OPINION

BEDSWORTH, J.

This is an appeal from a judgment that denied petitions for a peremptory writ of mandate to compel the County of Orange to set aside various approvals for a development project, along with other relief. The appellants are Endangered Habitats League, Sierra Club, Rural Canyons Conservation Fund, Sea and Sage Audubon, Inc., California Native Plant Society, California Oak Foundation, and Ray Chandos (collectively Endangered Habitats League). Rutter Development Corporation, Inc. (Rutter) is the developer.

Endangered Habitats League argues the project is inconsistent with the county's general plan, and an environmental impact report (EIR) fails to provide sufficient information to make an informed decision on the project. We agree and reverse.

* * *

The project at issue consists of two adjacent but noncontiguous sites in the Santa Ana Mountains, located near the intersection of Santiago Canyon Road and Live Oak Canyon Road. One is named Saddle Creek and the other Saddle Crest. They lie within the geographical area covered by the Foothill/Trabuco Specific Plan (specific plan).

On January 28, 2003, the Orange County Board of Supervisors (Board or Board of Supervisors) approved area plans for the two sites (AP 99-03 for Saddle Creek and AP 99-07 for Saddle Crest). It adopted Resolution 030-030, certifying the EIR as complying with the requirements of the California Environmental Quality Act (CEQA). (Pub. Res. Code, § 21000 et seq.) It also adopted Resolution 03-031, which found the project was consistent with the county's general plan. That resolution also found an amendment to the specific plan complied with CEQA, and approved the amendment (referred to in county terminology as a zone change). The actual amendment of the specific plan was made by Ordinance 03-009, passed at the same time. These actions followed.

Three separate petitions were filed seeking writs of both ordinary and administrative mandamus. (Code Civ. Proc., §§ 1085, 1094.5.) Since they raised similar questions of law and fact, the matters were consolidated. The specifics of the land use plans involved and the EIR will be set out in the course of our discussion, along with the details of Endangered Habitats' arguments. The trial judge denied each of the petitions and entered judgment on all for the county and Rutter.

I

(1) Endangered Habitats argues the project is inconsistent with the county's general plan because it will cause an impermissible increase in traffic on Santiago Canyon Road. We agree.

(2) We review decisions regarding consistency with a general plan under the arbitrary and capricious standard. These are quasi-legislative acts reviewed by ordinary mandamus, and the inquiry is whether the decision is arbitrary, capricious, entirely lacking in evidentiary support, unlawful, or procedurally unfair. ( Corona-Norco Unified School Dist. v. City of Corona (1993) 17 Cal.App.4th 985, 992 [ 21 Cal.Rptr.2d 803] [zone change]; Mitchell v. County of Orange (1985) 165 Cal.App.3d 1185, 1191-1192 [ 211 Cal.Rptr. 563] [specific plan].) Under this standard, we defer to an agency's factual finding of consistency unless no reasonable person could have reached the same conclusion on the evidence before it. ( Corona-Norco Unified School Dist., supra, at p. 992.) (3) All counties and cities must adopt a general plan for the physical development of their land. (Gov. Code, § 65300.) The general plan functions as a "'constitution for all future developments,'" and land use decisions must be consistent with the general plan and its elements. ( Citizens of Goleta Valley v. Board of Supervisors (1990) 52 Cal.3d 553, 570 [ 276 Cal.Rptr. 410, 801 P.2d 1161].) A "'project is consistent with the general plan if, considering all its aspects, it will further the objectives and policies of the general plan and not obstruct their attainment. [Citation.]'" ( Corona-Norco Unified School Dist. v. City of Corona, supra, 17 Cal.App.4th at p. 994.) Perfect conformity is not required, but a project must be compatible with the objectives and policies of the general plan. ( Families Unafraid to Uphold Rural etc. County v. Board of Supervisors, supra, 62 Cal.App.4th at p. 1336.) A project is inconsistent if it conflicts with a general plan policy that is fundamental, mandatory, and clear. ( Id. at pp. 1341-1342.)

The growth management element of the county's general plan establishes policies for traffic improvement and public facilities needed for orderly growth and development. Its declared purpose is, in part, "to mandate that growth and development be based upon the County's ability to provide an adequate circulation system. . . ." A "traffic level of service policy" addresses the need for highway improvements when development increases traffic. County policy is that improvements must be made within a stated time after issuance of various permits so as to achieve level of service (LOS) D at intersections, and LOS C on Santiago Canyon Road. Here is the relevant language: "LOS C shall . . . be maintained on Santiago Canyon Road links until such time as uninterrupted segments of roadway (i.e., no major intersections) are reduced to less than three miles." This policy requires compliance to be evaluated according to the county's Transportation Implementation Manual, which in turn provides that traffic analysis on Santiago Canyon Road "shall" use the methods described in the Highway Capacity Manual (HCM).

At issue is the approval of the area plans for the two sites that compose the project. The EIR reveals that under the HCM method, the project would cause the LOS on Santiago Canyon Road to fall to D/E in 2005 and to E by 2020. Under a different analysis, called the V/C method (for volume/capacity ratio), the results are acceptable — LOS B in both 2005 and 2020. The EIR concludes "the V/C methodology is considered more representative of actual conditions," and, relying on it, finds no significant impacts on Santiago Canyon Road.

The Board of Supervisors acknowledged the problem under the HCM method and likewise solved it by relying on the V/C analysis. In findings of fact in support of the EIR, the Board determined project impacts on transportation and circulation would be reduced to insignificance after mitigation. But the explanation for this finding reveals the Board's action was based upon uncritical adoption of the EIR's use of the V/C method: "Both short-range and long-range project and cumulative impacts to Santiago Canyon Road would be less than significant under the updated V/C analysis method. Existing and future LOS for this arterial exceed LOS 'C' and are therefore unacceptable . . . under the HCM methodology. The V/C methodology is considered more representative of actual conditions. . . . Project impacts to Santiago Canyon Road, therefore, are considered less than significant."

It is clear the project is inconsistent with the general plan's traffic service level policy. The general plan requires LOS C as determined under the HCM method, and the project does not comply. That it does so under the V/C method is of no import, since the general plan is unambiguous in demanding the evaluation be made by the HCM method. The approval of the area plans, AP 99-03 (Saddle Creek) and AP 99-07 (Saddle Crest), must be set aside.

Rutter's several arguments for consistency are all based on misreadings — and misrepresentations — of the record. It contends the EIR establishes the Santiago Canyon Road "intersection" will operate at an acceptable level of service. That is misleading. The issue is the level of service on Santiago Canyon Road, not one intersection along it. So this does not show consistency.

Rutter argues the HCM method was used because the EIR considers it along with the V/C method. This is semantic sleight of hand. Both the EIR and the Board of Supervisors relied on the V/C method to find traffic levels acceptable, after acknowledging they were not under the HCM method. You do not "use" something by disregarding it.

Equally disingenuous is an argument that the V/C method was not used because it is not mentioned in sections of the EIR setting out thresholds of significance and mitigation measures. But neither has anything to do with measuring levels of service before or after the project. Any suggestion that the V/C method was not used in the EIR is frivolous.

A bolder argument seems to say the Board of Supervisors did not rely on the V/C method, and all that happened was "clumsy wording of the Board's findings that mention 'the V/C method.'" That flat out misrepresents the record. Here is what the Board said explaining why there was no significant impact on Santiago Canyon Road: "The V/C methodology is considered more representative of actual conditions. . . . Project impacts to Santiago Canyon Road, therefore, are considered less than significant." So it is crystal clear that the Board of Supervisors relied on the V/C method, and the V/C method alone, in approving the project.

Finally, Rutter asserts that required mitigation measures will bring Santiago Canyon Road up to HCM service level C. We cannot agree.

The mitigation measures in question are contributions to two road improvement programs for Santiago Canyon Road. One is the Santiago Canyon Road Major Thoroughfare and Bridge Fee Program. The other is a supplemental program to fund improvements to Santiago Canyon Road. The EIR states that "upon implementation" of these mitigation measures (along with others not relevant here), "impacts to transportation and circulation will be less than significant." The findings of fact adopted by the Board of Supervisors in approving the EIR essentially state the same thing.

The problem we have is that the EIR does not state mitigated impacts will be less than significant under the required HCM method. Since the EIR uses the V/C method to analyze traffic impacts initially, we cannot reasonably assume the statement about mitigated impacts now referrs to the HCM method. (4) Moreover, the record citations offered by Rutter do not back up the claim that contributions to the fee programs will result in LOS C on Santiago Canyon Road. They show only the existence of "a fee program . . . to fund future improvements to Santiago Canyon Road," along with a study to identify the needed improvements. This falls short of a mitigation measure that requires improving the road to HCM level C. A fee-based mitigation program is sufficient under CEQA if there is evidence that mitigation will actually occur ( Save Our Peninsula Committee v. Monterey County Bd. of Supervisors (2001) 87 Cal.App.4th 99, 140-141 [ 104 Cal.Rptr.2d 326] [planned improvements included passing lane on one road segment and description of 12 proposed improvements for another stretch of highway].) But even where a developer's contribution to roadway improvements is reasonable, a fee program is insufficient mitigation where, even with that contribution, a county will not have sufficient funds to mitigate effects on traffic. ( Napa Citizens for Honest Government v. Napa County Bd. of Supervisors (2001) 91 Cal.App.4th 342, 364 [ 110 Cal.Rptr.2d 579].) Since there is no evidence here of what improvements will be funded by the fee programs, nor that it will achieve HCM service level C, we cannot find the mitigated project is consistent with the general plan. The inescapable conclusion is the project conflicts with the general plan because of the impact it will have on traffic along Santiago Canyon Road.

II

Endangered Habitats contends the specific plan amendment is inconsistent with the general plan in two other ways. First, it changes the rules to allow balancing of specific plan requirements, rather than compliance with all of them. Second, it exempts the project from otherwise mandatory specific plan requirements. We agree these are also fatal inconsistencies.

The growth management element of the general plan is again at issue. A policy entitled "transition areas for rural communities" provides, in relevant part: "New development within the . . . Foothill-Trabuco Specific Plan planning area shall be rural in character and shall comply with the policies of [that] plan in order to maintain a buffer between urban development and the Cleveland National Forest."

The specific plan does not identify any "policies." It does, however, distinguish between mandatory and permissive provisions. A "consistency checklist" in the appendix explains that "'shall' indicates a mandatory [r]egulation to which there are no exceptions, while 'should' indicates a non-mandatory [g]uideline." The checklist states projects must be consistent with all applicable regulations, but the planning commission may approve deviations from the guidelines if it finds "the project is in overall compliance with the Guidelines and consistent with the Goals and Objectives of the Specific Plan."

The specific plan amendment begins with the balancing provision: "In analyzing and considering applications for development and use within the Specific Plan area, the Director of Planning and Development Services, Planning Commission and Board of Supervisors shall not give precedence to one provision of the Specific Plan over another but shall balance consideration of Specific Plan development goals and policies so as to further overall Specific Plan goals and policies while not hindering their attainment."

The amendment goes on to enact new regulations on tree preservation, grading, and open space that apply only to this project. For example, in the case of oak woodlands, the amendment provides: "For the Saddle Creek and Saddle Crest Area Plan areas . . . the following shall apply." The new regulations follow. Then this: "Consistency with [the new regulations] shall exempt the Saddle Creek and Saddle Crest Area Plan areas from [the existing oak woodland regulations] herein and shall be regarded as satisfactorily preserving and protecting the oak woodlands on the subject properties."

The original tree preservation regulations required that oaks over 5 inches in trunk diameter, and sycamores over 35 inches, be transplanted unless the tree was "in poor health and would not survive transplantation." In that event, the tree was to be replaced according to a table that specified the size and number of replacement trees required. Under the amendment, oaks and sycamores may be replaced if they are "in poor health or would not likely survive transportation, or [were] determined to be infeasible for transportation as certified by an arborist. . . ." (Italics added.)

Grading standards in the specific plan set out the maximum volume of earth that may be moved and maximum permitted changes to slope and elevation of sites. They state each project "shall preserve a minimum of sixty-six (66) percent of the site in permanent, natural open space. . . ." Alternative grading standards permit approval of increased grading, slope, and elevation changes, but there is a trade-off: "The Alternative Grading Standards shall result in seventy (70) percent or more of the project site being preserved in natural open space." (Italics added.)

The amendment relaxes the open space prerequisite for use of the alternative grading standards and, under the relaxed standard, it approves alternative grading. The new standard is: "The Saddle Creek and Saddle Crest Grading standard shall result in seventy (70) percent or more of each [site] . . . being preserved in permanent open space." (Italics added.) According to the EIR, uses allowed in "permanent open space" include "natural undisturbed open space," "river rock walls," unpaved "emergency utility easements," "remedial grading," "fuel modification," and tree transplanting.

A

Endangered Habitats argues the amendment allows specific plan regulations to be balanced, and this conflicts with the general plan policy that new development must comply with all specific plan policies. It is right.

Preliminarily, we note the parties assume the specific plan policies are reflected in its mandatory provisions, the regulations. Without discussing the absence of any stated policies in the specific plan, they treat the general plan's direction regarding specific plan policies as applying to the latter's regulations. For purposes of this case, and this case only, we will assume that is correct.

The county's consistency finding recites the specific plan amendment (referred to as a zone change) is consistent with the general plan without explanation. The evidence before the Board when it made the finding was a staff report on the amendment and the testimony of Brian Speegle, assistant director of planning, given at the public hearing held to consider the approvals and resolutions for the project.

The staff report says the amendment allows for balancing of specific plan goals and objectives, since they cannot all be satisfied in every project. "In staff's view, this has been the process that is typically used at both the general and the specific levels of planning. . . . [T]he language proposed for addition . . . makes explicit what has been the practice. It would strengthen the F/TSP, although it does not represent a major change in approach to analyzing projects."

At the hearing, Speegle went further. Referring to the balancing provision, he said "there's a general amendment . . . that would be applied throughout the specific plan area, and it explicitly permits a balancing of the competing policies or requirements of the specific plan so that inconsistency with one policy does not automatically disqualify a project from consideration." (Italics added.) A little later, he put it this way: "[T]he specific plan already explicitly provides for some balancing of development and design guidelines, for example . . . the planning commission [can] find a project is in overall compliance with the guidelines without the project being consistent with each and every guideline. [¶] The specific plan amendment would extend this balancing to all specific plan policies and regulations." (Italics added.)

We cannot uphold the consistency finding, since it is not one a reasonable person could reach on this evidence. The staff report said the amendment applies only to balancing goals and policies, already the practice, yet it acknowledged the amendment would change the approval practice (the amendment does not represent "a major change" in approach). And, of course, if the amendment merely confirmed the extant practice, why was it needed? Speegle was much more direct when he said the amendment allows specific plan "requirements" and "regulations" to be balanced. As assistant planning director, presumably he was as well versed as anyone on the amendment, and no one got up to correct or challenge him. It may be that no balancing of regulations was intended, and the amendment might be clarified on that point. But as it stands, the balancing provision is inconsistent with the general plan.

Rutter contends Speegle's interpretation of the balancing amendment was an unfortunate and inadvertent misstatement, and every other discussion of the issue in the record states only goals and policies are to be balanced. We are not persuaded. Speegle never said he spoke carelessly or corrected his statements. As for the record, the developer points to the staff report, several places where the amendment is quoted or paraphrased, and an after-the-fact November 18, 2003 letter from its lawyer to the Board of Supervisors regarding a then-upcoming hearing on tract maps for the two sites. Of these, only the staff report is relevant, and it does not support the consistency finding. As we have said, that finding cannot be sustained.

B

The amendment gives the developer an unacceptable freebie. In effect, it exempts this project from specific plan regulations on tree preservation, grading, and open space, and substitutes new regulations that are less stringent. This, of course, directly contradicts the general plan policy that all new development must comply with all specific plan policies. Consistency requires more than incantation, and a county cannot articulate a policy in its general plan and then approve a conflicting project. ( Napa Citizens for Honest Government v. Napa County Bd. of Supervisors, supra, 91 Cal.App.4th at pp. 379-380.) Since no reasonable person could have made the consistency finding on the record before us, it must be set aside as arbitrary and capricious. ( Families Unafraid to Uphold Rural etc. County v. Board of Supervisors, supra, 62 Cal.App.4th at p. 1338.)

Rutter contends the consistency finding must be upheld because it is entitled to great weight and deference. But the developer makes no attempt to grapple with the evidence or provide us with a citation to something — anything — that reasonably supports the finding. Since no evidence in support of the finding has been identified, we cannot defer to it.

Rutter argues there is no inconsistency because the specific plan amendment need not comply with all specific plan policies, or all specific plan guidelines. The first point contends the general plan never says "all" specific plan policies must be met, so a conflict with some is acceptable. We cannot agree. The general plan states specific plan "policies" must be met, and that means all of them. The second argument is that the specific plan consistency checklist says a project need not be consistent with each and every guideline. That is true. The guidelines are permissive. But the checklist is unequivocal that projects must be consistent with all applicable regulations, which are mandatory. In the absence of that kind of consistency, these textual arguments are to no avail.

The developer relies on Corona-Norco Unified School Dist. v. City of Corona, supra, 17 Cal.App.4th at page 994, for the proposition that a specific plan need not comply with all general plan goals and policies, provided it furthers them and does not obstruct their attainment. That is an accurate statement of the law, but Rutter makes no attempt to show how it is met here. In any event, the case is distinguishable. In Corona-Norco, a school district challenged approval of a city zoning change that permitted new residential development, arguing the failure to include conditions to assure adequate schools was inconsistent with the general plan. The court held approval of the zoning amendment was not arbitrary or capricious. It explained the general plan did not contain any specific, mandatory requirement that projects guarantee adequate school facilities, and there was no evidence existing schools would not suffice. ( Id. at pp. 996-997.) Here, the general plan does articulate a specific and mandatory policy, and it has not been met, so Corona-Norco is of no avail.

Equally misplaced is citation of Napa Citizens for Honest Government v. Napa County Bd. of Supervisors, supra, 91 Cal.App.4th 342, for the proposition that specific plan amendments to benefit one project are well supported by case law. But that decision did not involve an amendment to allow one project to go forward, but rather an update of an entire specific plan for an airport industrial area. ( Id. at pp. 353-354.) Amending a specific plan for the entire area it covers is different from the present project-specific amendment.

Finally, Rutter argues there is no inconsistency because the amended regulations really do not make any important change to the specific plan. It insists the amended tree regulations promote an environmentally superior project, grading satisfies the alternative grading standards, and the amended open space requirement is essentially the same as the original. We cannot agree.

As to the amended tree regulations, whether this project is environmentally superior — in the developer's view — has no bearing on the issue at hand, which is consistency with the general plan. If the project does not comply with the general plan, neither we nor the developer can justify its approval on the basis that it appeals to us.

The project qualifies for alternative grading only because the threshold for it was changed. Under the changed standard, alternative grading standards could be invoked upon a showing of sufficient "permanent open space," rather than "natural open space." Rutter argues the two are the same, but they are not. There is a vast difference between "permanent open space" approved in the amended regulation and "natural open space" required by the original. It is true, as Rutter points out, that "natural" is not defined in the specific plan. But it has a recognized meaning — "produced or existing in nature; not artificial or manufactured" (Webster's New World Dict. (4th college ed. 2002) p. 959), or "occurring in conformity with the ordinary course of nature." (Webster's 3d New Internat. Dict. (1981) p. 1506.) Many of the uses permitted in permanent open space are artificial and do not occur in the ordinary course of nature, such as rock walls, easements, remedial grading, fuel modification, and tree planting. It is obvious that alternative grading would not be permitted under the specific plan, because the "permanent" open space allowed by the amendment does not qualify as "natural" open space required under the specific plan. Since the amendment materially loosened the three regulations at issue, it is inconsistent with the general plan.

C

Endangered Habitats contends an element of a plan may be changed for the entire geographical area, but a change in generally applicable requirements that affects only part of the area is impermissible. Conceding it cannot find any authority on point, Endangered Habitats asserts a specific plan would be meaningless if it could be changed for each project that comes along. Rutter responds the argument was not raised below, it fails for lack of authority and, in any event, the plain meaning of "in part" includes a site-specific amendment.

We decline to consider the argument because the issue was not raised below at the administrative level. (See A Local Regional Monitor v. City of Los Angeles (1993) 16 Cal.App.4th 630, 648 [ 20 Cal.Rptr.2d 228] [argument that EIR inconsistent with general plan waived because not raised below].) Endangered Habitats' contention it was brought up is not supported by the record. While numerous speakers at hearings before the Board of Supervisors and the Planning Commission asked them not to relax the specific plan, and some explained the amendment would render the specific plan internally inconsistent, no one broached the present statutory argument. On this record, the issue was waived.

III

Endangered Habitats next argues the EIR is insufficient because it uses an incorrect legal standard, improperly defers mitigation, fails to consider a significant impact, and fails to consider the specific plan amendment balancing provision. We conclude the legal standard and deferral of mitigation are improper, so the EIR is inadequate and must be set aside. (6) We review agency actions under CEQA to determine if there was a prejudicial abuse of discretion, which exists if the agency has not proceeded in a manner required by law or if its decision is not supported by substantial evidence. It is not our role to decide the correctness of an EIR's environmental conclusions, but only its sufficiency as an informational document. ( Citizens of Goleta Valley v. Board of Supervisors, supra, 52 Cal.3d at p. 564.) "The court must uphold an EIR if there is any substantial evidence in the record to support the agency's decision that the EIR is adequate and complies with CEQA. [Citation.] [¶] CEQA requires an EIR to reflect a good faith effort at full disclosure; it does not mandate perfection, nor does it require an analysis to be exhaustive. [Citation.]" ( Dry Creek Citizens Coalition v. County of Tulare (1999) 70 Cal.App.4th 20, 26 [ 82 Cal.Rptr.2d 398].)

A

The legal standard argument is that the EIR uses the wrong test for the threshold of significance of impacts on biological resources. It seems to.

(7) An EIR must identify and discuss "all significant effects on the environment" of a proposed project (Pub. Resources Code, § 21100, subd. (b)(1); CEQA Guidelines (Guidelines) § 15126 (a).) A significant environmental effect is "a substantial, or potentially substantial, adverse change in the environment." (Pub. Resources Code, § 21068.) A project has a significant effect on the environment if, among other things, it substantially reduces the habitat of a fish or wildlife species, causes a fish or wildlife population to drop below self-sustaining levels, threatens to eliminate a plant or animal community, or reduces the number or restricts the range of an endangered, rare, or threatened species. (Guidelines, § 15065 (a).)

The instant EIR sets out "thresholds of significance" used to determine whether the project caused significant environmental impacts on biological resources. Its test is that a significant impact would be identified there is a "substantial effect" on enumerated biological resources. This is qualified by the following definition: "Substantial effect means significant loss or harm of a magnitude which . . . 1) would cause species or a native plant [or] animal community to drop below self-perpetuating levels on a statewide or regional basis; or, 2) would cause a species to become threatened or endangered."

(8) The standard used in the EIR is therefore impermissibly lenient. The definition of substantial effect effectively limits significant environmental impact to reducing plant or animal communities below statewide or regional self-perpetuating levels, or making a species threatened or endangered. The proper standard, set out, ante, is considerably broader. The use of an erroneous legal standard is a failure to proceed in the manner required by law that requires reversal. ( No Oil, Inc. v. City of Los Angeles (1974) 13 Cal.3d 68, 88 [ 118 Cal.Rptr. 34, 529 P.2d 66].)

Rutter argues that focusing on the definition of substantial effect is misleading, because the listed criteria are broader and state a proper legal standard. It is wrong. The EIR states an impact would be significant only if it results in a "substantial effect" on the listed criteria. The definition is a limiting factor, and it amounts to an improper legal standard for identifying significant environmental impacts.

B

(9) On deferred mitigation, Endangered Habitats contends the EIR improperly defers analysis and mitigation of eleven project impacts. It is correct as to one.

"Deferral of the specifics of mitigation is permissible where the local entity commits itself to mitigation and lists the alternatives to be considered, analyzed and possibly incorporated in the mitigation plan. [Citation.] On the other hand, an agency goes too far when it simply requires a project applicant to obtain a biological report and then comply with any recommendations that may be made in the report. [Citation.]" ( Defend the Bay v. City of Irvine, supra, 119 Cal.App.4th at p. 1275.) If mitigation is feasible but impractical at the time of a general plan or zoning amendment, it is sufficient to articulate specific performance criteria and make further approvals contingent on finding a way to meet them. ( Id. at pp. 1275-1276.)

The only problem we see is mitigation of construction interference from noise, supply depots, and vehicle staging areas. The EIR provides that before a grading permit is issued, the developer must submit an acoustical analysis describing the "exterior noise environment" and "preliminary mitigation measures, if required." Before a building permit may be issued, another acoustical report must be submitted to demonstrate structures have been designed to meet "exterior and interior noise standards" satisfactory to the manager of the county's building permit division. That individual must also be satisfied the developer will place supply stockpiles and vehicle staging areas "as far [away] as practicable." This is inadequate. No criteria or alternatives to be considered are set out. Rather, this mitigation measure does no more than require a report be prepared and followed, or allow approval by a county department without setting any standards.

The remaining mitigation measures challenged by Endangered Habitats are sufficient, since they commit to mitigation and set out standards for a plan to follow. We consider each in turn.

Dirt hauling. During construction, there will be traffic disruption when dirt is hauled away from the sites. The EIR adequately says a construction vehicle plan must be developed and approved prior to issuing grading permits. It provides the plan must assure public safety, restrict the number of daily trips and limit or avoid them during peak hours, set up clearly marked no passing zones, and use "flaggers" at site entrances.

Drainage. Endangered Habitats contends mitigation of impacts on the drainage system is deferred because a study to determine the project's effect on existing drainage facilities is postponed. But the EIR states that impacts on hydrology and drainage are less than significant before mitigation, as well as after it, so we cannot see how waiting for this study makes any difference.

Fuel modification plans. Prior to the issuance of a grading permit, a fuel modification plan must be prepared. The plan has to comply with Orange County Fire Authority guidelines for such plans, and it must be approved by the Orange County Fire Authority. This is not improper deferral since, once again, there is a commitment to mitigate and adequate criteria to determine if the plan to be submitted is adequate.

Gnatcatcher habitat. The EIR states there will be impacts on the gnatcatcher from construction activities, apparently referring to the removal of its habitat in coastal sage scrub and scrub-chaparral ecozone. According to Endangered Habitats, mitigation is deferred because these impacts will be addressed sometime in the future. But the EIR sets out the possibilities — on-site or off-site preservation of similar habitat at a ratio of at least two to one, or one of several possible habitat loss permits from relevant agencies. We believe this enumeration of alternative mitigation measures saves the provision from improper deferral. Replanting of trees. One mitigation measure for the loss of trees is replanting new ones off site. The EIR allows the developer to seek approval for off-site replanting in the future, and it abandons this mitigation measure if approval cannot be obtained. But the EIR also says its finding of adequate mitigation does not rely on this mitigation measure, so there is no improper deferral here.

Tree restoration. Another mitigation measure for tree loss requires a tree restoration, maintenance, and monitoring plan to be prepared and approved prior to issuing grading permits. It provides the plan must "detail" long-term maintenance and monitoring, include requirements for replanting procedures, and include a contract with a certified arborist for at least 10 years. The arborist must make reports throughout the year and must be given decision-making power over tree care and maintenance. We find these standards sufficient.

Water quality issues. The problem is contaminated runoff from the project. A regional water quality control plan limits the pollutants that may be discharged into local waters. It also requires the county to develop its own plan to reduce pollutants from new development. At the time the EIR was prepared, the county's plan was being undated.

The EIR requires the developer to prepare a project water quality plan to reduce discharge into storm water runoff. It must incorporate "best management practices," which are a series of four traps and filters to remove various pollutants (one for trash, a second for biological matter, a third for nutrients and microbial contaminants, and a fourth to deal with fossil contaminants such as oil, grease, and hydrocarbons).

Endangered Habitats argues mitigation of runoff problems is improperly deferred based on four statements in the EIR. These concern water quality in general, fuel modification zones, Aliso Creek, and covenants, conditions, and restrictions (CCR's). The first statement appears in a prefatory section entitled "proposed water quality improvements." It says "[t]o the extent feasible and in accordance with County codes policies and practices, project design features shall be implemented to reduce runoff from the site." The other statements appear in the EIR's ensuing discussion of the project design features and other conditions imposed to reduce runoff. As to fuel modification zones and Aliso Creek, the EIR says the best management practices and irrigation practices "will be implemented" and incorporated in the final water quality management program to reduce runoff. The EIR also requires the developer to prepare CCR's that impose water quality restrictions on homeowner. Endangered Habitats does not find fault with the content of the CCR's, but with the fact that it does not have to be prepared until some unspecified time in the future.

There is no improper deferral on these issues. The general statement regarding proposed water quality improvements just describes what is to follow, and it is the later enumerated design features that count. The design features for fuel modification and Aliso Creek are adequate, since they require use of clearly identified standards in the form of the "best management practices." Nor is deferring preparation of CCR's a problem where the content is laid out and concededly adequate. So while final mitigation measures have not been adopted in these areas, deferral is permissible because the EIR commits to it and lists standards to be incorporated in the mitigation plan. The net result is the EIR improperly defers mitigation in only one area — construction noise and related impacts.

C

Endangered Habitats also faults the EIR for not identifying the traffic increase on Santiago Canyon Road as a significant environmental effect, and either adopting mitigation measures or finding mitigation is infeasible. We agree.

As we have explained, the EIR's conclusion of no significant impact on Santiago Canyon Road cannot be sustained, since it sidesteps the HCM method of analysis required by the general plan. Nor can we sustain the conclusion that contributions to road improvement fee programs are adequate mitigation, for there is no evidence the mitigation will take place. So the EIR is inadequate for failing to identify and discuss the impact of increased traffic on Santiago Canyon Road.

D

Finally, Endangered Habitats argues the EIR should have analyzed the impact of the balancing provision. The argument goes like this: Because the provision allows trading-off regulations, it will have an impact on the environment as some regulations are forfeited in favor of others.

We need not reach this point. It is moot in light of our holding that the balancing provision is inconsistent with the general plan, which will require us to direct issuance of the writ of mandate to set it aside.

In fine, approval of this project was flawed. Since the project and the specific plan amendment are both inconsistent with the general plan, and the EIR is inadequate, the petitions for a writ of mandate should have been granted.

The judgment appealed from is reversed. The matter is remanded to the trial court with directions to enter a writ of mandate to compel the Orange County Board of Supervisors to vacate approval of area plans AP 99-03 and AP 99-07, Resolution 03-030 certifying the EIR, Resolution 03-031 approving the specific plan amendment, and Ordinance 03-009 enacting the specific plan amendment. In addition, the trial court shall consider the additional requests for relief sought in the petitions in light of our holding that the writs must be granted. Appellants are entitled to costs on appeal.

Rylaarsdam, Acting P.J., and Moore, J., concurred.

Petitions for a rehearing were denied July 29, 2005, and the petition of real party in interest for review by the Supreme Court was denied October 12, 2005.