Southern Illinois Power Cooperative
i 3.545 Lake of Egypt Road Mario, IL 62959 (618) 964-1448 Fax. (618) 964-1867
Environmental Protection Agency 1200 Pennsylvania Avenue, N.W. Washington, D.C, 20460
May 12,2017
Re: Evaluation o f Existing Regulations, Docket ID No, EPA-HQ-OA-2017-0190, 82 Fed. Reg. 17,793 (Apr. 13, 2017), Comments of the Southern Illinois Power Cooperative
Dear Sir/Madam:
Southern Illinois Power Cooperative (SIPC) respectfully recommends that the Environmental Protection Agency (EPA) repeal the designation of Williamson County, Illinois as a nonattainment area under the sulfur dioxide (SO2) National Ambient Air Quality Standard (NAAQS), EPA finalized this action on January 18, 2017, two days before the previous administration took office,
The designation was based on faulty factual information on the extent to which the public has access to SIPC's coal-fired Marion Station and on unreliable air quality modeling provided by the Sierra Club. There is, in fact, no SO2 air quality problem in Williamson County. By contrast, EPA's nonattainment designation threatens severe economic consequences, including the loss o f hundreds of jobs that support operation o f the Marlon Station, in an area o f southern Illinois that is already economically stressed. As a regulation entailing large costs for no environmental benefit, EPA's nonatiainment designation is exactly the type o f regulation that should he repealed under Executive Order 13777,
We attach a one-page document to these comments that summarizes the issues, We also attach SIPC's Petition for Reconsideration o f the order designating Williamson County as nonattainment, which contains more background on this issue. We do not attach the voluminous exhibits to this Petition, but they are available in the docket (EPA-HQ-OAR-2014-0464). EPA denied this petition on January 28, 2017.
SIPC and Williamson County
SIPC is a small consumer-owned electric generation & transmission cooperative that selves approximately 250,000 homes in one o f the most economically depressed areas o f Illinois. We serve lour o f the top five, and nine o f the top fifteen, poorest counties in Illinois. As a cooperative, SIPC has no shareholders and makes no profit; any SIPC "profits" flow directly to ultimate consumers and any increased costs must be borne by those consumers as well.
SIPC's only solely-owned power facility (other than a 7.9% portion o f the Prairie State Generation Plant in'Washington County, IL) Is the Marion Station in Williamson County. The Station consists o f one 120 megawatt (MW) fluidized bed boiler that uses reclaimed and freshlymined Illinois coal, one 173 MW coal-fired cyclone boiler, and two 70 MW simple cycle
Touchstone Energy
Sierra Club v. EPA 18cv3472 NDCA
Tier 7
ED 002061 00070535-00001
combustion turbines. All o f the coal for the two-coal-fired units is acquired locally from southern Illinois mines.
Regulation at Issue
EPA made the nonattainment designation as a part o f "Air Quality Designations tor the 2010 Sulfur Dioxide (SO2) Primary National Ambient Air Quality Standard Round 2" (EPA-HQOAR-2014-0464), 81 Fed. Reg. 45,039 (July 12, 2016).1 This regulation resulted from EPA's June 22,2010 promulgation of a primary SEE NAAQS that for the first time was based on 1-hour daily maximum concentrations. When EPA issued this standard, however, it had not undertaken the work necessary to ensure that it could make atfainment/noiiattainment designations in time to meet Clean Air Act designation deadlines. When EPA was late in making these designations, it was sued by environmental groups. It then entered into a "sue-and-settle" consent decree requiring it to make designations on an artificial, accelerated timeline.
Historically, EPA has used actual air quality data obtained from air quality monitors to determine whether an area is in attainment or nonattainment o f a NAAQS, But under the accelerated deadline set by the consent decree, there was not enough time to collect actual air quality data and so the only wav to project whether areas like Williamson County would attain the new SO;; standard was to use notoriously fallible air quality models.
The Illinois Environmental Protection Agency (IE'PA) submitted air quality modeling to EPA showing that Williamson County attained the N AAQS, In contrast, the Sierra Club submitted air quality modeling showing nonattainment. EPA based its nonattainment determination on Sierra Club's modeling and rejected lEPA's modeling, SIPC's own modeling, performed by an independent consultant, also supported attainment.
There are two main areas of dispute concerning the air quality modeling. First, air quality modeling shows exceedances o f the SCE NAAQS in an area located to the northeast o f Lake Egypt Road (the Northern Property) that Is within the boundary o f the Marion Station. Because the public does not have access to this area, these modeled exceedances should not count in determining whether the plant will cause a violation o f the S 0 2 standard. Air quality standards apply only to the "ambient air," defined as air to which the public has access, and the public does not have access to the Northern Property,
Based on incomplete information, EPA concluded that the barrier around one part of the Northern Property fails to preclude public access to this area. EPA, however, does not dispute that, absent the alleged exceedances in this area, Williamson County would attain the SCE NAAQS. Hence, the Issue of whether the public does or does not have access to the Northern Property, by itself, is dispositive of whether Williamson County should be designated nonattainment.
A s afactual matter, the public does not have access to the Marion Station, including the Northern Property. Amazingly, in concluding that the public might have access, EPA relied on misleading Ooogie Earth photographs that lailed to show that the area EPA thought was inadequately fenced Is located at the foot o f a steeply-sloped embankment that leads up to a 55-
s Again, all o f our statements in this section are documented in the attached Petition for Reconsideration.
Sierra Club v. EPA 18cv3472 NDCA
Tier 7
ED 002061 00070535-00002
mph highway. For some reason, in deciding whether or not the public lias access to the Northern Property, EPA did not think to pick up the phone and call SIPC to discuss the matter or arrange a site visit. Instead, based on, at best, incomplete inform ation, EPA decided to m ake a nonattainment designation that threatens thejo b s o f hundreds and coaid economically devastate an already stressed area o f coal country. One phone call, one meeting, one site visit could have easily resolved this matter.
The second area o f dispute between SIPC and EPA concerns EPA's use o f faulty Sierra Club modeling. This issue is discussed in detail in the attached Petition for Reconsideration. We note here only that if these issues are resolved, as they must be, favorably for the IEPA and SIPC, Williamson County would be in attainment of the SO2 standard even if the public access issue were resolved against SIPC. In other words, there are two independent bases that justify reversal of EPA's decision.
Impact of NoiiattammeHt Designation
As a result o f the nonattainment designation, SIPC will be faced with the need to install additional controls on the M arion Station's coal units. This will result in higher dispatch costs, causing the units to be uneconomic. If the units cannot run, the plant's 130 employees wall lose their jobs, with a loss of S24 million in wages and benefits. Reductions hi coal purchases from local mines will result in the loss of another $60 million. We estimate that the indirect effects of closing the coal units at the Marion Station, including the cost to suppliers and trucking companies, to be an o th er 800 jobs and $800 million to the local economy.
Moreover, if we lose the Marion Station, we will have to replace it with higher cost sources o f power. This will result in rate increases to our member-customers, many o f whom have low incomes and cannot afford to pay the higher rates.
Beyond the impacts to the coal units, the nonattainment designation will require Illinois to develop an attainment plan and an attainment demonstration that could require additional local controls. The nonattainment designation also creates a stigma and a potentially significant impediment to economic development in Williamson County, making it difficult to attract new business and jobs to the area.
Economic impacts, however, do not tell the whole story. It is now widely understood that job losses create health and welfare problems for the unemployed person and his or her family. Less money is available to pay for the necessities o f life like food, energy and health care. Unemployment also increases drinking and alcohol abuse, depression and suicide. There is no evidence that EPA gave any thought to these outcomes when it designated Williamson County nonattainment.
Executive O rder 13777
In his February 24, 2017 Executive Order 13777, entitled "Enforcing the Regulatory Reform Agenda," the President directed agencies to implement a process that will lead to the repeal of regulations that meet any o f six criteria, including those that "(i) eliminate jobs, or inhibit job creation; (it) are ... unnecessary [and] (iii) impose costs that exceed benefits," Designating Williamson County as a nonattainment area meets each of these three factors. Jobs will be eliminated, the regulation is totally unnecessary as there is no real ambient air quality problem, and there are no benefits to offset tire significant costs.
Sierra Club v. EPA 18cv3472 NDCA
Tier 7
ED 002061 00070535-00003
This regulation truly is "Exhibit A" in the type o f regulation the President has vowed to eliminate. For the want o f a phone call to clarify whether the public does or does not have access to the Marion Station, hundreds may lose their jobs. Ready Solution
Fortunately, the Clean Air Act has a provision tailor-made for addressing this type of situation. Section. 110(k)(6) provides that, "Whenever the Administrator determines that the Administrator's action ... promulgating any ... area designation ... m is in error, the Administrator may in the same manner as the ... promulgation revise such action as appropriate." EPA's designation o f Williamson County as nonattainment was clearly erroneous and should be corrected.
We appreciate EPA's consideration o f this matter.
Respectfully,
Leonard F. Hopkins, P.E. Vice President o f Environmental & Safety
Sierra Club v. EPA 18cv3472 NDCA
Tier 7
ED 002061 00070535-00004
A Classic C ase of the Obama EFA 's Anti-Coal Regulatory Overreach
1 Misguided EPA action. On January 18, 2016 ju st before leaving office, EPA took final action to designate Williamson County a nonattainment area for the SQ2 NAAQS, The County is an economically challenged community in the heart o f coal country in rural southern Illinois.
2. 'The designation has no purpose other than to impose control burdens on the Marion Station, a small coal-fueled electric generating station, that could lead to plant closure. The Station is owned by the Southern Illinois Power Cooperative, a small consumer-owned cooperative utility that serves 4 o f the top 5 and 9 o f the top 15 poorest counties in Illinois.
3. The nonattainment designation is devastating:
Direct impacts on Non-Attainment D e s ig n a tio n
> Higher dispatch cost cause units to be u n e co n o m ic
in d ire c t impacts of N o n -A tta in m e n t D esignation > Higher rates to M em bers w ho least can afford it
> jeopardizes 130 em ployees jo bs at SIPC; $24M in wages and b en efits w ill be reduced.
> Loss o f econom ic d e ve lo p m e n t prospects fo r W illiam son County becoming "non-attainm ent"
> Reductions in coal purchases fro m local mines; $60M at risk,
> Job reductions in su p p lle rs/tru ckin g , SIPC regionally supports 800 jobs & $800M to local economy.
4. The nonattainment designation was erroneous. EPA used an air quality model supplied by the Sierra Club rather than the model used by the Illinois Environmental Protection Agency. Worse, the nonattainment designation depends on modeled air quality violations within the plant boundary even though it is well-established that modeled exceedances within facility boundaries do not count in making attaminent/nonattainment designations, because the public does not have access to the plant air.
5. EPA took the position that the plant boundary was Inadequately fenced and so the public might have access. But EPA never contacted SIPC about the matter and instead relied on misleading Google Earth, photos. In fact, the fencing around the plant is more than adequate and, along with other security measures, prevents public access. The reality is that the public does not have access and the plant is not causing an ambient air quality problem.
6. BPA's action is "Exhibit A" in the type o f regulations that this Administration has committed to pull back. In his February 24,2017 Executive Order 13777, entitled "Enforcing the Regulatory Reform Agenda," the President directed agencies to implement a process to review regulations that "eliminate jobs or" "inhibit job creation," that are "unnecessary" and that "impose costs that exceed benefits," EPA's action as to Williamson County meets these criteria: jobs will be eliminated, the regulation is totally unnecessary as there is no ambient air quality problem that needs solving, and there are no benefits to offset the significant costs.
7. BPA's action can easily be rectified. Section 110(k)(6) o f the Clean Air Act provides that, "Whenever the Administrator determines that the Administrator's action .,. promulgating any ,.. area designation .,. was in error, the Administrator may in the same manner as the .., promulgation revise such action as appropriate." EPA's designation o f Williamson County as nonattainment was clearly erroneous and should be corrected.
Sierra Club v. EPA 18cv3472 NDCA
Tier 7
ED 002061 00070535-00005
Hardin
Schiff Hardin LLP 233 South Wacker Drive Suite 6600 Chicago, IL 60606
312,258.5500 f 312,258.5600
sehiffhardin.com
September 12, 2016
Stephen J. Bonebrake 312.258.5646 sbonebrake@schifttiardin.com
VIA HAND DELIVERY AND EMAIL (a-and-r--Poeket@epa.gov1
Administrator McCarthy U.S. Environmental Protection Agency EPA Docket Center WJC West Building, Room 3334 1301 Constitution Avenue, NW, Washington, DG 20004
Re: Petition for Reconsideration of the United States Environ mental Protection Agency's Final Rule Designating Williamson County, Illinois, Nonattainment under Air Quality Designations for the 2010 Sulfur Dioxide (SOQ Primary National Ambient Air Quality Standard - Round 2
EPA Docket ID No, EPA-HQ-OAR<2014-O464
Dear Administrator McCarthy: On behalf of Southern Illinois Power Cooperative ("SEPC"), the undersigned petitions the
U.S. Environmental Protection Agency ("EPA" or the "Agency") to reconsider aspects o f the Air Quality Designations for the 2010 Sulfur Dioxide (S 02) Primary National Ambient Air Quality Standard - Round 2 (EPA-HQ-OAR-2014-0464) (hereinafter, the "2016 Designations"). 81 fed. Reg, 45,039 (Juiy 12, 2016). The rule becomes effective on September 12, 2016, 60 days after the date of publication in the Federal Register, which, was on July 12, 2016, Therefore, this petition for reconsideration is timely, 42 U.S.C, 7607(b)(1).*
1 Concurrent with the filing o f this petition for consideration, pursuant to 42 U.S.C. 7607(b), on September 9, 2016, SIPC filed for judicial review of EPA's nonattainment designation for Williamson
Sierra Club v. EPA 18cv3472 NDCA
Tier 7
ED 002061 00070535-00006
I. INTRODUCTION On July 12, 2016, EPA published the 2016 Designations. 81 Fed. Reg. 45,039. See
Exhibit 1. The rule established air quality designations under the revised Primary National Ambient Air Quality Standard for Sulfur Dioxide, 75 Fed. Reg. 35,520 (June 22, 2010) (`'2010 SO? NAAQS"), for certain areas subject to a March 2, 2015, consent decree entered by the U.S. District Court for the Northern District of California ("Consent Decree"). Sierra Club v. McCarthy, No. 13-CV-03953-SI, 2015 WL 889142 (N.D. Cal. Mar. 2, 2015), appeal docketed, 15- 15894 (9th Cir. May t, 2015). S1PC is the owner and operator o f the Marion Power Station located in Williamson County, Illinois ("Marion Station"). Si PC submits this petition for reconsideration ("Petition") pursuant to Section 307(d)(7)(B) (42 U.S.C. 7607) of the Clean Air Act ("CAA") o f EPAN designation o f Williamson County, Illinois, as nonattainment of the 2010 SCfc NAAQS.
This Petition focuses narrowly on the nonattainment designation for Williamson County, S1PC requests that EPA reconsider the nonattainment designation and instead designate Williamson County (which includes the area around Marion Station) as in attainment of the 2010 S02 NAAQS. In the alternative, if EPA believes that there is insufficient information to support aii attainment designation, SiPC requests that EPA designate Williamson County as unciassifiable given the absence of any valid and compelling data demonstrating nonattainment and the significant conflicts between the modeling submitted to EPA. In no event, however, does the available information support a finding of nonattainment.
Reconsideration and reversal of the Williamson County nonattainment designation for the 2010 S02 NAAQS is warranted for the following reasons:
County, Illinois in the United States Court of Appeals for the Seventh Circuit. SIPC v. EPA, Case Mo. 16- 3398 (filed September 9, 2016).
2
Sierra Club v. EPA 18cv3472 NDCA
Tier 7
ED 002061 00070535-00007
(!) EPA's nenattainment designation was based upon unreliable and non-representative data and inaccurate assumptions and inputs;
(2) Modeling data and supporting information submitted by SiPC during the public comment period demonstrates that Williamson County is attaining the 2010 SO2 NAAQS;
(3) Updated modeling data submitted with this Petition provides corroborative and compelling support that the third-party modeling relied upon by EPA - when corrected for known deficiencies - in fact demonstrates that Williamson County attains the 2010 SO2 NAAQS; and
(4) EPA may not base a nonattairsment designation on modeling data alone, II. BACKGROUND
A. Factual Background On June 22, 2010, EPA promulgated a primary SO? NAAQS at a one-hour SO2
concentration o f 75 parts per billion (ppb) (196.5 ug/m3) (established by a 3-year average of the annual 99th percentile of i-hour daily maximum concentrations), Consistent with Section 107(d) of the CAA, the Illinois Environmental Protection Agency ("Illinois EPA"') submitted its initial designation recommendations to EPA on June 2, 2011, including a recommended designation of unclassifable for Williamson County/ On July 25, 2013, EPA promulgated a final rule establishing air quality designations for 29 areas under the 2010 SO? NAAQS. Air Quality Designations for the 2010 Sulfur Dioxide (S02) Primary National Ambient Ah* Quality
2 In accordance with the CAA and EPA guidance, Illinois EPA designated all counties that lacked sufficient S02 ambient air quality monitoring data as unclassiftable. See Illinois EPA letter to EPA Region V re 2010 SO? NAAQS, at 1 (June 2, 2011), www.epa.gov/sUes/production/files/20l603/documents/i bree.pdf.
3
Sierra Club v. EPA 18cv3472 NDCA
Tier 7
ED 002061 00070535-00008
Standard, 78 Fed. Reg. 47,191 (Aug. 5, 2013), However, EPA did not issue a designation for Williamson County,
As a result o f EPA's failure to issue timely designations, three lawsuits were filed against EPA alleging that EPA had failed to perform a nondiscretionary duty under the CAA, indeed, some petitioners argued that EPA is required to designate areas unclassifiable with respect to the 2010 SCA NAAQS where EPA is not able to determine, based upon available information, whether the area is or is not meeting the I-hour SCA NAAQS. In an attempt to resolve EPA's failure to act, as noted above, in March 2015 EPA entered a Consent Decree with the Sierra Club and the Natural Resources Defense Council. See Sierra Club v. McCarthy, supra, Exhibit 2. Notably, the Consent Decree established an artificial, accelerated timeline that required that EPA designate by July 2, 2016 any undesignated area under the 2010 SO2 NAAQS that contains any large stationary source of SCA emissions. Id at 4. As discussed below, both the CAA and EPA historic practice mandate the use of actual, monitored, ambient air quality to designate regions as attainment, nonattainment, or unclassifiable of a NAAQS. The Consent Decree inappropriately precluded the ability to use monitoring data for areas containing large stationary sources without existing monitors, including SIPC's Marion Station in Williamson County.
As a consequence, EPA had but sixteen (16) months to solicit and attempt to evaluate any available data upon which to base its 2016 Designations. On September 18, 2015, Illinois EPA submitted modeling data to support Illinois EPA's recommendation that EPA designate the areain the vicinity of Marion Station (Williamson County) as attaining the 2010 SO2 NAAQS. See Illinois EPA letter to EPA Region V re: Updated Recommendations for 2010 SO2 NAAQS (September 18, 2015). and accompanying Illinois EPA Technical Support Document for SOi Designation Recommendations for Electric Power Facility Areas (September 18, 2015),
4
Sierra Club v. EPA 18cv3472 NDCA
Tier 7
ED 002061 00070535-00009
collectively Exhibit 3. Shortly thereafter, EPA Region 5 notified Illinois EPA directly that it had received public comments and modeling data from Sierra Club that supported a nonatiamment designation for the Marion Station area, and solicited comments from Illinois EPA on the reasonableness of Sierra Club's modeling.3 See EPA Region 5 letter to. Illinois EPA, re; Sierra Club modeling (October 20, 2015), Exhibit 4. On November 9, 2015, Illinois EPA concluded that Sierra Club's modeling could not be relied upon because it was flawed and not representative of ambient air quality around the Marion Station. See Illinois EPA letter to EPA Region V, re; Sierra Club Sulfur Dioxide Modeling (Nov. 9. 2015), Exhibit 6. in particular, Illinois EPA objected to Sierra Club's failure to: (i) utilize variable stack temperature and flue gas exit velocities required by EPA's Modeling Technical Assistance Document for the 2010 SO? NAAQS (``Modeling TAD") (Exhibit 7); (ii) utilize fence! me receptors to demarcate nonambient air areas (i.e,, those where public access is precluded); and (lit) utilize more representative background concentrations and inputs. Id,
On or about December 21, 2015, SIPC's consultant AECOM, submitted modeling to Illinois EPA to support an attainment finding for Marion Station. See AECOM, Marion Power Plant Southern Illinois Power Cooperative S02 Modeling Archive (Dec. 21, 2015), Exhibit 8. While AECOM's modeling results mirrored Illinois EPA's finding o f attainment, its modeling differed from Illinois EPA's modeling in several key respects. AECOM corrected source characterization and modeling inputs utilized in Illinois EPA's modeling submitted to EPA, including correcting stack diameters and building heights at the Marion Station and the handling of invalid data affecting stack temperature and exit velocity inputs. Most significantly,
3 Sierra Club's modeling was not made publicatly available for any entity other than Illinois EPA to review and evaluate. Indeed, SiPC did not obtain a copy of Sierra Club's modeling - upon which EPA based Its final nonatlainment designation for Marion Station, under the 2016 Designations - until August 24, 2016, in response to a Freedom of Information Act request submitted by S1PC. See Exhibit 5.5
5
Sierra Club v. EPA 18cv3472 NDCA
Tier 7
ED 002061 00070535-00010
consistent with EPA's Modeling TAD, AECOM corrected Illinois EPA's inclusion of modeling receptors in areas where public access Is prohibited on Marion Station's property. See Illinois EPA 120~Day Comment letter to EPA Region V on Proposed Designation, at 1 (April 19, 2016), Exhibit 9. Unfortunately, Illinois EPA did not submit AECOM's corrected modeling to EPA, See EPA Technical Support Document ("TSD") for Illinois* at 59,4 for EPA Region V, Proposed Designation for Illinois under 2016 Designations (February 16, 2016), collectively Exhibit 10,
On February 16, 2016, without the benefit of AECOM's corrected modeling, EPA issued its preliminary recommendation to designate the area around Marion Station in nonattainment of the 2010 SO? NAAQS based upon modeling data in EPA's possession at the time, which consisted only o f the uncorrected Illinois EPA modeling and the Sierra Club modeling that Illinois EPA had already determined was flawed for multiple reasons,5 id. While EPA admitted that Siena Club used "less reliable" data in their model, the Agency nonetheless based Us initial designation on Sierra Club's modeling because - according to EPA - it was "a more reliable assessment" of ambient conditions around Marion Station than the uncorrected Illinois EPA modeling. Id, Notably, the deficiencies identified by AECOM in Illinois EPA's modeling were similarly present in Sierra Club's modeling, including the placement of modeling receptors on non-public property within the property boundary of Marion Station, See Final TSD for Illinois for 2016 Designations, at 22, 31 (July 2, 2016) ("Final TSD"), Exhibit 11. EPA did not properly evaluate whether the modeling receptors were inappropriately located on non-public property. See, infra, at Section IV.C.i.
4 EPA makes no reference in the technical analysis for Williamson County that it knew of, received or evaluated the modeling AECOM submitted on behalf of SIPC to Illinois EPA in December 2015.
5 EPA did not perform any independent modeling of Marion Station. EPA's preliminary arid, later, final designation was based exclusively on EPA's review of Illinois EPA and third-party modeling data,
6
Sierra Club v. EPA 18cv3472 NDCA
Tier 7
ED 002061 00070535-00011
To ensure that EPA reviewed and considered SIPC's air quality modeling, which included necessary corrections to Illinois EPA's modeling, AECOM submitted modeling directly to EPA on behalf of S1PC on March 30, 2016. The modeling was largely identical to the modeling submitted by AECOM to Illinois EPA in December 2015, with the exception that AECOM relied upon more current emissions data from 2013-2015 and the most current version of AERMOD (the modeling software utilized by Illinois EPA and Sierra Club and recommended by EPA's Modeling TAD), Subsequently, on April 19, 2016, Illinois EPA submitted to EPA a 120-Day Comment letter in response to EPA's proposed nonattainment designation for Marion Station. See Exhibit 9. in that formal response, Illinois EPA found that AECOM's March 2016 modeling, which was submitted to EPA, corrected and improved upon Illinois EPA's modeling and provided substantia] support for designating the area around Marion Station in attainment of the 2010 SO2 NAAQS. Id. at 1,3 ("supporting] the methodology and inputs, and agreeing] with the results" of AECOM's "re-model" that utilized "refined inputs" from those used by Illinois EPA). In addition, Illinois EPA concluded that ACCOM's receptor placement "followed Federal guidance provided In the [Modeling TAD]." Id. at 2,
On July 2, 2016, EPA promulgated the 2016 Recommendations. EPA rejected Illinois EPA's recommendation and the supporting modeling independently provided by AECOM and designated Marion Station (and Williamson County) as nonattainment of the 2010 SO? NAAQS.
B. Regulatory Background EPA's directive under the CAA is unambiguous and unassailable. An area that "does not
meet" the NAAQS is designated nonattainment. An area that "meets the [NAAQS]" is designated attainment. And an area that "cannot be classified on the basis o f available information as meeting or not meeting the [NAAQS]" is designated unclassifiabie. 42 U.S.C.
7
Sierra Club v. EPA 18cv3472 NDCA
Tier 7
ED 002061 00070535-00012
740?{d)(l)(A)(i)~(iii) (emphasis added). EPA may not base a designation on just any available information. Rather, the data must be sufficiently compelling and reliable that it is representative of actual ambient air quality. See EPA Responses to Significant Comments on the 2016 Designations, at 78 (June 30, 2016) ("designations should be made based on representative data, including for both unclassifiable/attainment and nonattamment") (emphasis added), www.epa.gov/sites/productton/files/2016-07/docunients/so2d-r2-response-to-comments06302016.pdf {"Response to Comments"); see also EPA, Draft TSD Nebraska Area Designation for the 2010 SCK Primary National Ambient Air Quality Standard 47 n.6 (March 30, 2016), www3.epa.gov/so2designations/round2/07_NEjsd.pdf ("[area] [designations are intended to address current actual air quality,.,and, thus, are unlike attainment plan modeling, which must provide assurances that attainment will occur") (emphasis added).
An area is designated unclassifiable where there is no compelling data indicating attainment or nonattainment. By analogue, an attainment or nonattainment designation is justifiable only where compelling data exists. The need for compelling data requires that EPA not rely upon data that is "reasonably considered] to be unsound" in issuing a designation. Mississippi Comm'n Envtl. Quality v. EPA, 790 F.3d 138, 154 (D.C. Cir. 2015). As a result, "[i'Jn the absence of information clearly demonstrating a designation of 'attainment' or `nonattainment,'" EPA informed states that the CAA requires issuance of an unclassifiable designation. See EPA, "Updated Guidance for Area Designations for the 2010 Primary SO2 NAAQS," at 5, www.epa.gov/sites/production/files/2016-06/documents/20150320so2designations.pdf (March 20, 2015) (emphasis added) ("EPA 2015 Designation Guidance").6
" See also, supra, EPA, Draft TSD for Nebraska, at 33. 8
Sierra Club v. EPA 18cv3472 NDCA
Tier 7
ED 002061 00070535-00013
Most recently, EPA reversed its initial proposed nonattainment designation for Franklin County, Missouri based on the presence of conflicting modeling data results. See EPA Final TSD for Missouri under 2016 Designations (July 2, 2016), www.epa.gov/sites/prQduction/files/2016-07/doeuments/r7 mo final designationJsd_ 070!2016.pdf. Both Ameren Missouri (the owner of the primary SO? source in Franklin County) and the Missouri Department of Natural Resources submitted modeling supporting an attainment or uneiassifiahle designation. Sierra Club modeling - like here - supported a finding of nonattainment. In reviewing the available modeling data, EPA concluded that an uneiassifiahle designation was warranted because of flaws identified in each of the modeling runs submitted. See Response to Comments at 78 (unclasslfiable appropriate "[bjased on all the available and reliable data, including new information following the notification o f our intended nonattainment designation"),7
Taken as a whole, in order for a designation to be based on "actual air quality" and thus reasonable and lawful - the data (whether monitoring or modeling) relied upon must be considered:
(1) reliable - meaning the data must not be based on incorrect or questionable inputs or assumptions:
(2) supportable..meaning that the available data does not support a different designation: (3) representative - meaning that the data accurately reflects conditions that are, or have
in the past, occurred in the designation area; and
1See also EPA Final TSD for Missouri, at 26 ("the EPA's view is that the modeling results widely vary and greatly depend upon how the modeling was conducted, as discussed in this Technical Support Document. Because o f the issues present In the modeling methodologies, the EPA does not have a clear basis to determine whether the area currently meets or does not meet the 2010 S02 NAAQS based on all currently available information").
9
Sierra Club v. EPA 18cv3472 NDCA
Tier 7
ED 002061 00070535-00014
(4) inclusive - meaning that the data, takes into account all available information (provided it is reliable and representative).
As further discussed below, EPA relied upon Sierra Club and Illinois ERA modeling data that was neither reliable nor representative because the modeling contained incorrect data inputs and assumptions. Indeed, Illinois EPA acknowledged as much in Its November 9, 2015, letter to EPA. See Exhibit 6, At the same time, EPA improperly disregarded modeling data earlier submitted by AECGM that supported an attainment designation. In addition, modeling refinements submitted with this petition further support an attainment designation. Therefore, EPA's nonattainment designation for Marion Station was not based upon all available, reliable, information. When all reliable information is considered, modeling demonstrates that the area around the Marion Station Is attaining the 2010 SO2 NAAQS. The existence of compelling data requires that EPA designate the area in attainment. At most, if EPA finds that there are legitimate conflicts between the modeling submitted by the various parties, then an unclassifable designation is warranted. In no case does the available information support a nonattainment designation, HI. STANDARD FOR RECONS1 DERATION
Upon request, the CAA directs EPA to reconsider any issue that has been raised "with reasonable specificity" during the public comment period. 42 U.S.C. 7607(d)(7)(B). This does not mean that EPA is precluded from considering new information. Indeed, the CAA provides that EPA must consider issues that are o f "central relevance to the outcome of the rule" that were "impracticable to raise" or that "arose after the period for public comment." Id. An issue is of "central relevance" If it "provides substantial support for the argument that the promulgated regulation should be revised." EPA, Basis for Denial of Petitions to Reconsider the CAA
10
Sierra Club v. EPA 18cv3472 NDCA
Tier 7
ED 002061 00070535-00015
Section 111(b) Standards of Performance for Greenhouse Gas Emissions from New, Modified, and Reconstructed Fossil Fuel-Fired Electric Utility Generating Units 2 (April 2016), www.epa.gQv/siies/producttQn/files/20 i 6-04/documents/l 11b_recondocument aprl 12016.pdf.
Thus, on reconsideration, EPA must evaluate whether a reexamination o f existing information or consideration of new information substantially supports a different outcome. To the extent additional information can substantiate whether certain data relied upon or ignored by EPA impacted modeling results, ERA Is required to consider such information. See Catawba County v, EPA, 571 P.3d 20, 51 (D.C. Cir. 2009) (EPA calculations on reconsideration demonstrated that a correction in speeiation profile would not have impacted area designations for the 1997 particulate matter NAAQS); see also Response to Comments at 78 (finding "new information" sufficient to reverse a proposed designation under the 2010 S 0 2 NAAQS). Similarly, EPA must also consider information that evidences that data or information relied upon by the Agency to reach its designation was unreliable, non-representative or unsupportable. IV. ARGUMENT
A. EPA's Reliance on Third-Party Data to Discredit a State Attainment Designation Was Arbitrary anti Capricious. The NAAQS designation process is a cooperative effort between the EPA and the states.
See, e.g., Med. Advocates fo r Healthy Air v. U.S. E.P.A., No. C 11-3515 SI, 2012 WL 710352, at *1 (NT). Cal. Mar. 5, 2012); see also S. Coast Air Quality Mgnit Dist. v. E.P.A., 472 p,3d 882, 886 (D.C. Clr, 2006) decision clarified on denial o f reh'g, 489 F.3d 1245 (D.C. Cir. 2007). While the designation process allows for public comment and input, neither the CAA nor EPA regulations empower third-parties, like Sierra Club, to control the designation process. See Mississippi Comm'n on Emil. Quality y. E.P.A., 790 F,3d 138, 145 (D.C. Cir. 2015) ("EPA works collaborative))'- with the states to determine the NAAQS-atta foment status for all areas
11
Sierra Club v. EPA 18cv3472 NDCA
Tier 7
ED 002061 00070535-00016
within a respective state's borders,") (emphasis added). Indeed, it is the state that issues the NAAQS designation recommendation to EPA, Thus, EPA's task was to review Illinois EPA's designation recommendation, not to review Sierra Club modeling and base a nonattainment designation on such third-party data. See Pennsylvania, Dep't o f Envtl, Prot. v, E.P.A., 429 F.3d 1125, 126 (D.C, Cir. 2005) (Qajfter receiving state recommendations, EPA promulgates final designations").
But here, EPA effectively ignored both Illinois EPA's modeling and recommendations provided to EPA no less than three separate times - to designate Marion Station as attainment of the 2010 S02 NAAQS, As noted above, on September 18, 2015, Illinois E-PA submitted modeling supporting an attainment designation. On November 9, 2015, Illinois EPA again notified EPA of its recommendation to designate Williamson County as attainment, this lime by way of itemizing the deficiencies in Sierra Club's modeling that ERA later relied upon in its 2016 Designations, Lastly, on April 19, 2016, Illinois EPA agreed with AECOSVPs modeling (submitted on behalf o f SPC) and determined that it provided substantial evidence supporting an attainment designation for Marion Station,
As discussed in more detail below, EPA did - albeit briefly - evaluate the modeling provided by Illinois EPA. In the Final TSD, EPA concluded that both Illinois EPA and Sierra Club modeling were based on potentially significant deficiencies in modeling inputs, See Exhibit 11 at 30 (Illinois EPA and Sierra Club modeling relied on different meteorological data, emission average periods, Marion Station stack parameters and building heights; and modeling receptor placements). Yet, in substantiating its nonattainment designation, EPA relied upon Sierra Club modeling despite the Agency's acknowledgement that it used "modeling inputs in a less reliable fashion than Illinois (ERA]/' See Exhibit 10 at 59. Reliance on unreliable and non-
12
Sierra Club v. EPA 18cv3472 NDCA
Tier 7
ED 002061 00070535-00017
representative data is, on its own, impermissible in promulgating a rule. But the Agency's decision-making is even more flawed taking into consideration that ERA looked only towards Illinois EPA's modeling to substantiate the "credibility]'' of Sierra Club's modeling. See Exhibit 11 at 31. That is, EPA speculated that if Illinois EPA's modeling had used emission rates relied upon by Sierra Club, Illinois EPA would have modeled nonattainment. See id, But this ignored other errors in Illinois EPA's modeling. See Exhibit 9 ai 1 (acknowledging "discrepancies" in its modeling identified by AECOM). At best, EPA's misuse o f Illinois EPA's modeling can be characterized as an improper extraction and recharacterization of third-party data that ignored underlying flaws and inaccuracies. Indeed, EPA itself acknowledged that Sierra Club's modeling was based on faulty and non-representative modeling inputs. See Exhibit 11 at 30, 31; see also Exhibit 1(1 at 59, For all of these reasons, EPA's decision to rely on Sierra Club modeling was improper.
B, It was Arbitrary and Capricious for EPA to Designate Williamson County under the 2010 SO? NAAQS Based on Modeling Data Alone. If a designation may be based upon modeling data, then S1PC believes for the reasons set
forth in this petition that the area around the Marion Station should be designated attainment, However, EPA's reliance upon modeling data alone to make its designation is, by itself, impermissible and a separate basis for setting aside the nonattainment designation.
Neither the CAA nor EPA regulations explicitly authorize EPA to rely on modeling data alone to designate an area as attaining or not attaining a NAAQS. Indeed, initial area designations have historically always been based upon monitoring data rather than modeling data. See Exhibit 7 at 1,8 This is because, in part, modeling is generally conservative in contrast
8See also EPA Region V letter to Governor of Illinois Pat Quinn, re response to Illinois EPA air quality recommendation for 2010 S 0 2 NAAQS, at ! (designating as nonattainment only those areas where ''monitoring data indicates violations of the 1-hour SO2 standard," while leaving areas without
13
Sierra Club v. EPA 18cv3472 NDCA
Tier 7
ED 002061 00070535-00018
to monitoring, and tends to over-predict ambient concentrations when determining compliance with a NAAQS, See 40 C.F.R. Part 51, App. W, subsection 9,1.2 (Studies of Model Accuracy) ("[mjodels are more reliable for estimating longer time-averaged concentrations than for estimating short-term concentrations at specific locations"); subsection 10.2,2(b) (Use of Measured Data in Lieu of Model Estimates) (noting that reliance on monitoring data alone is acceptable "only in the ease of a NAAQS assessment"). Reliance on modeling alone is also inconsistent with the plain language of the CAA. See 42 U.S.C. 7407(d)(1)(A) (attainment, nonattainment and unclassifiable designations should be based on actual emissions); see also 42 U.S.C. 7407(d)(3) (redesignations are to be "on the basis of air quality data") (emphasis added); 42 U.S.C, 7407(d)(6)(A) (designations for the 1997 PM2.5 NAAQS must be "based on air quality monitoring data) (emphasis added).
With respect to the 2010 SO2 NAAQS, ERA initially declared that area designations must be determined by ambient SO2 monitoring data. See 74 Fed. Reg. 64,810, 64,846 (Dec. 8, 2009) ("monitoring data are used to determine whether an area is in violation of the SO2 NAAQS"). In EPA's final rule, ERA confirmed that compliance should be determined "based on 3 years of complete, quality assured, certified monitoring data." 75 Fed. Reg. at 35,569. However, for the first time ERA also discussed a hybrid modeling and monitoring approach under the 2010 SCK NAAQS where insufficient monitoring data was available. Id at 35,570-71, ERA based its change in position on recognition that many areas would initially be designated unclassifiable because of a lack of ambient monitoring data. Id, at 35,571. That did not occur in Williamson County, instead, the accelerated designation schedule under the Consent Decree precluded a designation based on monitoring data.
monitoring data for "future actions") (emphasis added), www.epa.gov/sites/production/files/201603/doeuments/il-epa-resp.pdf,
14
Sierra Club v. EPA 18cv3472 NDCA
Tier 7
ED 002061 00070535-00019
As noted above, ERA entered into a Consent Decree with Sierra Club that forced the accelerated 2016 Designations for certain yet-undesignaied areas under the 2010 S 0 2 NAAQS. Yet the Consent Decree, as a matter o f law, cannot authorize an otherwise prohibited action under the CAA. That is, the CAA (and EPA's own statements in the preamble to the 2010 S 0 2 NAAQS) directs ERA to designate areas that lack, sufficient available monitoring data as unclassifiable, See 42 U.S.C, 7407{d)(l)(A)(m). On EPA's own admission, the Consent Decree "affect[ed]...the extent to which the EPA will be able to use [monitored data]." See EPA, Data Requirements Rule for 2010 S02 NAAQS, 80 Fed. Reg. 51,052, 51,083 (Aug. 21, 2016). As a result because of the absence of actual monitored data to designate Williamson County under the 2010 S 0 2 NAAQS, EPA was required under the CAA to, at worse, designate the area as unclassifiable. EPA's reliance on modeling to designate Williamson County as nonattainment o f the 2010 SO2 NAAQS was inconsistent with the Clean Air Act and, thus, arbitrary and capricious.
C. EPA Improperly Relied Upon Flawed Modeling to Base Its Nonattainment Designation. EPA's nonattainment designation for Marion Station is fundamentally flawed because it
ignores and mischaraetenzes available data and is based on a number of inaccurate, unreliable and non-representative modeling inputs, as described and admitted by EPA in the Final TSD. This is not proper. See Motor Vehicle Mfrs. Ass'n ofU.S., Inc. v. State Farm Mat. Auto. Ins. Co., 463 U.S. 29, 43 (1983) (an agency "must examine the relevant data and articulate a satisfactory explanation for its action including a `rational connection between the facts found and the choice made'"). SIPC addresses the flaws in EPA's assessment and designation conclusion below,
i, EPA Improperly Included Receptors on Secured Non-Public Property. EPA relied upon insufficient, incomplete and inaccurate data to determine that the area
15
Sierra Club v. EPA 18cv3472 NDCA
Tier 7
ED 002061 00070535-00020
located to the northeast o f Lake Egypt Road ("Northern Property") constituted ambient air and should be included in any model for Marion Station, instead, available data demonstrates that public access is prevented on the Northern Property to the extent necessary to exclude modeling receptors on that property, in addition, to avoid any doubt on the issue SIPC has taken additional measures to ensure public access is prohibited.
NAAQS designations are intended to reflect actual ambient air quality conditions. Thus, because the CAA dictates the use of actual emissions, EPA designations are generally based on actual data collected from areas where ambient monitors can be feasibly located. As modeled data is intended to provide data representative of actual conditions, EPA policy directs sources to avoid modeling areas where public access is generally prohibited. See Exhibit ? at 8-9. EPA correctly noted that a "fey factor in in determining whether plant property is considered ambient air is the degree to which the public has access to the area." Exhibit 11 at 23 (emphasis added). Thus, public access does not need to be impossible. See Memorandum from Stephen D. Page, Director, Office of Air Quality Planning & Standards, EPA at 6 n.l (June 22, 2007). Exhibit 12 ("[pjreclude does not necessarily imply that public access is absolutely impossible, but rather that the likelihood of such access is small"). Rather, consistent with EPA guidance and past practice, EPA must consider the totality of measures taken to preclude public access to determine whether they are an "adequate barrier to preclude access to the public," EPA Model Clearinghouse Record, No, 99-V-03 (September 18, 1999) (discussing a "combination" o f measures used to restrict access) (emphasis added). Exhibit 13; see also EPA Model Clearinghouse Record, No. 99-V-02 (August 30, 1999). Exhibit 14,
EPA did not follow its own directive. Indeed, EPA admitted that it lacked sufficient information "as to the manner or degree to which public access is restricted" to determine
16
Sierra Club v. EPA 18cv3472 NDCA
Tier 7
ED 002061 00070535-00021
whether the public access to the Northern Property is adequately protected. See Exhibit 11 at 23. EPA's lack of information on public access is clue, in part, to Illinois EPA's failure to provide EPA with the modeling AECOM submitted on behalf of SI PC to Illinois EPA on December 21, 2015, SIPC excluded modeling receptors from the Northern Property in its December 2015 modeling. Neither Illinois EPA nor Sierra Club excluded receptors from the Northern Property, Id. at 22. As a consequence, this Petition is the first time that SIPC has had the opportunity to respond to EPA's apparent confusion about site control at Marion Station and to provide the information needed for EPA to properly determine that the general public does not, in fact, have access on the Northern Property, See id. at 23; see also 42 U.S.C, 7607(d)(7)(B) (issues or data that were impracticable to raise during public comment may be considered in a petition for reconsideration).
Nevertheless, rather than seeking information from SIPC about site access, EPA apparently looked for and relied upon just two (2) stock Google Maps images (a single Google "street view" image and a single Google satellite image) to conclude that the public had "easy access" to "most" of the Northern Property, Id. at 24-25, Figures 7 and 8, This is wholly deficient from both an evidentiary and a legal standpoint, EPA's historic practice is to rely on information provided by the source and, where appropriate, a site-visit to determine if "ambient air" exists. See Exhibit 13 (evidencing EPA's reliance on information provided about on-site security measures and a "tour of the,,.facility," which EPA described as being "useful and should help bring [the site access] issue to resolution"); Exhibit 14 (ERA noting that a "site visit will likely be necessary" to determine whether public access is adequately precluded), No such site-visit ever occurred at Marion Station,9 EPA's Final TSD admits that it based its evaluation
<J SIPC has no record o f any EPA site visit to evaluate whether public access was precluded at or in the vicinity of Marion Station,
17
Sierra Club v. EPA 18cv3472 NDCA
Tier 7
ED 002061 00070535-00022
on the degree to which public access is available on the Northern Property entirely upon the two Google images.
As a result, EPA's erroneous conclusion about public access is based upon the Agency's superficial review o f images that appear to depict a relatively low guard rail separating the Northern Property from Lake Egypt Road. However, images are often deceiving. Had EPA visited Marion Station, the Agency would have recognized that the satellite imagery wholly fails to reflect that public access is more than adequately prohibited for the following reasons:
(1) Access to the Northern Property along the 300-yard section depicted in Figure 7 of EPA's Final TSD is prohibited by man-made barriers: a. A barbed-wire fence prohibits access to the Northern Property along either side o f the brief 300-yard guard rail section. See Exhibit 15 and Exhibit 16. The barbed-wire fence abutting this section is not visible in either image relied upon by EPA. b. The 300-yard guard rail section was located on top of a spillway servicing the Lake of Egypt. See Exhibit 16 (illustrating Lake Egypt Road sitting atop of spillway, with Lake of Egypt to the left and the Northern Property to the right); see also Exhibit 17 (security camera image depicting Lake Egypt Road separating the Northern Property and the Lake o f Egypt). The Lake Egypt Road runs parallel immediately to the South of this 300-yard section. No pedestrian crossing or walkway is present alongside the Lake of Egypt road at this location. See id. Vehicles travel at or above the 55 mph listed speed limit, restricting the likelihood of public incursion onto the Northern Property, c. The spillway is elevated over thirty (30) feet above the Northern Property,
18
Sierra Club v. EPA 18cv3472 NDCA
Tier 7
ED 002061 00070535-00023
See Exhibit 18 (photograph taken at the base of the 30 foot elevation drop). A steep gradient descends onto the Northern Property immediately to the North of the guardrail depicted in Exhibits 15 and 16. See Exhibit 17. (2) Natural barriers further restrict public access. a. The Lake o f Egypt to die South of the Northern Property restricts public access points. See Exhibit 16 and Exhibit 17. b. The significant gradient between the guard rail and the Northern Property is a deterrent for public access. See Exhibit 17. (3) SIPC has always - and will continue - to take additional reasonable measures to secure against public access to the Northern Property. Specifically; a. Warning signs are posted along the entire Northern Property boundary notifying the public that the area is restricted. See Exhibit 16 and Exhibit 19. b. SIPC operates and maintains a security camera that is regularly monitored by the on-duty boiler operator at Marion Station. See Affidavit of Robert Todd Gallenbach, Vice President of Operations for SIPC at Marion Station, at Jf 57, Exhibit 20. The security camera feed provides a complete, unobstructed, view of all access points to the Northern Property (i.e., the area depicted in Figure 8 of EPA's Final TSD). Id at Attachment I; see also Exhibit I7.i0 In the unlikely event that an unauthorized member o f the public gains access to the Northern Property, the boiler operator monitoring the security camera will notify security to promptly remove such individuals from the property. See Affidavit a t f 8.
m SIPC also operates another security camera which covers other portions of the Marion Station property (areas where public access is also restricted through natural and man-made barriers),
19
Sierra Club v. EPA 18cv3472 NDCA
Tier 7
ED 002061 00070535-00024
(4) In an abundance o f caution, SIPC recently added a barbed-wire security fence behind the existing guardrail located on the 300-yard stretch of spillway with a barbed-wire security fence,n In addition, new "restricted area" warning signs were posted on the new barbed-wire fencing. Images of the fencing and warning signs are provided in Exhibit 21, Exhibit 22, Exhibit 23 and Exhibit 24, While public access has been adequately precluded by existing, long-standing, site control measures, the new fence in this limited area provides still further assurance that the Northern Property is secure from public access,
SIPC assumes that USEPA was unaware of these significant site control measures at the time it made its final designation. Based upon these measures, collectively, it is clear that public access to the Northern Property is more than adequately prohibited.
EPA erred by concluding that public access is available merely because of two historic images that showed the absence of a fence along a brief 300-yard section. See In the Matter o f Hibbing Taconite Company, PSD Appeal No. 87-3, at 17-18 (July 19, 1989) ("The test for ambient air exclusion does not require a continuous fence around the perimeter o f the property. Other types o fphysical harriers can effectively preclude access,") (emphasis added), Exhibit 25. EPA has previously found that physical barriers like those described above, in conjunction with security sign postings and 24-hour security camera surveillance is more than adequate to preclude public access, Exhibit 13 ("[a] combination of fencing and other physical barriers, posting, and use of the 24-hour security camera surveillance and truck patrolling system should provide an adequate barrier to preclude access to the public"). In fact, EPA has found the use of video surveillance alone adequate to restrict access to the general public. See Exhibit 13,
!1 See Exhibit 13 (finding the installation of "more fencing, and signage,,,appropriate" to ensure site access is restricted),
20
Sierra Club v. EPA 18cv3472 NDCA
Tier 7
ED 002061 00070535-00025
BPA tsef acknowledged that the "removal of receptors from the area northeast of the plant may represent the most significant difference between [AECOM's] modeling and the modeling by Illinois [EPA] and Sierra Club, Exhibit 11 at 30 (emphasis added).12 Thus, It Is irrefutable that the proper selection of receptor locations is of "central relevance to the outcome" of EPA's 2016 Designations for Marion Station, indeed, in Illinois EPA's 120-Day Comment Letter to ERA advocating for an attainment designation for Marion Station, Illinois EPA informed EPA that it supported AECOM's modeling in part because it properly excluded receptors from areas (like the Northern Property) where public access is prohibited. See Exhibit 9. EPA's finding that the public has access to the Northern Property both ignored contrary information and was based upon insufficient information. As a result, EPA's determination was improper, When all available information is considered, the evidence substantially supports the finding that Illinois EPA and Sierra Club modeling improperly included receptors on the Northern Property,
When the receptors on the Northern Property are excluded, as required by EPA's guidance given SIPC's substantial site control measures, modeling conclusively demonstrates attainment of the 2010 SO2 NAAQS. See Exhibit 11 at 28, Table 4 (AECOM's March 2016 Modeling demonstrates default modeling of 190,4 ug/m3); see also, infra, Section 1V,D, AECOM's Updated Modeling for Marion Station, September 2016, at Section 5.0, Table 5-1 (demonstrating default modeling of 191,0 ug/m3).
ii> It was improper fo r EPA to Rely on Modeling Based on Inaccurate or NonRepresentative Modeling inputs,
Aside from reliance upon the improper inclusion of receptors on the Northern Property,
u SPC agrees. As evidenced by the modeling submitted by AECOM on March 30, 2016 and with this Petition, the exclusion of receptors in the Northern Property has a significant Impact on both the total ambient levels around Marlon Station and the area of peak concentration (which occurs within the Northern Property /'receptors are not excluded). See infra, Section IV.D,
21
Sierra Club v. EPA 18cv3472 NDCA
Tier 7
ED 002061 00070535-00026
EFA's nonattainment conclusion is flawed because it relies upon modeling that contains other significant flaws while essentially ignoring the ABCGM modeling that supports an attainment designation and that Illinois EPA supported.
There is no dispute that modeling provided by Illinois EPA and Sierra Club rely upon modeling parameters and inputs that differ significantly from those used by AECOM. Both Illinois EPA and EPA highlight five significant mistakes (in addition to receptor placement) in the modeling provided by Illinois EPA and Sierra Club: (1) background data: (2) emission rates; (3) stack parameters; (4) building heights: and (5) hourly stack temperature and exit velocity. See Exhibit 9; Exhibit 11 at 21-30. Significantly, because EPA did not perform an independent modeling analysis, EPA was "not dear how significant [of an impact].. The revisions to stack parameters and building heights'5 (as well as other modeling parameters) would have on modeling results. Exhibit 11 at 30. Nonetheless, EPA issued a nonattainment designation based upon these flawed modeling submissions even in the face of Illinois EPA's information that made clear that the modeling submissions by it and Sierra Club were flawed and did not reflect the most accurate data..
Modeling must "'dearly demomtratfie] that an area and its associated boundary are properly designated `attainment' or `nonattainment'" to warrant anything but an unclassrfiable designation. See EPA 2015 Designation Guidance at Attachment 2, pg. 3 (emphasis added). ERA made no such demonstration here, Instead, in contradictory fashion, EPA on the one hand wholly disavowed ACCOM's modeling because it `inappropriately removed [modeling] receptors" (it did not, as discussed above) while simultaneously accepting Sierra Club (and Illinois EPA) modeling as "credible" despite the above-noted mistakes and inaccuracies in modeling inputs. See Exhibit 11 at 31. EPA may not rely upon "unsound" data in issuing a
22
Sierra Club v. EPA 18cv3472 NDCA
Tier 7
ED 002061 00070535-00027
designation. See Mississippi Comm'n, 790 F.3d at 154. And it certainly may not do so knowingly without significant justification or substantiation.
AECOM analyzed the modeling inputs relied upon by Sierra Club and determined that the differences have a material impact on modeling results. See AECOM, Updated 3-Hour SQi NAAQS Compliance Modeling Demonstration for the Marion Power Plant, at Section 2.1 (September 2016) ("2016 AECOM Report"), attached as Exhibit 26.ij
(1) Background data. Sierra Club used constant, rather than hourly and seasonal variations, in background concentrations. This does not conform with EPA's Modeling TAD. See Exhibit 7 at Section 5.2. Sierra Club also utilized a non geograph leally appropriate - and, thus, non-representative - background concentration derived from an ambient monitor in Oglesby, Illinois using data from 2011 through 2013, despite the lact that Sierra Club modeling used emissions data from 2012 through 2014, Finally, contrary to EPA's Modeling TAD, Sierra Club inappropriately included emissions from Ameren Missouri's Joppa Power Plant located over 48 kilometers from Marion Station. Id. at Section 8,1,
(2) Emission Rates. Sierra Club (and Illinois EPA) relied upon emission data that is not representative of current (and future) emission rates at Marion Station. In 2015, emission rates at Marion Station decreased significantly as a result of measures taken by SIPC to comply with the federal Mercury and Air Toxics Standard (MAI'S) and the Cross-State Air Pollution Rule (CSAPR). See Exhibit 11 at 26, Table 2 (illustrating that actual S02 emissions decreased by approximately 50% in 2015); see also Exhibit 26, at Section 5,0. As a result, AECOM's modeling, which relied upon
13 Notably, both Illinois EPA and EPA reached many of these same conclusions; yet, EPA ignored Illinois EPA's recommendations and relied upon Sierra Club's modeling for designation.23
23
Sierra Club v. EPA 18cv3472 NDCA
Tier 7
ED 002061 00070535-00028
emissions data from 2013-15, is far more representative than Sierra Club modeiing based on 2012-2014 emissions data, EPA's rationale - that the use of 2012-14 data is credible because it represents only a 7% increase from 2013-15 data utilized by AECOM - is irrelevant and fundamentally unsound. See Exhibit 11 at 30. A 7% difference could clearly make the difference between attainment and nonattainment,14 As noted above, area designations are intended to reflect actual ambient conditions. The use of emissions data that does not reflect federally mandated reductions generates results that are not representative of actual ambient conditions and is improper, (3) Corrected Source Characterization -- Stack Parameters and Building Heights. AECOM modeling used more representative building heights and stack parameters (stack diameter and stack height) for Unit 123 at Marion Station. Illinois EPA acknowledged that AECOM's source characterization parameters were more reliable and accurate inputs for modeiing. See Exhibit 9. As demonstrated in Section IV I) of this Petition, infra, even minor corrections to building dimensions can have a significant impact on attainment modeling. EPA acknowledged that Sierra Club's modeling included incorrect source characterization inputs but, in conclusory fashion, relied upon the absence o f information quantifying the impact of those deficiencies to substantiate EPA's reliance on Sierra Club's modeling, See Exhibit 11 at 3Q. EPA's reliance on insufficient and non-compelling data is contrary to both EPA's own Modeling TAD and the CAA,
14 Indeed, 7% of the 2010 S02 NAAQS limit of 196.4 iig/rrG amounts to an undeniably significant 14 ug/m3 difference,24
24
Sierra Club v. EPA 18cv3472 NDCA
Tier 7
ED 002061 00070535-00029
(4) Stack Temperature and Flow. AECOM modeling relied upon hourly varying stack temperatures, velocities and flow rate data. Not only is this approach consistent with EPA's Modeling TAD (as opposed to Sierra Club's use of constant values), but it allowed AECOM'5s modeling to substitute representative stack temperature data for periods without valid data. Critically, EPA agreed that such "input data improved the representation o f source characteristics.'' Exhibit ! I at 25 (emphasis added). As a result, EPA's reliance on modeling that used non-representative stack temperature and flow inputs was inconsistent with EPA's Modeling TAD and, therefore, not appropriate.
S1PC does not contend that EPA's designation was improper only because it relied on modeling that contained a single non-representative or inaccurate data input. Instead, SIPC contends that EPA's nonattainment designation was wholly based upon a plethora o f errors and inconsistencies in the modeling upon which EPA relied, while simultaneously ignoring or discrediting better Information provided by SIPC and AECOM, EPA's mandate Is to evaluate all available data and to designate an area as attainment or nonattainment only where compelling data supports that designation. Based on the totality of inaccurate, faulty and non-representative data in Sierra Club's modeling, that modeling cannot - and does not - provide compelling data that the area around the Marion Station is in nonattainment of the 2010 SQj NAAQ5. Therefore, EPA's reliance on Sierra Club's modeling was improper,
iii. EPA Improperly Ignored AERMOD Modeling that Utilized More Representative Source Characterization Inputs,
AECOM's March 2016 modeling included modeling runs that adjusted stack temperature inputs to more accurately represent emission characteristics from Marion Station stacks. See AECOM, Characterization of 1-Hour SO2 Concentrations for the Marion Power Plant (March25
25
Sierra Club v. EPA 18cv3472 NDCA
Tier 7
ED 002061 _00070535-00030
2016), Exhibit 27; see also Exhibit 8, EPA rejected these modeling rims without any evaluation because EPA argued they constituted an alternative model that required pre-approval under EPA's modeling regulations, 40 C.F.R. Part 53, Appendix W, See Exhibit 11 at 22. EPA is mistaken. While EPA correctly noted that the use of alternative modeling requires EPA approval under Appendix W, AECOM's adjustment o f stack temperature inputs to account for moisture in the flue gas (referred to as the "AERMOIST formulation") did not use or implement an alternate model. See Exhibit 8 (discussing corrections to various stack inputs). Appendix W does not require EPA approval to use more representative source inputs. Therefore, EPA must reconsider and evaluate AECOM's modeling runs utilizing AERMOIST.
Appendix W is unambiguous. Section 3.2.2 sets forth specific criteria and conditions to evaluate the "acceptability of an alternative m odels See 40 C.F.R. Part 51, Appendix W, subsection 3.2.1(a) (emphasis added); subsection 3.2.2(a) ("where "an alternative model is more appropriate than a preferred model, that model may be used subject to the recommendations of this subsection"); subsection 3.2.2(e) (identifying the 5 conditions for approval)ility of an alternative model).b Wholly absent from the rule is any requirement that conditions the use of an alternate input to an existing mode! upon EPA approval. This makes sense. Each emission source generally possesses unique source characteristics that must be input into a model to
iS EPA recently issued guidance clarifying that the use of so-called "beta options" in AERMOD modeling requires "adherejnce] to the requirements of Section 3.2 in the current 2005 version of the Guideline on Air Quality Models (Appendix W)." See EPA Memorandum, Clarification of the Approval Process for Regulatory Application of the AERMOD Modeling System Beta Options (Dec. 10, 2015), https;//www3.epa,gov/ttn/scram/guidance/clarification/AERMOD__Beta_Options Memo-20151210.pdf. Beta options such as "LOWW1ND3" and " ADJJU*" utilize changes to EPA's preferred modeling approaches under AERMOD. As a result, use o f beta option alternative models requires EPA approval under Appendix W. Both AECOM and Illinois EPA submitted data and information supporting the use of LOWW1ND3 at Marion Station in accordance with Appendix W. See Exhibit 8; Exhibit 9 ("Illinois ERA supports allowing the use of such beta options such as LOWW1ND3").26
26
Sierra Club v. EPA 18cv3472 NDCA
Tier 7
ED 002061 00070535-00031
generate a source-specific result. Changes in source characterization inputs are an inherent part of site-specific modeling and, therefore, do not require EPA approval."'
Unit 4 at Marion Station operates a wet limestone flue gas desulfurization device, which causes high stack moisture content in the flue gas and increased buoyancy in the resulting plume. See Exhibit 26 at Section 4,6. AERMOD can effectively model this unique source characteristic by merely changing the hourly input stack temperature of the model's default dry plume (/>., the AERMOIST formulation). Id. No modification of the underlying mode! occurs. Id,
As discussed above, EPA must rely on modeling that is representative of actual conditions. The incorporation of AERMOIST into the default AERMOD modeling more accurately reflects actual ambient conditions around Marion Station. The impact on modeling results is significant, hi AECOM's March 2016 modeling, AECOM demonstrated an 8% decrease (approximately 15 ug/m3) in modeled emissions as compared to modeling without AERMOIST. See Exhibit 27, Table 4-1 (comparing modeled concentrations in "current default" to "current default with AERMOIST"), Similarly, AECOM's updated September 2016 modeling indicated a 4% decrease (approximately 8 ug/m3). See Exhibit 26, Table 5-1 (comparing modeled concentrations in Scenario 1 to Scenario 2). in both runs, use of AERMOIST demonstrated that the area around Marion Station is in attainment of the 2010 SO? NAAQS.
D. Even Absent the Exclusion of Modeling Receptors from the Northern Property, Marion Station Models Attainment of the 2010 SQ2 NAAQS. The exclusion of modeling receptors from the Northern Property is consistent with EPA
16 See EPA Region 4 Response to Eastman Chemical's Additional EASTMOD information, attached as Appendix B to AECOM March 2016 modeling (Exhibit 27) submission to EPA ("EPA Region 4 has determined that the proposed AERLIFT component of EASTMOD is a source characterization procedure and is not an integral part of the AERMOD Modeling system").27
27
Sierra Club v. EPA 18cv3472 NDCA
Tier 7
ED 002061 00070535-00032
guidance and the CAA, EPA's principal objection to AECOM'$ modeling, however, is the exclusion of receptors from this area. Conversely. SIPC's principal objection to EPA's determination - apart from EPA's improper inclusion of the receptors at the Northern Property is EPA's reliance on non-representative modeling provided by Sierra Club, in order to address these discrepancies, AECOM ran additional modeling that conservatively (i) included receptors57 in the Northern Property while (ii) utilizing more representative source characterization modeling inputs. The modeling demonstrates that even when including receptors In the nonpublic Northern Property, Marion Station models attainment of the 2010 SO,? NAAQS.
The results (and additional analysis) of AECOM !s updated modeling are included in the 2016 AECOM' Report (Exhibit 26) and AECOM's supporting modeling files provided thereto.1s AECOM utilized more accurate (/.e., more representative) building dimensions based upon sitespeeific surveys taken in 2016 for two buildings located in close proximity to the Unit 123 stack.
' As further discussed in the 2016 AECOM Report, AECOM conservatively added additional receptors
along Lake Egypt Road to address EPA's concerns that modeling should include roadways. SPC notes
that EPA's Modeling TAD effective at the date EPA promulgated the 2016 Designations did not require
receptors on roadways. See Exhibit 7 at Section 4,2 ("[i]n areas where it is not feasible to place a
monitor...receptors can be ignored or not placed"). Indeed, this makes particular sense with respect to
Lake Egypt Road where the listed speed limit is 55 mph with shoulder, making the placement of monitors
virtually impossible, In August 2016, after EPA designated Williamson County nonattainment, EPA
issued an updated modeling TAD, in which the Agency advised that it: may be appropriate to locate
receptors "near" roadways.
See https://www.epa.gov/sites/production/files/2016-
0/documents/so2modelmgad.pd, Thus, although not required, SPC directed AECOM to update its
default model to include receptors along Lake Egypt Road.
Si PC further notes that the inclusion o f receptors on Lake Egypt Road addresses the concert? EPA expressed in the Final TSD regarding the location of peak 3-year average 99>hpercentile 1-hour average concentrations. EPA noted that AECOM's modeling indicated peek concentrations occurred to the South-Southwest of Marion Station, while both Illinois EPA's and Sierra Club's modeling indicated peak concentrations to the North. See Exhibit 11 at 28-29. In the 2016 AECOM Report, default modeling now indicates that peak concentrations occur to the Northeast of the Marion Station stacks. Exhibit 26 at Section 5.0. Nevertheless, even with peak concentrations to the North of Marion Station, default modeling demonstrated attainment o f the 2010 S02 NAAQS.
!g Note; Due to email size restrictions, a CD containing the modeling tiles for the 2016 AECOM Report was hand-delivered to EPA along with a copy of this Petition. Modeling flies were not provided in the email submission of this Petition.28
28
Sierra Club v. EPA 18cv3472 NDCA
Tier 7
ED 002061 00070535-00033
See 2016 AECOM Report at Sections 4,1, 5.0. The use of more representative building dimensions in conjunction with AECOM's default modeling inputs relied upon in AECOM's March 2016 modeling demonstrates that the highest ambient concentration of SCh is 194,9 ug/m3, below the 2010 S02 NAAQS. Id. at Table 5-1. Further, when receptors are properly excluded from the Northern Property, modeling indicates that the highest ambient concentrations at receptor locations are even lower (191.0 ug/m3). Id.
In sum, modeling provides substantial support that Marion Station is in attainment o f the 2010 SOz NAAQS, regardless of the use or exclusion of receptors in the Northern Property.
CONCLUSION For the reasons stated above, SIPC requests that EPA reconsider its 2016 Designation for Williamson County, Illinois and designate the area around Marion Station as in attainment of the 2010 SOz NAAQS. In the alternative, SIPC requests that EPA designate the area around Marion Station as unclassifiable of the 2010 SO2 NAAQS if EPA determines that the existence of conflicting modeling data precludes an attainment designation at this time.
Respectfully submitted, Southern Illinois Power Cooperative
Sierra Club v. EPA 18cv3472 NDCA
J. Michael Showalter David M. Coring Schiff Hardin LLP 233 South Wacker Drive, Suite 6600 Chicago, Illinois 60606 Counselfo r Southern Illinois Power Cooperative29
29
Tier 7
ED 002061 00070535-00034
EXHIBIT LIST
Exhibit 1 EPA Final Rule - 2016 Designations, 81 Fed. Reg. 45039 (July 12, 2016) Exhibit 2 Sierra Club v. McCarthy (N.D. Cal, Mar. 2, 2 0 1 5 )- Consent Decree Exhibit 3 Illinois ERA September 18, 2015 2010 SO2 NAAQS Recommendations Exhibit 4 EPA Region 5 October 20, 2015 letter to Illinois EPA re Sierra Club Exhibit 5 Illinois EPA August 24, 2016 letter re SIPC FOIA Request Exhibit 6 Illinois EPA November 9, 2015 letter to EPA Region 5 re Sierra Club Exhibit 7 EPA Modeling TAD, December 2013 Exhibit 8 AECOM SO2 Modeling on behalf of SIPC, December 21,2015 Exhibit 9 Illinois EPA 120-Day letter to EPA re Proposed Designations, April 19, 2016 Exhibit 10 EPA Region 5 Preliminary Recommendations to Illinois, February 16, 2016 Exhibit 11 EPA Final TSD, July 2, 2016 Exhibit 12 EPA Memorandum, Interpretation of Ambient Air, June 22, 2007 Exhibit 13 EPA Model Clearinghouse, Record No, 99-V-03, September 18, 1999 Exhibit 14 EPA Model Clearinghouse, Record No, 99-V-02, August 18, 1999 Exhibit 15 Photograph - Guardrail Northern Property Exhibit 16 Photograph..Lake Egypt Road; Guardrail: Fencing; Spillway; Northern Property Exhibit 17 Photograph - Security Camera; Northern Property Exhibit 18 Photograph Northern Property; slope Exhibit 19 'Photograph - Northern Property; signage Exhibit 20 Affidavit of Robert Todd Gailenbach, Vice President o f Operations, Marion Station Exhibit 21 Photograph - New fencing; new signage: Northern Property Exhibit 22 Photograph - New fencing: new signage; Northern Property Exhibit 23 Photograph New fencing: new signage; Northern Property Exhibit 24 Photograph - New fencing; new signage; Northern Property Exhibit 25 EPA, In the Matter o f nibbing Taconite Co, Exhibit 26 AECOM Updated Modeling Report for Marion Station, September 2016
(CD o f Modeling Data Provided to EPA by hand-delivery) Exhibit 27 AECOM Modd lug Report for Marion Station, March 2016
Sierra Club v. EPA 18cv3472 NDCA
30 Tier 7
ED 002061 00070535-00035