 500 permit application is a perfect illustration of the danger that the EAB identified as inherent in the concept of a “redefining the source” exemption – Westmoreland has not taken a “hard look” at whether IGCC might be an appropriate consideration under the BACT analysis here and, failing to do so, threatens to “circumvent the purpose of BACT.” 45 The PSD permit application for Nueces Syngas, LLC for example, includes emission limits for the IGCC turbines (in lb/MMBTU) of 0.018 for NOx, 0.017 for SO2, 0.037 for CO, 0.003 for VOC, 0.006 for PM and PM10, and 0.001 for H2SO4. There are other recent permit applications in the record that also demonstrate the tremendous opportunities for emission reductions with IGCC. Moreover, this technology is now a viable and ready option for electric power production, as evidenced by among other things the 25 or so proposed IGCC plants around the country. See the Department of Energy’s document: Tracking New Coal-Fired Power Plants, available at: http://www.netl.doe.gov/coal/refshelf/ncp.pdf. 46 Moreover, to the extent that Westmoreland is concerned about cost implications of IGCC, the technological availability or reliability of the technology, or other technological or economic considerations, the appropriate mechanism to address those concerns is the BACT top-down analysis – not through up-front exclusion of the technology from consideration.

25

Indeed, EPA itself has publicly recognized IGCC as an “inherently low-polluting process/practice,”47 and has reaffirmed its view that IGCC is an available method for cleaning and treating coal to remove air pollutants prior to combustion:

One approach to controlling SO2 emissions from steam generating units is to limit the maximum sulfur content in the fuel. This can be accomplished by burning… a fuel that has been pre-treated to remove sulfur from the fuel… There are two ways to pre-treat coal before combustion to lower sulfur emissions: Physical coal cleaning and gasification… Coal gasification breaks coal apart into its chemical constituents (typically a mixture of carbon monoxide, hydrogen, and other gaseous compounds) prior to combustion. The product gas is then cleaned of contaminants prior to combustion. Gasification reduces SO2 emissions by over 99 percent.48

As a result of fuel cleaning, IGCC units “will inherently have only trace SO2 emissions because over 99 percent of the sulfur associated with the coal is removed by the coal gasification process.” 70 Fed. Reg. at 9715. 49

Because the Clean Air Act and implementing regulations clearly require evaluation of technologies like IGCC which can achieve the statutory intent of reducing emissions through process changes, available methods and systems and techniques, innovative combustion techniques, and fuel cleaning, and because NDDH and Westmoreland failed entirely to conduct an analysis of IGCC as a possible control option, the Gascoyne 500 permit application is woefully incomplete and the proposed permit inadequate.

47 See, e.g., Robert J. Wayland, U.S. EPA Office of Air and Radiation, OAQPS, “U.S. EPA’s Clean Air Gasification Activities”, Presentation to the Gasification Technologies Council Winter Meeting, January 26, 2006, slide 4; and “U.S. EPA’s Clean Air Gasification Initiative”, Presentation at the Platts IGCC Symposium, June 2, 2005, slide 11 (citing the “inherently lower emissions of nitrogen oxides, sulfur dioxides, and mercury,” as among the “fundamental advantages” of IGCC). Mr. Wayland also correctly notes that IGCC units use less water, and produce fewer global warming pollutants than conventional pulverized coal units, another point relevant to the statutory directive to “take into account environmental . . . impacts” in determining BACT limits. Wayland January 26, 2005 Presentation, Slide 4; 42 U.S.C. § 7479(3). 48 U.S. EPA, Standards of Performance for Electric Utility Steam Generating Units for Which Construction is Commenced After September 18, 1978, 70 Fed. Reg. 9706, 9710-11 (February 28, 2005). 49 Indeed, IGCC is a prime example of “fuel cleaning” (which also is a required BACT consideration under the Act) – involving the pre-combustion transformation of otherwise dirty coal into a fuel (syngas) that can be more cleanly burned in a combined-cycle power block.

26

Recent State Actions Requiring Consideration of Clean Coal Technology Establish Irrefutable Precedence for the Consideration of IGCC. In March 2003, the State of Illinois required the applicant for a proposed CFB coal-fired electric generation facility to conduct a robust analysis of IGCC as a core element of its BACT analysis:

Additional material must be provided in the BACT demonstration to address Integrated Gasification Coal Combustion (IGCC) as it is a `production process’ that can be used to produce electricity from coal. In this regard, the Illinois EPA has determined that IGCC qualifies as an alternative emission control technique that must be addressed in the BACT demonstration for the proposed plant. In addition, based on the various demonstration projects that have been completed for IGCC, the Illinois EPA believes that IGCC constitutes a technically feasible production process.

Accordingly, Indeck must provide detailed information addressing the emission performance levels of IGCC, in terms of expected emissions rates and possible emission reductions, and the economic, environmental and/or energy impacts that would accompany application of IGCC to the proposed plant. This information must be accompanied by copies of relevant documents that are the basis of or otherwise substantiate the facts, statements and representations about IGCC provided by Indeck. In this regard, Indeck as the permit applicant is generally under an obligation to undertake a significant effort to provide data and analysis in its application to support the determination of BACT for the proposed plant.50

In an ensuing letter, the State of Illinois then formally informed EPA that Illinois has “concluded that it is appropriate for applicants for [proposed coal-fired power plants] to consider IGCC as part of their BACT demonstrations.”51 Similarly, the Georgia Department of Natural Resources, in a March 2002 letter regarding the permit application of Longleaf Energy Station, also relied, in part, on the failure of the permit applicant to consider clean coal combustion technology in finding the application deficient. In making its determination of deficiency, Georgia stated that the applicant did not “discuss any other methods from generating electricity from the combustion of coal, such as pressurized fluidized bed combustion or integrated gasification combined cycle.” 52 Georgia further stated that the applicant “should discuss these technologies and explain why you elected to propose a pulverized coal-fired steam electric power plant instead.”53

50 Letter from Illinois Division of Air Pollution Control to Jim Schneider, Indeck-Elwood, LLC (March 8, 2003), Attachment 8A. 51 Letter from Illinois EPA Director to EPA Regional Administrator, Region V (March 19, 2003), Attachment 8B. 52 Letter from James A. Capp, Manager, Stationary Source Permitting Program, Georgia DNR, to D. Blake Wheatley, Assistant Vice President, Longleaf Energy Associates, LLC (March 6, 2002). Attachment 8C. 53 Id.

27

Reflecting the viability of IGCC, the State of New Mexico issued a letter on December 23, 2002 requiring the permit applicant for a new coal-fired power plant to conduct a site-specific analysis of IGCC as well as CFB as part of the BACT analysis for the proposed facility: “The Department requires a site-specific analysis of IGCC and CFB in order to make a determination regarding BACT for the proposed facility.” The New Mexico determination goes on to provide: “The analysis must include a discussion of the technical feasibility and availability of IGCC and CFB for the proposed site in McKinley County, including a discussion of existing IGCC and CFB systems.”54 On August 29, 2003, New Mexico issued its evaluation of the applicant’s response. New Mexico found that the applicant’s BACT analysis had in fact indicated that IGCC is commercially available but that the applicant had improperly relied on cost to find that the technology was infeasible:

Mustang concludes that neither IGCC nor CFB are technically feasible control options for the Mustang site. After careful review of the revised BACT analysis, as well as information gathered from independent sources, the Department determines that Mustang’s conclusion is not supported by the evidence. Accordingly, the Department finds that Mustang has not demonstrated the technical infeasibility of IGCC and CFB. Moreover, applying the criteria in the NSR Manual, the Department determines that IGCC and CFB are technically feasible at the Mustang site, and must be evaluated in the remaining steps of the top down BACT methodology.

(a) IGCC and CFB are technically feasible at the Mustang site. A technology is considered to be technically feasible if it is commercially available and applicable to the source under consideration. See NSR Manual at B.17-18. A technology is commercially available if it has reached a licensing and commercial sales stage of development. Id. A technology is applicable if it has been specified in a permit for the same or a similar source type. Id.

Mustang’s revised BACT analysis indicates that IGCC is commercially available, and IGCC has been specified in air quality permits for coal-fired power plants. See, e.g., Lima Energy Facility, 580 megawatt coal-fired power plant. Similarly, CFB is commercially available and has been specified in air quality permits for coal-fired power plants. See, e.g., AES Puerto Rico 454 megawatt coal-fired power plant; Reliant Energy Seward 584 megawatt coal-fired power plant.

(b) For both IGCC and CFB, Mustang improperly relies on cost to

determine technical infeasibility. A technology is technically feasible when the resolution of technical difficulties is a matter of cost. See NSR Manual at B.19-20. Mustang’s revised BACT analysis indicates

54 Letter from New Mexico Environment Department to Larry Messinger, Mustang Energy Corporation (Dec. 23, 2002). Attachment 8D.

28

that the resolution of technical difficulties for both IGCC and CFB are a matter of cost. These costs do not support a finding of technical infeasibility, but may be considered during Step 4 of the top down BACT methodology. See NSR Manual at B.26.55

In addition, the Montana Board of Environmental Review found that the Montana Department of Environmental Quality must consider IGCC as an available technology in the BACT review for a coal-fired power plant. Specifically, the Board of Environmental Review stated “. . .the Department should require applicants to consider innovative fuel combustion techniques in their BACT analysis and the Department should evaluate such techniques in its BACT determination in accordance with the top-down five-step method.”56 It is important to note that, while some of these states were operating under SIP-approved PSD programs, the definition of BACT that applied in all cases is virtually identical to the federal definition of BACT (as is North Dakota’s) with respect to consideration of inherently lower emitting processes. It is noteworthy that these states determined it was entirely appropriate, indeed necessary, to require consideration of IGCC in the BACT review for a coal-fired power plant. The aforementioned state determinations are attached hereto. Westmoreland Failed to Address IGCC in the BACT Analysis and the Permit Must be Denied. IGCC is an available method, system and technique for curbing air pollutants from Gascoyne 500 consistent with North Dakota’s definition of BACT. Indeed, as discussed below, the National Energy Technology Laboratory’s June 2006 examination of proposed coal-fired power plants nationwide reveals over 20 IGCC facilities proposed throughout the nation. The Gascoyne 500 application should have reviewed two IGCC options:

1. Entrained-Flow gasification systems such as those offered by GE, Shell and ConocoPhillips, and

2. Fluidized-Bed gasification processes that are optimized for low-rank coals such as

GTI’s U-Gas process.

55 Letter from New Mexico Environment Department to Larry Messinger, Mustang Energy Company (Aug. 29, 2003), at p. 3, Attachment 8E. 56 Montana Board of Environmental Review, Findings of Fact, Conclusions of Law, and Order In the Matter of the Air Quality Permit for the Roundup Power Project (Permit No, 3182-00), Case No. 2003-04 AQ (June 23, 2003) at 18-19. See Attachment 8F for a copy of this finding.

29

Entrained-Flow Gasification

Electricity generation from coal using IGCC technology is a commercially available and proven process. IGCC units generate electricity by integrating a coal gasifier with combined cycle (combustion turbine and steam turbine) electricity generation equipment (see figure below).

Two full scale commercial IGCC electric generating units are in operation in the United States: Tampa Electric Company’s 262 MW unit at the Polk plant in Florida and Cinergy’s 192 MW unit at the Wabash River plant in Indiana, which both rely on coal as a fuel source.57 Two other coal-based IGCC plants operate in Europe, NUON/Demkolec is a 253 MW plant in the Netherlands, and ELCOGAS in Spain is 298 MW.58 IGCC units can be constructed with multiple gasifiers to achieve unit availability at levels comparable to those of conventional baseload facilities. For instance, the Eastman Chemical plant in Kingsport, Tennessee has utilized a dual-gasifier design to produce chemicals from syngas and has experienced 98 percent availability since 1986.59

57 Resource Systems Group, Inc., EPIndex. See www.epindex.com 58 Major Environmental Aspects of Gasification-Based Power Generation Technologies, Dec 2002, Table 1-7, page 1-26, Attachment 9. 59 Smith, R.G., “Eastman Chemical Plant Kingsport Plant Chemicals from Coal Operations, 1983-2000,” 2000 Gasification Technologies Conference, Attachment 10.

30

ChevronTexaco claims that its new Standard Project Initiative Reference IGCC Plant achieves greater than 90% availability by using multiple gas trains.60 Worldwide there are 131 gasification projects in operation with a combined capacity equivalent to 23,750 MW of IGCC units.61 Although not all of these projects produce electricity from coal, they demonstrate widespread commercial application of gasification technology for fuel processing, one of two key components of an IGCC plant. Moreover, the June 2006 National Energy Technology Laboratory’s review of proposed coal-fired power plants nationwide reveals over 20 new IGCC facilities in various stages of development nationwide.62 The second component is a combined cycle electricity generating system, which is now commonplace for new natural gas fired power plants. IGCC units are available from major well-known vendors. Coal gasification equipment is available from GE63, Shell, and Global Energy, while major turbine manufacturers, including GE and Siemens-Westinghouse, provide combined cycle generators designed to run on the synthesis gas produced by coal gasifiers. Engineers from Texaco, Jacobs Engineering, and GE have teamed up to offer a standardized IGCC design.64 James Childress, the Executive Director of the Gasification Technology Council, provided testimony to the U.S. Senate Environment and Public Works Committee stating, “[g]asification is a widely used commercially proven technology.”65 At the same hearing, Edward Lowe, Gas Turbine-Combined Cycle Product Line Manager for General Electric Power Systems, stated that, “IGCC is inherently less polluting and more efficient than any other coal power generation technology.”66 Likewise, the National Coal Council, in a May 2001 report, confirms that IGCC is "viable, commercially available technology."67 ChevronTexaco, in an October 2002 presentation, states that, “IGCC is a current viable choice for clean coal capacity.”68 And the Center for Energy and Economic Development (CEED) states that, “IGCC technology is available for deployment today.”69

60 O’Keefe, L. and Sturm, K., “Clean Coal Technology Options – A Comparison of IGCC vs. Pulverized Coal Boilers,” presentation to the 2002 Gasification Technologies Conference, October 2002. Attachment 11. 61 Simbeck, Dale, SFA Pacific Inc. Gasification Technology Update, presented to the European Gasification Conference, April 8-10, 2002. The total capacity is based on output of synthesis gas. Many of these projects produce chemicals in addition to or instead of electricity. 62 NETL, Tracking New Coal-Fired Power Plants Coal’s Resurgence in Electric Power Generation, June 21, 2006. 63 On June 30, 2004, GE acquired the gasification business of ChevronTexaco 64 O’Keefe, Luke, et al. A Single IGCC Design for Variable CO2 Capture. Attachment 12. 65 Childress, James M. Statement Submitted for the Record, Senate Environment and Public Works Subcommittee on Clean Air, Wetlands and Climate Change, January 29, 2002. 66 Lowe, Edward. Outlook on Integrated Gasification Combined Cycle (IGCC) Technology. Senate Environment and Public Works Subcommittee on Clean Air, Wetlands and Climate Change, January 29, 2002. 67 National Coal Council, Increasing Electricity Availability from Coal-Fired Power Plants in the Near Term, p. 20 (May 2001). Attachment 13. 68 “Clean Coal Technology Options – A Comparison of IGCC vs. Pulverized Coal Boilers,” Luke O’Keefe and Karl Sturm (ChevronTexaco), October 28, 2002, p. 8. Attachment 11. 69 See www.ceednet.org/fueling/investing.asp

31

Westmoreland proposes to use lignite coal from an adjacent mine. Lignite was one of the earliest fuels to be demonstrated on gasfication systems (beginning with Dow Chemical’s work in gasification at LTWI facility). Therefore, the design fuel poses no technical barriers for an IGCC plant at the Gascoyne 500 site with an Entrained-Flow design.

An Entrained-Flow IGCC plant uses approximately one-half to two-third less water than a CFB plant.70 Therefore, water use poses no technical barriers for IGCC use at the Gascoyne 500 site. Fluidized-Bed IGCC Systems

Fluidized-Bed gasification systems differ from Entrained-Flow gasifiers in several key respects:

� They are especially effective on low-rank coals such as lignites � They operate at moderate temperature (1700 F- 1900 F) instead of more than

2300 F � They create more “char” which must be recycled. � The capital cost and operating costs are lower on lignite than for Entrained-Flow

gasifiers. GTI markets a Fluidized-Bed technology called U-Gas. The largest U-Gas gasifier operated between 1994 and 2001 in Shanghai, China. The 800 ton/day facility provided industrial fuel. The coal gasification fuel-processing step in IGCC power plants results in superior environmental performance and lower emissions compared to the CFB technology that is proposed for the Gascoyne 500 power plant. Gasifying coal at high pressure prior to combustion facilitates removal of pollutants that would otherwise be released into the air. According to James Childress, “…criteria pollutant emissions for a coal-based IGCC plant are well below those of even the most modern pulverized coal plants with post combustion cleanup.”71 Mercury removal rates of greater than 90 percent can also be achieved using currently available control technologies with IGCC. DOE states that “an IGCC power plant has the potential of achieving very high mercury removal performance with established technology” and mercury removal in an IGCC power plant can be expected to be very high in removal effectiveness, low in cost, and reliable in design.”72

Table 2 summarizes the Gascoyne 500 proposed permit emission rates with permit emission rates for a recently proposed permit for an IGCC plant. For each of the important pollutants in the BACT analysis, IGCC is the top ranked technology.

70 Major Environmental Aspects of Gasification-Based Power Generation Technologies, U.S. DOE/NETL,

December 2002 at page 2-61. Attachment 14. 71 Childress, James M. Statement Submitted for the Record, Senate Environment and Public Works Subcommittee on Clean Air, Wetlands and Climate Change, January 29, 2002. 72 “The Cost of Mercury Removal in an IGCC Plant,” US DOE, NETL, September 2002 at 1-2. Attachment 15.

32

Table 2: Comparison of Emission Rates for Gascoyne 500 to a Recently Proposed

Permit for an IGCC Plant

Gascoyne 500 CFB

Proposed Emission

Rates

Christian County Generation IGCC

Emission Rates*

(lb/MMBtu) (lb/MMBtu)

NOx 0.09 (30-day average) 0.0246 (24-hour average)

VOC 0.005 0.006

PM10 0.012 (filterable) 0.0063

CO 0.154 0.036

Sulfuric Acid Mist 0.0054 0.0026

SO2 0.06 (30-day average) 0.0117 (3-hour average)

Hg 0.0000131273 0.000001974

* All IGCC emission rates for the BACT analysis are based on the November 26, 2007 proposed permit to Christian County Generation for an IGCC facility to be located in Taylorville, Illinois with the exception of the mercury emission rate. As discussed in footnote 74, the mercury emission rate is from the Christian County Generation permit application for the Taylorville IGCC facility. A copy of the proposed permit is included as Attachment 16 to this letter. The limits in the proposed permit are in terms of heat input of the syngas. We converted those limits to be in terms of heat input to the coal for a direct comparison to the proposed Gascoyne 500 emission limits in this table.

For the limits found in Table 2 under baseload conditions, IGCC would yield significantly lower amounts of virtually all pollutants, as well as significantly lower amounts of the climate changing emissions of CO2. We did not address CO2 BACT in the above table but IGCC technology offers opportunity for significant CO2 reduction over conventional coal-fired power plants and at much lower costs. IGCC facilities are typically more efficient than circulating fluidized bedboilers, thus producing less CO2 compared to a CFB boiler producing the same amount of electricity.75 Furthermore, IGCC allows for an option to make even deeper cuts in carbon dioxide that conventional coal plants cannot do. The CO2 in the syngas can be captured and sequestered at a fraction of the cost of post-combustion carbon capture and sequestration at other coal

73 This mercury emission rate was calculated from the Westmoreland’s estimate of controlled mercury emissions divided by the maximum heat input capacity of the Gascoyne 500 facility from data provided in Appendix B-6 of the Gascoyne 500 permit application. 74 This is the mercury emission rate provided in the Christian County Generation permit application for the Taylorville IGCC facility. The facility will be equipped with a mercury removal system (see page 4 of the proposed permit (Attachment 16 to this letter). The Illinois EPA has only included the mercury limit of the New Source Performance Standards in the proposed permit, i.e., 20 x 10-6 lb/MWh. 75 The Gascoyne 500 net heat rate of 10,720 Btu/net kWh is higher than the heat rates for current subcritical pulverized coal boilers, as shown by a comparison of the Gascoyne 500 net heat rate to the net heat rates of current and state-of-the-art facilities producing electricity from coal. See EPA’s Final Report entitled “Environmental Footprints and Costs of Coal-Based Integrated Gasification Combined Cycle and Pulverized Coal Technologies,” at ES-7. Attachment 17.

33

plants. According to the Department of Energy, IGCC has two major advantages over conventional coal combustion technology for more readily capturing the CO2 emissions:

First, the syngas . . .has a very high CO2 concentration, which can be made much higher by further converting the CO to CO2 prior to combustion. Second, IGCC gasifiers typically operate under relatively high pressure (~400 psig in the Wabash plant). Both of these conditions make recovery of the CO2 from the syngas much easier than capture from the flue gas.76

The high concentration of CO2 in the syngas along with the high pressure allows for the use of physical absorption to capture CO2 rather than the more energy intensive chemical absorption processes that need to be used in lower pressure situations.77 Ninety percent of the CO2 can be removed from an IGCC plant with using only 5% of the power from the plant, as compared to a pulverized coal facility that would use 28% of its power to achieve similar levels of CO2 capture.78 In addition, the waste leaving an IGCC plant is vitrified, thereby potentially reducing some of the solid waste disposal issues associated with coal combustion. Indeed, IGCC plants produce 30-50% less solid waste than CFB plants.79 Again, North Dakota has a duty under federal and state law to consider the environmental impacts of the solid waste associated with different technology options. IGCC is clearly an available method, system and technique for producing electricity from lignite coal and thus should have been fully and fairly evaluated in the BACT analysis by NDDH and Westmoreland. For both the Entrained-Flow and Fluidized-Bed gasification systems, NDDH must develop average and incremental costs for each pollutant removed including CO2 and compare these costs to the proposed configuration of the Gascoyne 500 facility. Issuance of a permit for Gascoyne 500 without an adequate analysis of IGCC would be unlawful. V. WESTMORELAND DID NOT EVALUATE A SUPERCRITICAL CFB

BOILER IN THE BACT ANALYSIS Westmoreland should have also considered the construction of a supercritical CFB boiler. Supercritical CFB boilers are more efficient and thus use less fuel and emit less carbon dioxide emissions. This technology is discussed in the Western Governor’s Association Technology Working Group’s report on advanced clean coal technologies (Attachment

76 Major Environmental Aspects of Gasification-Based Power Generation Technologies, US DOE, December 2002, at 2-45 to 2-46. Attachment 14. 77 Id. at 2-46. 78 Id. at 2-48. 79 Id., Table 1-7, page 1-27.

34

18). This technology can achieve net plant efficiencies of 40%.80, 81 NDDH must require evaluation of this inherently lower emitting technology in its BACT review for Gascoyne 500. VI. THE PROPOSED BACT EMISSION LIMITS FAIL TO REFLECT THE

MOST STRINGENT EMISSION LIMITS FOR OTHER CIRCULATING

FLUIDIZED BED POWER PLANTS.

The Proposed NOx Limit Does Not Reflect BACT The proposed NOx BACT emission limit of 0.09 lb/MMBtu (30-day average) does not reflect the maximum degree of reduction that can be achieved at CFB boilers. A review of EPA’s spreadsheet of national coal-fired utility projects indicates most recently submitted applications for CFB boilers are proposing NOx emission limits of 0.07 lb/MMBtu. According to the EPA spreadsheet, these limits are based on CFB boilers with either SNCR or SCR. The EPA spreadsheet on new coal-fired utility projects is also included at Attachment 20, as downloaded from http://www.epa.gov/ttn/catc/products.html#misc. NDDH’s proposed BACT limit for Gascoyne 500 of 0.09 lb/MMBtu reflects only 40% control of NOx from the selective noncatalytic reduction (SNCR) system, considering the stated 0.15 lb/MMBtu NOx emission rate to be emitted from the boiler. NDDH Permit at 59; Gascoyne 500 permit application at 5-9. Yet, NDDH indicated that NOx removal efficiencies as high as 75% could be achieved with SNCR. NDDH Permit Analysis at 60. Westmoreland indicated in their NOx BACT analysis that SNCR systems have been designed to achieve 40-60% removal of NOx. Gascoyne 500 permit application at 5-10. Both NDDH and Westmoreland have stated higher levels of control could be achieved with a SNCR system, but NDDH and Westmoreland only evaluated SNCR at a level of control of 40% NOx removal. At 60% removal, a NOx emission rate of 0.06 lb/MMBtu could be met. Neither NDDH or Westmoreland provided information to adequately support why higher percent removal efficiencies could not be met at the Gascoyne 500 plant with SNCR. For example, could higher percent removal efficiencies/lower emission rates be achieved with a higher ammonia injection rate or with the addition of other chemicals to improve the performance of the SNCR system?82 In fact, there are several proposed or final permits for coal-fired CFB boilers that will be equipped with SNCR and will be required to meet lower NOx emission limits than what has been proposed for Gascoyne 500 which indicates the likelihood that higher NOx removal efficiencies could be achieved with SNCR at Gascoyne 500. Westmoreland 80 Goidich, Stephen J. et al, Design Aspects of the Ultra-Supercritical Boiler, presented at the International Pittsburgh Coal Conference, Pittsburgh, PA, September 12-15, 2005. Attachment 19. 81 See also http://www.power-technology.com/contractors/boilers/foster_wheeler/; http://www.iea.org/Textbase/techno/iaresults.asp?id_ia=18 for further information on efficiencies of supercritical fluidized bed boilers. 82 See Institute of Clean Air Companies website, which states that SNCR performance can be enhanced with the addition of other chemicals which can “improve performance, reduce equipment maintenance, and expand the temperature window within which SNCR is effective.” See discussion of SNCR at http://www.icac.com/.

35

pointed out three recently permitted waste coal-fired CFB facilities in Pennsylvania that will be equipped with SNCR systems, and two of the three Pennsylvania facilities have lower NOx emission limits Gascoyne 500 Permit Application at 5-27. See also NDDH Permit Analysis at 62. Specifically, the Greene Energy and Beech Hollow CFB plants are required to meet a NOx BACT limit of 0.08 lb/MMBtu. While Westmoreland attempted to discount these lower NOx limits based on the fact that these facilities are subject to lowest achievable emission rate (LAER) requirements or that the facilities are burning waste coal, neither of Westmoreland’s claims are sufficient to justify the exclusion of such lower emission limits from the NOx BACT review for Gascoyne 500. Indeed, EPA’s New Source Review Workshop Manual specifically indicates that technologies and levels of control required as LAER should be reviewed in the BACT analysis.83 Further, rather than trying to discount these lower NOx limits by showing that North Dakota lignite is different from the waste coal to be utilized at these Pennsylvania facilities, NDDH and Westmoreland must evaluate the degree of NOx reduction that is required of the SNCR systems at these Pennsylvania facilities to meet the applicable NOx emissions limits. If such facilities will need to operate their SNCR systems to achieve greater than 40% control, then that provides strong justification for Westmoreland to evaluate such higher levels of control in its BACT analysis for Gascoyne 500. The proposed Bonanza waste coal-fired unit is another new waste coal-fired CFB boiler with a lower NOx BACT emission limit. EPA has proposed a permit for this facility with a NOx BACT limit of 0.08 lb/MMBtu, and the assumed NOx emission rate exiting the boiler was 0.15 lb/MMBtu similar to that expected from Gascoyne 500.84 Thus, the proposed NOx BACT limit at the Bonanza Waste Coal-Fired Unit reflects a 47% reduction in NOx emissions from the SNCR. In its draft Statement of Basis for the Bonanza Waste Coal-Fired Unit, EPA identifies other CFB facilities that are or will be equipped with SNCR systems with proposed BACT limits or emission rates lower than 0.09 lb/MMBtu. This includes the Highwood Generating Station which is subject to a 0.07 lb/MMBtu BACT limit on an annual average. The projected NOx emissions exiting the Highwood CFB boiler were identified as 0.14 lb/MMBtu, which means that the 0.07 lb/MMBtu BACT limit reflects a 50% reduction in emissions across the SNCR.85 EPA also identified several CFB units equipped with SNCR systems that are actually meeting NOx emission rates lower than 0.07 lb/MMBtu.86 Thus, all of the information provided above and in the documents attached to this letter as well as in the Gascoyne 500 permit application and NDDH’s Permit Analysis provide sufficient documentation to show that lower NOx emission rates can be met with an SNCR system and thus such rates should have been evaluated by NDDH and

83 See EPA’s October 1990 draft New Source Review Workshop Manual at B.5., B.6. 84 See EPA’s June 14, 2006 Draft Statement of Basis for Bonanza Power Plant, Waste Coal-Fired Unit, at 34-35, 39. Attachment 21. 85 See 11/30/05 Highwood Generating Station Permit Application at 5-32. Attachment 22 86 See EPA’s June 14, 2006 Draft Statement of Basis for Bonanza Power Plant, Waste Coal-Fired Unit, at 39.

36

Westmoreland in a top-down NOx BACT analysis for Gascoyne 500. It must be noted that the cost of a particular control technology cannot be considered as unreasonable if other facilities have been required to bear similar costs as part of a BACT determination.87 Considering that other recently proposed CFB power plants have proposed or have been required to meet higher levels of NOx control and lower NOx emission limits, meeting a similar lower NOx limit at Gascoyne 500 must be considered economically reasonable. In evaluating SCR at Gascoyne 500, an emission rate of 0.07 lb/MMBtu was assumed. That only reflects an emission reduction from SCR of 53%, when SCR can achieve 80-90% NOx removal efficiencies. Thus, the cost analysis must be redone based on an emission limit that reflects the maximum emissions reduction likely from SCR. Had Westmoreland assumed a more appropriate NOx emission rate that reflects the maximum degree of emission reduction achievable with an SCR system, the cost of the SCR system would have been significantly less (30-50% lower). Further, more evaluation must be done by Westmoreland and NDDH regarding high dust and/or low temperature catalysts as part of the BACT analysis. For example, the CleanAIR ENDURE™ works in exhaust gas temperatures as low as 300° F (which is not significantly higher than the 270º F temperature expected of the exhaust gas exiting the particulate control device (Gascoyne 500 Permit Application at 5-13)). This catalyst also can be installed at the tail end (low dust) position after the particulate matter control equipment. If such a catalyst could be used at Gascoyne 500, there would be more options for reheating the exhaust gas from 270º F to 300º F, and the associated costs could be much lower than assumed in the SCR cost analysis (for which Westmoreland assumed the gas would have to be reheated from the 270º F to 650º F). A copy of the brochure for CleanAIR ENDURE™ is available at www.cleanairsys.com/products/scr/ENDURE-Power-Plant.pdf. There are also catalyst options available for high dust gas streams.88 Westmoreland and/or NDDH should investigate these and other concepts further with vendors to evaluate the various options that are available to successfully transfer the highly effective SCR NOx removal systems to a CFB boiler such as what is planned at Gascoyne 500. The BACT analysis cannot be considered complete without a thorough evaluation of SCR. Thus, additional evaluations as discussed above are needed of both SNCR and SCR for reducing NOx emissions at Gascoyne 500 before NDDH can properly determine NOx BACT for Gascoyne 500. Such additional research and analysis must be done to ensure that the NOx BACT limit reflects the maximum degree of reduction of NOx emissions that can be achieved at Gascoyne 500.

87 See EPA’s October 1990 Draft New Source Review Workshop Manual at B.44. 88 See, e.g., Scot Pritchard et al., Optimizing SCR Catalyst Design and Performance for Coal-fired Boilers, Presented at EPA/EPRI 1995 Joint Symposium, Stationary Combustion NOx Control, available at www.cormetech.com/brochures/OptimizingSCRCatalystDesignPerformanceCoalFiredBoilers.pdf.

37

The Proposed SO2 Limit Does Not Reflect BACT NDDH and Westmoreland eliminated coal washing and fuel switching because they claimed none of these controls would result in a significant reduction in SO2 emissions and that the controls were not economical. See NDDH Permit Analysis at 32-33. However, coal washing and fuel switching or fuel blending would reduce SO2 emissions to some extent. The uncontrolled SO2 emission rate from the Gascoyne 500 lignite is quite high at 4.55 lb/MMBtu, as compared to other nearby western coals such as the low sulfur coals of the Powder River Basin. Considering the problems with complying with the SO2 PSD increment in the region and the facility’s proximity to the Theodore Roosevelt National Park Class I area, every reduction measure for SO2 should be analyzed in the BACT analysis. Thus, NDDH must evaluate coal washing, coal switching as well as coal blending, along with add-on SO2 controls, in its top down BACT process to ensure that the SO2 BACT limit for Gascoyne 500 truly reflects the maximum degree of reduction in SO2 emissions that can be achieved. As discussed in the Gascoyne 500 permit application, Westmoreland assumed a much higher maximum uncontrolled SO2 emission rate at Gascoyne 500 than the maximum uncontrolled emission rate expected at the Gascoyne 175 MW plant, even though both mine-mouth facilities will obtain their coal from the same mine. Gascoyne 500 permit application at 5-55. The maximum uncontrolled SO2 emission rate for the Gascoyne 175 MW facility was identified as 3.48 lb/MMBtu whereas the maximum uncontrolled SO2 emission rate for Gascoyne 500 was identified as 4.55 lb/MMBtu89. Apparently this is because “[t]he additional fuel needed for the larger [Gascoyne 500] power plant will limit the mine’s ability to bypass certain coal seams with slightly higher sulfur." Id. These discrepancies in coal characteristics of the mine clearly indicate that the average sulfur content of the Gascoyne 500 coal will be much lower than what was assumed for the maximum sulfur content. These wide variations in the uncontrolled SO2 emissions expected from the Gascoyne lignite simply point to the need for NDDH to impose a percent SO2 reduction requirement, in addition to an SO2 BACT emission limit, to ensure that the SO2 control equipment is consistently operated to achieve the maximum degree of reduction in SO2 emissions that is achievable. Such a SO2 control efficiency requirement would ensure proper operation and maintenance of the spray dryer regardless of the sulfur content in the coal. The compliance issues with the SO2 increments at Theodore Roosevelt National Park and other nearby Class I areas as well as the significant visibility impacts of Gascoyne at nearby Class I areas provide further basis for such a removal efficiency requirement. EPA Region VIII made this same comment to the Montana Department of Environmental Quality pertaining to the recently issued Roundup Power Plant PSD permit. A copy of the EPA’s December 18, 2002 comment letter is included at Attachment 23 to this letter. Further, the worst case coal quality should not be used to establish a long term averaging time (such as 30-day average) BACT limit. EPA commented on this in the attached December 2002 letter to the Montana Department of Environmental Quality, stating “[w]hile use of the worst-case coal scenario might be appropriate for establishing a short-

89 See Gascoyne 500 permit application at 5-55.

38

term (3-hour or 24-hour) SO2 emission limit, we consider it inappropriate for establishing a 30-day average emission limit, especially considering that coal blending can be used at minimal additional cost (and is routinely used in the power plant industry) to eliminate or reduce the effect of coal sulfur ‘spikes.’”90 In addition, in evaluating the SO2 removal efficiency expected from the CFB boiler, Westmoreland improperly assumed only 88% of the SO2 would be removed in the boiler. This does not reflect the maximum degree of SO2 reduction that can be achieved in the CFB boiler. For the Gascoyne 175 MW project, it was assumed that the CFB boiler would remove 90% of the SO2 in the coal91, and the maximum expected uncontrolled SO2 emission rate of the coal was 30% lower than claimed for Gascoyne 500 as discussed above. Similarly, 90% SO2 removal was assumed to occur in the South Heart boiler with an expected maximum uncontrolled SO2 emission rate of 3.92 lb/MMBtu.92 These two permit applications (and the final permit for Gascoyne) provide strong support that 90% SO2 removal is achievable across the CFB at the Gascoyne 500 facility. Therefore, the baseline for evaluating BACT and ultimately setting the BACT limit must reflect 90% SO2 removal across the CFB boiler and a worst case maximum SO2 emission rate exiting the boiler of 0.455 lb/MMBtu. Further, NDDH and Westmoreland failed to evaluate the maximum degree of SO2 emission reduction that can be achieved with either the wet flue gas desulfurization (FGD) system or the dry FGD system that were evaluated as BACT at the Gascoyne 500. Westmoreland only evaluated 91% SO2 removal with a wet FGD system.93 Yet, information submitted to the Utah Division of Air Quality by Sargent & Lundy indicates that between 92-93% SO2 removal efficiency could be achieved in a wet FGD system with an SO2 inlet concentration of 0.564 lb/MMBtu (i.e., Westmoreland’s stated SO2 emission rate expected from the Gascoyne 500 CFB boiler).94 That maximum expected SO2 removal efficiency would result in SO2 emission rates ranging from 0.04 to 0.045 lb/MMBtu. Such emission rates must be evaluated by Westmoreland in its BACT analysis of the wet FGD system for Gascoyne 500. With respect to a dry FGD system, Westmoreland evaluated an SO2 emission limit that only reflected an 86.8% SO2 removal efficiency across a spray dry absorber.95 This does not reflect the maximum degree of emission reduction that can be achieved across a spray

90 See December 18, 2002 letter from Richard R. Long, EPA Region 8, to Steve Welch, Montana Department of Environmental Quality, at 2. (Attachment 23). 91 See May 14, 2004 Gascoyne 175 permit application at 5-11, Attachment 24. 92 See August 2005 South Heart Power Plant permit application at 4-36, Attachment 25. 93 This percent removal efficiency was calculated based on Westmoreland’s claimed SO2 emission rate exiting the Gascoyne 500 CFB boiler of 0.564 lb/MMBtu and its proposed SO2 emission rate achievable with a wet FGD system of 0.05 lb/MMBtu. See Gascoyne 500 permit application at 5-34 and 5-41. 94 See attached excerpt from the November 18, 2003 Sargent & Lundy Technical Memorandum submitted to the Utah Division of Air Quality entitled “Chart 1 WFGD Control Efficiency as a Function of Uncontrolled SO2 Emission Rate” regarding the IPP Unit 3 Air Permit Application SO2 Control Efficiency. Attachment 26. 95 This percent removal efficiency was calculated based on Westmoreland’s claimed SO2 emission rate exiting the Gascoyne 500 CFB boiler of 0.564 lb/MMBtu and its proposed SO2 emission rate achievable with a spray dry absorber of 0.06 lb/MMBtu. See Gascoyne 500 permit application at 5-34 and 5-44.

39

dry absorber at Gascoyne 500. Ninety percent removal across a spray dryer was evaluated as BACT in the South Heart Power Plant permit application, with an inlet SO2 concentration to the spray dry absorber of 0.392 lb/MMBtu.96 In the Gascoyne 175 MW permit application, it was assumed that the spray dry absorber would remove 89.1% of the SO2 emissions with an inlet SO2 emission rate of 0.348 lb/MMBtu.97 There is ample indication in PSD permit applications that spray dry absorbers can achieve at least 90 percent SO2 removal even with low sulfur coals. Indeed, as found by EPA in its review of the New Source Performance Standards for coal-fired electric utility boilers, a spray dryer/absorber can generally achieve greater than 90 percent SO2 removal. See 70 Fed.Reg. 9711 (February 28, 2005). Even Westmoreland admitted that 90 percent SO2 removal could be achieved with a spray dry absorber. See Gascoyne 500 permit application at 5-59. Thus, NDDH must evaluate an SO2 BACT limit that reflects at least 90% SO2 removal across the spray dry absorber. Assuming 90% SO2 removal across the CFB boiler and an additional 90% SO2 removal efficiency across the spray dry absorber (which is what was assumed in the South Heart Power Plant SO2 BACT evaluation) would equate to a BACT emission limit for this control configuration of 0.046 lb/MMBtu. NDDH and Westmoreland also did not evaluate the maximum degree of SO2 reduction achievable with a circulating dry scrubber. Westmoreland assumed that a circulating dry scrubber would achieve the same 86.8% SO2 removal efficiency as assumed for the spray dry absorber (Gascoyne 500 permit application at 5-45). Yet, Westmoreland provided no documentation to support this assumed SO2 control efficiency for a circulating dry scrubber and even stated that vendors have indicated that circulating dry scrubbers can achieve higher SO2 removal efficiencies than spray dry absorbers. Indeed, the Mustang Power Company plans to use a circulating dry scrubber at its proposed New Mexico power plant, and it has indicated SO2 removal of 97% can be achieved. Thus, NDDH must evaluate the maximum degree of SO2 emission reduction that can be achieved across a circulating dry scrubber in its SO2 BACT analysis for Gascoyne 500. NDDH and Westmoreland failed to justify the proposed SO2 BACT emission limit of 0.06 lb/MMBtu in light of the lowest SO2 emission rate required at any CFB boiler – i.e., the 0.022 lb/MMBtu SO2 emission rate that has been required as BACT at the AES-Puerto Rico plant, which is equipped with a circulating dry scrubber. This facility is required to burn low sulfur coal (1% or less) and meet a 0.022 lb/MMBtu SO2 limit on a

three-hour average. Based on the worst-case coal quality to be used at AES-Puerto Rico (0.8% and 12,000 BTU/lb), the uncontrolled SO2 emission rate of AES-Puerto Rico is 1.6 lb/MMBtu, which is almost one-third of the worst case 4.55 lb/MMBtu uncontrolled SO2 emission rate expected from the Gascoyne coal. Yet, the AES-PR facility is required to meet a limit that requires an overall 98.6% reduction of SO2 based on worst case coal, on a three-hour averaging time. Thus, the AES-PR facility provide a pertinent example that high levels of SO2 control can be achieved even with low sulfur coal and on a short term averaging time. The SO2 BACT limit of the AES-PR permit also supports the need for NDDH and Westmoreland to fully evaluate coal washing, switching and blending in its

96 See August 2005 South Heart Power Plant permit application at 4-36. Attachment 25. 97 See May 14, 2004 permit application for the original Gascoyne 175 MW unit at 5-11. Attachment 24.

40

BACT analysis). A copy of the AES-PR permit and other relevant documents are including in Attachment 27 of this letter. Westmoreland apparently took into account a “safety margin” in proposing its 0.06 lb/MMBtu 30-day average SO2 BACT emission limit, which NDDH is apparently accepting. Gascoyne 500 permit application at 5-59. However, because other similar facilities have proposed or been required to meet lower SO2 emission limits and meet higher levels of SO2 control even with shorter averaging times, NDDH and Westmoreland do not have adequate justification for including a safety margin and/or for proposing a higher SO2 BACT emission limit without providing a reasoned technical basis as to why lower limits/higher levels of SO2 removal cannot be achieved at Gascoyne 500. NDDH has failed to address these requirements in its proposed SO2 BACT determination. Thus, the SO2 BACT analysis is flawed and NDDH’s proposed SO2 BACT limit fails to reflect the maximum degree of SO2 emission reduction that can be achieved at Gascoyne 500. The PM10 Emission Limits Also Do Not Reflect BACT The proposed filterable PM10 BACT limit for Gascoyne 500 is 0.012 lb/MMBtu. This limit does not reflect the maximum emission reductions that can be achieved with a fabric filter baghouse. First, NDDH and Westmoreland had failed to provide adequate justification that Gascoyne 500 could not meet the lowest filterable PM10 BACT limit currently proposed or required for a CFB boiler – 0.01 lb/MMBtu. In its permit application, South Heart proposed a 0.01 lb/MMBtu PM10 limit as BACT. See August 2005 South Heart permit application at 4-52 (Attachment 25). South Heart will be a 500 MW CFB facility burning North Dakota lignite equipped with SNCR, a spray dry absorber, and a fabric filter baghouse – virtually identical to Gascoyne 500. South Heart projects a higher uncontrolled PM10 emission rate than predicted for Gascoyne 500 (14.82 lb/MMBtu (see South Heart permit application at 4-46)) as compared to the uncontrolled PM10 emissions expected at Gascoyne 500 (11.54 lb/MMBtu (see Gascoyne 500 permit application at 5-70)), and South Heart proposed a PM10 BACT emission limit based on a higher PM10 removal efficiency than assumed for Gascoyne 500 - i.e., 99.93% (see South Heart permit application at 4-49) as compared to 99.89% (see Gascoyne 500 permit application at 5-67)). Thus, if South Heart claims that a PM10 BACT limit of 0.010 lb/MMBtu is achievable, it must also be considered achievable for the very similar Gascoyne 500 facility. Further, NDDH’s decision to eliminate a baghouse with specialty bags that could achieve 99.93% PM10 reduction based on cost is unjustified when South Heart has proposed such a level of PM10 control as BACT. In addition, it must be noted that Westmoreland did not provide adequate information to distinguish the Gascoyne 500 facility’s PM10 emissions and achievable BACT emission limits from the PM10 emissions at Northampton, JEA Northside, and York County facilities, all of which have lower PM10 BACT limits than currently proposed for Gascoyne 500. NDDH and Westmoreland should have evaluated the degree of PM10 reduction that must be achieved at these facilities before it can conclude that Gascoyne 500 is sufficiently different from

41

these facilities to warrant a higher PM10 BACT limit. In any case, given what has been proposed for South Heart, there does not appear to be any adequate justification for the Gascoyne 500 PM10 BACT limit to be any higher than 0.01 lb/MMBtu. Further, even lower PM10 emission rates can be achieved in practice. As discussed in detail in Environmental Defense et al.’s April 29, 2005 comments to EPA on its proposed revisions to the coal-fired utility boiler new source performance standards (NSPS), many facilities are achieving PM10 emission rates much lower than 0.01 lb/MMBtu. Those comments and the relevant attachments are included in Attachment 28 of this letter. Importantly, stack test data for the Northampton Generating facility, the only other facility listed in the Gascoyne 500 permit application as having a BACT limit of 0.010 lb/MMBtu, is actually achieving a total PM (filterable plus condensable) emission rate of 0.0045 lb/MMBtu. This information on the emission rates actually being achieved in practice must be evaluated and considered by NDDH in the Gascoyne 500 BACT analysis for PM10. Thus for the all of the above reasons, NDDH’s proposed filterable PM10 BACT limit for Gascoyne 500 PSD does not reflect BACT and NDDH’s proposed PM10 BACT determination is unjustified. The H2SO4 Limit Does Not Reflect BACT NDDH and Westmoreland proposed a BACT limit for H2SO4 BACT of 0.0054 lb/MMBtu that is reflective of an overall 93% control of this pollutant. NDDH Permit Analysis at 72; Gascoyne 500 permit application at 5-90. Yet, 95% H2SO4 control was indicated as the maximum reduction achievable with a CFB boiler and a spray dry absorber at the Gascoyne 175 MW facility (see January 2005 Gascoyne Power Project 175 MW CFB BACT Determination at C-59, Attachment 29). Further, there are several facilities with lower H2SO4 BACT limits as discussed in the Gascoyne 500 permit application (at 5-90) and in the NDDH Permit Analysis (at 71). Thus, these levels of H2SO4 emissions and control should have been evaluated for the Gascoyne 500 facility. If the most appropriate H2SO4 BACT limit is below the detection limit of the methodology used to measure H2SO4 as claimed by Westmoreland (Gascoyne 500 permit application at 5-90), then NDDH must consider other operational limitations in addition to (and not in lieu of) an H2SO4 emission limit to ensure the maximum degree of H2SO4 emission reduction is achieved. The best way to ensure this may be through an SO2 removal efficiency requirement across the spray dryer, which we contend must be imposed as part of the SO2 BACT requirements to ensure that the maximum achievable degree of SO2 emission reduction is continuously met. The Fluorides Limit Does Not Reflect BACT NDDH and Westmoreland proposed a BACT limit for fluorides (as HF) BACT of 0.0017 lb/MMBtu that is reflective of an overall 92% control of this pollutant. NDDH Permit Analysis at 73; Gascoyne 500 permit application at 5-91. Yet, Westmoreland indicates that HF removal efficiencies “in the range of 95%” can be met with a dry FGD system

42

and a fabric filter baghouse. Gascoyne 500 permit application at 5-90 to 5-91. Further, the fluorides BACT limit for the original Gascoyne 175 MW unit is 0.00053 lb/MMBtu (3-hour average), which is less than half of the fluorides BACT limit proposed by Westmoreland for Gascoyne 500. (MDU/Westmoreland Power Permit to Construct for Gascoyne Generating Station, Condition II.A.1. (at 5), Attachment 30 to this letter). Thus, the proposed fluorides limit does not appear to reflect the maximum degree of fluoride reduction that is achievable at Gascoyne 500. The Total PM10 Limit Does Not Reflect BACT NDDH and Westmoreland proposed a total PM10 limit (filterable plus condensibles) of 0.025 lb/MMBtu as BACT. See NDDH Permit Analysis at 74; Gascoyne 500 permit application at 5-94. However, this proposed limit is too high and fails to reflect the maximum degree of total PM10 emission reduction achievable at Gascoyne 500. Specifically, as discussed above, the proposed BACT limits for filterable PM10, H2SO4 and fluorides are all too high and do not reflect BACT. If more appropriate BACT limits were proposed for these pollutants (no higher than 0.010 lb/MMBtu for filterable PM10, 0.0039 lb/MMBtu (95% control) for H2SO4, and 0.0011 lb/MMBtu (95% control) for fluorides), then total the total PM10 limit would be reduced to 0.02 lb/MMBtu. It is also likely that the assumed levels of control and emission rates for the other condensable particulates (HCl, (NH4)2SO4, and condensable organics) should be lower. The actual total PM10 emissions data at Northampton that is attached to this letter (in Attachment 28) also makes the proposed total PM10 limit of 0.025 lb/MMBtu appear incredibly high. As stated above, the Northampton waste coal facility is actually achieving a total PM10 rate of 0.0045 lb/MMBtu. Neither NDDH or Westmoreland provided any data on total PM10 BACT limits at other facilities, but the proposed total PM10 BACT limit is clearly much higher than recently issued permits for coal-fired power plants. For example, the October 12, 2004 PSD permit for the Sevier power plant in Utah, a coal-fired CFB boiler with similar PM controls as proposed at Gascoyne 500, has a total PM10 emission limit of 0.0154 lb/MMBtu. See Attachment 31 to this letter. The AES-Puerto Rico permit, also a coal-fired CFB boiler with similar PM controls, requires that a total PM10 limit of 0.015 lb/MMBtu be met. See Attachment 27 to this letter. Westmoreland needs to evaluate these lower limits in a proper total PM10 BACT analysis such that the total PM10 BACT limit reflects the maximum emission reduction achievable. It must be noted that a total PM10 BACT limit is a requirement for all PSD permits. Because total PM10 emissions must be modeled to assess a facility’s compliance with the PM10 national ambient air quality standards (NAAQS) and prevention of significant deterioration increments, as well as to assess a facility’s impact on visibility, it is imperative that an enforceable total PM10 BACT limit be required. See EPA’s March 31, 1994 letter to the Iowa Department of Natural Resources on this matter (Attachment 32). NDDH must ensure that the total PM10 BACT limit reflects the maximum degree of reduction in all PM10 pollutants that can be achieved at Gascoyne 500. For all of the

43

reasons discussed above, NDDH’s proposed total PM10 BACT limit for Gascoyne 500 does not satisfy BACT requirements for PM10. The BACT Limits Must Be Met on a Continuous Basis and Meet Enforceability Criteria All BACT limits must be met on a continuous basis and must meet enforceability criteria, but NDDH’s proposed permit for Gascoyne 500 does not adequately address EPA requirements regarding these requirements. Specifically, as discussed in EPA's October 1990 Draft New Source Review Workshop Manual, "BACT emission limits or conditions must be met on a continual basis at all levels of operation (e.g., limits written in lb/MMBtu or percent reduction achieved), among other things. (NSR Workshop Manual at B.56). Yet, NDDH has proposed to allow the lb/MMBtu BACT limits to not apply during startup and shutdown. Proposed Gascoyne 500 permit condition II.A.1. EPA’s January 28, 1993 guidance memo entitled “Automatic or Blanket Exemptions for Excess Emissions During Startup, and Shutdowns Under PSD” (available at http://www.epa.gov/region07/programs/artd/air/nsr/nsrpg.htm) specifically disallows automatic exemptions from BACT emission limits and instead informs states to use enforcement discretion in determining whether to enforce for violations of BACT emission limits. EPA’s policy also indicates that alternative emission limits for startup and shutdown “could effectively shield excess emissions arising from poor operation and maintenance or design, thus precluding attainment.” EPA’s January 28, 1993 guidance memo at 3. Thus, NDDH must require that the BACT limits for Gascoyne 500 be met at all times. With respect to SO2, a percent SO2 reduction requirement is necessary in addition to the proposed SO2 emission limits to ensure that high levels of SO2 control reflective of BACT are continually met when the facility is burning coal with sulfur content lower than the maximum expected uncontrolled SO2 emission rate upon which the proposed BACT limit is based. Such a percent removal efficiency may also be necessary to ensure that the maximum degree of H2SO4 reduction is continually achieved. Both of these issues were discussed in our comments above. The permit application must also specify appropriate compliance methods and recordkeeping requirements to show compliance with these emission limits. As discussed in the NSR Workshop Manual, “the construction permit should state how compliance with each limitation will be determined.” (See NSR Workshop Manual at H.6.). Further, the test methods must provide for continuous compliance where feasible. NDDH has not met these requirements in its proposed permit for Gascoyne 500, because the permit does not specifically indicate that a continuous opacity monitoring system (COMS) must be used to demonstrate compliance with the 10% opacity BACT limit (in Table 3 of the proposed permit). Further, it is not appropriate to exempt Gascoyne 500 from utilizing COMS if Westmoreland decides to utilize a PM continuous emission monitoring systems (CEMS). Condition II.A.14.a.1) of the proposed permit. Both a visible emissions limit and a PM10 limit are required under the definition of BACT, and

44

compliance method must provide for continuous compliance if possible. Clearly it is possible for Gascoyne 500 to be equipped with both COMS and PM CEMS. In addition, the permit must also require that Westmoreland monitor heat input to the Gascoyne 500 boilers. Further, the process description in Section I.D. of the proposed permit must be more definitive in limiting maximum heat input capacity to 5,799 MMBtu/hr, to ensure that Westmoreland cannot take any action that would increase the maximum heat input capacity of the boilers without a permit modification. NDDH must address all of the above issues in the Gascoyne 500 permit to ensure enforceability of the permit and to ensure BACT limits are met on a continuous basis. VII. THE GASCOYNE 500 PERMIT APPLICATION FAILED TO ADDRESS PM2.5 AS A PSD POLLUTANT. Under 40 C.F.R. § 52.21(b)(2), a major modification is any physical change in or change in the method of operation of a major stationary source that would result in: a significant emissions increase and a significant net emissions increase of any regulated NSR

pollutant. The regulations, 40 C.F.R. § 52.21(b)(50), define “regulated NSR pollutant” to mean, among other things, “[a]ny pollutant for which a national ambient air quality standard has been promulgated and any constituents or precursors for such pollutants identified by the Administrator (e.g., volatile organic compounds and NO[x] are precursors for ozone).” EPA has promulgated a NAAQS for PM2.5. 62 Fed. Reg. 38,652 (July 18, 1997). The regulations list significance levels for a number of “regulated NSR pollutants,” but not PM2.5. 40 C.F.R. § 52.21(b)(23)(i). When a significance level has not been identified for a regulated NSR pollutant, the significance level is any emission rate over zero. 40 C.F.R. § 52.21(b)(23)(ii). These federal regulations have all been incorporated by reference into North Dakota’s regulations at NDAC §33-15-15-01.2. Although there is no analysis in the application regarding PM2.5, a facility of this size will undoubtedly be emitting it in substantial amounts. Consequently, Westmoreland is required to comply with all PSD requirements, including monitoring, modeling, and BACT regarding PM2.5, and NDDH cannot issue a PSD permit for this facility unless this pollutant is properly addressed. We are aware that EPA issued guidance providing that sources would be allowed to use implementation of a PM10 program as a surrogate for meeting PM2.5 NSR requirements. John Seitz, “Interim Implementation for the New Source Review Requirements for PM[2.5],” (October 23, 1997). The purpose of that guidance was to provide time for the development of necessary tools to calculate the emissions of PM2.5 and related precursors, adequate modeling techniques to project ambient impacts, and PM2.5 monitoring sites. 70 Fed. Reg. 65984, 66043 (Nov. 1, 2005). EPA has resolved most of these issues. Id. More importantly, the guidance clearly contravenes the regulations. In order to protect public health and the environment, the regulations must be implemented as written.

45

VIII. NO PRECONSTRUCTION AMBIENT MONITORING DATA WAS

PROVIDED IN THE PERMIT APPLICATION OR UTILIZED IN THE

MODELING Although the Gascoyne 500 facility’s predicted ambient impacts are greater than the PSD monitoring significance levels for PM10 and SO2 (Gascoyne 500 permit application at 7-22 and 7-24), NDDH did not require any preconstruction monitoring. NDDH Permit Analysis at 180. NDDH apparently allowed data from Dickinson, Beulah, Bismarck, and Theodore Roosevelt National Park to be used to assessing background concentrations, but no information was provided in NDDH’s Permit Analysis on how the background concentrations were derived, the dates of the monitoring or the monitoring results. Both federal and state PSD permitting regulations require such preconstruction monitoring, and NDDH must not exempt Westmoreland from these substantive requirements. With respect to the existing PM10 monitoring, there must be a thorough demonstration showing that the existing PM10 monitoring data is reflective of current PM10 concentrations in the location of the proposed Gascoyne 500 power plant in order for preconstruction monitoring to be waived. Westmoreland must be required to conduct preconstruction ambient monitoring for all pollutants for which it will have a significant ambient impact pursuant to 40 C.F.R. §52.21(i)(8) as incorporated by reference into North Dakota regulations at NDAC 33-15-15-01.2. This includes ozone monitoring. As required by 40 C.F.R. §52.21(i)(5)(i)(e), footnote 1, (as revised November 29, 2005, 70 Fed.Reg.71611), Westmoreland is required to conduct preconstruction monitoring for ozone because it would have a net emissions increase of NOx greater than 100 tons per year. Consequently, Westmoreland must be required to conduct one year of preconstruction ozone monitoring. Such monitoring is necessary in part for the NAAQS modeling to accurately reflect background concentrations. The modeling analyses for Gascoyne 500 cannot be considered complete until the monitoring has been conducted and appropriate background concentrations determined. IX. THE GROWTH ANALYSIS IS INADEQUATE Pursuant to 40 C.F.R. §52.21(o) (incorporated by reference into North Dakota’s regulations at NDAC 33-15-15-01.2), Westmoreland is required to provide an air quality impact analysis as a result of the general commercial, residential, industrial and other growth associated with the Gascoyne 500 facility including the mine. 40 C.F.R. §52.21(o)(2). Westmoreland is also required to assess impacts on visibility and soils and vegetation as a result of the source and expected growth associated with the facility and operation of the facility. 40 C.F.R. §52.21(o)(1). As discussed in the U.S. EPA’s October 1990 draft New Source Review Workshop Manual (at D.3 to D.4), this analysis must include an assessment of the amount of residential growth the source will bring to

46

the area, which depends on the number of new employees and the available workforce in the area as well as housing. The analysis must then evaluate the associated commercial and industrial growth associated with the new employees as well as the Gascoyne 500 facility. Further, the analysis must evaluate the growth due to construction activities and mobile sources, both permanent and temporary. Westmoreland did not adequately meet these criteria in its analysis of the growth associated with the proposed Gascoyne 500 power plant. Westmoreland provided a few sentences in its Gascoyne 500 permit application to meet this requirement. Gascoyne 500 permit application at 9-1. Without providing any estimate of the workforce needed for either construction or operation, Westmoreland concluded that the project “is only expected to induce a small amount of growth in the air basin.” Id. NDDH proposes to accept Westmoreland’s very limited description of expected growth based on its own review of the growth that has occurred in Beulah, North Dakota near which other large power plants and mines have located since the 1980’s. NDDH Permit Analysis at 161. However, NDDH’s analysis of changes in population at Beulah does not meet the requirement for Westmoreland to estimate the expected growth from the Gascoyne facility and mine under the PSD regulations. Westmoreland must provide more detailed information on the number of people needed for construction and for operation of the Gascoyne 500 facility (including the mine) and must project the expected short term and long increase in local population, the associated new residences, mobile source growth, and other expected commercial or industrial growth.

Westmoreland’s statements conflict with previous statements made regarding the Gascoyne project. In a March 21, 2001 news release of the Lignite Vision 21 Program about the original Gascoyne power plant (which was, at that time, originally planned to be a 500 MW power plant), it is stated that “a 500 megawatt power plant would result in the production of an additional 3 million tons of coal, 1,300 new jobs and $6 million annually in tax revenue.” This news release further states that “[a]ccording to Westmoreland Power, Inc., the Gascoyne site offers a number of attractive features for a $700+ million, 500 megawatt power plant, ‘including low-cost fuel supply, favorable environmental characteristics, and an opportunity for economic development in an area

of the state that currently does not enjoy the economic benefits of the lignite industry.’” [Emphasis added.] See http://www.lignitevision21.com/030101news.htm. A copy of this press release is included as Attachment 36 to this letter. A May 14, 2004 press release by Governor Hoeven’s office about the Gascoyne 175 MW facility indicated that the facility and associated mine “would create 100 full-time jobs as well as numerous ancillary jobs, because of the services needed for a facility of its size.” See http://governor.state.nd.us/media/news-releases/2004/05/040514b.html, copy of press release in Attachment 37 to this letter. At the June 21, 2007 public hearing on the proposed Gascoyne 500 permit, it was indicated that there will be about 120 jobs in mining and 84 plant jobs and that most of the plant jobs will be “highly skilled.”

Based on these public statements, the Gascoyne 500 facility and mine would necessitate somewhere between 100-1300 employees for operation. The population of Gascoyne in

47

2000 was 23. Clearly, any increase in population of the area would be significant. According to the Bowman County internet site, there was an available workforce in 2003 of 1,763 people but the unemployment rate was only 1.8%. See http://www.bowmannd.com/demographics.html. Thus, the number of people available in the county for work at the Gascoyne 500 power plant and mine would be roughly 32. Clearly, then, there will be significant population growth as a result of people moving to, or commuting to, the area to work at the Gascoyne facility. And there will be associated development and growth.

Not only must Westmoreland provide a more detailed estimate of the growth and associated air emissions expected due to operation of the facility, but it must also provide a more detailed estimate of the temporary growth and associated air emissions for the construction of the Gascoyne facility. Then, Westmoreland must include these emissions along with emissions from operation of the facility and other existing and permitted sources in the area in an analysis on ambient air quality, on visibility in the impact area of the source, and on soils and vegetation in the impact area of the source. Westmoreland failed to conduct the first step of these analyses, much less any analyses – that is, to adequately assess the emissions associated with the residential, commercial and industrial growth expected to result from the Gascoyne 500 power plant and mine. Thus, the Gascoyne 500 permit application is severely incomplete in this regard.

X. A CLASS II VISIBILITY IMPACTS ANALYSIS MUST BE COMPLETED Westmoreland failed to conduct a visibility impacts assessment in the vicinity of the Gascoyne 500 facility. Such an analysis is required under the “additional impacts analysis” requirement of the PSD regulations. (See 40 C.F.R. §52.21(o) incorporated by reference into North Dakota’s regulations at NDAC §33-15-15-01.2.) As discussed in the October 1990 Draft New Source Review Workshop Manual (at D.5), the visibility impairment analysis pertains to the visibility impacts that occur in the impact area of the proposed new source, and that this visibility analysis “is distinct from the Class I area visibility analysis requirement.” NDDH appears to accept Westmoreland’s Class I area visibility impacts analysis as meeting this requirement. NDDH Permit Analysis at 156-7. Westmoreland must be required to conduct such an analysis to estimate the visibility impacts of the Gascoyne facility and the associated mine in the impact area of the facility along with the expected residential, commercial and industrial growth in the area as a result of the Gascoyne facility. NDDH cannot propose to issue a permit to Gascoyne 500 until this PSD requirement is addressed.

XI. THE SOILS AND VEGETATION ANALYSIS IS INADEQUATE Westmoreland did not provide an analysis of the Gascoyne 500 facility/mine and associated residential, commercial and industrial growth on soils and vegetation in the impact area of the Gascoyne facility. Instead, Westmoreland presented its analysis of the Gascoyne 500 facility’s impacts on sulfur and nitrogen deposition in nearby Class I areas. See Gascoyne 500 permit application at 9-2 to 9-3. While the sulfur and nitrogen deposition analysis is an important part of a Class I area analysis of impacts on air quality

48

related values (AQRVs), it does not satisfy the requirements for an analysis of impacts on soils and vegetation in the vicinity of the facility which is required under 40 C.F.R. §52.21(o). NDDH has proposed to find that there would be no adverse impacts on soils or vegetation around Gascoyne 500 because compliance with the secondary NAAQS was shown and the secondary NAAQS are intended to protect welfare. NDDH Permit Analysis at 159. However, secondary NAAQS are set on a national basis, and do not reflect the effects of air pollutants at all localities. Further, the soils and vegetation analysis is a separate and distinct requirement from the analysis of compliance with the NAAQS that is required under the PSD program. The soils and vegetation analysis for Gascoyne 500 should have included an inventory of the soil and vegetation types with commercial or recreational value found in the Gascoyne 500 impact area. EPA’s New Source Review Workshop Manual at D.4. Then, Westmoreland should have determined if there are sensitive soils or species that could be impacted by the air pollutants to be emitted (or formed as a result of, such as ozone) by the Gascoyne 500 facility, mine, and associated growth in the area. The EPA’s New Source Review Workshop Manual identifies some references for determining whether harmful effects to the soils and vegetation will occur, including the EPA Air Quality Criteria Documents for the various NAAQS that have been promulgated, various Department of Interior and US Forest Service documents including the National Park Service’s Air Quality in the National Parks which lists numerous studies on the biological effects of air pollution on vegetation. Certainly, additional studies and documents have been made available in the last 15 years since EPA’s New Source Review Workshop Manual was published. Westmoreland should have provided analyses to predict the impacts that could occur on soils and vegetation in the area. NDDH should not have proposed to issue the permit for Gascoyne 500 until this requirement was properly met. XII. THE MODELING ANALYSIS FOR GASCOYNE 500 MUST ALSO

INCLUDE THE ALLOWABLE EMISSIONS OF THE GASCOYNE 175

FACILITY

NDDH must require Westmoreland to include the Gascoyne 175 MW facility and mine concurrently with the Gascoyne 500 facility in all air quality analyses required for the PSD permit. The initial permit to construct the Gascoyne 175 MW facility was issued to Westmoreland and Montana Dakota Utilities in June of 2005. One year later, Westmoreland submitted an application to construct the Gascoyne 500 facility. Construction has not yet commenced on the Gascoyne 175 MW facility. Thus, although the permit applications were submitted separately for these facilities, the facilities must be evaluated as one entire source for the purposes of the PSD permit for Gascoyne 500. This means that in evaluating whether the Gascoyne source’s impacts will be over the regulatory ambient significance levels, both facilities must be modeled together. Further, in determining the Gascoyne source’s impact area for each pollutant, both facilities must be modeled. Also, in determining Gascoyne’s impacts on visibility and other air quality

49

related values of Class I areas, both the Gascoyne 175 MW facility as well as the Gascoyne 500 facility and the associated mining activities with all Gascoyne power plants must be modeled simultaneously to predict the overall impacts from the Gascoyne source. While it appears that the Gascoyne 175 MW facility was included in the cumulative air analyses, it is not clear whether the 175 MW facility and associated mining operations were included in determining significance or the significant impact area of the Gascoyne source. It does not appear that the Gascoyne 175 MW facility was included in the Class I visibility or AQRV analysis. Any attempt to only model impacts from the Gascoyne 500 MW facility and not the entire Gascoyne source as currently permitted or proposed for permit issuance must be considered circumvention of the PSD permitting regulations and must not be allowed by NDDH.

XIII. NDDH MUST REQUIRE THE MODELING OF THE MAXIMUM

ALLOWABLE EMISSION RATES OF THE GASCOYNE 175 MW FACILITY

OVER THE AVERAGING TIME OF THE POLLUTANT IN QUESTION IN THE

CLASS I ANALYSIS Westmoreland must be required to model the maximum allowable emission rates of the Gascoyne 175 MW facility and associated mining operations in conjunction with the maximum allowable emission rates of the Gascoyne 500 facility (as proposed by Westmoreland) over the averaging time of each pollutant in question in determining significance of the Gascoyne source and in evaluating NAAQS, PSD increment, visibility and AQRV impacts. A review of the data in Table 7-6 of the Gascoyne 500 permit application (at page 7-15) indicates that Westmoreland did model the maximum allowable short term average emission rates of the Gascoyne 175 MW facility in its Class II analysis (although it appears that Gascoyne 175 was modeled as a nearby source rather than being modeled with the Gascoyne 500 facility as part of the overall Gascoyne source). However, for the Class I increment analysis, Westmoreland failed to model the maximum allowable short term average emission rates of the Gascoyne 175 MW facility and Westmoreland failed to include the allowable emissions from the mining operations associated with the Gascoyne 175 MW facility. Specifically, Westmoreland lists the following emission rates in Table 8-5 (page 8-12) of the Gascoyne 500 permit application that were modeled for the Gascoyne 175 MW facility in the Class I PSD increment modeling: SO2: 138.3 lb/hr PM10: 62.5 lb/hr

50

The modeled SO2 emission rate is lower than the allowable 3-hour average emission limit of 169 lb/hr and the allowable 24-hr average SO2 emission limit of 140 lb/hr at the Gascoyne 175 MW facility.

Westmoreland must model the maximum allowable emission rates of the Gascoyne 175 MW facility over the averaging time of the pollutant/standard being modeled along with the Gascoyne 500 facility emissions. Thus, for the 3-hour average SO2 increment, the Gascoyne 175 MW facility must be modeled at its 3-hour average allowable SO2 emission limit of 169 lb/hr, and the facility must be modeled at its 24-hour allowable SO2 emission limit of 140 lb/hr for the modeling of the 24-hour SO2 increment. In addition, the allowable emission limits of the other emission units/sources associated with the Gascoyne 175 MW facility must be modeled over the averaging time of the pollutant in question, including the PM10 emissions from the mining operations. Further, the maximum allowable 24-hour average SO2, NOx, and total PM10 emission rates of the Gascoyne 175 MW facility must be modeled along with the maximum 24-hour allowable emission rates of the Gascoyne 500 facility in the Class I area visibility and AQRV modeling assessments. NDDH cannot determine whether the Gascoyne facility will cause or contribute to Class I increment violations or whether the Gascoyne facility will adversely impact AQRVs in Class I areas until the Gascoyne 175 MW facility is properly modeled with the Gascoyne 500 facility. XIV. THE CUMULATIVE NAAQS/CLASS II INCREMENT INVENTORY IS

INCOMPLETE The cumulative NAAQS and Class II increment analysis is incomplete for numerous reasons. First, as discussed above, Westmoreland should have modeled all of the Gascoyne facilities together as one source in determining the significant impact area of the Gascoyne source. When Gascoyne 175 is properly included in the modeling, it is probable that the significant impact area will extend father than what was modeled by Westmoreland for Gascoyne 500. Then, the area to be inventoried for a NAAQS and Class II increment analysis includes, at a minimum, all sources within the significant impact area of the Gascoyne source plus an area that extends 50 km beyond the significant impact area. (See pages C.32. and C.35 of the October 1990 Draft New Source Review Workshop Manual). Further, sources located outside 50 km from the impact area should also be included if such source would cause a significant concentration gradient within the significant impact area of the Gascoyne source. (See page C.32 of New Source Review Workshop Manual). Table 7-6 of the Gascoyne 500 permit application includes the sources considered “nearby” to Gascoyne 500 for inclusion in a cumulative NAAQS and/or Class II increment analysis. The list of sources in Table 7-6 fails to include all of the SO2 and

51

NOx sources that should be included in such an analysis. Clearly there are sources that will likely have a significant concentration gradient in the vicinity of Gascoyne that should be included in Class II increment and NAAQS modeling. Those include but are not limited to all of the seven coal-fired power plants that are within roughly 200 kilometers of the Gascoyne source, Dakota Gasification, the Little Knife Gas Plant, and the Mandan Refinery. According to EPA’s AirData website, the combined NOx emissions from these facilities were over 82,000 tpy in 1999 and the combined SO2 emissions from these facilities were over 236,000 tpy in 1999. See http://www.epa.gov/air/data/netemis.html?st~ND~North%20Dakota. Further, Westmoreland should also have included the projected emissions of sources which have been issued PSD permits but which are not yet operating (see page C.34 of the New Source Review Workshop Manual). Thus Westmoreland should have included the maximum allowable emission rates of the Red Trail Energy Ethanol plant to be located in Richardton, and the maximum allowable short term average emission rates must be evaluated in determining compliance with short-term average standards or increments. Also, if the permit application for the South Heart power plant is complete, that source’s proposed allowable emission rates should have been modeled as well. Westmoreland should also have included emissions from oil and gas wells in the vicinity of the project. The mobile source and fugitive emissions associated with the roads for oil and gas development must also be included in the inventory of sources for a cumulative analysis. Thus, NDDH cannot adequately assess whether the Gascoyne source will cause or contribute to a violation of the NAAQS or Class II increments based on the analysis provided in the Gascoyne 500 permit application. NDDH must require Westmoreland to conduct a complete NAAQS and Class II increment by modeling the Gascoyne 175 facility and mine and the Gascoyne 500 facility and mine together as one source and by requiring the emissions inventory for the cumulative NAAQS and Class II increment analyses to be expanded to include all of the above sources and any other sources of air pollution, including minor and area sources, within the vicinity of the Gascoyne source. XV. THE PM10 CLASS II INCREMENT MODELING IS FLAWED Westmoreland’s Class II PM10 increment modeling is flawed and NDDH cannot properly assess whether the Gascoyne source (including the mine) will cause or contribute to a violation of the Class II PM10 increment. First, as discussed above, Westmoreland should have modeled all of the Gascoyne facilities and mining operations (i.e., those associated with the 175 MW facility and those associated with the 500 MW facility) together as one source in defining the significance impact area and the sources to be included in the cumulative analysis. Second, Westmoreland should have modeled a more complete cumulative PM10 emissions inventory including those sources discussed in the above comment and nearby fugitive dusts sources or other minor PM10 sources in the area.

52

Westmoreland’s permit application shows that the PM10 impacts of Gascoyne 500 facility and mine are well in excess of the Class II PM10 increment. See Table 7-12 of Westmoreland’s permit application for Gascoyne 500. The maximum predicted 24-hour average PM10 impact is 81.51 µg/m3 compared to the 24-hour average PM10 Class II increment of 30 µg/m3. However, in the cumulative PM10 Class II increment analysis, Westmoreland assumed that PM10 emission reductions from the Gascoyne mine since its operations at the time of the PM10 minor source baseline date of 1978 expanded the increment. However, it seems that the baseline PM10 emissions have been overestimated and also the projective fugitive PM10 emissions may have been underestimated. Specifically, Westmoreland assumed baseline emissions of the mine to be 50 tons per year (tpy) of PM10 for handling and processing, 149.3 tpy for mine pit source, and 278.2 tpy for haul roads, for a total of 477.5 tpy. Gascoyne 500 Permit Application at 7-20. These calculations were based on a lignite production rate of the mine of approximately 3,000,000 tons per year. NDDH Permit Analysis at 89. The total PM10 emissions expected from the Gascoyne 500 mining operations are approximately 179 tpy – less than half of the PM10 emissions calculated as baseline emissions for the mine. Gascoyne 500 permit application, Table 3-1. Yet, to support Gascoyne 500, the current mine will have a lignite production rate of approximately 4,600,000 tons per year.98 Adding the coal to be utilized by the Gascoyne 175 MW facility will likely increase the production rate of the mine by 35% above the lignite needed to support the Gascoyne 500 facility. Thus, the projected Gascoyne mining operation will have a production rate 1.5 – 2 times greater than the production rate of the Gascoyne mine at the time of the PM minor source baseline date (1976-1977), and yet the fugitive PM10 emissions from the future Gascoyne mining operations will be less than half of the fugitive PM10 emissions from the Gascoyne mine in 1976-1977? Thus, something appears seriously amiss in either the PM10 emissions calculations for the baseline emissions of the mine and/or in the projected PM10 emissions of the mining operations to support the Gascoyne source. It is imperative that NDDH ensure that the PM10 emissions estimated for the baseline mining operations be as accurate as possible to ensure that baseline PM10 emissions are not wrongfully inflated. In addition, it is wrong for Westmoreland to assume that all PM10 emissions associated with the mining operation stopped once production of lignite ceased. It is most probable that the haul roads that were part of the mine still exist, and may be traveled by motor vehicles. Further, such roads are still a source of windblown dust, as is the entire mine. NDDH must also ensure the PM10 projections of the new Gascoyne mining operations were properly calculated, and, importantly, NDDH must also ensure that that all assumptions used in estimating PM10 emissions are made enforceable through permit conditions and that Westmoreland is held accountable to meet such requirements. While NDDH has proposed as a condition of the permit that Westmoreland subsequently submit a fugitive dust control plan (with no specified deadline for such submittal) (see Condition II.A.8. of the proposed permit for Gascoyne 500), NDDH must include specific, enforceable fugitive dust control provisions now in the proposed permit. Given that

98 This was calculated based on the primary fuel feed rate of 527 tons of lignite per hour (from Gascoyne 500 Permit Application, Appendix B-1, assuming continual operation throughout the year.

53

without such enforceable PM10 control requirements, it is likely that the PM10 increment could be violated by the Gascoyne source and mine (especially given the likely miscalculation of baseline emissions for the mine), the public has the right to review and comment on the fugitive dust control measures that are being relied upon to avoid a violation of a PSD standard. For the American Colloid facility, Westmoreland appears to have only modeled point source emissions of PM10 from this facility. Yet, there are fugitive PM10 emissions sources associated with the surface mining operations of this facility and none of those sources of emissions appear to have been included in the cumulative Class II PM10 increment analysis. For all of the above reasons, the cumulative PM10 Class II increment analysis for the Gascoyne facility is flawed, and NDDH cannot rely on the modeling analysis presented by Westmoreland to ensure that the Gascoyne source won’t cause or contribute to a PM10 increment violation until NDDH addresses all of the flaws in the increment analysis. XVI. NO MODELING ANALYSIS WAS DONE FOR OZONE As required by 40 C.F.R. §52.21(k)(1), Westmoreland is required to demonstrate that it won’t cause or contribute to a violation of the ozone NAAQS, yet no ozone analysis was provided for Gascoyne 500 either in the permit application or in NDDH’s analysis. The Gascoyne source is considered major for ozone because it has the potential to emit more than 100 tons per year of NOx and thus is required to conduct an ambient air analysis of ozone impacts.99 The PSD permitting regulations do not provide for any exemption from modeling ozone impacts for a major source of ozone precursors such as Gascoyne. Therefore, NDDH cannot issue the permit for Gascoyne until it is adequately demonstrated that the Gascoyne source won’t cause or contribute to a violation of the ozone NAAQS. XVII. THE CLASS I INCREMENT ANALYSIS IS FLAWED AND CANNOT BE

RELIED UPON TO DETERMINE WHETHER THE GASCOYNE SOURCE

WILL CAUSE OR CONTRIBUTE TO A PSD INCREMENT VIOLATION The Assumptions For The Class I Cumulative PSD Increment Analysis Must Not Rely On The Unlawful And Scientifically Unsound Assumptions Of North Dakota’s August 2005 SO2 PSD Air Quality Modeling Protocol Westmoreland’s Class I increment analyses for SO2 and PM10 appear to be based on a similar modeling approach as used by NDDH in its SO2 increment modeling study100 (North Dakota’s SO2 PSD Air Quality Modeling Report, August 19, 2005 (final),

99 See definition of major stationary source at 40 C.F.R. §52.21(b)(1)(ii) as revised on November 29, 2005, 70 Fed.Reg.71611. See also 40 C.F.R. §52.21(i)(5)(i)(e), footnote 1 (revised November 29, 2005), which states that “. . . any net emissions increase of 100 tons per year or more of. . .nitrogen oxides subject to PSD would be required to perform an ambient impact analysis. . . .” 100 See, e.g., NDDH Permit Analysis at 124; Gascoyne 500 permit application at 8-4 to 8-5.

54

Prepared by NDDH-Environmental Health Section, Attachment 38 to this letter). However, numerous aspects of the North Dakota SO2 PSD Air Quality Modeling Report are not only scientifically unsound but also inconsistent with the Clean Air Act. Further, EPA has not yet taken any final action approving or authorizing the modeling approach and methodology used in the state’s SO2 PSD Air Quality Modeling Report.101 Although EPA recently proposed revisions to the federal PSD requirements pertaining to increment consumption analyses (72 Fed.Reg. 31372-31399, June 6, 2007), it must be noted that EPA has not yet promulgated these revisions. Further, the EPA’s proposed rulemaking makes clear that the majority, if not all, of the methodologies used by NDDH in its Class I SO2 increment analysis are not legal under the current PSD regulations and such methodologies will not be legal unless and until EPA promulgates the proposed revisions to the PSD regulations. Thus, NDDH must not rely on the methodology of its 2005 North Dakota SO2 PSD Air Quality Modeling Report in assessing whether the Gascoyne source would cause or contribute to a violation of any Class I PSD increment. Instead, NDDH and Westmoreland must properly assess whether it would cause or contribute to a violation of the PSD increments consistent with current federal statute, regulations and guidance. The following details the illegalities with the NDDH methodology for assessing increment consumption and thus explains why Westmoreland’s cumulative Class I increment analyses are fatally flawed. As discussed in our comment above, first and foremost is that Westmoreland should have modeled all of the Gascoyne sources together, including the 175 MW unit, the 500 MW unit, the mining operations and all other sources associated with the facility. Following are the other significant flaws in the Class I increment analyses: a) Use of Annual Average Emission Rates Drastically Underestimates Consumption

of 3-Hour and 24-hour Average PSD Increments Westmoreland’s cumulative Class I increment analysis drastically underestimated the contribution of existing sources to increment consumption by modeling the existing sources at their annual average emission rate. Specifically, Table 8-5 of the Gascoyne 500 permit application shows that the 2003-2004 average SO2 and PM10 emission rates of the existing sources in the region were based on the annual emissions from each unit divided by the annual operating hours, to come up with an annual average emission rate. As discussed in North Dakota’s SO2 PSD Air Quality Modeling Report (August 19, 2005 FINAL) at 10, NDDH found it appropriate to determine current SO2 emission rates of existing sources in this manner because of the definition of “actual emissions” in the PSD regulations which is defined “in general” as the average emission rate in tons per year at which a facility emitted during the two year period prior to a particular date. North

101 To date, EPA has never issued a final approval of the North Dakota modeling protocol or of the final North Dakota’s SO2 PSD Air Quality Modeling Report. Even North Dakota’s August 19, 2005 report indicates that EPA has only verbally approved the state’s protocol. See North Dakota’s SO2 PSD Air Quality Modeling Report, August 19, 2005, at 23. Such “verbal” approval does not constitute final agency action by EPA.

55

Dakota has adopted the federal definition of “actual emissions” at 40 C.F.R. §52.21(b)(21) by reference at NDAC §33-15-15-01.2. However, the term “actual emissions” was not included in the federal PSD regulations to define the averaging time of emissions to be used to assess compliance with the PSD increments. The PSD increments have been in effect in EPA’s PSD regulations since 1975, but it wasn’t until after the court decision, Alabama Power Company v. Costle (636 F.2d 323 (1979)), that EPA added a definition of “actual emissions” to the PSD regulations. Specifically, EPA added the “actual emissions” definition to the PSD regulations on August 7, 1980, and the preamble to this rulemaking makes clear that its purpose was to define whether a major modification occurred at an existing stationary source. See 45 Fed.Reg. 52680, 52698-9 (August 7, 1980). Further, EPA’s modeling guidelines, incorporated by reference into North Dakota’s regulations at NDAC §33-15-15-01.2, make clear that determinations of compliance with short term ambient standards require that averaging times for emission rates modeled reflect the averaging time of the standard being protected. Specifically, Section 11.2.3.3 of 40 C.F.R. Part 51, Appendix W, provides as follows:

[S]equential modeling must demonstrate that the allowable increments are not exceeded temporally and spatially, i.e., for all receptors for each time

period throughout the year(s) (time period means the appropriate PSD averaging time, e.g., 3-hour, 24-hour, etc.).

This means that the averaging times for emission rates used in PSD modeling must reflect the averaging time of the PSD increments in order to ensure protection of both the annual average and the short term average increments. For SO2 and PM10, EPA set different increments for different averaging times consistent with the averaging times of the NAAQS for these pollutants. Further, EPA policy has made clear that PSD permits must contain short term limits to ensure protection of the short term average NAAQS and increments.102 There is no valid reason why existing sources can use annual average emission rates to show compliance with 3-hour and 24-hour average NAAQS and PSD increments but new sources must be subject to short term average emission limits to ensure compliance with the short term average NAAQS and PSD increments. Indeed, as a practical matter, it does not make sense to allow compliance with 3-hour average and 24-hour average emission rates to be based on annual average emissions. This is similar to allowing a person to average all the variations in his driving speed over an entire year to see whether he is complying with a 55 mile per hour speed limit. It makes the short term average standards meaningless and calls into question why such short term standards are even necessary. Yet, Congress included 3-hour average SO2 increments and 24-hour average SO2 and particulate matter increments in the Clean Air Act when amended in 1977.103

102 See November 24, 1986 EPA Memorandum from Gerald A. Emison, Director, OAQPS, to David Kee, Director, Air Management Division, EPA Region VIII. See also EPA’s New Source Review Workshop Manual (October 1990 Draft) at B.56. 103 Section 163(b)(1) of Clean Air Act.

56

Numerically, the annual average increments are more stringent than the short term increments and no exceedances are allowed. If compliance with the short term increments could be based on annual average emission rates, then why would Congress have also enacted 3-hour and 24-hour average increments that are numerically higher than the annual increments and for which one exceedance per year is allowed? Westmoreland’s approach in using annual average emission rates of existing sources is also inconsistent with EPA’s New Source Review Workshop Manual, which is EPA’s definitive guidance on how to write and review PSD permit applications. Specifically, EPA’s guidance includes a discussion of how a PSD applicant is to perform a PSD increment analysis, as follows:

For a PSD increment analysis, an estimate of the amount of increment consumed by existing point sources generally is based on increases in actual emissions occurring since the minor source baseline date. . . For any increment-consuming (or increment-expanding) emissions unit, the actual emission limit, operating level, and operating factor may all be determined from source records and other information (e.g., State emissions files). For the annual averaging period, the change in the actual emissions rate should be calculated as the difference between:

• the current average actual emissions rate ,and

• the average actual emissions rate as of the minor source baseline date (or

major source baseline date for major stationary sources).

In each case, the average rate is calculated as the average over previous 2-year period (unless the permitting agency determines that a different time period is more representative of normal source operation). For each short-term averaging period (24 hours and less), the change in the actual emissions rate for the particular averaging period is calculated as the difference between:

• the current maximum actual emissions rate, and

• the maximum actual emissions rate as of the minor source baseline date

(or major source baseline date for applicable major stationary sources

undergoing construction before the minor source baseline date).

In each case, the maximum rate is the highest occurrence for that

averaging period during the previous 2 years of operation.

EPA’s New Source Review Workshop Manual (October 1990 Draft ) at C.48-49 [Emphasis Added.]

57

Thus, the New Source Review Workshop Manual makes clear that, to determine consumption of the short term average increments from existing sources, a source’s current emissions are based on the highest emission rate for that averaging time during the previous two years of operation. Westmoreland’s use of annual average emission rates to reflect current emissions of existing sources is acceptable for determining compliance with annual average increments, but not for assessing compliance with 3-hour and 24-hour average SO2 and PM10 PSD increments. For all of the above reasons, Westmoreland’s use of annual average emission rates from existing sources to determine increment consumption is inconsistent with state and federal requirements and is scientifically unsound. As a result, Westmoreland’s analysis significant underestimated the amount of SO2 and PM10 increment consumption from existing sources. Thus, Westmoreland’s increment analysis cannot be relied upon to determine whether the Gascoyne source will cause or contribute to violations of the 24-hour average SO2 or PM10 increments or the 3-hour average SO2 increment. Consequently, the cumulative increment analysis must be revised to ensure that the current emissions of existing sources are based on the maximum 3-hour average SO2 emission rates (for the 3-hour average SO2 increment) and the maximum 24-hour average SO2 and PM10 emission rates (for the 24-hour average SO2 and PM10 increments). b) The Baseline Emissions for Sources Existing at the Time of the Baseline Date

Fail to Reflect the Emissions in the Baseline Concentration It appears that, for most sources, Westmoreland relied on the baseline inventory developed by NDDH in its SO2 increment analysis (August 19, 2005 (FINAL)). However, in its increment analysis, NDDH unlawfully interpreted “baseline concentration” to include emissions that would not be considered as contributing to the baseline concentration under the statutory definition at §169(4) of the Clean Air Act, rather than counting those emissions as consuming the increment. Westmoreland’s reliance on the NDDH’s SO2 baseline inventory meant that the baseline emissions in its cumulative SO2 increment analysis were unlawfully inflated. It is not clear what baseline inventory Westmoreland used for PM10 sources in its Class I PM10 increment analysis. We assume that, at least for the sources that were part of both the SO2 and PM10 cumulative increment inventory, similar issues exist with the baseline emissions data for the PM10 sources modeled by Westmoreland in its cumulative Class I PM10 increment analysis. The Clean Air Act definition of “baseline concentration” reads as follows:

. . . with respect to a pollutant, the ambient concentration levels which exist at the time of the first application for a permit in an area subject to this part, based on air quality data available in the Environmental Protection Agency or a State air pollution control agency and on such monitoring data as the permit applicant is required to submit. Such

58

ambient concentration levels shall take into account all projected emissions in, or which may affect, such area from any major emitting facility on which construction commenced prior to January 6, 1975, but which has not begun operation by the date of the baseline air quality concentration determination. Emissions of sulfur oxides and particulate matter from any major emitting facility on which construction commenced after January 6, 1975, shall not be included in the baseline and shall be counted against the maximum allowable increases in pollutant concentrations established under this part.

Clean Air Act, § 169(4). Federal and state regulations require the use of actual emissions in determining those emissions that are to be considered part of the baseline concentration:

Baseline concentration means that ambient concentration level which exists in the baseline area at the time of the applicable minor source baseline date. A baseline concentration is determined for each pollutant for which a minor source baseline date is established and shall include: (a) The actual emissions, as defined in paragraph (b)(21) of this section, representative of sources in existence on the applicable minor source baseline date. . . .

40 CFR §52.21(b)(13)(i), incorporated by reference into North Dakota’s rules at NDAC §33-15-15-01.2.

In turn, the state and federal definition of actual emissions provides:

In general, actual emissions as of a particular date shall equal the average rate, in tons per year, at which the unit actually emitted the pollutant during a consecutive 24-month period which precedes the particular date and which is representative of normal source operation. The Administrator shall allow the use of a different time period upon a determination that it is more representative of normal source operation. Actual emissions shall be calculated using the unit’s actual operating hours, production rates, and types of materials stored, combusted, or processed during the selected time period.

40 C.F.R. §52.21(b)(21)(ii) [Emphasis added], incorporated by reference into North Dakota’s rules at NDAC §33-15-15-01.2. As discussed in EPA’s New Source Review Workshop Manual, “In certain limited situations. . . a different (2 year) time period may be used upon a determination by the reviewing agency that it is more representative of normal source operation. Normal source operations may be affected by strikes, retooling, major industrial accidents, and other catastrophic occurrences.”104 Thus, based on these definitions, the baseline inventory should generally be based on the actual emissions that existed at the time of the first PSD permit application (i.e., the time

104 US EPA, New Source Review Workshop Manual, October 1990 Draft, at A.39.

59

of the minor source baseline date, which for SO2 in western North Dakota, is December 19, 1977105) unless a source can show that an event such as a strike, retooling, major industrial accident or other catastrophic event occurred during that time.

However, for some sources (e.g., the Stanton Unit 1, M.R. Young Units 1 and 2, and Leland Olds Units 1 and 2 power plants and the Royal Oak Plant106), NDDH used emissions data from after the minor source baseline date to establish baseline emissions. NDDH essentially claimed that the more recent years were more reflective of normal source operation. While that may be acceptable in some cases, overall it appears that NDDH chose years with the highest heat input and highest SO2 emissions for these sources, rather fairly evaluating whether the two years immediately prior to the December 1977 baseline date were reflective of normal source operations. Indeed, in looking at the heat input data for the power plants for which NDDH used 1978-1979 data for the baseline inventory, it is clear that the heat input for these plants (and thus SO2 emissions) was the greatest in 1978 or 1979 as compared to any other year of the eight year period reviewed for each plant with the exception of Stanton Unit 1.107 For the Royal Oak plant which is an increment-expanding source because it shut down after the minor source baseline date, it appears that NDDH used 1978-1979 as the baseline period to capture increased emissions due to two furnaces that were added in 1976 and 1978.108 NDDH found that coal usage increased dramatically during 1978 and 1979 –clearly this was a result of the modifications. Indeed, NDDH states that the addition of these furnaces were major modifications to the plant109. Because these modifications were made after the SO2 major source baseline date of January 6, 1975, the increased emissions associated with these modifications consume the available increment and must not be considered part of the baseline emissions inventory.110

EPA stated in the preamble to its August 7, 1980 PSD regulations that, due to the Court upholding the concept of an actual emissions baseline in Alabama Power Company v.

Costle, its current policy was that increases in emissions due to capacity utilization, hours of operation, or that were otherwise reasonably anticipated to occur would generally be

excluded from the baseline concentration and thus would consume the available increment. EPA did state that, if a source can demonstrate that its emissions after the baseline date were more representative of normal source operation, then such emissions could be included in the baseline concentration. However, EPA’s intent was that the baseline emissions should be representative of normal source operations at the time of the

baseline date not over the life of the source. See 45 Fed. Reg.52713, 52714. Thus,

105 NDDH’s “A proposed alternative air quality modeling protocol to examine the status of attainment of PSD Class I increments (MOU Protocol),” August 18, 2005, at 4. Attachment 39. 106 Id. at 19. 107 Id. at 37-38 (Appendix D). 108 Id. at 36. 109 Id. 110 As discussed in the preamble to EPA’s August 7, 1980 PSD rulemaking, emissions from major source construction “including modification” that commenced after the major source baseline date consume increment. 45 Fed.Reg. 52717. See also 40 C.F.R. §52.21(b)(13)(ii)(a), incorporated by reference into North Dakota’s rules at NDAC §33-15-15-01.2.

60

NDDH’s methodology of determining baseline emissions does not reflect EPA policies on normal source operation. The result is that baseline emissions from these sources were inflated, which ultimately reduces the amount of increment considered to be consumed by these sources in the increment modeling analysis by shrinking the total amount of increment consuming emissions. Westmoreland apparently relied on NDDH’s SO2 baseline inventory in its cumulative PSD increment analysis, and the PM10 baseline inventory was likely based on similar assumptions. Consequently, its cumulative PSD Class I increment analyses for SO2 and PM10 are flawed and cannot be relied upon to show that the Gascoyne source won’t cause or contribute to a violation of the PM10 or SO2 Class I increments.

c) Westmoreland’s Modeled Cumulative Increment Concentrations May Not Have

Been Properly Paired in Space and Time It is not clear whether Westmoreland’s Class I modeling determined increment consumption by comparing current concentrations to one baseline concentration at each receptor (i.e., paired in space only) or by comparing current concentrations to baseline concentrations for each averaging period modeled (i.e., paired in space and time). NDDH’s increment analysis used both approaches to determine increment consumption111, but Westmoreland’s Class I Modeling Report does not clarify which approach was used. If Westmoreland used the paired in space only method of estimating increment consumption, this approach is not scientifically valid and it would not ensure compliance with the PSD increments. The paired in space only method disregards the variability in concentrations of pollutants that occurs over time due to changes in meteorology. This method also does not provide an accurate assessment of degradation of air quality. Specifically, the paired in space only method would essentially establish a 24-hour PSD SO2 increment of 5 ug/m3 on two days per year (the days with the highest and second highest baseline concentrations112), and would allow greater than 5 µg/m3 deterioration on the other 363 days each year with lower baseline concentrations. Thus, to the extent Westmoreland’s cumulative increment analyses are based on the paired in space only methodology, the analyses are flawed and cannot be relied upon to determine whether the Gascoyne facility will contribute to a SO2 or PM10 increment violation in nearby Class I areas.

d) Westmoreland Should Have Modeled as Long a Period of Meteorological Data as

Possible to Determine Whether the Gascoyne Source Will Cause or Contribute to a Violation of the Class I Increments for SO2

Westmoreland’s Class I increment modeling relied on 2000-2002 “RUC-2” meteorological data in its assessment of compliance with the Class I increments.113 This is the same meteorological data that was used by NDDH in its August 19, 2005 SO2

111 North Dakota’s SO2 PSD Air Quality Modeling Report, August 19, 2005, at 19. 112 NDDH used the second high baseline concentration in its paired in space only increment assessments. See NDDH, “A proposed alternative air quality modeling protocol to examine the status of attainment of PSD Class I increments,” August 18, 2005, at 26-27. 113 Gascoyne 500 permit application at 8-5.

61

increment analysis. However, NDDH previously used a meteorological dataset that spanned 1990-1994.114 According to EPA’s Guideline on Air Quality Models, “[a]s long a period of record as possible should be used in making estimates to determine design values and PSD increments.” 40 C.F.R. Part 51, Appendix W, Section 8.2.1.1(e), incorporated by reference into North Dakota’s regulations at NDAC §33-15-15-01.2 . Since this data is readily available and since compliance with the Class I SO2 increments in North Dakota is so tenuous, this other 5 year period of meteorological data must also be used to determine whether the Gascoyne facility will cause or contribute to a violation of the Class I SO2 increments. e) Westmoreland Should Have Included the SO2 Emissions from the Sources that

Received Variances in the Cumulative Class I Increment Analysis As was discussed in EPA’s May 24, 2002 comments to NDDH on its Proposed Finding of Adequacy, NDDH has illegally excluded from its increment consumption analysis emissions from the Little Knife Gas Plant and the Dakota Gasification Company. These two facilities received variances from the Federal Land Managers which allowed for the construction of the plants in spite of the fact that the facilities each would contribute to SO2 increment violations in North Dakota’s Class I areas. As discussed in EPA’s May 24, 2002 letter, the FLM variances allowing for construction of these two facilities does not mean that their emissions no longer consume increment. See Alabama Power Co. v.

Costle, 636 F.2d 323, 361 (D.C. Cir., 1979). North Dakota should have revised its state implementation plan to address the increment violations including the contribution by the Little Knife Gas Plant and the Dakota Gasification Company. Thus, in determining whether the Gascoyne source would contribute to Class I SO2 increment violations, the emissions of the Little Knife Gas Plant and the Dakota Gasification Company must be included. For all of the above reasons, Westmoreland’s methodology and assumptions used to assess compliance with the PSD Class I increments for SO2 and PM10 is flawed. Westmoreland has failed to definitively show that the Gascoyne source will not cause or contribute to a violation of the SO2 or PM10 increment in nearby Class I areas and therefore no permit can be issued for construction of this facility. Accordingly, NDDH cannot issue a permit for the Gascoyne source until a proper cumulative Class I increment analysis is provided that shows the Gascoyne source will not cause or contribute to a violation of the Class I PSD increments. The Class I Cumulative Source Inventory Must Include the Impact of Permitted Sources Not Yet Operating As discussed in our comment regarding the Class II area impacts assessment, Westmoreland should have also included the emissions of sources that are permitted but

114 See Calpuff Analysis of Current PSD Class I Increment Consumption in North Dakota and Eastern Montana Using Actual Annual Average SO2 Emission Rates, NDDH, May 2003, at 10. Attachment 40.

62

not yet operating in its Class I area impacts assessment. (See page C.34 of the New Source Review Workshop Manual). Thus the Class I increment analysis for the Gascoyne source should have included the maximum allowable emission rates of the Red Trail Energy Ethanol plant to be located in Richardton, and the maximum allowable short term average emission rates must be evaluated in determining compliance with short-term average standards or increments. Also, if the permit application for the South Heart power plant is complete, that source’s proposed allowable emission rates should be modeled as well.

The Cumulative Class I PM10 Increment Analysis Must Also Include the Increment Consuming Emissions from Roads/Mobile Sources

It appears that Westmoreland did not include any emissions increases due to roads or mobile sources in its cumulative Class I PM10 increment analysis. Any growth in PM10 emissions from mobile sources and roads that has occurred since the minor source baseline date consumes the available increment and thus must be included in the modeling.115 Westmoreland should also have included the increased PM10 emissions from new paved and unpaved roads near the Class I areas due to activities such as oil and gas development that have been constructed since the minor source baseline date. Further, if the South Heart permit application has been deemed complete, Westmoreland should also have included the PM10 emissions from the mining operations associated with the South Heart plant. Without inclusion of nearby roads and mobile source emissions in its cumulative PM10 analysis, Westmoreland’s cumulative Class I PM10 increment analysis cannot be relied upon to verify that the facility will not cause or contribute to a violation of the PM10 NAAQS.

XVIII. THE CLASS I ANALYSIS MUST ANALYZE IMPACTS ON OTHER

NEARBY CLASS I AREAS In addition to the deficiencies in the modeling discussed in this letter, the Class I analysis for the Gascoyne source is incomplete because Westmoreland limited the modeling analysis to North Dakota Class I areas. There are several other Class I areas within 250-300 km of the proposed power plant that could be impacted by the source, including Medicine Lakes Wilderness Area, the Fort Peck Indian Reservation Class I area, the Northern Cheyenne Indian Reservation Class I area, Wind Cave National Park, and Badlands National Park. Yet, the Gascoyne 500 permit application does not even mention these other Class I areas. The Gascoyne 500 permit cannot be issued without an analysis of whether the source would cause or contribute to a violation of the PSD increments or cause or contribute to adverse impacts on visibility or other air quality related values in these Class I areas.

115 See EPA’s New Source Review Workshop Manual, October 1990 Draft, at C.10. and C.36.

63

XIX. THE GASCOYNE 500 FACILITY WILL ADVERSELY IMPACT

VISIBILITY AT THEODORE ROOSEVELT NATIONAL PARK AND THE

LOSTWOOD WILDERNESS AREA

Westmoreland’s modeling for impacts to visibility indicates that the Gascoyne 500 facility will adversely impact visibility at Theodore Roosevelt National Park and the Lostwood Wilderness Area. See Gascoyne 500 permit application at 9-10. Even without modeling the combined impact of all Gascoyne facilities (which should have been done as we discussed above), Westmoreland’s modeling using Federal Land Managers’ required modeling procedures (i.e., following the FLAG guidance116) showed that the Gascoyne 500 facility could cause greater than a 10% change in visibility at Theodore Roosevelt National Park (all three units) and at the Lostwood Wilderness Area. While Westmoreland attempted to discount these significant visibility impacts by conducting “Tier 2” and “Tier 3” visibility analyses that they admitted had only been proposed by the National Park Service, it must be noted that the Federal Land Managers have not yet adopted any revisions to their 2000 FLAG guidance. Indeed, the National Park Service pointed this out in their June 26, 2006 letter to you on the Gascoyne 500 permit application.117 Thus, Westmoreland’s Class I area visibility analysis must be based on FLAG procedures. It is also important to note that Westmoreland must be required to model all of the Gascoyne facilities together in evaluating visibility impacts – meaning the Gascoyne 175 MW facility as well as the Gascoyne 500 facility and the mining activities. Westmoreland must not be allowed to evaluate these two facilities separately when they are clearly part of the same new major source. Based on the current policies of the Federal Land Manager, the Gascoyne 500 facility alone will adversely impact visibility in Theodore Roosevelt National Park and the Lostwood Wilderness Area. Thus, according to the Federal Land Managers’ FLAG guidance, if a source would cause a 5% change in visibility or greater at a Class I area, then a cumulative visibility analysis should be completed. Further, under North Dakota’s regulations, Westmoreland must demonstrate that the emissions from the Gascoyne facility will not cause or contribute to an adverse impact on visibility in any Federal Class I area. NDAC §33-15-19-02.1. Protection of Class I areas such as Theodore Roosevelt National Park is a core purpose of the PSD program, CAA § 160(2), 42 U.S.C. § 7470(2), and must be safeguarded. Westmoreland clearly has not met this requirement. In addition, as discussed above, Westmoreland has improperly limited its Class I analyses to only evaluate impacts at North Dakota Class I areas when the Gascoyne source could impact Class I areas in Montana and South Dakota as well. Thus, the Gascoyne 500 permit application cannot be considered complete without such a cumulative visibility analysis. If the visibility analyses at any other Class I areas shows a

116 The 2000 “FLAG” guidance is the Federal Land Managers Air Quality Related Values Workgroup (FLAG) Phase I Report, December 2000, which is included as Attachment 41 to this letter. 117 June 26, 2006 letter from John Bunyak, National Park Service Air Resources Division, and Sandra Silva, Fish & Wildlife Service Air Quality Branch, to Terry O’Clair, NDDH, Attachment 42.

64

greater than 5% change in visibility will occur, Westmoreland must be required to conduct cumulative visibility impacts for those Class I areas as well. In any case, the visibility analysis that Westmoreland presented in its PSD permit application clearly indicates that, under FLAG procedures, Gascoyne 500 will adversely impact visibility at Theodore Roosevelt National Park and Lostwood Wilderness Area. Accordingly, NDDH must find that Gascoyne 500 will adversely affect visibility in North Dakota’s Class I areas, and thus NDDH must not issue the permit for Gascoyne 500. XX. NDDH FAILED TO LIMIT MERCURY EMISSIONS FROM THE

GASCOYNE FACILITY NDDH has failed to eliminate or limit mercury emissions to be emitted from the Gascoyne facility under the North Dakota regulations. Specifically, NDAC §33-15-01-15.1 provides that no person shall permit or cause air pollution. “Air pollution” is defined in NDAC §33-15-01-04.3 as “the presence in the outdoor atmosphere of one or more air contaminants in such quantities and duration as is or may be injurious to human health, welfare, or property or animal or plant life, or which unreasonably interferes with the enjoyment of life or property.” “Air contaminant” is defined in NDAC §33-15-01-04.2 in a very broad manner, similar to the definition of “air pollutant” in the Clean Air Act.118 At the June 21, 2007 public hearing on the proposed Gascoyne 500 permit, Terry O’Clair stated that the Gascoyne facility could emit 0.33 tons per year of mercury and that no mercury-specific pollution controls would need to be installed at Gascoyne 500 to meet federal or state mercury regulations. Gascoyne’s mercury emissions would be higher than the 1999 actual emissions of any of the existing North Dakota coal-fired power plants.119 Mercury is a very hazardous neurotoxin that is dangerous even at low levels. Citizens of North Dakota and surrounding states are already subject to unacceptable mercury levels due to mercury emissions from existing power plants in North Dakota. In fact, all of the waterbodies in the state of North Dakota are under fish consumption advisories due to mercury.120 Any new source of mercury emissions should be subject to the strictest standards possible to avoid exacerbating an already severe problem.

Coal-fired power plants are the single largest source of mercury air emissions in the nation, and deposition of these air emissions causes an accumulation of mercury in water bodies. Mercury emitted from coal plants becomes methylmercury in the environment, where it becomes toxic in even minute amounts. Readily absorbed by living tissues,

118 NDAC §33-15-01-04.2 defines “air contaminant” as “any solid, liquid, gas, or odorous substance, or any combination thereof.” 119 1999 mercury emissions from EPA’s Information Collection Request. This data is available at http://www.epa.gov/ttnatw01/combust/utiltox/utoxpg.html. 120 2003 Advisory for Human Consumption of Fish Caught in North Dakota.

65

methylmercury can cause serious birth defects, central nervous system and brain damage, diminished intelligence, and, recent evidence suggests, autism. According to the FDA standard, it would only take one pound of methylmercury to contaminate 500,000 pounds of fish, which, when consumed by humans and wildlife, increases their mercury levels. NDDH would allow Gascoyne 500 to emit up to 660 pounds of mercury each year. Yet, there is a commercially available and proven technology that could significantly reduce the mercury emissions from Gascoyne 500 at reasonable costs – activated carbon injection. There is a wealth of information available on this technology at vendor websites such as ADA-Environmental Services (www.adaes.com), at the Institute of Clean Air Companies website (www.icac.com), at the National Energy Technology Laboratory website (http://www.netl.doe.gov/technologies/coalpower/ewr/mercury/index.html) and on EPA’s utility air toxics website (http://www.epa.gov/ttn/atw/combust/utiltox/utoxpg.html), to name just a few information sources. Considering the significant impacts that such levels of mercury emissions would have on public health and the environment and considering that there are commercially available mercury control technologies that can achieve high levels of mercury control at very reasonable cost, NDDH cannot lawfully refuse to exercise its authority under NDAC §33-15-01-15.1 to limit mercury emissions in taking action on the proposed Gascoyne 500 permit.

66

Thank you for considering our comments. We request that you include Dakota Resource Council, the Rocky Mountain Office of Environmental Defense, and the Sierra Club on your mailing list for any activities associated with the Gascoyne 500 PSD permit application, including the submittal of additional information from Westmoreland and any NDDH proposed PSD permit for Gascoyne 500. Sincerely, Mark Trechock, Vickie Patton, Executive Director Senior Attorney Dakota Resource Council Environmental Defense P.O. Box 1095 2334 North Broadway Dickinson, ND 58602 Boulder, CO 80304 (701) 483-2851 (303) 440-4901 Bruce Nilles, Midwest Regional Representative Sierra Club Midwest Clean Energy Campaign 122 W. Washington Avenue, Suite 830 Madison, WI 53703 Attachments (all on attached CD) cc: Callie Videtich, U.S. EPA Region VIII (8P-AR)