Center for Biological Diversity v. Department of Fish & Wildlife

*213Opinion

This case presents three issues regarding the adequacy of an environmental impact report for a large land development in northwest Los Angeles County, each issue arising under the California Environmental Quality Act (CEQA; Pub. Resources Code, § 21000 et seq.): (1) Does the environmental impact report validly determine the development would not significantly impact the environment by its discharge of greenhouse gases? (2) Are mitigation measures adopted for protection of a freshwater fish, the unarmored threespine stickleback, improper because they involve taking of the fish prohibited by the Fish and Game Code? (3) Were plaintiffs’ comments on two other areas of disputed impact submitted too late in the environmental review process to exhaust their administrative remedies under Public Resources Code section 21177?

We conclude, first, that as to greenhouse gas emissions the environmental impact report employs a legally permissible criterion of significance— whether the project was consistent with meeting statewide emission reduction goals—but the report’s finding that the project’s emissions would not be significant under that criterion is not supported by a reasoned explanation based on substantial evidence. Second, we conclude the report’s mitigation measures calling for capture and relocation of the stickleback, a fully protected species under Fish and Game Code section 5515, subdivision (b)(9), themselves constitute a taking prohibited under subdivision (a) of the same statute. Finally, we hold that under the circumstances of this case plaintiffs exhausted their administrative remedies regarding certain claims of deficiency by raising them during an optional comment period on the final report.

I. Factual and Procedural Background

The Department of Fish and Wildlife (DFW, formerly the Department of Fish and Game) and the United States Army Corps of Engineers prepared a joint environmental impact statement/environmental impact report (the EIR)1 for two natural resource plans (the “Resource Management and Development Plan” and the “Spineflower Conservation Plan”) related to a proposed land development called Newhall Ranch. To be developed over about 20 years on *214almost 12,000 acres along the Santa Clara River west of the City of Santa Clarita, the proposed Newhall Ranch would consist of up to 20,885 dwelling units housing nearly 58,000 residents as well as commercial and business uses, schools, golf courses, parks and other community facilities. The project applicant and owner of Newhall Ranch is real party in interest The Newhall Land and Farming Company (Newhall).

Newhall Ranch’s potential environmental impacts were previously studied by the County of Los Angeles in connection with the county’s 2003 approval of a land use plan for the proposed development; the present EIR draws on but is independent of the environmental documentation for that approval. DFW acted as the lead state agency in preparing the EIR because the project (i.e., the Resource Management and Development Plan and the Spine flower Conservation Plan) called for DFW’s concurrence in a streambed alteration agreement and issuance of incidental take permits for protected species. Although DFW has direct authority only over biological resource impacts from the project, the agency attempts in the EIR to evaluate all environmental impacts from the project and the Newhall Ranch development that would be facilitated by project approval.

DFW and the United States Army Corps of Engineers (the Corps), the lead federal agency, issued a draft EIR in April 2009 and a final EIR in June 2010. In December 2010, DFW certified the EIR, made the findings required by CEQA as to significant impacts, mitigation, alternatives and overriding considerations, and approved the project. Of relevance here, DFW found that the project could significantly impact the unarmored threespine stickleback but that adopted mitigation measures would avoid or substantially lessen that impact, and that “taking into account the applicant’s design commitments and existing regulatory standards,” Newhall Ranch’s emissions of greenhouse gases would have a less than significant impact on the global climate.

Plaintiffs challenged DFW’s actions by a petition for writ of mandate.2 The superior court granted the petition on several grounds. The Court of Appeal reversed, rejecting all of plaintiffs’ CEQA claims. We granted plaintiffs’ petition for review.

II. Discussion

The general principles governing our review of DFW’s actions can be simply stated. In reviewing an agency’s nonadjudicative determination or *215decision for compliance with CEQA, we ask whether the agency has prejudi-cially abused its discretion; such an abuse is established “if the agency has not proceeded in a manner required by law or if the determination or decision is not supported by substantial evidence.” (Pub. Resources Code, § 21168.5.)3 In determining whether there has been an abuse of discretion, we review the agency’s action, not the trial court’s decision. “[I]n that sense appellate judicial review under CEQA is de novo.” (Vineyard Area Citizens for Responsible Growth, Inc. v. City of Rancho Cordova (2007) 40 Cal.4th 412, 427 [53 Cal.Rptr.3d 821, 150 P.3d 709] (Vineyard Area Citizens))

On particular questions of CEQA compliance, however, the standard of review depends on “whether the claim is predominantly one of improper procedure or a dispute over the facts.” (Vineyard Area Citizens, supra, 40 Cal.4th at p. 435.) “While we determine de novo whether the agency has employed the correct procedures, ... we accord greater deference to the agency’s substantive factual conclusions. In reviewing for substantial evidence, the reviewing court ‘may not set aside an agency’s approval of an EIR on the ground that an opposite conclusion would have been equally or more reasonable,’ for, on factual questions, our task ‘is not to weigh conflicting evidence and determine who has the better argument.’ (Laurel Heights [Improvement Assn. v. Regents of University of California (1988)] 47 Cal.3d [376,] 393 [253 Cal.Rptr. 426, 764 P.2d 278].)” (Ibid.)

A. The EIR’s Determination the Project’s Greenhouse Gas Emissions Will Not Have a Significant Environmental Impact

1. Background

In California’s landmark legislation addressing global climate change, the California Global Warming Solutions Act of 2006 (Health & Saf. Code, § 38500 et seq.), Statutes 2006, chapter 488, page 3419 (enacting Assem. Bill No. 32 (2005-2006 Reg. Sess.) (Assembly Bill 32)), our Legislature emphatically established as state policy the achievement of a substantial reduction in the emission of gases contributing to global warming. (Health & Saf. Code, §§ 38500, 38501.) More specifically, Assembly Bill 32 calls for reduction of such emissions to 1990 levels by the year 2020. (Health & Saf. Code, § 38550.) The law designates the State Air Resources Board (the Air Board) as the state agency charged with regulating greenhouse gas emissions {id., § 38510) and calls for the Air Board to coordinate with other state agencies to implement the state’s reduction goal (id.. § 38501, subd. (1)).

Under Assembly Bill 32, the Air Board was required to determine as accurately as possible the statewide level of greenhouse gas emissions in *2161990 and to approve on that basis a statewide emissions limit to be achieved by 2020. (Health & Saf. Code, § 38550.) The Air Board was required to prepare and approve by January 1, 2009, a “scoping plan” for achieving the “maximum technologically feasible and cost-effective” reductions in greenhouse gas emissions by 2020. {Id., § 38561, subd. (a).)

In its 2008 Climate Change Scoping Plan, the Air Board explained that “[r]educing greenhouse gas emissions to 1990 levels means cutting approximately 30 percent from business-as-usual emission levels projected for 2020, or about 15 percent from today’s levels.” (Air Bd., Climate Change Scoping Plan (Dec. 2008) Executive Summary, p. ES-1 (Scoping Plan).) The Scoping Plan then set out a “comprehensive array of emissions reduction approaches and tools” to meet the goal, including expanding energy efficiency programs, achieving a statewide renewable energy mix of 33 percent, developing with our regional partners a cap-and-trade program for greenhouse gases, establishing targets and policies for emissions in transportation and implementing existing clean transportation programs, and creating targeted fees on certain activities affecting emissions. {Id., pp. ES-3 to ES-4.)

The Scoping Plan’s “business-as-usual” model is important here, as it formed the basis for the present EIR’s greenhouse gas significance analysis. The Air Board had previously identified a year 2020 annual emissions limit, equal to its estimate of statewide 1990 emissions, of 427 million metric tons of carbon dioxide equivalent (MMTC02E). (Scoping Plan, supra, at p. 5.) In the Scoping Plan, the board estimated emissions by economic sector in the period 2002 to 2004, finding they totaled 469 MMTC02E annually. Those annual emissions were then projected forward to the year 2020, employing population and economic growth estimates, yielding a business-as-usual figure of 596 MMTC02E. {Id., p. 13.) The target of 427 MMTC02E is about 29 percent below the 2020 forecast of 596 MMTC02E, giving the Air Board the 30 percent reduction goal quoted earlier.

The Scoping Plan’s 2020 forecast is referred to as a “business-as-usual” projection because it assumes no conservation or regulatory efforts beyond what was in place when the forecast was made. It “represent[s] the emissions that would be expected to occur in the absence of any GHG [greenhouse gas] reductions actions.” (Scoping Plan, supra, appen. F, California’s Greenhouse Gas Emissions Inventory, p. F-3.) For example, the emissions forecast for electricity generation assumes “all growth in electricity demand by 2020 will be met by in-state natural gas-fired power plants” and the estimate for on-road vehicle emissions “assumes no change in vehicle fleet mix over time.” {Id., p. F-4.)

Neither Assembly Bill 32 nor the Air Board’s Scoping Plan set out a mandate or method for CEQA analysis of greenhouse gas emissions from a *217proposed project. A 2007 CEQA amendment, however, required the preparation, adoption and periodic update of guidelines for mitigation of greenhouse gas impacts. (Stats. 2007, ch. 185, § 1, p. 2330, adding Pub. Resources Code, § 21083.05.) In 2010, the Natural Resources Agency adopted a new CEQA guideline on determining the significance of impacts from greenhouse gas emissions. (Cal. Code Regs., tit. 14, § 15064.4.)4

The new guideline provides that a lead agency should attempt to “describe, calculate or estimate” the amount of greenhouse gases a project will emit, but recognizes that agencies have discretion in how to do so. (Guidelines, § 15064.4, subd. (a).) It goes on to provide that when assessing the significance of greenhouse gas emissions, the agency should consider these factors among others: “(1) The extent to which the project may increase or reduce greenhouse gas emissions as compared to the existing environmental setting; [¶] (2) Whether the project emissions exceed a threshold of significance that the lead agency determines applies to the project[;] [¶] (3) The extent to which the project complies with regulations or requirements adopted to implement a statewide, regional, or local plan for the reduction or mitigation of greenhouse gas emissions. Such requirements must be adopted by the relevant public agency through a public review process and must reduce or mitigate the project’s incremental contribution of greenhouse gas emissions. If there is substantial evidence that the possible effects of a particular project are still cumulatively considerable notwithstanding compliance with the adopted regulations or requirements, an EIR must be prepared for the project.” (Id., subd. (b).)

2. The EIR’s Significance Determination

In order to evaluate the project’s greenhouse gas emissions impact, the EIR attempts to quantify the emissions currently generated on the project site in its existing uses and the emissions that would be generated by full development of the Newhall Ranch community. Annual emissions from the existing uses (primarily oil wells and agriculture) are estimated at 10,272 metric tons of C02, which the EIR conservatively treats as zero for purposes of the impact analysis. The annual greenhouse gas emissions from Newhall Ranch at full build-out are projected to be 269,053 metric tons of C09 equivalent (MTC02E).

*218The EIR asserts that while this annual emissions increase of 269,053 MTC02E is “an obvious change to existing, on-site conditions,” the global nature of climate change and the “absence of scientific and factual information” on the significance of particular amounts of greenhouse gas emissions make the change “[in]sufficient to support a significance determination.” The EIR accordingly goes on to consider “whether the proposed Project’s emissions . . . would impede the State of California’s compliance with the statutory emissions reduction mandate established by AB 32.”

The EIR’s method for determining whether the project would impede achievement of Assembly Bill 32’s goals is modeled on the Air Board’s use, in its Scoping Plan, of comparison to a “business-as-usual” projection as a measure of the emission reductions needed to meet the 2020 goal (determined to be a reduction of 29 percent from business as usual). As explained earlier, the Scoping Plan forecasted statewide greenhouse gas emissions under a business-as-usual scenario in which no additional regulatory actions were taken to reduce emissions. The EIR does the same for Newhall Ranch, estimating at 390,046 MTC02E per year the emissions “if the proposed Project and resulting development were constructed consistent with [the Air Board’s] assumptions for the CARB 2020 NAT [no action taken, or business as usual] scenario.” Because the EIR’s estimate of actual annual project emissions (269,053 MTC02E) is 31 percent below its business-as-usual estimate (390,046 MTC02E), exceeding the Air Board’s determination of a 29 percent reduction from business as usual needed statewide, the EIR concludes the project’s likely greenhouse gas emissions will not impede achievement of Assembly Bill 32’s goals and are therefore less than significant for CEQA purposes.

3. Analysis

We consider whether DFW abused its discretion in determining the project’s greenhouse gas emissions would not have a significant environmental impact, either because it failed to proceed in the manner required by CEQA or because it made the no significant impact determination without the support of substantial evidence in the administrative record. (§ 21168.5.)

Plaintiffs contend the EIR’s no significant impact conclusion resulted from use of a legally improper baseline for comparison. Relying on this court’s decision in Communities for a Better Environment v. South Coast Air Quality Management Dist. (2010) 48 Cal.4th 310 [106 Cal.Rptr.3d 502, 226 P.3d 985] (Communities for a Better Environment), in which we disapproved the defendant district’s use of pollutant emission levels allowed under prior permits—but not reflecting actual existing conditions—as a comparative baseline for a CEQA significance evaluation, plaintiffs argue DFW erred in *219determining significance by comparison to the hypothetical business-as-usual scenario rather than by comparison to existing greenhouse gas emissions on the project site.

DFW contends it properly relied on methodology devised by the Air Board, the state agency with greatest expertise on climate change. Newhall defends the EIR’s approach and conclusion extensively, arguing that DFW acted within its discretion under Guidelines section 15064.4 in adopting compliance with Assembly Bill 32’s goals as its significance criterion and that both DFW’s choice of methodology and its conclusion of no significant impact should be reviewed only for support by substantial evidence.

We begin with the broadest question posed: Did DFW abuse its discretion in adopting consistency with Assembly Bill 32’ s reduction goals as its significance criterion for the project’s greenhouse gas emissions? We review this issue de novo, as it is predominantly a legal question of correct CEQA procedure. (Communities for a Better Environment, supra, 48 Cal.4th at p. 319; Vineyard Area Citizens, supra, 40 Cal.4th at p. 435.)

Before considering the principal statutory and regulatory provisions governing CEQA analysis of greenhouse gas emissions (§ 21083.05; Guidelines, § 15064.4), we address two related aspects of the greenhouse gas problem that inform our discussion of CEQA significance.

First, because of the global scale of climate change, any one project’s contribution is unlikely to be significant by itself. The challenge for CEQA purposes is to determine whether the impact of the project’s emissions of greenhouse gases is cumulatively considerable, in the sense that “the incremental effects of [the] individual project are considerable when viewed in connection with the effects of past projects, the effects of other current projects, and the effects of probable future projects.” (§ 21083, subd. (b)(2); see § Guidelines, 15064, subd. (h)(1).) “With respect to climate change, an individual project’s emissions will most likely not have any appreciable impact on the global problem by themselves, but they will contribute to the significant cumulative impact caused by greenhouse gas emissions from other sources around the globe. The question therefore becomes whether the project’s incremental addition of greenhouse gases is ‘cumulatively considerable’ in light of the global problem, and thus significant.” (Crockett, Addressing the Significance of Greenhouse Gas Emissions Under CEQA: California’s Search for Regulatory Certainty in an Uncertain World (July 2011) 4 Golden Gate U. Envtl. L.J. 203, 207-208 (hereafter Addressing the Significance of Greenhouse Gas Emissions).)

Second, the global scope of climate change and the fact that carbon dioxide and other greenhouse gases, once released into the atmosphere, are *220not contained in the local area of their emission means that the impacts to be evaluated are also global rather than local. For many air pollutants, the significance of their environmental impact may depend greatly on where they are emitted; for greenhouse gases, it does not. For projects, like the present residential and commercial development, which are designed to accommodate long-term growth in California’s population and economic activity, this fact gives rise to an argument that a certain amount of greenhouse gas emissions is as inevitable as population growth. Under this view, a significance criterion framed in terms of efficiency is superior to a simple numerical threshold because CEQA is not intended as a population control measure.

The EIR makes this point in its response to plaintiff Center for Biological Diversity’s comments on the greenhouse gas significance analysis: “[Wjhen location does not matter (such as in the case of GHG emissions), evaluation of project significance via an efficiency metric is appropriate. [¶] . . . [F]or a global environmental issue (such as climate change), utilizing an absolute number as a significance criterion equates to attempting to use CEQA to discourage population growth. Of note, the future residents and occupants of development enabled by Project approval would exist and live somewhere else if this Project is not approved. Whether ‘here or there,’ GHG emissions associated with such population growth will occur.”

These considerations militate in favor of consistency with meeting Assembly Bill 32 ’s statewide goals as a permissible significance criterion for project emissions. Meeting our statewide reduction goals does not preclude all new development. Rather, the Scoping Plan—the state’s roadmap for meeting Assembly Bill 32’s target—assumes continued growth and depends on increased efficiency and conservation in land use and transportation from all Californians. (See Scoping Plan, supra, pp. ES-1 [meeting the Assembly Bill 32 goal ‘“means reducing our annual emissions of 14 tons of carbon dioxide equivalent for every man, woman and child in California down to about 10 tons per person by 2020”]; Scoping Plan, at pp. 15 [‘“Every part of California’s economy needs to play a role in reducing greenhouse gas emissions.”], 42 [outlining energy efficiency measures for both new and existing buildings].) To the extent a project incorporates efficiency and conservation measures sufficient to contribute its portion of the overall greenhouse gas reductions necessary, one can reasonably argue that the project’s impact ‘“is not ‘cumulatively considerable,’ because it is helping to solve the cumulative problem of greenhouse gas emissions as envisioned by California law.” (Addressing the Significance of Greenhouse Gas Emissions, supra, 4 Golden Gate U. Envtl. L.J. at p. 210.)

Given the reality of growth, some greenhouse gas emissions from new housing and commercial developments are inevitable. The critical CEQA *221question is the cumulative significance of a project’s greenhouse gas emissions, and from a climate change point of view it does not matter where in the state those emissions are produced. Under these circumstances, evaluating the significance of a residential or mixed-use project’s greenhouse gas emissions by their effect on the state’s efforts to meet its long-term goals makes at least as much sense as measuring them against an absolute numerical threshold.

Using consistency with Assembly Bill 32’s statewide goal for greenhouse gas reduction, rather than a numerical threshold, as a significance criterion is also consistent with the broad guidance provided by section 15064.4 of the CEQA Guidelines. As the issuing agency explained, section 15064.4 was drafted to reflect “the existing CEQA principle that there is no iron-clad definition of ‘significance.’ ” (Natural Resources Agency, Final Statement of Reasons for Regulatory Action: Amendments to the State CEQA Guidelines Addressing Analysis and Mitigation of Greenhouse Gas Emissions Pursuant to SB 97 (Dec. 2009) p. 20 (Final Statement of Reasons); cf. § 21083.05 [requiring periodic update of CEQA Guidelines for mitigation of greenhouse gas emissions to reflect new information or criteria established by Air Board].) Section 15064.4 was not intended to closely restrict agency discretion in choosing a method for assessing greenhouse gas emissions, but rather “to assist lead agencies” in investigating and disclosing “all that they reasonably can” regarding a project’s greenhouse gas emissions impacts. (Final Statement of Reasons, supra, at p. 20.)5

While Guidelines section 15064.4 states a lead agency “should consider,” among other factors, “[t]he extent to which the project may increase or reduce greenhouse gas emissions as compared to the existing environmental setting” (id., subd. (b)(1)) and “[wjhether the project emissions exceed a threshold of significance that the lead agency determines applies to the project” (id., subd. (b)(2)), the section does not mandate the use of absolute numerical thresholds to measure the significance of greenhouse gas emissions. The factors listed in subdivision (b) are not exclusive. They are rather intended “to assist lead agencies in collecting and considering information *222relevant to a project’s incremental contribution of GHG emissions and the overall context of such emissions.” (Final Statement of Reasons, supra, at p. 24.)

The present EIR discloses the project’s likely increase in emissions over the existing environment, informing the reader that the project will increase greenhouse gas emissions by 269,053 MTC02E compared to the existing environmental setting (Guidelines, § 15064.4, subd. (b)(1)), but declines to consider the impact significant based on the size of that increase alone ‘“because of the absence of scientific and factual information regarding when particular quantities of greenhouse gas emissions become significant.” As for a significance threshold (id., subd. (b)(2)), the EIR asserts that no agency had adopted an applicable threshold.

Plaintiffs challenge these statements as insufficient to justify the EIR’s choice of methodology, noting that California air pollution control officials and air quality districts have made several proposals for numerical thresholds. But given that multiple agencies’ efforts at framing greenhouse gas significance issues have not yet coalesced into any widely accepted set of numerical significance thresholds, but have produced ‘“a certain level of consensus” on the value of Assembly Bill 32 consistency as a criterion (Addressing the Significance of Greenhouse Gas Emissions, supra, 4 Golden Gate U. Envtl. L.J. at p. 209), we cannot conclude DFW’s discretionary choice of Assembly Bill 32 consistency as a significance criterion for this project violated Guidelines section 15064.4, subdivision (b)(1) or (2).

Subdivision (b)(3) of Guidelines section 15064.4 states the lead agency should also consider “[t]he extent to which the project complies with regulations or requirements adopted to implement a statewide, regional, or local plan for the reduction or mitigation of greenhouse gas emissions.” Assembly Bill 32 did not create a set of ‘“regulations or requirements” implementing a ‘“plan” (Guidelines, § 15064.4, subd. (b)(3)); indeed, it is not a plan but rather a statement of policies and objectives. The Scoping Plan adopted pursuant to Assembly Bill 32 is a plan for reducing greenhouse gas emissions, but does not itself establish the regulations by which it is to be implemented; rather, it sets out how existing regulations, and new ones yet to be adopted at the time of the Scoping Plan, will be used to reach Assembly Bill 32’s emission reduction goal. At the time the Natural Resources Agency promulgated Guidelines section 15064.4, the agency explained that the Scoping Plan “may not be appropriate for use in determining the significance of individual projects . . . because it is conceptual at this stage and relies on the future development of regulations to implement the strategies identified in the Scoping Plan.” (Final Statement of Reasons, supra, at pp. 26-27.)

*223In short, neither Assembly Bill 32 nor the Scoping Plan establishes regulations implementing, for specific projects, the Legislature’s statewide goals for reducing greenhouse gas emissions. Neither constitutes a set of “regulations or requirements adopted to implement” a statewide reduction plan within the meaning of Guidelines section 15064.4, subdivision (b)(3). That guideline, however, does not expressly or impliedly prohibit a lead agency from using the Assembly Bill 32 goals themselves to determine whether the project’s projected greenhouse gas emissions are significant. As noted by the Natural Resources Agency in its amicus curiae brief, “a discussion of a project’s consistency with the State’s long-term climate stabilization objectives . . . will often be appropriate . . . under CEQA,” provided the analysis is “tailored . . . specifically to a particular project.” Indeed, to proceed in this manner is consistent with CEQA’s “inherent recognition . . . that if a plan is in place to address a cumulative problem, a new project’s incremental addition to the problem will not be ‘cumulatively considerable’ if it is consistent with the plan and is doing its fair share to achieve the plan’s goals.” (Addressing the Significance of Greenhouse Gas Emissions, supra, 4 Golden Gate U. Envtl. L.J. at pp. 210-211.) For this reason as well, we conclude DFW’s choice to use that criterion does not violate CEQA. The only published Court of Appeal decisions to consider this question have reached the same conclusion, albeit with little discussion. (Friends of Oroville v. City of Oroville (2013) 219 Cal.App.4th 832, 841 [164 Cal.Rptr.3d 1]; Citizens far Responsible Equitable Environmental Development v. City of Chula Vista (2011) 197 Cal.App.4th 327, 335-336 [127 Cal.Rptr.3d 435].)

A qualification regarding the passage of time is in order here. Plaintiffs do not claim it was improper for this EIR, issued in 2010, to look forward only to 2020 for a guidepost on reductions in greenhouse gas emissions, and we therefore do not consider the question whether CEQA required the EIR to address the state’s goals beyond 2020. Nevertheless, over time consistency with year 2020 goals will become a less definitive guide, especially for long-term projects that will not begin operations for several years. An EIR taking a goal-consistency approach to CEQA significance may in the near future need to consider the project’s effects on meeting longer term emissions reduction targets.6

*224Having concluded DFW did not proceed in violation of CEQA by its choice of Assembly Bill 32 consistency as a significance criterion, we proceed to plaintiffs’ contention that the agency violated CEQA by comparing the project’s expected emissions to a hypothetical business-as-usual scenario rather than to a baseline of emissions in the existing physical environment.

In Communities for a Better Environment, supra, 48 Cal.4th 310, a refinery sought a permit to conduct a new process using some new and some existing equipment, including existing boilers used for steam generation, each of which was subject to an existing permit setting its maximum rate of operation. (Id. at pp. 317-318.) The negative declaration the regional air district prepared for the project, in determining the significance of the project’s nitrogen oxide emissions, treated emissions that could be generated by the existing boilers operating together at their maximum permitted capacity (a condition that did not occur in normal operation) as part of the baseline for environmental review rather than as part of the project. (Id. at p. 318.) Although the negative declaration acknowledged that actual nitrogen oxide emissions would increase under the project by an amount that would normally be considered significant, the declaration determined the emissions were not significant because they were below what could have been emitted by the refinery’s boilers under the existing permits. (Ibid.)

We held the air district’s approach violated the rule expressed in Guidelines section 15125, subdivision (a), as well as in case law, that the comparative baseline for a significance determination should normally be the existing physical conditions in the project’s vicinity. (Communities for a Better Environment, supra, 48 Cal.4th at pp. 320-322.) “By comparing the proposed project to what could happen, rather than to what was actually happening, the District set the baseline not according to ‘established levels of a particular use,’ but by ‘merely hypothetical conditions allowable’ under the permits. [Citation.] Like an EIR, an initial study or negative declaration ‘must focus on impacts to the existing environment, not hypothetical situations.’ [Citation.]” (Id. at p. 322.)

Contrary to plaintiffs’ arguments, we do not see the EIR’s approach here as comparable to that of the negative declaration in Communities for a Better Environment. Unlike the air district in Communities for a Better Environment, DFW does not claim its business-as-usual model represented “the physical environmental conditions ... as they exist” at the time of environmental *225analysis. (Guidelines, § 15125, subd. (a).) Rather, it employs a hypothetical business-as-usual emissions model merely as a means of comparing the project’s projected emissions to the statewide target set under the Scoping Plan. The business-as-usual emissions model is used here as a comparative tool for evaluating efficiency and conservation efforts, not as a significance baseline.

The percentage reduction from business as usual identified by the Scoping Plan is a measure of the reduction effort needed to meet the 2020 goal, not an attempt to describe the existing level of greenhouse gas emissions. Similarly, the EIR employs its calculation of project reductions from business-as-usual emissions in an attempt to show the project incorporates efficiency and conservation measures sufficient to make it consistent with achievement of Assembly Bill 32’s reduction goal, not to show the project will not increase greenhouse gas emissions over those in the existing environment. As discussed earlier, distinctive aspects of the greenhouse gas problem make consistency with statewide reduction goals a permissible significance criterion for such emissions. Using a hypothetical scenario as a method of evaluating the proposed project’s efficiency and conservation measures does not violate Guidelines section 15125 or contravene our decision in Communities for a Better Environment.

Notwithstanding this conclusion, we agree with plaintiffs that DFW abused its discretion in finding, on the basis of the EIR’s business-as-usual comparison, that the project’s greenhouse gas emissions would have no cumulatively significant impact on the environment. We reach this conclusion because the administrative record discloses no substantial evidence that Newhall Ranch’s project-level reduction of 31 percent in comparison to business as usual is consistent with achieving Assembly Bill 32’s statewide goal of a 29 percent reduction from business as usual, a lacuna both dissenting opinions fail to address. Even using the EIR’s own significance criterion, the EIR’s analysis fails to support its conclusion of no significant impact.

The Scoping Plan set out a statewide reduction goal and a framework for reaching it—a set of broadly drawn regulatory approaches covering all sectors of the California economy and projected, if implemented and followed, to result in a reduction to 1990-level greenhouse gas emissions by the year 2020. The plan expressed the overall level of conservation and efficiency improvements required as, among other measures, a percentage reduction from a hypothetical scenario in which no additional regulatory actions were taken. But the Scoping Plan nowhere related that statewide level of reduction effort to the percentage of reduction that would or should be required from *226individual projects, and nothing DFW or Newhall has cited in the administrative record indicates the required percentage reduction from business as usual is the same for an individual project as for the entire state population and economy.

Plaintiffs put forward one ready reason to suspect that the percent reduction is not the same, and that in fact a greater degree of reduction may be needed from new land use projects than from the economy as a whole: Designing new buildings and infrastructure for maximum energy efficiency and renewable energy use is likely to be easier, and is more likely to occur, than achieving the same savings by retrofitting of older structures and systems. The California Attorney General’s Office made this point while commenting on an air district’s greenhouse gas emissions reduction plan, in a letter one of the plaintiffs brought to DFW’s attention in a comment on the EIR: “The [air district] Staff Report seems to assume that if new development projects reduce emissions by 29 percent compared to ‘business as usual,’ the 2020 statewide target of 29 percent below ‘business as usual’ will also be achieved, but it does not supply evidence of this. Indeed, it seems that new development must be more GHG-efficient than this average, given that past and current sources of emissions, which are substantially less efficient than this average, will continue to exist and emit.” In its administrative response to this comment, DFW observed that the Scoping Plan did call for emissions reductions from existing buildings (though these are not separately quantified) and that one air district’s analysis of the Scoping Plan indicated the “land-use driven” economic sector would be required to make only a 26.2 percent reduction from business as usual.

DFW’s responses to comments on the EIR do not suffice to demonstrate that a 31 percent reduction from business as usual at the project level corresponds to the statewide reductions called for in the Scoping Plan. In its brief, Newhall characterizes this question as one of competing expert opinions, on which the courts must defer to the lead agency. But Newhall points to no expert opinion stating generally that the Scoping Plan contemplates the same emission reductions from new buildings as from existing ones, or more particularly that the Scoping Plan’s statewide standard of a 29 percent reduction from business as usual applies without modification to a new residential or mixed-use development project.

Even if the statewide and economy-wide percentage reduction set out in the Scoping Plan were shown to be generally appropriate for use as a criterion of significance for individual projects, the EIR’s conclusion that greenhouse gas emissions will be less than significant would still lack substantial supporting evidence. This is because the EIR makes an unsupported assumption regarding statewide density averages used in the Scoping *227Plan, an assumption that if incorrect could result in a misleading business-as-usual comparison. As plaintiffs point out, the EIR’s business-as-usual scenario assumes residential density equal to that currently found in the Santa Clarita Valley. Because Newhall Ranch as designed would have greater residential density than the existing average for the Santa Clarita Valley, the EIR makes a downward adjustment from business as usual in projected vehicle miles traveled, and consequently in greenhouse gas emissions from mobile sources (a substantial part of the total emissions). As far as the EIR reveals, however, the Scoping Plan’s statewide business-as-usual model is not necessarily based on residential densities equal to the Santa Clarita Valley average.

The Scoping Plan’s business-as-usual projection of vehicle miles traveled in 2020 was derived using an established growth model for such projections. (Scoping Plan, supra, appen. F, at pp. F-3 to F-4.) But nothing DFW or Newhall points to in the administrative record shows the statewide density assumptions used in that model mirror conditions in the Santa Clarita Valley. To the extent the Scoping Plan’s business-as-usual scenario assumes population densities greater than the Santa Clarita Valley density assumed in the EIR’s business-as-usual projection, the EIR’s comparison of project reductions from business as usual to reductions demanded in the Scoping Plan will be misleading. The administrative record does not establish a firm ground for the efficiency comparison the EIR makes and thus, for this reason as well, does not substantially support the EIR’s conclusion that Newhall Ranch’s 31 percent emissions savings over business as usual satisfies the report’s significance criterion of consistency with the Scoping Plan’s 29 percent statewide savings by 2020.

At bottom, the EIR’s deficiency stems from taking a quantitative comparison method developed by the Scoping Plan as a measure of the greenhouse gas emissions reduction effort required by the state as a whole, and attempting to use that method, without consideration of any changes or adjustments, for a purpose very different from its original design: to measure the efficiency and conservation measures incorporated in a specific land use development proposed for a specific location. The EIR simply assumes that the level of effort required in one context, a 29 percent reduction from business as usual statewide, will suffice in the other, a specific land use development. From the information in the administrative record, we cannot say that conclusion is wrong, but neither can we discern the contours of a logical argument that it is right. The analytical gap left by the EIR’s failure to establish, through substantial evidence and reasoned explanation, a quantitative equivalence between the Scoping Plan’s statewide comparison and the EIR’s own project-level comparison deprived the EIR of its “ ‘sufficiency as an informative document.’ ” (Laurel Heights Improvement Assn. v. Regents of University of California, supra, 47 Cal.3d at p. 392.)

*228Justice Corrigan argues our conclusion on this point, requiring DFW to support its chosen quantitative method for analyzing significance with evidence and reasoned argument, is inconsistent with the deferential nature of our review. (Cone. & dis. opn. of Corrigan, J., post, at p. 243.) We disagree. A lead agency enjoys substantial discretion in its choice of methodology. But when the agency chooses to rely completely on a single quantitative method to justify a no-significance finding, CEQA demands the agency research and document the quantitative parameters essential to that method. Otherwise, decision makers and the public are left with only an unsubstantiated assertion that the impacts—here, the cumulative impact of the project on global warming—will not be significant. (See Guidelines, § 15064, subd. (f)(5) [substantial evidence to support a finding on significance includes “facts, reasonable assumptions predicated upon facts, and expert opinion supported by facts,” but not “[ajrgument, speculation, [or] unsubstantiated opinion”].)

Nor is Justice Corrigan correct that our analysis “assumes project-level reductions in greenhouse gas emissions must be greater than the reductions California is seeking to achieve statewide.” (Conc. & dis. opn. of Corrigan, J., post, at pp. 241-242.) As discussed just above (ante, at p. 227), we hold only that DFW erred in failing to substantiate its assumption that the Scoping Plan’s statewide measure of emissions reduction can also serve as the criterion for an individual land use project.

We further agree with plaintiffs that DFW’s failure to provide substantial evidentiary support for its no significant impact conclusion was prejudicial, in that it deprived decision makers and the public of substantial relevant information about the project’s likely impacts. (Neighbors for Smart Rail v. Exposition Metro Line Construction Authority (2013) 57 Cal.4th 439, 463 [160 Cal.Rptr.3d 1, 304 P.3d 499] (lead opn. of Werdegar, J.); Environmental Protection Information Center v. California Dept. of Forestry & Fire Protection (2008) 44 Cal.4th 459, 485-486 [80 Cal.Rptr.3d 28, 187 P.3d 888]; Sierra Club v. State Bd. of Forestry (1994) 7 Cal.4th 1215, 1236-1237 [32 Cal.Rptr.2d 19, 876 P.2d 505].) 'in this EIR, DFW employed the business-as-usual comparison as its sole criterion of significance. In the absence of substantial evidence to support the EIR’s no-significance finding, as noted above, the EIR’s readers have no way of knowing whether the project’s likely greenhouse gas emissions impacts will indeed be significant and, if so, what mitigation measures will be required to reduce them. This is not the sort of “[insubstantial or merely technical omission[]” that can be overlooked in deciding whether to grant relief. (Neighbors for Smart Rail v. Exposition Metro Line Construction Authority, supra, at p. 463.)

We briefly address some of the potential options for DFW on remand and for other lead agencies faced with evaluating the cumulative significance of a *229proposed land use development’s greenhouse gas emissions. While the burden of CEQA’s mandate in this context can be substantial, methods for complying with CEQA do exist. We do not, of course, guarantee that any of these approaches will be found to satisfy CEQA’s demands as to any particular project; what follows is merely a description of potential pathways to compliance, depending on the circumstances of a given project.

First, although we have found the particular comparison made here lacking in support, and although doubt has been cast on the Scoping Plan’s project-level appropriateness (see Final Statement of Reasons, supra, at pp. 24-25), a business-as-usual comparison based on the Scoping Plan’s methodology may be possible. On an examination of the data behind the Scoping Plan’s business-as-usual model, a lead agency might be able to determine what level of reduction from business as usual a new land use development at the proposed location must contribute in order to comply with statewide goals.

Second, a lead agency might assess consistency with Assembly Bill 32’s goal in whole or in part by looking to compliance with regulatory programs designed to reduce greenhouse gas emissions from particular activities. (See Final Statement of Reasons, supra, at p. 64 [greenhouse gas emissions “may be best analyzed and mitigated at a programmatic level”].) To the extent a project’s design features comply with or exceed the regulations outlined in the Scoping Plan and adopted by the Air Board or other state agencies, a lead agency could appropriately rely on their use as showing compliance with “performance based standards” adopted to fulfill “a statewide . . . plan for the reduction or mitigation of greenhouse gas emissions.” (Guidelines, § 15064.4, subds. (a)(2), (b)(3); see id., § 15064, subd. (h)(3) [determination that impact is not cumulatively considerable may rest on compliance with previously adopted plans or regulations, including “plans or regulations for the reduction of greenhouse gas emissions”].)

A significance analysis based on compliance with such statewide regulations, however, only goes to impacts within the area governed by the regulations. That a project is designed to meet high building efficiency and conservation standards, for example, does not establish that its greenhouse gas emissions from transportation activities lack significant impacts. (Final Statement of Reasons, supra, at p. 23.) Although transportation accounts for almost 40 percent of the state’s greenhouse gas emissions, and transportation emissions are affected by the location and density of residential and commercial development, the Scoping Plan does not propose statewide regulation of land use planning but relies instead on local governments. (Scoping Plan, supra, at pp. 11, 27.)

*230Local governments thus bear the primary burden of evaluating a land use project’s impact on greenhouse gas emissions. Some of this burden can be relieved by using geographically specific greenhouse gas emission reduction plans to provide a basis for the tiering or streamlining of project-level CEQA analysis. Guidelines section 15183.5, added in 2010 along with section 15064.4, explains in detail how a programmatic effort such as “a general plan, a long range development plan, or a separate plan to reduce greenhouse gas emissions” (id., § 15183.5, subd. (a)) may, if sufficiently detailed and adequately supported, be used in later project-specific CEQA documents to simplify the evaluation of the project’s cumulative contribution to the effects of greenhouse gas emissions (§ 15183.5, subd. (b)). The Scoping Plan encourages local jurisdictions to develop “ ‘climate action plans’ ” or greenhouse gas “ ‘emissions reduction plans’ ” for their geographic areas, and several jurisdictions have adopted or proposed such plans as tools for CEQA streamlining. (Final Statement of Reasons, supra, at p. 65; see, e.g., City of Milpitas, Climate Action Plan: A Qualified Greenhouse Gas Reduction Strategy (May 2013) p. 1-1; City of San Bernardino, Sustainability Master Plan (Public Review Draft, Aug. 2012) p. 4.)

In addition, CEQA expressly allows streamlining of transportation impacts analysis for certain land use projects based on metropolitan regional ‘“sustainable communities strategies.” Under follow-up legislation to Assembly Bill 32 (Stats. 2008, ch. 728, p. 5065, commonly known as Senate Bill 375) each metropolitan planning organization in the state is to prepare a ‘“sustainable communities strategy” or alternative plan to meet regional targets set by the Air Board for greenhouse gas emissions from cars and light trucks. (Gov. Code, § 65080, subd. (b)(2).) CEQA documents for certain residential, mixed-use and transit priority projects that are consistent with the limits and policies specified in an applicable sustainable communities strategy need not additionally analyze greenhouse gas emissions from cars and light trucks. (§§ 21155.2, 21159.28; Guidelines, § 15183.5, subd. (c).)

Third, a lead agency may rely on existing numerical thresholds of significance for greenhouse gas emissions, though as we have explained (ante, at p. 221), use of such thresholds is not required. (Guidelines, § 15064.4, subd. (b)(2); see, e.g., Bay Area Air Quality Management Dist. (BAAQMD), California Environmental Quality Act Guidelines Update: Proposed Thresholds of Significance (May 3, 2010) pp. 8-21 [regional air quality district for the San Francisco Bay Area proposes a threshold of 1,100 MTC02E in annual emissions as one alternative agencies may use in determining CEQA significance for new land use projects].)7 Thresholds, it *231should be noted, only define the level at which an environmental effect “normally” is considered significant; they do not relieve the lead agency of its duty to determine the significance of an impact independently. (Guidelines, § 15064.7, subd. (a); Mejia v. City of Los Angeles (2005) 130 Cal.App.4th 322, 342 [29 Cal.Rptr.3d 788].)

For a large land use project such as Newhall Ranch, using a numerical threshold may result in a determination of significant greenhouse gas emission impacts. In that circumstance, the lead agency must adopt feasible mitigation measures or project alternatives to reduce the effect to insignificance; to the extent significant impacts remain after mitigation, the agency may still approve the project with a statement of overriding considerations. (§§ 21002, 21002.1, subd. (b), 21081; Guidelines, §§ 15091, 15093, 15126.6.) Were DFW to determine on remand that adding hundreds of thousands of tons of greenhouse gasses to the atmosphere has a cumulatively significant effect, therefore, it would not necessarily be required to disapprove the project on that basis. The agency could instead adopt whatever feasible alternatives and mitigation measures exist beyond the efficiency and conservation features already incorporated in the project design and, to the extent those measures do not reduce the cumulative impact of the project below the chosen threshold of significance, DFW could add a discussion of these impacts, and the countervailing benefits of the project, to the statement of overriding considerations the agency previously adopted in approving the project.

B. The EIR’s Mitigation Measures for Protection of Unarmored Threespine Stickleback

Finding that infrastructure construction and building of Newhall Ranch could result in significant impacts to special status wildlife and plant species, DFW adopted numerous biological impact mitigation measures. Mitigation measures BIO-44 and BIO-46 provide for collection and relocation of special status fish, including the unarmored threespine stickleback, during construction in, or diversion of, the Santa Clara River. Such actions would be performed by United States Fish and Wildlife Service personnel or their agents.

*232We agree with plaintiffs that specifying these actions as mitigation in an EIR violates the Fish and Game Code section 5515’s prohibition on authorizing the taking or possession of fully protected fish in mitigation of project impacts under CEQA. DFW may conduct or authorize capture and relocation of the stickleback as a conservation measure to protect the fish and aid in its recovery, but the agency may not rely in a CEQA document on the prospect of capture and relocation as mitigating a project’s adverse impacts.

Fish and Game Code section 5515 lists 10 species of “fully protected” fish, including the unarmored threespine stickleback, Gasterosteus aculeatus wil-liamsoni. (Id., subd. (b)(9).) Subdivision (a) of that statute provides in pertinent part: “(1) Except as provided in Section . . . 2081.7, or 2835, fully protected fish or parts thereof may not be taken or possessed at any time. . . . However, the department may authorize the taking of those species for necessary scientific research, including efforts to recover fully protected, threatened, or endangered species. . . . [¶] (2) As used in this subdivision, ‘scientific research ’ does not include any actions taken as part of specified mitigation for a project, as defined in Section 21065 of the Public Resources Code.” (Fish & G. Code, § 5515, subd. (a), italics added.)8

Fish and Game Code section 86 defines “take” as to “hunt, pursue, catch, capture, or kill, or attempt to hunt, pursue, catch, capture, or kill.” (Italics added.) This definition governs construction of the Fish and Game Code generally unless particular provisions or context require otherwise. (Id., §2.)

In light of the definition of “take” in section 86 as including an animal’s “pursuit],” “catch,” or “capture,” the capture and relocation of stickleback contemplated by mitigation measures BIO-44 and BIO-46 violates Fish and Game Code section 5515. Although trapping and transplantation are defined as possible conservation measures for endangered species under Fish and Game Code section 2061,9 the stickleback, as a fully protected species, is *233subject to the stricter prohibitions against taking set forth in Fish and Game Code section 5515, including an express prohibition on taking as mitigation for a project under CEQA. (§ 5515, subd. (a)(2).)

DFW and Newhall argue the references to “pursue,” “catch” and “capture” in Fish and Game Code section 86 should be understood to exclude trapping and transplantation done for conservation purposes. Because the stickleback is listed as an endangered species (Cal. Code Regs., tit. 14, § 670.5, subd. (a)(2)(L)) as well as a fully protected one, they argue, the prohibition on taking stickleback as a fully protected species must be harmonized with the endangered species act’s permission to trap and transport endangered species for protective purposes. (Fish & G. Code, § 2061.) In the context of the Fish and Game Code’s solicitude for conservation of endangered and threatened species, the prohibition on taking should, DFW maintains, be understood as referring to “[ajctivities . . . that adversely affect fish and wildlife—not . . . activities intended to move fish and wildlife out of harm’s way.”

We must reject the claim DFW may authorize, as CEQA mitigation, actions to protect a fully protected species from harm when, as here, those actions are otherwise prohibited as takings. The Legislature has expressly precluded this interpretation of the statutes by providing, in Fish and Game Code section 5515, subdivision (a), that permitted taking of a fully protected species for “scientific research” may include “efforts to recover” the species but that such “scientific research” does not include “any actions taken as part of specified mitigation for a project” as defined in CEQA. We cannot give effect to this provision and at the same time hold that DFW may, as CEQA mitigation, authorize the trapping and transplantation of stickleback—actions that plainly call for the fish’s “catch,” or “capture” (Fish & G. Code, § 86). That such catch or capture is intended to protect the stickleback from harm caused by the project’s construction is inherent in its adoption as CEQA mitigation and is expressly barred under section 5515.

Legislative history supports our conclusion. The language allowing taking for recovery efforts but not for CEQA mitigation was added to Fish and Game Code section 5515, subdivision (a) in 2003. (Stats. 2003, ch. 735, §4, pp. 5521-5522.) As introduced on February 20, 2003, the bill simply defined “scientific research” to include recovery efforts for fully protected species. (Sen. Bill No. 412 (2003-2004 Reg. Sess.) as introduced Feb. 20, 2013.) An Assembly committee analysis of the bill as introduced, explained that the Natural Resources Agency secretary had testified that the fully protected *234species law’s absolute prohibition on taking had led to certain problems: “1) Fully protected status conflicts with recovery efforts because there is no allowance for management pursuant to a recovery effort. For example, the fully protected species statute is in direct conflict with regional, multi-species conservation planning, such as the Natural Community Conservation Planning Program. [¶] 2) Fully protected status does not allow for incidental take of species due to otherwise lawful activities. [¶] 3) The law does not provide for mitigation of fully protected species. Because mitigation is not an option, the Department’s only recourse is to initiate legal proceedings to address conflicts with fully protected species.” (Assem. Com. on Water, Parks and Wildlife, Analysis of Sen. Bill No. 412 (2003-2004 Reg. Sess.) as introduced Feb. 20, 2003, p. 2.) The analysis continued: ‘“According to the author this measure is intended to address the problem identified by Secretary Nichols in #1 above. In order to ensure broader recovery planning efforts can take place some take may be necessary.” {Ibid.)

The bill was subsequently amended in the Assembly to add the proviso that ‘“scientific research” does not include ‘“any actions taken as part of specified mitigation for a project, as defined in Section 21065 of the Public Resources Code.” (Sen. Bill No. 412 (2003-2004 Reg. Sess.) as amended Aug. 28, 2003.) A new committee analysis noted that the bill now ‘“[ejxcludes, from ‘scientific research,’ any actions taken to mitigate a project under the California Environmental Quality Act (CEQA).” (Assem. Com. on Appropriations, Analysis of Sen. Bill No. 412 (2003-2004 Reg. Sess.) as amended Aug. 28, 2003, pp. 1-2.)

Though not explicitly noted in the legislative history, the August 28, 2003, amendment was consistent with the earlier report’s observation that, of the three problems identified by Secretary Nichols, the bill was intended to address only the first problem: the prohibition on taking members of a fully protected species tended to hinder management programs for the species’ recovery. (Assem. Com. on Water, Parks and Wildlife, Analysis of Sen. Bill No. 412 (2003-2004 Reg. Sess.) as introduced Feb. 20, 2003, p. 2.) It was not aimed at the separate asserted problem of mitigation of the effects other actions would have on a fully protected species. {Ibid.) The August 28 amendment, by reaffirming the taking prohibition as to CEQA mitigation measures, effectuated this distinction in legislative intent.

Consistent with this history and the statutory language, we read Fish and Game Code section 5515, subdivision (a) as allowing the trapping and transplantation of fully protected fish species as part of a species recovery program, but not as mitigation for a project. Mitigating the adverse effect of a land development project on a species is not the same as undertaking positive *235efforts for the species’ recovery, a distinction recognized in the 2003 legislation by its explicit exclusion of CEQA mitigation measures from the definition of “scientific research.” The Legislature evidently believed the prohibition on taking or possessing fully protected species should be relaxed to permit the use of wildlife management techniques needed for species recovery, but that agencies should not be allowed to rely on the availability of such techniques in approving or carrying out projects that would have significant adverse effects on a fully protected species. We therefore say nothing to preclude DFW’s use or authorization of trapping and transplantation to protect the stickleback from threats to its survival and recovery, as expressly allowed under Fish and Game Code section 5515, subdivision (a)(1); based on subdivision (a)(2) of that statute, we hold only that such actions may not be relied on or “specified” as project mitigation measures pursuant to CEQA.

In the context of Fish and Game Code section 5515, limiting the definition of “taking”—which includes but is not limited to hunting and killing animals (Fish and G. Code, § 86)—to actions intended to harm a fully protected animal, as DFW urges, would also render unnecessary, or at least very puzzling, the Legislature’s proviso that taking is not permitted as CEQA project mitigation. (§ 86, subd. (a)(2).) Hunting and killing animals might sometimes be necessary as a conservation measure, for example, to obtain biological samples or to relieve a dangerous local population pressure, but one struggles to imagine the circumstances in which a CEQA document would propose mitigating a project’s adverse impacts on a fully protected species by killing or otherwise intentionally harming members of the species. If Fish and Game Code section 5515, subdivision (a)(l)’s prohibition on “tak[ing] or possess[ing]” a fully protected fish referred only to intentionally harmful acts, the Legislature would not likely have thought it necessary to specify in section 5515, subdivision (a)(2) that such taking or possession could not be proposed as a means of mitigating adverse project effects.

In addition, narrowing Fish and Game Code section 86’s definition of “take” to actions intended to harm an animal could in theory allow unauthorized persons found pursuing and catching a protected species to assert as a complete defense that their intent was not to harm the animal but to restore or transplant it to a safe habitat, a result we doubt very much the Legislature intended. We are loath to adopt a construction that would, for example, sanction an amateur conservationist capturing and moving a southern sea otter (fully protected under Fish & G. Code, § 4700, subd. (b)(8)) from its established habitat to a cove where the person believes it will be safer and healthier. On this point, Justice Chin observes that the Legislature did not intend such a result for endangered species any more than for fully protected ones. (Dis. opn. of Chin, J., post, at p. 253.) We agree: the broad definition of “take” in Fish and Game Code section 86 ensures that DFW can maintain legal control over actions interfering with threatened, endangered *236and fully protected animals even where those actions may not have been intended to kill or hurt the animal.

DFW urges deference to its interpretation of Fish and Game Code provisions, an area in which it has both expertise and substantial administrative responsibility. We consider an agency’s interpretation of statutes and regulations in light of the circumstances, giving greater weight where the interpretation concerns technical and complex matters within the scope of the agency’s expertise. (Yamaha Corp. of America v. State Bd. of Equalization (1998) 19 Cal.4th 1, 12 [78 Cal.Rptr.2d 1, 960 P.2d 1031].) Even in substantive areas of the agency’s expertise, however, our deference to an agency’s statutory interpretation is limited; determining statutes’ meaning and effect is a matter “lying within the constitutional domain of the courts.” (Id. at p. 11.) That said, we acknowledge DFW’s superior expertise in the administration of the Fish and Game Code, and we would not lightly adopt an interpretation of that code’s provisions the department persuasively argued would defeat its ability to pursue species conservation and recovery. Again, however, we do not hold trapping and transplantation of fully protected fish species is prohibited as part of a species recovery effort. We hold only that such actions may not be specified as project mitigation measures in an EIR or other CEQA document. Nothing we say precludes DFW from using its expertise and judgment in determining, at any time, how best to protect a fully protected species from an imminent threat to its habitat.

Justice Chin points out that Fish and Game Code section 2061, relating to endangered species, refers separately to “taking,” “live trapping,” and “transplantation,” implying these actions differ from one another. (Dis. opn. of Chin, J., post, at p. 251.) That this provision uses “taking” in a limited sense denoting mortality or other permanent removal from the ecosystem, a meaning far narrower than the generally applicable definition of Fish and Game Code section 86, does not compel or even suggest the same limited meaning was intended in Fish and Game Code section 5515, relating to fully protected fish species. Indeed, we observe that a closely analogous statute, Fish and Game Code section 3511, while prohibiting the taking or possession of fully protected birds, provides an exception allowing permits for “live capture and relocation” of such birds to protect livestock, suggesting those actions would otherwise be within the statutory prohibition on taking or possession, the same prohibition contained in Fish and Game Code section 5515.

Justice Chin further argues our interpretation of Fish and Game Code section 5515 as distinguishing between capture and transplantation performed for conservation purposes and the same actions specified as CEQA mitigation measures has “little substance.” (Dis. opn. of Chin, J., post, at p. 250.) To the *237contrary, we see a significant distinction between discussing in an EIR measures that might be taken as part of an ongoing species recovery effort and specifying those actions as binding mitigation measures upon which project approval is conditioned. (See Guidelines § 15126.4, subd. (a)(2) [“Mitigation measures must be fully enforceable through permit conditions, agreements, or other legally-binding instruments.”].) Decision makers and the public could well be influenced in their evaluation of a project by the existence or nonexistence of such enforceable mitigation measures.

C. Timeliness of Plaintiffs’ Comments on Cultural Resources and Steelhead Smolt Impacts

The Court of Appeal held two of plaintiffs’ challenges to the EIR, regarding impacts on Native American cultural resources and on steelhead smolt (juveniles), were not preserved because they were not timely brought to DFW’s attention in the administrative process. The issue turns on plaintiffs’ compliance with section 21177, which sets out the requirement that a CEQA claim be administratively exhausted before forming the basis for a judicial challenge to the agency’s actions.

Section 21177, subdivision (a) provides that before an alleged ground for noncompliance with CEQA may be brought to court it must have been “presented to the public agency orally or in writing by any person during the public comment period provided by this division or prior to the close of the public hearing on the project before the issuance of the notice of determination.” DFW held no public hearing on final approval of the present project (the Resource Management and Development Plan and the Spineflower Conservation Plan); the question is therefore whether plaintiffs’ claims regarding Native American cultural resources and steelhead smolt were presented to DFW “during the public comment period provided by [CEQA].” (§ 21177, subd. (a).)

As noted earlier, what we have referred to as the EIR was actually a combined environmental impact statement and environmental impact report (EIS/EIR) prepared jointly under NEPA and CEQA by the Corps and DFW, the lead federal and state agencies, respectively. (Ante, at pp. 213-214.) CEQA requires a public comment period on the draft EIR, but not on the final EIR; a comment period on the final EIR before project approval is optional with the lead agency. (§ 21091, subd. (a); Guidelines, § 15089, subd. (b).) NEPA regulations, in contrast, allow agencies and members of the public to submit comments on a final EIS at any time before the final agency decision, which ordinarily may not be issued earlier than 30 days after notice *238of the final EIS. (40 C.F.R. §§ 1503.1(b), 1506.10(b)(2) (2015).) In compliance with its federal obligations, the Corps published a notice of availability of the final EIS/EIR, inviting public comments during the period June 18, 2010, through July 19, 2010, later extended through August 3, 2010. Plaintiffs raised the disputed issues regarding Native American cultural resources and steelhead smolt impacts in comment letters during this period.

Because plaintiffs’ comments were made during the Corps-noticed comment period for the final EIS/EIR, rather than during the earlier CEQA-mandated period for comments on the draft EIS/EIR, DFW and Newhall contend they came too late to preserve plaintiffs’ claims under section 21177, subdivision (a). Under the circumstances of this case, we disagree.

In the final EIR, DFW stated that while CEQA did not require a comment period on it, DFW would make the final EIR available to the public “at the time the Corps begins its required 30-day public review.” In its findings on project approval, DFW noted that “CEQA allows, but does not require, public review of a Final EIR” and that the Corps’ 45-day comment period (extended from 30 days) is “equivalent” to the 45-day period required by CEQA for draft EIR’s submitted for review by other agencies. The findings further explained that comments on the final EIS/EIR were given to the applicant (Newhall) for preparation of draft responses, that DFW “coordinated with the Corps and the applicant during the initial discussions” regarding these comments, and that “[bjased on the input received from both DFG [now DFW] and the Corps, the applicant and its consultant team completed responses to the comments.” In sum, “DFG has provided input and coordinated with the Corps and the applicant with respect to the draft responses on the Final EIS/EIR.”

On completion of the response and revision process, the lead agencies together prepared an addendum containing portions of the final EIS/EIR that had been modified in response to comments on that document. The agencies included that addendum, together with the final EIS/EIR itself and the comments and responses to comments, in their final decision documents. This addendum adopted a new mitigation measure for Native American cultural resources, and the responses by DFW to plaintiffs’ comments on the final EIR include responses on impacts on steelhead.

We need not decide whether every federally mandated comment period on a final combined EIS/EIR also constitutes a CEQA comment period for purposes of section 21177, subdivision (a). In this case, the lead state *239agency, DFW, participated fully in the post-final EIS/EIR process, helping to prepare responses to the comments received and including those comments and responsive changes in the version of the final EIR it certified as compliant with CEQA when approving the project. Where the lead agency under CEQA has treated a federal comment period on a final EIS/EIR as an opportunity to receive additional comments on CEQA issues as well and has responded to those comments and included the responses in its final decision document, the lead agency has effectively treated the federal period as an optional comment period on the final EIR under Guidelines section 15089, subdivision (b). Such an optional comment period is “provided by” CEQA for purposes of section 21177. (See Environmental Protection Information Center v. California Dept. of Forestry & Fire Protection, supra, 44 Cal.4th at p. 484 [lead federal agency’s notice of availability of final EIS/EIR, which also invited comments to be sent to lead state agency, reopened public comment period for CEQA purposes]; Galante Vineyards v. Monterey Peninsula Water Management Dist. (1997) 60 Cal.App.4th 1109, 1120 [71 Cal.Rptr.2d 1] [“the phrase ‘during the public comment period provided by this division,’ . . . includes optional comment periods . . .”].)

The purposes of requiring exhaustion of administrative remedies, as section 21177 does, are to lighten the judicial burden by providing a remedy at the administrative level and, where a judicial remedy is nonetheless sought, facilitating a complete record that draws on the administrative agency’s expertise and has already been sifted for relevant evidence. (Tomlinson v. County of Alameda (2012) 54 Cal.4th 281, 291 [142 Cal.Rptr.3d 539, 278 P.3d 803].) In this case, where DFW independently reviewed plaintiffs’ comments on the final EIS/EIR, contributed its expertise to the drafting of responses and revisions based on those comments, and included those responses and revisions in the final version of the EIR it certified and relied on in making its approval decision, the statute’s purpose has been served. We conclude the disputed comments were timely under section 21177, subdivision (a) because they were submitted during a public comment period provided by CEQA.

The Court of Appeal, after holding plaintiffs had not administratively exhausted their claims on these topics, went on to reject those claims on the merits, finding the EIR’s determinations to be supported by substantial evidence. DFW and Newhall argue the Court of Appeal’s judgment may be upheld on this alternative ground, whereas plaintiffs insist the merits must be revisited because the Court of Appeal’s disregard for information presented in the comments it deemed untimely tainted its evaluation of the merits. We leave for the appellate court the question of whether its determinations on the merits require reexamination.

*240III. Conclusion

We conclude, contrary to the holdings of the Court of Appeal, that DFW abused its discretion by making the determination, without the support of substantial evidence, that the project’s greenhouse gas emissions would have no significant impact, and in imposing biological resource mitigation measures that call for the trapping and transplantation of a fully protected fish species. We further conclude the Court of Appeal erred in holding plaintiffs failed to preserve their claims regarding Native American cultural resource and steelhead smolt impacts. On remand, the Court of Appeal shall decide whether, in light of our exhaustion holding, the Native American cultural resource and steelhead smolt claims warrant reexamination on the merits. The Court of Appeal shall further decide, or remand for the superior court to decide, the parameters of the writ of mandate to be issued. (See § 21168.9.)

Justice Chin suggests that by reversing and remanding in this case, we inordinately delay the construction of Newhall Ranch and push its thousands of potential residents into housing that “will undoubtedly be far less green than this project promises to be.” (Dis. opn. of Chin, J., post, at p. 254.) It is not the courts’ role, of course, to decide where in the state new housing should be built, and our review of a lower court’s CEQA ruling does not turn on our independent assessment of the project’s environmental merits. Even if Newhall Ranch offered the environmentally best means of housing this part of California’s growing population, CEQA’s requirements for informing the public and decision makers of adverse impacts, and for imposition of valid, feasible mitigation measures, would still need to be enforced.

Nor is Justice Chin’s assumption regarding the project’s superlative environmental profile necessarily supported by the record. As plaintiffs point out, the hypothetical business-as-usual model used in the EIR to assess greenhouse gas emissions counterfactually assumes the continuation of building and vehicle efficiency standards and an electricity generation source mixture that have, in actuality, been superseded by stricter standards and practices. The EIR’s calculation of a 31 percent reduction in comparison to this model therefore does not mean Newhall Ranch would emit 31 percent fewer greenhouse gasses than other mixed-use projects that could actually be built under current standards. Finally, one should not assume a sizeable new housing development, to be built largely on undeveloped land with habitat for several sensitive species, will have comparatively minor impacts either on greenhouse gas emissions or on fish and wildlife. The dissent’s claim that today’s decision threatens the “subversion” of CEQA into a tool for delay of a uniquely meritorious project (dis. opn. of Chin, J., post, at p. 253) is neither warranted by the facts nor consonant with the scope of judicial review under CEQA.

*241IV. Disposition

The judgment of the Court of Appeal is reversed and the matter is remanded to that court for further proceedings consistent with our opinion.