NOT FOR OFFICIAL PUBLICATION.

UNDER ARIZONA RULE OF THE SUPREME COURT 111(c), THIS DECISION IS NOT PRECEDENTIAL AND MAY BE CITED ONLY AS AUTHORIZED BY RULE.

MEMORANDUM DECISION

JOHNSEN, Judge.

¶1 The Director of the Arizona Department of Environmental Quality ("ADEQ") and Rosemont Copper Company appeal the superior court's order reversing ADEQ's decision to grant an air-quality permit allowing Rosemont to build and operate a new open-pit copper mine in Southern Arizona. For the following reasons, we reverse the superior court's order and affirm the decision to grant the permit.

FACTS AND PROCEDURAL BACKGROUND

¶2 Rosemont requires an air-quality permit to proceed with its plan to build an open-pit copper mine and processing plant some 30 miles southeast of Tucson, on the northeast face of the Santa Rita Mountains. The pit is projected to be about 4,000 feet in diameter. At its peak over the mine's projected 20-year life, Rosemont expects the mine to produce as much as 221 million pounds of copper and 4.7 million pounds of molybdenum a year, along with smaller amounts of gold and silver. Rosemont applied for a permit in November 2011. Its application described its planned mining processes and included information about emissions expected from those processes, along with control mechanisms it intended to install to reduce those emissions. Rosemont also submitted the results of computer modeling its consultants performed using a standard air-quality program called AERMOD (short for American Meteorological Society/Environmental Protection Agency Regulatory Model).1

¶3 After ADEQ reviewed Rosemont's application, it issued a draft permit and took public comments for nearly three months. In response to some of the comments, the Department modified the permit conditions to require Rosemont to install a monitor to track particulate matter smaller than ten microns.

¶4 In January 2013, ADEQ issued a Class II Air Quality Permit to Rosemont, and an organization called Save the Scenic Santa Ritas ("SSSR") and others appealed. An administrative law judge ("ALJ") heard 13 days of testimony. Fourteen witnesses testified, including several ADEQ employees who participated in reviewing Rosemont's application, Rosemont employees and consultants who were involved in the proposed mine project and in preparing the permit application, and expert witnesses retained by SSSR. The evidence primarily focused on the adequacy of the computer modeling Rosemont submitted with its permit application, ADEQ's review of that modeling, and SSSR's critiques. The witnesses included chemical, civil, mining, metallurgical and mechanical engineers, meteorologists, geologists, and atmospheric and environmental scientists.

¶5 In April 2014, the ALJ issued a 49-page decision containing detailed findings of fact and conclusions of law, recommending that the Director dismiss the appeals and affirm the Department's decision to issue the permit. After reviewing the ALJ's findings and conclusions, the Director dismissed the appeals. SSSR appealed to the superior court pursuant to Arizona Revised Statutes ("A.R.S.") sections 12-904(A) (2016) and 49-443(A) (2016).2 After briefing and oral argument, the superior court reversed the Department's decision to issue the permit. ADEQ and Rosemont timely appealed. We have jurisdiction pursuant to A.R.S. §§ 49-443(C) (2016) and 12-2101(A)(1) (2016).

DISCUSSION

A. Appeal of an Agency Decision: Legal Principles.

¶6 We review de novo a superior court order reviewing an agency action. Carlson v. Ariz. State Pers. Bd., 214 Ariz. 426, 430, ¶ 13 (App. 2007). The court must affirm the agency action "unless after reviewing the administrative record and supplementing evidence . . . the court concludes that the action is not supported by substantial evidence, is contrary to law, is arbitrary and capricious or is an abuse of discretion." A.R.S. § 12-910(E) (2016).3 Although this court applies its own independent judgment to questions of law, we defer to the agency's factual findings and will not substitute our judgment for that of the agency on matters of agency expertise. Webb v. State ex rel. Ariz. Bd. of Med. Exam'rs, 202 Ariz. 555, 557, ¶ 7 (App. 2002). We will not re-weigh the evidence to resolve any perceived conflicts, but must affirm the agency's decision if it is supported by substantial evidence. DeGroot v. Ariz. Racing Comm'n, 141 Ariz. 331, 335-36 (App. 1984).

B. ADEQ's Power to Deny the Permit.

¶7 By state law, a new mine may not start up without an air permit from ADEQ:

A.R.S. § 49-401(B) (2016).

¶8 In challenging ADEQ's decision to grant the Rosemont permit, SSSR cited A.R.S. § 49-427(A) (2016), which provides:

SSSR argued that in considering Rosemont's application, ADEQ needed to apply the National Ambient Air Quality Standards ("NAAQS"), which are established by the Environmental Protection Agency ("EPA") and which ADEQ has adopted by rule. A.A.C. R18-2-201 et seq.4 SSSR contended ADEQ should have rejected Rosemont's permit because the modeling Rosemont submitted was flawed for a variety of reasons, and that a proper model would have demonstrated a violation of the NAAQS.

¶9 On appeal, ADEQ argues that even if Rosemont's model had predicted air-quality violations, another statute, A.R.S. § 41-1030 (2016), deprived the Department of the power to reject the permit for that reason. Subpart (B) of that statute provides:

ADEQ argues that when Rosemont applied for its permit, no statute or rule specifically authorized the Department to reject an application for failure to demonstrate compliance with NAAQS.5 ADEQ argues that, without such authority, under § 41-1030(B), it had to grant Rosemont the permit if the company promised to install emission controls "designed to meet applicable standards" and the Department's "experience or calculations show[ed] that the proposed facility can meet" those standards.

¶10 SSSR vigorously argues in response that § 41-1030(B), a statute generally referring to administrative licensing decisions, does not limit the Department's power under § 49-427, a statute referring specifically to ADEQ. We need not resolve whether a predicted air-quality violation by Rosemont would have required the Department to deny its application, however, because we conclude that substantial evidence supported the Department's determination that the proposed Rosemont mine will not exceed air-quality standards.

C. The Validity of Rosemont's Air-Dispersion Model.

1. Modeling and the effects of the Tucson airshed.

¶11 The customary way to assess whether emissions from a new industrial source will degrade air quality is an air-dispersion model. Such a model calculates emissions the new source is expected to generate and predicts whether those emissions, when added to the reported background levels of pollutants already existing in the area ("ambient conditions"), and taking into account atmospheric and meteorological conditions, may violate air-quality standards. When used for regulatory purposes, air-dispersion models are designed to be conservative, meaning that they over-predict the polluting effects of the new source on ambient air quality.

¶12 In its appeal of the Department's decision, SSSR mounted a broad attack on the Rosemont model, arguing that in several key respects it violated ADEQ and EPA guidelines by failing to use sufficiently conservative calculated, estimated and reported data about emissions. As Rosemont and the Department argue, however, EPA regulations do not demand use of the most conservative data. The primary goal is accuracy: "The model that most accurately estimates concentrations in the area of interest is always sought." 40 C.F.R. pt. 51, App. W, at § 1.0(d). "In all cases, the model applied to a given situation should be the one that provides the most accurate representation of atmospheric transport, dispersion, and chemical transformations in the area of interest." 40 C.F.R. pt. 51, App. W, at § 1.0(e).

¶13 The validity of air-dispersion modeling hinges on, among other things, the validity of the model's governing assumptions. One of the principal issues SSSR raised in its appeal was whether the Rosemont model used properly representative data about background conditions at the mine site. SSSR argued the Rosemont model was invalid because it did not use background data that reflected emissions from Tucson, 30 miles northwest of the Rosemont site.

¶14 The Department operates several monitoring stations around the state that gather air-quality data that can be used to show ambient background conditions for purposes of an air-dispersion model. In the absence of historical data from monitor stations located near the Rosemont site, the company used background data gathered at rural locations elsewhere in the state. SSSR argued those data did not fairly represent conditions at the planned mine site, asserting that conditions at the Rosemont site are bound to be significantly influenced by emissions from the Tucson urban area.

¶15 There is no dispute that the prevailing winds in the area of the planned mine do not blow from the north to the southeast (in the direction of the Rosemont mine from Tucson). SSSR argued, however, that Rosemont should have performed more sophisticated analyses to confirm that winds would not cause emissions from Tucson and the Interstate 10 corridor to be "connected meteorologically at certain times of the day and on certain days of the year" with the Rosemont location. SSSR's expert witness, Dr. Eric Betterton, testified that a computer program called HYSPLIT showed a connection between Tucson emissions and the Rosemont site.6 He testified that a HYSPLIT model run for one day in July 2012 showed that a particle released at the Saguaro National Park, located just east of Tucson, moved toward the south, suggesting that winds sometimes may carry Tucson emissions toward the Rosemont site. He urged the Department to require Rosemont to further analyze whether air quality near the proposed mine site is affected by emissions from Tucson.

¶16 David Strohm, a project manager at JBR Environmental, testified on behalf of Rosemont that it was unlikely that Tucson emissions would have a significant and consistent influence on the Rosemont site. He testified the HYSPLIT program tracks the movement of a distinct "parcel" of air from one place to another, but cannot be used to analyze the interplay among air parcels and or to show the magnitude one "parcel" of air might have on another.

¶17 Dr. Feng Mao, an environmental engineer with ADEQ who reviewed Rosemont's model, testified that meteorological data Rosemont collected from its site and from a nearby monitoring station indicated Tucson pollution would not affect ambient conditions at the Rosemont site. Although Tucson is northwest of the Rosemont site, data from the Rosemont monitor reflected three percent winds coming from the east and 18 percent coming from the west, while data from the other monitor showed two percent of eastern winds and 20 percent of western winds. Mao also testified that the Santa Rita Mountains create a barrier that further reduces the likelihood that emissions from Tucson would significantly affect the Rosemont site.

¶18 Mao also reviewed ozone data from a monitor in Green Valley, located 15 miles west of the Rosemont site, but, like Rosemont, about 30 miles south of Tucson. That monitor did not show ozone concentrations coming from Tucson, from which Mao inferred that Tucson emissions do not significantly affect the Rosemont site.

¶19 Despite Betterton's suggestion that emissions from Tucson sometimes may affect ambient conditions near the Rosemont site, we cannot say the Department abused its discretion or acted arbitrarily and capriciously in accepting Rosemont's use of data from monitoring locations that do not reflect emissions from in and around Tucson. See, e.g., United States v. Alpine Land & Reservoir Co., 887 F.2d 207, 213 (9th Cir. 1989) ("Deference to an agency's technical expertise and experience is particularly warranted with respect to questions involving engineering and scientific matters.").

2. Use of ozone and particulate data from Chiricahua National Monument.

¶20 Rosemont used data from a monitor in the Chiricahua National Monument, about 100 miles south of the Rosemont site, to derive estimated background concentrations of particulate matter and ozone near the mine site.7 SSSR argued Rosemont should have used data from another monitor located in the Saguaro National Monument, just east of Tucson and 32 miles north of the Rosemont site.

¶21 Although SSSR pointed out that the Saguaro monitor is much closer to the Rosemont site than the Chiricahua monitor, under EPA rules, proximity is not the sole basis on which to select a monitoring location. Regulations explain that whether a location is "representative" of another depends on whether the two are similar in terms of terrain, distance from the source of the monitored pollutant, elevation and proximity to other emission sources. EPA Guidance W, at § 8.3(a). When no background monitors are "in the vicinity of the source," the rules allow use of data from a monitor located elsewhere that is "impacted by similar natural and distant man-made sources." 40 C.F.R. pt. 51, App. W, at § 8.2.2(c).

¶22 Mao, the ADEQ reviewer, testified Rosemont demonstrated that the Chiricahua location and the Rosemont site share similar elevations and are similarly removed from urban centers. Mao explained that although the Saguaro monitor is closer to the proposed Rosemont site, the Saguaro monitor is affected by emissions from Tucson that are not present at the Rosemont location. Mao also testified that the Chiricahua monitor showed higher concentrations of ozone than reflected at the Green Valley monitor, which is closer to Rosemont than the Chiricahua monitor. Mao concluded based on that comparison that, consistent with EPA guidance, use of data from the Chiricahua monitor resulted in a conservative modeling result, despite its more distant location.

¶23 Although two SSSR expert witnesses, Betterton and D. Howard Gebhart, testified the Saguaro monitor would be more appropriate because it is closer to Rosemont and more likely to reflect Tucson emissions, given the evidence that Tucson emissions are unlikely to significantly affect ambient conditions at the Rosemont site, and the similarities between conditions at the Rosemont site and at the Chiricahua monitor, the Department did not act arbitrarily or capriciously in accepting Rosemont's decision to use background ozone and particulate-matter data from the Chiricahua monitor.

3. Use of nitrogen dioxide data from Alamo Lake.

¶24 SSSR also argued Rosemont's modeling was flawed because it used background nitrogen dioxide data gathered from a monitor located at Alamo Lake, located near Wenden, in western Arizona, rather than data from a monitor at Children's Park, located in downtown Tucson. SSSR argued that Alamo Lake is not a representative background monitor because it is 200 miles away from Rosemont, it sits at a lower elevation than the Rosemont site, and some periods of time are missing from the data set at that location. Rosemont argued Alamo Lake is the most appropriate monitor for nitrogen dioxide because it is the only nitrogen dioxide monitor located in a rural location in Arizona not influenced by a coal-fired power plant.

¶25 The Rosemont site ranges from 4500 to 6300 feet above sea level; the elevation of Alamo Lake is between 1200 and 1300 feet. Moreover, Rosemont is just 15 miles away from a traffic corridor; Alamo Lake sits 50 miles away from the nearest traffic corridor. Mao testified that nevertheless, under the circumstances, Rosemont's use of Alamo Lake for background levels of nitrogen dioxide was appropriately conservative. By way of explanation, Mao testified ADEQ collects nitrogen dioxide data at only two rural locations that are relatively unaffected by urban vehicle emissions — Alamo Lake and Tonto National Monument, located about 60 miles east/northeast of Phoenix. Based on wind patterns flowing from Phoenix to Tonto National Monument, Mao knew that the Tonto data would be strongly affected by Phoenix's urban plume. Despite this connection to Phoenix, concentrations of nitrogen dioxide at the Tonto monitor were lower than those at Alamo Lake. From that, Mao concluded that Rosemont's use of Alamo Lake's comparably higher nitrogen dioxide background concentrations was appropriately conservative.

¶26 Although SSSR argued that the Alamo Lake data were incomplete because they did not span a full three years, ADEQ guidelines only say that three years of data are preferred, and EPA guidance does not establish a minimum monitoring period. Ariz. Dep't of Envtl. Quality, Air Dispersion Modeling Guidelines for Arizona Air Quality Permits ("ADEQ Guidelines"), at § 3.9 (2004); 40 C.F.R. pt. 51, App. W, at § 8.22.

¶27 For reasons discussed above in connection with the use of the Chiricahua monitor for ozone and particulate-matter data, substantial evidence supported the Department's decision to accept use of Alamo Lake as the background monitor for nitrogen dioxide concentrations. Although SSSR argued Rosemont should have used a monitor located in Children's Park in Tucson, that monitor showed higher concentrations of nitrogen dioxide, precisely due to its proximity to Tucson vehicle emissions. Evidence in the record supported the Department's determination that Tucson emissions generally do not affect ambient conditions at the Rosemont site. Because evidence in the record supported use of the Alamo Lake monitor data, the Department did not abuse its discretion when it declined to require Rosemont to use data from the Children's Park monitor.

4. The proper in-stack ratio.

¶28 SSSR also criticized the Rosemont model's assumptions concerning emissions to be produced by the huge haulage trucks that will transport ore from the Rosemont pit for processing. The Rosemont model used a calculation called the NO2:NOx ratio to estimate nitrogen dioxide that will result from emissions from the "stack," or tailpipe, of the trucks.8

¶29 As planned, Rosemont will mine as much as 110,000,000 tons a year of mineral ore and waste material. Working 24 hours a day, 365 days a year, haulage trucks will carry ore to associated processing facilities and waste to tailings piles. Accordingly, the Rosemont model results are very sensitive to the choice of an emissions ratio (the "in-stack ratio") to apply to those trucks. Rosemont's model used a ratio of five percent; SSSR pointed out that EPA rules establish 50 percent as the default ratio to be used in the absence of more precise calculations, and argued the Department should have required Rosemont to use a ratio of 10-50 percent. See 40 C.F.R. pt. 51, App. W, at ¶ 5.2.4(d); U.S. Envtl. Prot. Agency, memorandum, "Additional Clarification Regarding Application of Appendix W Modeling Guidance for the 1-hour NO2 National Ambient Air Quality Standard" (Mar. 1, 2011).

¶30 In support of its decision to use a ratio of only five percent, Rosemont submitted a report created by Leonard Montenegro, an air-quality modeling expert who reviewed literature concerning tests and estimates of vehicle in-stack ratios. The report explained the difficulty of obtaining direct (in-stack or in-pipe) measurements from mobile sources (by contrast to measurements of emissions from industrial smoke-stacks, for which EPA requires actual testing). The report summarized six studies that reported ratios ranging from two percent to 15 percent. Although Montenegro also cited estimated ratios ranging up to 30 percent, he stressed the reliability of two studies that reflected ratios ranging from two percent to six percent. The report also asserted that EPA tests, which take place after exhaust has been allowed to mix and react with the ambient air, are inappropriate for model input, which requires in-stack measurements before combustion with the ambient air has occurred.

¶31 SSSR's expert, Gebhart, pointed to the studies described in the Montenegro report that reflected ratios higher than five percent, including a 30 percent ratio calculated from diesel vehicles fitted with after-treatment devices. He also noted that under EPA guidance, in the absence of more appropriate source-specific information on in-stack ratios, a permittee should apply a default ratio of 50 percent.

¶32 Mao testified Rosemont's use of the five percent ratio for haulage truck emissions was appropriate and reasonable. He testified Rosemont was not required to use the EPA default ratio of 50 percent because the Montenegro report provided more appropriate source-specific information. Mao also testified that the diesel vehicle study that Gebhart cited used data from ambient air, not from measurements taken in the "stack." Although Gebhart also pointed to ratios of five to 15 percent in a letter by a truck manufacturer cited in the Montenegro report, Montenegro explained it was not clear if the manufacturer's "estimates are for engine-out or tailpipe" concentrations of nitrogen dioxide.

¶33 SSSR also argued that Rosemont submitted modeling results to the United States Forest Service pursuant to the National Environmental Policy Act that used a ten percent in-stack ratio, and which as a result predicted the new mine was likely to violate air-quality standards. Rosemont argued to the ALJ that it performed a modeling run using the ten percent ratio figure because the Forest Service required it, not because that ratio figure is more accurate. Mao testified there were no test results requiring use of ten percent rather than the five percent ratio Rosemont used in the model it submitted to ADEQ. Because the Department's approval of Rosemont's use of a ratio of five percent was supported by substantial evidence, the agency did not act arbitrarily or capriciously in accepting that figure.

5. Adequacy of on-site data.

¶34 SSSR also raised several issues with Rosemont's collection of data to show ambient conditions at its mine site. It first argued the Rosemont meteorological data were flawed because, due to an equipment failure, three months of data were missing, out of three years collected. On-site meteorological data is an air-dispersion model's source of information about air flow at the source of the emissions. EPA guidelines suggest the use of data from "consecutive years from the most recent," and SSSR argued that because of the three-month gap, Rosemont's data were unacceptable.

¶35 ADEQ modeling guidelines require "one year of meteorological parameters from a representative on-site location." ADEQ Guidelines, at § 3.8. EPA guidelines likewise require one year of site-specific data. 40 C.F.R. pt. 51, App. W at § 8.3.1.2(b). Mao testified he reviewed and validated the site-specific meteorology data Rosemont submitted. To fill in the three-month gap, Rosemont substituted data collected during the same season the following year. Mao reviewed the data and found there were no substantial differences in the data from year to year, which led him to believe that the use of the substituted data would not undermine the representativeness of the meteorology data. Because Rosemont only was required to submit one year of data, and because sufficient evidence supported the use of the substituted data, ADEQ did not act arbitrarily or capriciously in accepting Rosemont's meteorological data for purposes of modeling.

¶36 SSSR also took issue with Rosemont's use of a single meteorological monitoring site at its mine, arguing that, due to the site's complex terrain, multiple monitors would have provided better information.

¶37 EPA guidelines state:

40 C.F.R. pt. 51, App. W, at § 8.3(a), (c). With respect to site-specific meteorological data, EPA guidelines state, "Spatial or geographical representativeness is best achieved by collection of all of the needed model input data in close proximity to the actual site of the source(s)." Id. at § 8.3.3.1(a). ADEQ air dispersion modeling guidelines similarly accord priority to the representativeness of meteorological data, requiring "one year of meteorological parameters from a representative on-site location." ADEQ Guidelines, at § 3.8.

¶38 Rosemont placed its meteorological station at the center of its planned open pit. Michael Sundblom, who supervises the ADEQ evaluation unit responsible for processing Rosemont's application, testified that was an appropriate location because it is at the spot where the emissions would originate. He explained that dispersion modeling performs best when the meteorology data are derived from the point where maximum concentrations are expected. Mao testified his primary concern in reviewing the model was with the maximum possible predicted concentrations that might result. Mao explained that, based on the specific topography of the Rosemont site, the most significant emissions effects would be close to the ground and close to the emissions source, which is where he focused his review.

¶39 Neither EPA nor ADEQ guidelines require the use of multiple meteorological monitoring sites such as SSSR advocates. The AERMOD model can use data from only one meteorological station at a time; multiple model runs would have to be performed with additional meteorological data. Based on this record, sufficient evidence existed on which ADEQ could decide not to require Rosemont to collect meteorological data from more than one location at the site.

¶40 Finally, SSSR took issue with Rosemont's placement of its monitor for particulate matter at a location two miles away from its meteorological monitor. SSSR admitted there is no requirement that the monitors be placed near each other. Shantanu Kongara testified for Rosemont that there was no reason why the monitors should be placed near each other, because the particulate-matter monitor collected data concerning background concentrations, while the meteorological monitor collected data concerning wind patterns. Kongara testified that placement of the particulate monitor conformed to EPA regulations.

¶41 Because SSSR showed no requirement that the monitors be placed near one another, and because sufficient evidence supported Rosemont's placement of the monitors, the Department did not abuse its discretion or act arbitrarily in accepting Rosemont's decisions of where to place its on-site monitor locations.

CONCLUSION

¶42 For the foregoing reasons, sufficient evidence supported ADEQ's decision to issue the permit to Rosemont. The record does not demonstrate the Department's decision was contrary to law, arbitrary and capricious or an abuse of discretion. Accordingly, we reverse the superior court's order and affirm the Department's decision to issue the air-quality permit.