Opinion

In this opinion, we hold that rule 4570, promulgated by the San Joaquin Valley Unified Air Pollution Control District (district), as mandated by Health and Safety Code section 40724.6, was adopted without conducting an adequate assessment of its impact on public health. We also conclude that section 40724.6 is intended to address the district’s failure to meet federal and state ambient air quality standards for ozone and does not regulate ammonia emissions produced by large confined animal facilities. Finally, we determine that the district’s findings were not arbitrary and capricious.

PROCEDURAL AND FACTUAL SUMMARIES

This appeal is from the denial of a writ of mandate sought by petitioner Association of Irritated Residents (association) to compel the district to comply with section 40724.6. Section 40724.6 initially was passed in 2003 as part of Senate Bill No. 700 (2003-2004 Reg. Sess.) (Sen. Bill 700). (Stats. 2003, ch. 479, § 6.) The statute mandates that those state air pollution control districts designated federal nonattainment areas for ozone, as of January 1, 2004, adopt and implement a rule requiring confined animal facilities (facilities) to reduce emissions of air contaminants. The rule is to apply to existing facilities and does not apply to expanded or new facilities. Prior to the passage of section 40724.6, agricultural sources of air contaminants were exempted from air pollution control district regulation. (Assem. Com. on Appropriations, Rep. on Sen. Bill No. 700 (2002-2003 Reg. Sess.) as amended Aug. 21, 2003, p. 4.) The Legislature declared its intent when enacting Senate Bill 700: “The purpose of the act adding this section is to establish a new set of programs at the state and regional levels to reduce air emissions from agricultural sources in order to protect public health and the environment.” (Stats. 2003, ch. 479, § 1, subd. (a)(7).) The agricultural industry, including the many facilities located in the San Joaquin Valley (valley), contribute large quantities of emissions and particulates (PM10 and PM25) that significantly contribute to the air pollution impacting the valley and its residents. (Stats. 2003, ch. 479, § 1.) The valley’s air basin, home to one of the largest agricultural industries in the world, is a nonattainment area according to federal and state standards for three pollutants: ozone, PM10, and PM2 5. (Stats. 2003, ch. 479, § 1, subd. (a)(1), (2), (3).)

The district, in compliance with section 40724.6, and after a series of public hearings and studies, adopted its rule 4570. Rule 4570 establishes a permit process for large confined animal facilities and is intended to control volatile organic compounds (VOC’s) by controlling animal enteric, feed, and waste emissions with improved management practices. VOC’s combine with nitrogen oxides, heat, and sunlight to form ozone. Rule 4570 requires that facilities choose from a variety of mitigation measures—generally feed and waste management practices—with the goal of reducing VOC emissions. The rule also provides for compliance testing, sets up a compliance schedule, and establishes recordkeeping requirements for all facilities.

According to the association, the district did not comply with the statutory mandates when it passed rule 4570 because the district (1) failed to perform a health-effects analysis before adopting rule 4570; (2) failed to adopt a rule that reduces all air contaminants; (3) failed to regulate ammonia as an air contaminant; (4) failed to adopt a rale that actually reduces VOC emissions; and (5) acted arbitrarily and capriciously by claiming emission reductions that lack evidentiary support. The Community Alliance for Responsible Environmental Stewardship, the California Dairy Campaign, the Milk Producers Counsel, and the Western United Dairymen have intervened on behalf of the district and in support of rule 4570.

The trial court denied the petition deciding that (1) the district did consider the public health impact of rule 4570 and that all issues related to public health raised by the association were addressed; (2) the intent and language of section 40724.6 do not require rule 4570 to include within its parameters air contaminants other than those directly related to the district’s nonattainment of federal and state ozone standards; (3) there is no requirement that rule 4570 address ammonia emissions; and (4) the claimed emission reductions are not lacking in evidentiary support.

DISCUSSION

I. Standard of review

The adoption of rule 4570 was a quasi-legislative act reviewable by a petition for traditional mandamus. (Code Civ. Proc., § 1085; Muzzy Ranch Co. v. Solano County Airport Land Use Com. (2008) 164 Cal.App.4th 1,13 [ 78 Cal.Rptr.3d 691 ]; Santa Margarita Area Residents Together v. San Luis Obispo County Bd. of Supervisors (2000) 84 Cal.App.4th 221, 227-228 [ 100 Cal.Rptr.2d 740 ].) A writ of mandate under Code of Civil Procedure section 1085 is a method for compelling an agency to perform a legal duty. (Pomona Police Officers’ Assn. v. City of Pomona (1997) 58 Cal.App.4th 578, 583-584 [ 68 Cal.Rptr.2d 205 ].)

The trial court’s role in a traditional mandamus proceeding is a limited one. It must determine whether the agency’s action was arbitrary, capricious, or without evidentiary support, and/or whether it failed to conform to the law. The trial court may not substitute its judgment for that of the agency or force the agency to exercise its discretion in a certain way. (Neighbors in Support of Appropriate Land Use v. County of Tuolumne (2007) 157 Cal.App.4th 997, 1004 [ 68 Cal.Rptr.3d 882 ].)

The reviewing court exercises independent judgment in determining whether the agency action was “consistent with applicable law.” (Associated Builders & Contractors, Inc. v. San Francisco Airports Com. (1999) 21 Cal.4th 352, 361 [ 87 Cal.Rptr.2d 654 , 981 P.2d 499 ].) Where the issue is one of statutory interpretation, the question is one of law for the courts, which are the “ ‘ultimate arbiters’ ” of statutory construction. (San Francisco Fire Fighters Local 798 v. City and County of San Francisco (2006) 38 Cal.4th 653, 667-668 [ 42 Cal.Rptr.3d 868 , 133 P.3d 1028 ]; see Katosh v. Sonoma County Employees’ Retirement Assn. (2008) 163 Cal.App.4th 56, 62, fn. 4 [ 77 Cal.Rptr.3d 324 ].) Since we apply the same standard as the trial court, its determination is not binding on us. (Personnel Com. v. Board of Education (1990) 223 Cal.App.3d 1463, 1466 [ 273 Cal.Rptr. 288 ].)

II. Public health impacts

The association contends that the district failed to consider the public health impacts of rule 4570 prior to its adoption. According to the association, the district’s final staff report (staff report) and the appendices prepared in connection with rule 4570 do not adequately assess public health issues as required by the statute. Although the association acknowledges that information concerning public health was submitted during the rule-making process, it claims the district must do more than “abstractly consider information in the record”; it must actively perform an assessment equivalent to the assessment provided for the other eight factors listed in section 40724.6, subdivision (e)(1) through (8).

Section 40724.6, subdivision (e), reads:

The language of the statute unambiguously requires that, prior to adopting the rule, the district must perform an assessment of the rule’s impacts and that assessment must include the impact of the regulated emissions on the public health. (§ 40724.6, subd. (e)(2).) The association claims that a public health assessment should have been done in the same manner as was done in other assessments provided by the district as attachments to the staff report. These include an analysis of emission reductions, cost effectiveness, socioeconomic impacts, ammonia reduction, environmental impacts, and rule consistency. These are independent analyses done by experts that provide more than conclusory or summary information. We agree with the district, however, that the statute does not require that the assessment take any particular form or be performed by independent experts. The statute requires only that an assessment be done prior to the adoption of the rule and that it include an analysis of the identified factors. (§ 40724.6, subd. (e).) The question here is whether the district’s staff report is sufficient to meet the statutory requirements. On this point, we agree with the association and conclude that the staff report is insufficient because it does not assess the rule’s impact on public health.

The staff report acknowledges the statutory requirement that certain items be assessed in the rulemaking process. The report provides a table identifying each required assessment and states where in the report the required analysis appears. According to the report, the impact of the regulated emissions on public health is done in “Section IB and Section II of the Staff Report.” However, these sections do not speak to the impact of the regulated emissions on public health. Section II begins at page 9 and ends at page 22 of the staff report. It provides general background information on the agricultural animal industry, without any reference to public health. Section IB is a general overview of the purpose of rule 4570 and offers no analysis of the rule’s impact on public health. The staff report does contain section VIII, which is labeled “Health Effects.” The section consists of one paragraph and reads in its entirety as follows; “Respiratory diseases associated with agriculture were one of the first-recognized occupational hazards. According to the Institute for Agriculture and Trade Policy (Wallinga 2004), studies at beef, swine, and poultry facilities show increased occurrences of asthma, sinusitis, bronchitis, decreased lung function, and depression among workers at concentrated animal feeding operations. Of the 331 VOC and gaseous compounds found in odorous samples from North Carolina swine facilities, 157 are known airway irritants; chronic irritation can permanently scar lungs and lead to respiratory problems. (Wallinga 2004.)”

This single paragraph is inadequate and does not comply with the statutory directive. The district conceded at argument that this single paragraph was not intended to comply with section 40724.6, subdivision (e)(2), although it argued it does not matter. The paragraph addresses occupational health, not general public health concerns, and does not touch upon the required analysis, i.e., the impact of the rule on public health. The district suggests that there is no need to discuss the impact on public health during the rulemaking process because the rule is intended to assure compliance with health-based federal ambient air quality standards. If this were true, there would be no need for the Legislature to mandate such an assessment. To the contrary, an “assessment” is an evaluation, a determination of significance. (Webster’s New World Dict. (2d college ed. 1982) p. 83.) We decline to construe a statute in a way that makes meaningless the words chosen by the Legislature. (People v. Western Air Lines, Inc. (1954) 42 Cal.2d 621, 638 [ 268 P.2d 723 ].)

The statute requires that the rule be assessed in light of the identified problem and that the actual words of the statute be given their plain and commonsense meaning. (Mercer v. Department of Motor Vehicles (1991) 53 Cal.3d 753, 763 [ 280 Cal.Rptr. 745 , 809 P.2d 404 ].) It is not enough to say that VOC emissions from large confined animal facilities add to poor air quality, which in turn impacts public health. This fact was acknowledged in the legislative history for Senate Bill 700. (Stats. 2003, ch. 479, § 1, subd. (a)(7).) In addition, it is insufficient to discuss the adverse health impact of ozone generally, for example, the general discussion in the district’s Extreme Ozone Attainment Demonstration Plan included in the district’s rule-adoption materials. Section 40724.6, subdivision (e)(2), expressly requires that the district undertake an assessment of the impact of rule 4570 on public health. (Sen. Rules Com., Off. of Sen. Floor Analyses, Sen. Bill No. 700 (2002-2003 Reg. Sess.) as amended Sept. 9, 2003, p. 3 [bill requires district to perform assessment of rule’s impact prior to adoption].) The statute requires that “[t]he nature and quantity of emissions from the category, and the significance of those emissions in adversely affecting public health and the environment,” be assessed in connection with “the impact of the rule or regulation.” (§ 40724.6, subd. (e)(2).)

There is nothing in the administrative record that provides this analysis. Without this information, important competing policy issues cannot adequately be addressed or weighed. When adopting rule 4570, the district’s governing board stated that VOC emission reductions would benefit the valley’s public health by reducing unhealthful concentrations of ambient ozone. This, however, is a conclusory statement offered at the end of the rulemaking process and is unsupported by any analysis of the true impact of the rule on public health. It is not an adequate assessment of the issue.

We conclude that the subject was not adequately covered by comments raised or reports submitted during the public hearing process. During the public hearing process, the association provided an analysis of the cost impact the ill health effects of poor air quality has on the valley. This report, however, did not address the impact of rule 4570 on public health and is insufficient to meet the district’s duty to assess the impact of its rule on public health before its adoption. (§ 40724.6, subd. (e).) Likewise, the air quality report prepared by the American Lung Association, which was provided during the public comment period and which addresses the impact of air pollution (including ozone) on public health nationally, does not address the impact of rule 4570 on public health in the valley.

At the public hearing, the district’s executive officer told those present that the district was “all about” public health. However, he offered no insight regarding what the district expected to achieve concerning public health benefits with the adoption of rule 4570. He simply stated that the district had reduced emissions significantly in the recent past, even though he acknowledged that asthma rates continue to climb. This explanation misses the point. Section 40724.6, subdivision (e)(2), requires an assessment that answers the question of how this rule will impact human health or, at the very least, why the rule will have no impact.

Similarly, the district claims that the statutory assessment requirement is “a proportionality requirement,” and requires that the district “be mindful of imposing large economic burdens for de minimis air quality benefits.” The district is correct and this consideration is a part of the required assessment. Subdivision (e)(5) of section 40724.6 expressly requires assessment of the cost of implementing rule 4570 to the affected sources and businesses that are regulated. By the same token, an equal part of the required assessment is whether the controls imposed are sufficient to move the valley toward cleaner air.

Senate Bill 700 was enacted in recognition that valley air must be improved to protect public health and that the past practice of exempting agricultural sources of air pollutants was not justified by potential adverse economic impacts on the agricultural industry. (See Stats. 2003, ch. 479, § 1, subd. (a)(1), (4), (7).) If, as the district claims, the rule is intended to assist in meeting federal health-based ambient air quality standards, it is not surprising that the Legislature has required the district to evaluate how its rule will impact the health concerns which led to the Clean Air Act in the first place. (Wooley & Morss, Clean Air Act Handbook, supra, § 1:1, p. 4 [Clean Air Act requires identification of air pollutants believed to cause/contribute to air pollution levels that may endanger public health; national ambient air quality standards set at levels adequate to protect public health].) After all, the Clean Air Act’s primary consideration is the public welfare of the nation. (See 42 U.S.C. §§ 7602 , 7408(a) & 7409(b)(1); Whitman v. American Trucking Assns., Inc., supra, 531 U.S. at pp. 465-467 [air quality standards set to protect public health; cost not a factor to be considered in setting standards].)

From this record, we cannot tell whether rule 4570’s planned reduction in VOC emissions will have any impact on public health at all. The association suggests there will be no significant impact, claiming that the district has simply required the status quo. Whether the association is correct is something the district should have considered in its public health analysis and that should have been contained in the rulemaking process record. To the contrary, the district obviously knows how to provide an appropriate assessment. When faced with other assessment requirements, like those found in section 40727.2, subdivision (a) (requires rule-consistency analysis) and section 40920.6 (requires best-available retrofit control technology), it responded with a full-blown independent assessment of the identified factor.

The district provides no insight into how it balanced public health concerns against other interests during the rulemaking process. The statute requires that the rule be assessed prior to adoption and submitted to a public-hearing process. (§ 40724.6, subd. (e).) As with any environmental regulation, a delicate balancing of competing policies and interests must occur. For example, the need for jobs, economic viability in the valley, a consistent high-quality food source, plentiful clean water, and many other interests must compete with each other when government seeks to regulate human activity to protect the environment. (See San Francisco Ecology Center v. City and County of San Francisco (1975) 48 Cal.App.3d 584, 596 [ 122 Cal.Rptr. 100 ] [court recognizes agency must balance competing interests; key is whether agency’s “mode of analysis” is exposed].)

In addition, the purpose of requiring that proposed regulations be submitted to a public hearing process is to ensure that every interest is represented, that the rule makers are well informed, and that an equally well-informed public is able to persuade and monitor government through the democratic process. (In re Bay-Delta et al. (2008) 43 Cal.4th 1143, 1162 [ 77 Cal.Rptr.3d 578 184 P.3d 709 ] [purpose of California’s environmental impact report is to give public and government agencies information needed to make informed decisions, protecting not only environment but also providing informed self-government]; Schoen v. Department of Forestry & Fire Protection (1997) 58 Cal.App.4th 556, 573-574 [ 68 Cal.Rptr.2d 343 ] [purpose of requiring public review is to demonstrate to an apprehensive citizenry that agency has analyzed and considered ecological implications of its action; public review permits accountability and informed self-government]; Salmon River Concerned Citizens v. Robertson (9th Cir. 1994) 32 F.3d 1346, 1356 [informed decisionmaking and informed public participation are purpose of required public hearing process].) It is for this reason that the interveners’ argument (that for the association to be successful on appeal, it must prove that the rule would not have been adopted had the assessment been provided) fails. The prejudice is not that the rule was adopted, but that it was adopted without informed and transparent decisionmaking.

For example, the district claims that rule 4570 will reduce VOC’s by 7,563 tons per year; however, it makes no statement about how this will impact public health concerns. The report discusses how much the changes in feed and waste management will cost facilities and identifies a number of possible controls which have been rejected because of their high cost. If costs are going to justify mandating lesser controls instead of tougher ones, the public is entitled to know what the cost of this decision will be to public health. If the available science is insufficient to justify more expensive, tougher environmental controls, the public is entitled to know this as well.

We have no quarrel with the district’s legislatively granted authority to apply its expertise to these questions. (San Francisco Fire Fighters Local 798 v. City and County of San Francisco, supra, 38 Cal.4th at p. 667 [court gives deference to presumed expertise of agency within its scope of authority].) However, although an agency is free to make this policy call, it can do so only when the balancing process is exposed to public view. (Schoen v. Department of Forestry & Fire Protection, supra, 58 Cal.App.4th at pp. 573-574; § 40724.6, subd. (e).) Neither the staff report nor the district representatives during the review process addressed rule 4570’s impact on public health, despite the expressed concern for public health. Will this rule make a significant measurable difference in air quality (ozone levels) and bring improvement to public health concerns related to poor air quality? If not, are higher costs justified based on the cost of public health issues that are tied to air pollution? These are questions that remain unanswered because the required assessment was never done. If the goal is healthier air, the district has not shown whether it has taken steps toward reaching that goal.

As with any environmental regulation, there are difficult choices to be made. On this record, it was impossible for the public or the district’s governing body to know how the acknowledged public health interest fared when balanced against the other competing interests. For this reason, the trial court erred when it denied the petition for writ of mandate.

III. Air contaminants besides ozone

The association claims the district’s decision to address only VOC emissions in rule 4570 violates the statutory requirement that the rule reduce “emissions of air contaminants” from facilities. (§ 40724.6, subd. (a).) According to the association, the statutory reference to ozone nonattainment in section 40724.6, subdivision (b), is simply a trigger for application of the statute. Air pollution control districts that are designated as federal nonattainment areas of ozone as of January 1, 2004, are those districts that must comply with this statute, but no other district. (Ibid.) The association also contends that the remaining language of the statute suggests an intent to address all air contaminants, including ammonia emissions. The district, on the other hand, claims section 40724.6 applies only to ozone and ozone precursors. It argues that ammonia is not an ozone precursor and, as a result, regulation of ammonia is not mandated or permitted under the statute.

A complete reading of the statutory changes accomplished by Senate Bill 700, and the statement of legislative intent provided in Senate Bill 700, compel the conclusion that the legislative intent is to control agricultural emission sources as they impact state and federal ambient air quality standards, particularly those for which the state is classified as being in nonattainment. This means that section 40724.6 is, as the district contends, an ozone-specific statute.

The legislative preamble for Senate Bill 700 begins, “[a]gricultural operations necessary for growing crops or raising animals are a significant source of directly emitted particulates, and precursors of ozone and fine particulate matter. These emissions have a significant adverse effect on the ability of areas of the state, including, but not limited to, the San Joaquin Valley, to achieve health-based state and federal ambient air quality standards.” (Stats. 2003, ch. 479, § 1, subd. (a)(1), italics added.) The very next statement refers to nitrogen oxides, particulate matter, and VOC’s, which are expressly identified as contributing “twenty-six percent of the smog-forming emissions in the San Joaquin Valley.” (Id., § 1, subd. (a)(2).) Ozone is the prime ingredient of smog. (U.S. EPA Green Book Criteria Pollutants [as of Nov. 19, 2008].) These statements unquestionably link the ambient air quality standards to the pollutants identified, particulate matter, nitrogen oxides, and VOC’s. The latter two are ozone precursors. Along with particulate matter, these are the pollutants with which California struggles to reach attainment with national ambient air quality standards. (Ibid.)

Senate Bill 700, consistent with its statement of intent, includes rule-mandating provisions addressing both ozone and particulate matter in sections 40724, 40724.5, 40724.6, and 40724.7. These sections address the two identified problems, ozone and particulate matter, depending on whether the district has achieved attainment of state and federal ambient air quality standards related to these pollutants. Section 40724, using language similar to section 40724.6’s, applies to districts designated as serious federal nonattainment areas for the applicable ambient air quality standard for particulate matter. Similar to section 40724.6, the statute requires that the district adopt a rule regulating agricultural sources of particulate matter (including large confined animal facilities) using the best-available retrofit control technology. Section 40724.5 applies to districts designated as moderate federal nonattainment areas for applicable ambient air quality standards for particulate matter. It requires the adoption of control measures unless the district determines that agricultural practices do not significantly cause or contribute to a violation of state or federal standards. Section 40724.6 parallels section 40724 in that it requires a rule regulating large confined animal facilities where the district is classified as severe or extreme nonattainment for ozone. Section 40724.7 parallels section 40724.5 in that it requires a rule controlling large confined animal facilities unless the district finds that the facilities will not contribute to a violation of state or federal standards.

All four of these sections contain triggering language related to attainment or nonattainment of either ozone or particulate matter. The trigger is not capricious: It relates directly to the level of attainment or nonattainment achieved by the district related to the triggering pollutant. The Legislature has determined that if a district has attained federal and state ambient air quality standards for either particulate matter or ozone, it will be given greater deference in addressing the problem pollutant. If not, the statute mandates more stringent controls of sources for the identified pollutant. We appreciate the fact that the words chosen in these four sections are not identical, but they must be read in harmony with one another. (Dyna-Med, Inc. v. Fair Employment & Housing Com. (1987) 43 Cal.3d 1379, 1387 [ 241 Cal.Rptr. 67 , 743 P.2d 1323 ]; Mannheim v. Superior Court (1970) 3 Cal.3d 678, 687 [ 91 Cal.Rptr. 585 , 478 P.2d 17 ].)

We reach the same conclusion when we apply the statutory construction canons ejusdem generis and noscitur a sociis. Under the rule of noscitur a sociis, the meaning of a word may be enlarged or restrained by reference to the intent of the whole clause in which it is used. (People v. Stout (1971) 18 Cal.App.3d 172, 177 [ 95 Cal.Rptr. 593 ].) Under the rule of ejusdem generis, the general term or category is restricted to those things that are similar to the more specific term or category enumerated. (Kraus v. Trinity Management Services, Inc. (2000) 23 Cal.4th 116, 141 [ 96 Cal.Rptr.2d 485 , 999 P.2d 718 ].) The legislative use of the general term “air contaminants” is restricted by the more specific references to the ambient air quality standards for ozone in sections 40724.6 and 40724.7 and for particulate matter in sections 40724 and 40724.5. (§ 40724.6, subd. (a).) As the district argues, the express references in section 40724.6, subdivision (d)(1)(B), to “severe and extreme nonattainment areas” can only be read as a reference to ozone, for only nonattainment to ozone ambient air quality standards is classified as “severe” or “extreme.” (See 40 C.F.R. § 51.903 (2008).) These classifications do not exist for other pollutants regulated by national ambient air quality standards. (See 40 C.F.R. §§ 50.4-50.12 (2008); U.S. EPA Green Book, supra, [as of Nov. 19, 2008]; U.S. EPA Air Quality Maps, Attainment Designations [as of Nov. 19, 2008].)

The language used in section 40724.6 is not as broad as the association contends. Although only the “trigger” language is limited specifically to ozone, the statutory language also generally refers to “applicable emissions threshold[s],,‘’ “emissions of air pollutants that . . . cause or contribute to a violation of a state or federal ambient air quality standard,” “all regulated air pollutants,” and “pollutants that contribute to the nonattainment of any ambient air quality standard . . . .” This language consistently refers to pollutants regulated by ambient air quality standards, such as ozone and particulate matter. Even if we were to read section 40724.6 as applying to all or any “ambient air quality standard[s],” a term of art under the federal Clean Air Act, ammonia is not a pollutant that is subject to a national ambient air quality standard. National standards have been set for only six pollutants: particulate matter, sulfur dioxide, ozone, nitrogen dioxide, carbon monoxide, and lead. Other hazardous air pollutants are not regulated through ambient air quality standards. (Wooley & Morss, Clean Air Act Handbook, supra, § 1:3, pp. 10—11.) In any event, the district determined that, when adopting its rule 4570 pursuant to section 40724, the regulation of ammonia was not necessary for attainment of any air quality standard, and the association has not shown otherwise.

The primary role of the court when interpreting a statute is to ascertain the intent of the lawmakers to effectuate the purpose of the law. (People v. Carron (1995) 37 Cal.App.4th 1230, 1236 [ 44 Cal.Rptr.2d 328 ].) We believe a fair reading of Senate Bill 700 and section 40724.6 is consistent with the district’s interpretation that section 40724.6 addresses ozone and ozone precursors. There is nothing in the statute or the legislative history to suggest that the rule was intended to address all other air pollutants.

IV. Arbitrary, capricious, and lack of evidentiary support

The association’s last contention is that the district’s findings are arbitrary, capricious, and lack evidentiary support. We have a limited standard of review in mandamus actions and may reverse only where the action taken is “so palpably unreasonable and arbitrary as to show an abuse of discretion as a matter of law.” (Carrancho v. California Air Resources Board (2003) 111 Cal.App.4th 1255, 1265 [ 4 Cal.Rptr.3d 536 ].) We grant deference to the agency’s decision and may not substitute our judgment for that of the public agency granted authority over a particular area of subject matter expertise. (See Neighbors in Support of Appropriate Land Use v. County of Tuolumne, supra, 157 Cal.App.4th at p. 1004 .)

The association claims that the district (1) relied upon unsupported reduction assumptions; (2) double counted reduction estimates; (3) assumed a 50 percent control efficiency for all liquid manure-handling mitigation options based on the control efficiency of only phototropic lagoons; (4) relied on emission-reduction estimates for applying manure to croplands without evidentiary support; and (5) failed to analyze more expensive mitigation measures based on an assumption that facilities would choose the cheapest mitigation measure. We reject each of these contentions because the district applied its expertise to the scientific and policy considerations before it and reached a conclusion that was neither arbitrary nor capricious given the record evidence. Further, we agree with the district that section 40724.6 does not require that the rule result in any specific numeric reduction of VOC’s; it simply mandates that the district adopt a rule to control VOC emissions generated by large confined animal facilities based on the available science and, to the extent feasible, balance all statutory concerns. As a result, any numerical error in the assumptions used by the district fails to undermine the validity of its rule.

A. Reduction assumptions

The association claims that the district arbitrarily relied on default assumptions to reach its conclusion that rule 4570 will reduce VOC emissions by 9,594,705 pounds per year, a 65 percent reduction in VOC’s. According to the association, this assumption was arrived at by assigning a 10 percent reduction for 13 of the 19 mitigation measures offered to facilities although there is no scientific data to support a conclusion that a 10 percent reduction would occur upon implementation of any of these 13 mitigation measures. The district acknowledged that many of its assumptions were based on studies and data showing that these practices would result in some reduction, but that there was insufficient science to know exactly what the reductions would be. The district assigned a conservative percentage to the anticipated reduction to be achieved and stated that, as science becomes available, a more exact percentage will be attached to each of these measures. As the district repeatedly stated in the rulemaking process, this is a newly regulated industry and the supporting science is still developing. In the absence of more specific science, the district relied on the available studies and industry expertise to assign its numbers. It was not overly optimistic, assigning conservative numbers to estimated reductions. The association has not provided any science to refute the estimates assigned by the district.

B. Double counting

The association also claims that the district acted arbitrarily when it “double counted” reduction estimates. Rule 4570 allows facilities to feed their animals according to National Research Council (Council) guidelines as a chosen mitigation measure. It is estimated that this option will reduce valley VOC emissions by 10 percent. However, the association argues that this assumption is based on scientific studies in which the animals already were being fed according to the guidelines. As a result, the district’s estimated reduction of 10 percent arbitrarily double counts the benefits of feeding according to the guidelines and is unlikely to bring an actual reduction in VOC emissions. We disagree with the association’s position. A number of studies considered by the district suggested that dietary manipulations cause a reduction in animal VOC emissions. According to one study, changing diet from dry com to high-moisture com potentially reduces VOC emissions by 63 percent. Other studies suggest reductions to a lesser extent, but there appears to be agreement that the type of feed an animal receives impacts its enteric emissions. The district chose the lowest reduction reported as a conservative estimate of what might be accomplished if a facility chose to feed according to Council guidelines. Although there is a conflict in the record regarding whether the studies being done involved animals fed according to Council guidelines, any dispute must be resolved in favor of the district. (Gantt v. Sentry Insurance (1992) 1 Cal.4th 1083, 1087 [ 4 Cal.Rptr.2d 874 , 824 P.2d 680 ] [on appeal, we view record most strongly in favor of judgment], overruled on other grounds in Green v. Ralee Engineering Co. (1998) 19 Cal.4th 66, 80, fn. 6 [ 78 Cal.Rptr.2d 16 , 960 P.2d 1046 ].)

Additionally, the district staff has determined that a significant number of facilities do not consistently feed according to Council guidelines and that adopting this as a mitigation measure will result in real reductions of VOC’s. Even if the 10 percent is an overly optimistic number, the statute does not mandate a quantitative reduction in VOC emissions. It leaves the number crunching and the science to the agency, which has the expertise to make these types of decisions. Based on this record, we cannot say that the district’s assumptions are arbitrary and capricious.

C. Liquid manure mitigation measures

According to the association, the district improperly assumed a 50 percent control efficiency for each of the liquid manure mitigation measures based on the control efficiency of phototropic lagoons, which are managed to encourage the growth of photosynthetic bacteria. These bacteria get their energy from the sun and generate about 80 to 90 percent fewer VOC’s in the decomposition of organic materials. Since only one of the mitigation measures identified is the use of what would be a deliberately managed phototropic lagoon, the association claims the assumed reduction of 50 percent is arbitrary.

While we agree that not all the mitigation measures for liquid manure management are linked directly to evidence of a 50 percent reduction in VOC emissions, we cannot say that the district’s conservative assignment of a 50 percent reduction in emissions for this category of mitigation measures is arbitrary. VOC’s are formed as intermediate metabolites in the degradation of organic material, e.g., manure. VOC’s will be higher where there are storage tanks, ponds, overloaded anaerobic lagoons, and land application sites allowing the escape of VOC’s as they form. Anaerobic systems are the most common liquid waste management systems used. Effective anaerobic systems minimize VOC emissions. The management practices listed for this category of mitigation measures is designed to make existing anaerobic systems more effective. VOC’s are minimized when open storage ponds are eliminated; where existing lagoons are managed under properly stabilized anaerobic conditions, causing photosynthetic bacteria to flourish; where solids are eliminated or reduced; and where lagoon pH is maintained between 6.5 and 7.5.

According to the staff report, loading rates are the same for phototropic lagoons, for anaerobic lagoons managed according to Council guidelines, and for facilities using a solids separator. Also according to the staff report, this justifies an assumption that control efficiencies are similar when facilities use either a phototropic lagoon or a solids separator. The rulemaking materials corroborate the staff’s analysis of the management practices listed. In the absence of scientific data proving the contrary, and in the absence of any requirement that rule 4570 achieve a specific quantitative reduction in emissions, it is not arbitrary for the district to exercise its subject matter expertise and assign a conservative estimate to the category of mitigation measures for all liquid waste management by using the studies related to the use of phototrophic lagoons as a starting point.

D. Application of manure to cropland

Another available mitigation measure challenged by the association is the application of manure to cropland. According to the association, the district, without evidentiary support, assigned a 50 percent reduction in emissions to this measure, despite wide variation in reported reduced emissions achieved by applying manure to cropland. According to the district, rapid incorporation of animal waste into cropland by tilling, injection, and other methods, is an effective control measure. One study by the South Coast Air Quality Management District staff shows a reduction of 23 percent in VOC emissions when manure quickly is incorporated into soil. According to the district, this occurs because soil acts as a biofilter, defined as a medium of organic material with microbial activity. Additional studies show that use of biofilters generally can reduce emissions by 10 to 90 percent, depending on how the biofilters are maintained. The association has not provided persuasive authority for its contention that soil is not a biofilter, and we must defer to the district’s expertise in the absence of any showing to the contrary. Although we question why the district rejected the 23 percent number reported under the same conditions, i.e., when actual application to cropland was studied, we cannot say the number used by the district (50 percent) is arbitrary in light of the studies showing a much higher reduction obtained with use of properly maintained biofilters.

E. Failure to assess more expensive mitigation measures

Finally, according to the association, the district arbitrarily assumed that facilities would choose only the cheapest mitigation measures when complying with rule 4570 and therefore failed to attach emission-reduction estimates to the more expensive measures. First, there is no requirement that the district provide quantitative emission estimates for measures considered outside the best-available retrofit control technology. The statute requires an analysis of the emission-reduction potential of the rule and an assessment of the availability and cost effectiveness of alternatives. (§ 40724.6, subd. (e)(3), (6).) It then limits implementation to feasibility. (§ 40724.6, subd. (b).) The various alternative technologies were discussed in the staff report, and those outside the best-available retrofit control technology criteria were identified and evaluated.

Although no number was assigned, the superior control efficiency of most of the identified mitigation measures was discussed. For example, at pages 30 and 31 of the staff report, the district identified housing animals in an enclosed structure with emissions vented through a biofilter as an alternative mitigation measure. According to the district “[t]his option alone may achieve highest VOC reductions of all the management practices proposed combined.” This measure was considered outside best-available retrofit control technology because of its high cost and because it has not yet “been achieve[d] in practices at facilities similar in size to those defined as large [confined animal facilities] by the [Air Resources Board].” Likewise, the staff report identifies complete aeration of lagoons for liquid manure management as a superior management practice because complete aeration allows decomposition of the organic materials with minimal VOC emissions. The district notes, however, that the technology requires significant energy consumption and has not been achieved in practice at facilities within the valley air basin. As a result, the district concluded that the measure is beyond best-available retrofit control technology. The district’s conclusion is not arbitrary, especially in light of the statutory requirement that the emission reductions be assessed to determine the rule’s “potential” for emission reduction. (§ 40724.6, subd. (e)(3).) These mitigation measures were included in the control measure menus in order to allow those facilities willing to implement the more expensive measures credit for their economically challenging choice, in other words, to encourage voluntary practices exceeding the scope of the statutory mandate. It is reasonable to assume that the facilities would opt not to use those measures so expensive that the cost far outweighs annual profits. It is not unreasonable for the district to calculate the potential emission reduction of its rule based on the likelihood that only the cheapest options will be chosen. This is a conservative approach, not an arbitrary one.

DISPOSITION

The order denying the writ of mandate is reversed and the trial court is ordered to grant the writ, instructing the district to complete an assessment on the public health impacts of the rule. Costs on appeal are awarded to the association.

Levy, J., and Dawson, J., concurred.

All further statutory references are to the Health and Safety Code unless otherwise noted. The Clean Air Act requires that each state adopt a state implementation plan to address air pollution problems and identify how the state will achieve and maintain national air quality standards for identified pollutants. (El Comite Para El Bienestar v. Warmerdam (9th Cir. 2008) 539 F.3d 1062 ; see generally Wooley & Morss, Clean Air Act Handbook (17th ed. 2007) § 1:16, pp. 34—35.) The state in turn has created air quality control districts to address regional air quality problems stemming from sources other than motor vehicles, of which the district is one. (§ 40000.) The district includes within its jurisdiction the counties of Fresno, Kern, Kings, Madera, Merced, San Joaquin, Stanislaus, and Tulare. (§ 40600, subd. (a).) For a helpful overview of the federal and state legislative and regulatory scheme designed to combat national air pollution, see Association of Irritated Residents v. Fred Schakel Dairy (E.D.Cal., Dec. 2, 2005, No. CIV F 05-00707 AWI SMS) 2005 WL 3299508 , 61 ERC (BNA) 1801 . Ozone is an air pollutant that is regulated by the Clean Air Act, 42 United States Code section 7401 et seq., through the adoption of national ambient air quality standards. National ambient air quality standards are intended to protect public health from the adverse effects of ambient air pollutants based on the latest scientific knowledge. (Whitman v. American Trucking Assns., Inc. (2001) 531 U.S. 457, 465 [ 149 L.Ed.2d 1 , 121 S.Ct. 903 ]; Lead Industries Ass’n v. Environmental Protection (D.C. Cir. 1980) 208 U.S. App.D.C. 1 [ 647 F.2d 1130, 1152 ].) Those areas of the country not meeting the national standards are classified depending on the severity of their nonattainment. Ozone is not controlled directly, but by limiting emission of volatile organic compounds and nitrogen oxides. (See generally Wooley & Morss, Clean Air Act Handbook, supra, § 1:3, pp. 9-11.) A confined animal facility is one in which animals raised for commercial purposes are confined in large pens or corrals and are fed primarily by a method other than grazing. (§ 39011.5, subd. (a)(1).) The quoted language is found in section 40724.6, subdivisions (c), (c)(1), (d)(1)(A), and (B), respectively, with italics added.