I concur in the majority’s conclusion that the final environmental impact report (FEIR) for the Sunrise Douglas project adequately assessed the near-term environmental impacts of supplying water to the proposed development. This conclusion rests in large part on the majority’s finding of a reasonable likelihood that groundwater from the North Vineyard Well Field (Well Field) would be available to supply the project’s near-term needs. I agree in particular that substantial evidence supports the FEIR’s reliance on the Well Field even though Well Field water had not been reserved “ ‘for any specific user’ ” and would be made available “ ‘on a “first-come, first served” basis’ ” (maj. opn., ante, at p. 436), even though existing demand and new demand in the region “might also be satisfied from the Well Field” (ibid..), even though serving that demand and the initial phase of the Sunrise Douglas project “would require much more water than . . . [the Well Field] can safely provide” {ibid.), and even though “much uncertainty remains” as to the Well Field’s ability to supply water to the project in the near term (id. at p. 437). As the majority explains, nothing in the administrative record demonstrates “that these competing demands can be satisfied only from the Well Field or that they will all materialize in full in the near term and have priority over the Sunrise Douglas project.” (Id. at p. 436.) Indeed, as the majority subsequently explains, there is more than enough water that can be drawn from the Well Field to satisfy this project’s near-term demand even after one subtracts the expected demand for “already entitled development.” (Id. at p. 437, italics added.)
Like the majority, I further agree that the FEIR need not provide “firm assurances” of long-term water supplies at the early stages of the land use planning and approval process, inasmuch as the “ultimate question” under the California Environmental Quality Act (Pub. Resources Code, § 21000 et seq.; CEQA) “is not whether an EIR establishes a likely source of water, but whether it adequately addresses the reasonably foreseeable impacts of supplying water to the project.” (Maj. opn., ante, at p. 434.) The requisite level of specificity in identifying water supplies thus increases “ ‘at each step as land *451use planning and water supply planning move forward from general phases to more specific phases.’ ” (Id. at p. 434.) For example, because the SunRidge Specific Plan is further along the planning process than is the Sunrise Douglas Community Plan (id. at pp. 437-438), CEQA imposes a greater level of specificity in identifying water supplies for the specific plan than it does for file community plan. What is sufficiently specific for the specific plan in the near term should therefore prove more than sufficient for the community plan in the long term, inasmuch as “CEQA should not be understood to require assurances of certainty regarding long-term future water supplies at an early phase of planning for large land development projects.” (Id. at p. 432.)
The surprising thing, though, is that the majority has adopted precisely the opposite rule in analyzing the sufficiency of the FEIR for this project in the long term. The FEIR estimates the average water demand of the entire Sunrise Douglas Community Plan at full build-out will be 22,103 acre-feet annually (afa). The sources identified in the record to meet this demand are more than ample: at least 5,500 afa from the Well Field, with a possibility of up to 10,000 afa; 15,000 afa of American River water under the Sacramento County Water Agency’s existing contract with the federal Bureau of Reclamation (an allocation known as Fazio water); 15,000 afa of American River water under the water agency’s agreement in principle with the Sacramento Municipal Utility District (SMUD); an additional 15,000 afa as to which the water agency and SMUD are in negotiations; and 33,000 afa of intermittent water consisting of excess flows on the American and Sacramento Rivers for which the water agency is applying. In other words, the FEIR has identified sufficient water for this project three or four times over.
Why the majority nonetheless holds that the FEIR has insufficiently identified long-term water supplies for Sunrise Douglas—and, in doing so, reverses both the trial court and the Court of Appeal—is thus difficult to comprehend. There does not appear to be a problem with the likelihood that the identified water supplies will come to fruition. Although these supplies “are not certain to materialize,” the majority correctly points out that “CEQA does not require this level of certainty at planning stages prior to approval of permits, subdivision maps or other development entitlements.” (Maj. opn., ante, at p. 438.) There also does not appear to be a problem with the analysis of the reasonably foreseeable impacts of supplying water to the project in the long term, inasmuch as the FEIR for the Water Forum proposal “extensively analyzed the environmental impacts of the participants’ planned increases in surface water diversion”—indeed, a summary of these impacts and the proposed mitigation measures occupies 85 pages of that FEIR—and the FEIR for this project analyzed “[t]he impacts of groundwater withdrawals at the Well Field.” (Maj. opn., ante, at pp. 423, 424.)
*452The majority’s rejection of the Sunrise Douglas FEIR rests instead on the FEIR’s failure to balance total long-term water supply and demand in the entirety of the Sacramento County Water Agency’s zone 40, an area comprising the southern and eastern regions of the county that is almost 10 times as large as the Sunrise Douglas project. The majority simply asserts, without explanation, that while substantial evidence “support[s] the conclusion that some part of the planned new surface water supplies will be developed and made available to the Water Agency for use in its zone 40” (maj. opn., ante, at p. 438), there is “too great a degree of uncertainty regarding the long-term availability of water for this project.” (Id. at p. 439, italics added.) The distinction is an elusive one. The Fazio water for the long term, like the Well Field water in the short term, will be made available to users on a first-come, first-served basis, and, as with the Well Field water, there is no indication in the record that capacity for these long-term supplies has been “ ‘reserved . . . for any specific user,’ ” that these other “competing demands” can be satisfied only from the identified supplies, or that the potential demand from other sources will all “materialize in full” in the relevant period and “have priority over the Sunrise Douglas project.” (Id. at p. 436.) The only significant distinction I can see is that, in contrast to its discussion of the Well Field water, the majority does not identify any portion of the project’s long-term supplies that has been “already allocated to other entitled uses.” (Id. at p. 437.) But that distinction, of course, would favor the FEIR’s analysis of the project’s long-term supplies. Thus, if the majority’s analysis of the two situations had been consistent, the majority should have found substantial evidence that these long-term supplies will be available at least in substantial part to supply the Sunrise Douglas project. The majority finds otherwise only by assuming that other users will have priority on all of the identified supplies—or, to put it another way, by speculating that there is evidence outside the record that would rebut the county board of supervisors’ finding, sustained by both the trial court and the Court of Appeal below, that the supplies will be adequate. (See maj. opn., ante, at pp. 445-446, fn. 15.)
The path the majority pursues to reverse the lower court judgments is a curious one. What dooms the FEIR here, according to the majority, is the potential for increased long-term demand from other, purely hypothetical projects that could be developed under the 1993 general plan for the zone 40 area—even if, so far as the record discloses, those projects have not yet been entitled, approved, or even proposed. In other words, Sunrise Douglas must be held hostage to a balancing of supply and demand for all conceivable development that is not prohibited by the county’s general plan—even if no one has yet stepped forward to propose such development.
*453Until today, this was not the law in California.1 The majority can find no support for its new rule in the statute for, as the majority concedes (maj. opn., ante, at p. 428), neither CEQA itself nor this court’s decisions have ever before required a project environmental impact report (EIR) not only to demonstrate a reasonable likelihood that there is water for the project at issue but also that there is water for all hypothetical future projects nearby, including those no entity has yet planned to build. Thus, as the majority elsewhere observes, “Decision makers must, under the law, be presented with sufficient facts to ‘evaluate the pros and cons of supplying the amount of water that the [project] will need.’ ” (Maj. opn., ante, at p. 431, quoting Santiago County Water Dist. v. County of Orange (1981) 118 Cal.App.3d 818, 829 [173 Cal.Rptr. 602].) An EIR “must analyze, to the extent reasonably possible, the impacts of providing water to the entire proposed project.” (Maj. opn., ante, at p. 431.) An EIR, in particular, need not analyze a “ ‘worst case scenario’ ” and “need not identify and analyze all possible resources that might serve the Project should the anticipated resources fail to materialize.” (Napa Citizens for Honest Government v. Napa County Bd. of Supervisors (2001) 91 Cal.App.4th 342, 373 [110 Cal.Rptr.2d 579].) None of these cases requires an EIR to identify a water supply sufficient to meet the demands of all development envisioned by the project, together with all hypothetical future development that might look to the same supplies.
The majority suggests that a balancing of total supply and demand in the zone 40 region is required by the CEQA Guidelines (Cal. Code Regs., tit. 14, § 15000 et seq.) in order to evaluate the long-term cumulative impact of development on water supply. (Maj. opn., ante, at p. 441, citing CEQA Guidelines, Cal. Code Regs., tit. 14, § 15130, subd. (b)(1)(B).) But a “cumulative impact” consists of “the change in the environment which results from the incremental impact of the project when added to other closely related past, present, and reasonably foreseeable probable future projects” (CEQA Guidelines, tit. 14, § 15355, subd. (b), italics added), not (as the majority apparently assumes) all possible future projects. Under the majority’s newly minted rule, no project could ever be approved in the zone 40 area until the entire region’s projected long-term water supply and demand are in balance.
This is essentially the rule that the Legislature considered—and rejected—in amending the Water Code in 1995. The initial versions of Senate Bill No. 901 (1995-1996 Reg. Sess.), which (among other things) added sections 10910 to 10915 to the Water Code, directed the lead agency for a *454project EIR to request a water supply and demand assessment from the appropriate public water system, and stated that the lead agency “shall consider a project to have a significant effect on the environment” if, based on that assessment, “water supplies are, or will be, insufficient to meet the reasonable needs of the proposed project in addition to existing and planned future uses.” (Sen. Bill No. 901 (1995-1996 Reg. Sess.) § 2, as amended July 5, 1995, italics omitted, proposed Wat. Code, § 10915.) The bill as enacted, however, deleted the requirement that the lead agency make a finding of a significant environmental impact under such circumstances and directed the lead agency, if it determined that water supplies will not be sufficient to meet existing and planned future uses, instead simply to “include that determination in its findings.”2 (Sen. Bill No. 901 (1995-1996 Reg. Sess.) § 4, as amended Sept. 7, 1995, italics omitted; Stats. 1995, ch. 881, § 4, p. 6705, adding Wat. Code, § 10911.) This sequence of events makes me confident that the Legislature did not intend to require a project EIR to balance water supply with water demand not only for the project itself but also for the entire region. (Cf. Hess v. Ford Motor Co. (2002) 27 Cal.4th 516, 532 [117 Cal.Rptr.2d 220, 41 R3d 46] [“ ‘Generally the Legislature’s rejection of a specific provision which appeared in the original version of an act supports the conclusion that the act should not be construed to include the omitted provision’ ”]; accord, INS v. Cardoza-Fonseca (1987) 480 U.S. 421, 442-443 [94 L.Ed.2d 434, 107 S.Ct. 1207] [“ ‘Few principles of statutory construction are more compelling than the proposition that Congress does not intend sub silentio to enact statutory language that it has earlier discarded in favor of other language’ ”].) The majority offers no justification for effectively reinserting what the Legislature has rejected.
Indeed, the legislative history leading to the elimination of the Senate Bill No. 901 (1995-1996 Reg. Sess.) stricter requirement explains why this court ought not itself resurrect it. One legislative analysis warned that the required finding of a significant environmental impact due to an imbalance between water supply and demand on a regional basis “could be a severe roadblock to housing development as it is the [Department of Housing and Community Development]’s experience that many areas of the State cannot demonstrate water supply availability for all potential development which could be permitted under their general plan land use designations within the next five years. Also, it would be infeasible for many cities or counties to demonstrate water supply availability for all potential development over the 10 to 20 year timeframes of general plan updates.” (Dept. of Housing and Community Development, analysis of Sen. Bill No. 901 (1995-1996 Reg. Sess.) Aug. 7, 1995, p. 5.) The Department of Housing and Community Development’s analysis further warned that “[w]here there may be an adequate water supply *455for a housing project and the project may have no significant effect on the environment, but an inadequate water supply exists for long term future uses, mitigation measures in the form of fees are likely to be assessed to buy water or develop new supplies. These are likely to significantly increase costs for new housing development.” (Id. at p. 6.) Moreover, “[u]sing the complex and bureaucratic CEQA process to assure local water planning is likely to result in significant administrative costs which will, in every likelihood, be charged to new development because there is no other pocket to pay.” (Id. at p. 8.) Finally, such an approach would supply “new opportunities for court challenges of new housing and job-creating development. From the perspective of possible environmental litigation, the bill would create great uncertainty.” (Id. at p. 7.)3
I also find it interesting that neither plaintiffs nor the Attorney General as amicus curiae, when offered the opportunity at oral argument to embrace the majority’s new rule, chose to do so. Plaintiffs stated instead that “the EIR must address the water supply essential for the scope of the project that is approved,” not for the entire general plan. The Attorney General similarly explained that the general rule under CEQA is that an agency must consider “all the significant environmental impacts for the project that it is approving,” distinguishing the SunRidge Specific Plan and Sunrise Douglas Community Plan from the entire zone 40 area, and that considering the entire general plan was thus “too far out from where this court needs to go.”
By recognizing that CEQA does not require a project EIR to balance water supply and demand on a regional basis, I do not intend to diminish the significance of a finding in a project FEIR that projected supply will not be able to satisfy the entirety of projected demand contemplated by a general plan. Obviously, if new supplies are not found, then a decision to approve one project means that projects proposed later in time may be unable to identify adequate water supplies and therefore may not be built. If not all of the development contemplated by the general plan can be built, cities and counties must ensure that the projects that are approved are of the highest priority, in order to prevent the negative economic or social effects from haphazard development. However, one must also remember that “[ejconomic or social effects of a project shall not be treated as significant effects on the environment” (CEQA Guidelines, tit. 14, § 15131, subd. (a)) and therefore are beyond the scope of CEQA. Under the majority’s new rule, however, once a city or county approves a general plan, it could not approve a project in *456furtherance of that plan unless or until it had secured water sources for build-out of the entire general plan. Nothing in CEQA requires such a result. (Atherton v. Board of Supervisors (1983) 146 Cal.App.3d 346, 351 [194 Cal.Rptr. 203] [“ ‘where future development is unspecified and uncertain, no purpose can be served by requiring an EIR to engage in sheer speculation as to foture environmental consequences’ ”].)
It is no answer to suggest, as the majority does, that the FEIR for the Sunrise Douglas Community Plan might have been adequate if it instead had disclosed “concrete plans for new surface water diversion, treatment and transmission facilities that would tend to tie the new water particularly to Sunrise Douglas,” akin to those included in the SunRidge Specific Plan’s discussion of water from the Well Field. (Maj. opn., ante, p. 445.) The majority seems to forget that “[t]o interpret CEQA itself as requiring such firm assurances of future water supplies at relatively early stages of the land use planning and approval process would put CEQA in tension with . . . more specific water planning statutes.” (Maj. opn., ante, at p. 434.) Indeed, it is precisely because “full build-out of the Community Plan may take 15 or 20 years” (id. at p. 445) that the analysis of water supplies for the Community Plan did not need to be as detailed as the analysis for water supplies for the Specific Plan, which would begin to draw water “within about 18 months of project approval.” (Id. at p. 445.) The majority’s insistence that the analysis of zone 40 water supplies in the long-term must be as concrete as that for the Well Field in the near-term completely inverts its earlier assertion that “ ‘water supplies must be identified with more specificity at each step as land use planning and water supply planning move forward from general phases to more specific phases.’ ” (Id. at pp. 433-434.)
The reader might likewise be forgiven for looking with skepticism at the majority’s assurance that “CEQA does not necessarily require that an EIR show that total water supply and demand are or will be in balance in an area,” inasmuch as the majority elsewhere condemns this FEIR because it “could not demonstrate a likelihood of adequate long-term supply for Sunrise Douglas without showing that plans for the Zone 40 area call for at least a rough balance between water supply and demand, a showing the FEIR fails to make.” (Compare maj. opn., ante, at p. 445 with id. at p. 446.) And if, as the majority belatedly states, it would be enough for the FEIR, as to future water supplies needed for the project, to “include only the public water system’s plans for acquiring the additional supplies, including cost and time estimates and regulatory approvals the system anticipates needing” (maj. opn., ante, at p. 433; see id. at pp. 446-447), one wonders why the majority goes on at length to discuss far more burdensome requirements—and what authority it has to do so.
*457In sum, the majority’s insistence that the FEIR should have identified sufficient water not merely for the project itself but also for all conceivable future development in the region suffers from a number of serious defects. It is not supported by any statute or guideline—or, indeed, by any party to this litigation. It is inconsistent with the legislative history of Water Code section 10911. It is inconsistent as well with the majority’s own analysis of the environmental effects of drawing on this project’s near-term water supplies. And, as the Legislature recognized in rejecting such an approach in 1995, it will discourage new housing development, increase its cost, create uncertainty, and trigger more litigation. For all these reasons, I respectfully dissent.
Appellants’ petition for a rehearing was denied April 18, 2007, and the opinion was modified to read as printed above. George, C. J., did not participate therein. Baxter, J., was of the opinion that the petition should be granted.
It also, quite obviously, is not the test by which the majority has approved the adequacy of the FEIR’s analysis of water supplies in the near term. The majority finds that analysis adequate, notwithstanding the fact that supplying existing and new demand in the area as well as a significant portion of the Sunrise Douglas project from the Well Field in the near term “would require much more water than the 10,000 afa that source can safely provide.” (Maj. opn., ante, at p. 436.)
As the majority concedes, the county’s compliance with these Water Code provisions is not at issue in this case. (Maj. opn., ante, at p. 433, fn. 8.)
The Governor’s Office of Planning and Research also cautioned that an early version of the bill made no provision for measures that may act to reduce overall demand by requiring “new development to retrofit old, existing development in order to free sufficient ‘wasted’ water to serve the new project.” (Governor’s Off. of Planning and Research, analysis of Sen. Bill No. 901 (1995-1996 Reg. Sess.) Apr. 3, 1995, p. 6.)