Opinion by Judge NOONAN; Concurrence by Judge NOONAN.
ORDER
The opinion filed on May 14, 2008 is amended as follows: Replace the final paragraph with “For the reasons stated, we REVERSE the district court’s denial of a preliminary injunction and REMAND with instructions to grant immediately a preliminary injunction on the three proposed projects to the extent that they are inconsistent with the 2001 FEIS.”
OPINION
NOONAN, Circuit Judge:Sierra Forest Legacy (Sierra Forest) appeals the decision of the district court denying a preliminary injunction against the United States Forest Service (the USFS or the Forest Service) in a suit challenging its decision to permit logging in accordance with changes made in 2004 by the USFS in the relevant forest plan. Other parties, noted in the caption, have intervened on each side. The Attorney General of California, Edmund G. Brown, Jr., has filed an amicus brief in support of Sierra Forest.
We hold that the district court abused its discretion. We reverse and remand.
PROCEEDINGS
Sierra Forest is comprised of the Sierra Nevada Forest Protection Campaign, Center for Biological Diversity, Natural Resources Defense Council, Sierra Club, and The Wilderness Society, many of whose members enjoy and are educated by the affected forests and the wildlife dependent on habitats within them. This suit was begun in 2005 in response to the Supplemental Environmental Impact Statement (SEIS) issued by the USFS in January of 2004 as a supplement to the Final Environmental Impact Statement (FEIS), issued by the USFS in 2001 in implementation of the Sierra Nevada Forest Plan Amendment.
Under the SEIS, the USFS approved logging in three specific sites: Basin, Empire, and Slapjack. On September 10, 2007, the USFS announced that it intended to advertise and award logging contracts for these sites. On September 21, 2007, Sierra Forest moved for a preliminary injunction. On October 15, 2007, the district court denied the motion.
*1231Sierra Forest appeals, raising several claims under the National Forest Management Act (NFMA), 16 U.S.C. §§ 1600-1614, and the National Environmental Policy Act (NEPA), 42 U.S.C. §§ 4321-4370f. In light of our disposition, we do not reach all of the arguments raised by Sierra Forest.
ANALYSIS
The Standard
Our review is a review of a motion preliminary to a trial. As the district court’s decision is preliminary, so must our decision be preliminary. It is not on the merits. We need not address all aspects of the projects. Our decision must defer to the discretion of the district judge who has had to act with some dispatch. See Lands Council v. Mantin, 479 F.3d 636, 639 (9th Cir.2007) (citation omitted). When a preliminary injunction is sought, there is a sense of urgency on each side — to go ahead expeditiously with the project; to stop what is seen as harm that cannot be undone. Deferential as we are, we cannot default in reviewing de novo the law binding on the judge who has discretion but not carte blanche. See Sports Form, Inc. v. United Press Int’l, Inc., 686 F.2d 750, 752 (9th Cir.1982). We state only the facts relevant to the result.
A district court abuses its discretion if it bases its decision on an erroneous legal standard or clearly erroneous finding of fact. See Earth Island Inst. v. U.S. Forest Serv., 351 F.3d 1291, 1298 (9th Cir.2003) (citation omitted). The familiar criteria to be met to obtain the issuance of an injunction before the, trial are a strong likelihood of success on the merits; the possibility of irreparable harm; a balance of hardships favoring the plaintiffs; and advancement of the public interest. See id. at 1297-98 (citation omitted).
Probability of Success on the Merits
There is no disagreement that USFS is authorized to take action to prevent the occurrence of forest fires. One necessary step is the clearing of brush, including the removal of small trees. Doing so involves the expenditure of funds. The USFS does not assert, however, that it is necessary as a preventive measure to cut down the larger trees that provide the habitat in which various species thrive. These trees constitute a desirable prize for loggers who seek to convert them into lumber for commercial purposes. The USFS acknowledges that its reason for selling the forest trees to commercial loggers is to raise funds to carry on its fire prevention duties. Sierra Forest and the State of California seek to preserve the larger trees and so to preserve the habitat that supports various species. We need decide here a limited and narrow issue: Does the 2004 SEIS prepared by USFS regarding its plans to sell off the forest trees comply with the requirements of NEPA?
Sierra Forest argues that USFS violated NEPA’s requirement to “[rjigorously explore and objectively evaluate all reasonable alternatives” to a proposed plan that has significant environmental effects. 40 C.F.R. § 1502.14(a) (2000). USFS cannot rely on its discussion of alternatives in the 2001 FEIS to satisfy this requirement for the 2004 SEIS. “[W]here changed circumstances affect the factors relevant to the development and evaluation of alternatives,” USFS “must account for such change in the alternatives it considers.” Natural Res. Def. Council v. U.S. Forest Serv., 421 F.3d 797, 813-14 (9th Cir.2005) (citation omitted).
Such changed circumstances plainly exist here. First, USFS altered its modeling techniques between the issuance of the 2001 FEIS and the 2004 SEIS and failed *1232to update its analysis of the 2001 FEIS alternatives under these new techniques. Second, the 2004 SEIS introduced substantively new objectives from those contained within the 2001 FEIS. A primary purpose of the new framework adopted by the SEIS is the provision of funds for the reduction of fuel, that is, for the reduction of the risk of fire in the forests. This goal has become an imperative after the catastrophic fires that have devastated forests in the northwest. Severe wildfires have increased dramatically in the Sierra Nevada from an average of 43,000 acres per year ten years ago to an average of 63,000 acres per year. Control of wildfires is an imperative for the inhabitants of land bordering the forests. It is an imperative for defenders of the habitat and the wildlife within them. Fire is a force that must be managed if the environment is to be protected.
The SEIS proposes a simple solution:
Opportunities for Leveraging Appropriated Funds to Accomplish Fuels Treatments
Under Alternative S2, revenues from the sale of commercial forest products could be obtained from some fuels treatments. This would increase the likelihood of accomplishing the projected acres of treatment, an essential first step in achieving the desired reductions in acres burned. Where consistent with desired conditions, area treatments would be designed to be economically efficient and meet multiple objectives.
Timber sale contracts provide a mechanism for the efficient removal of eommercially-valuable sawtimber. Contracts that have sufficient value offer capabilities for funding the accomplishment of additional resource management goals. Records from recent timber offerings indicate that sales with higher volumes per acre attract higher bids. Sales yielding an average 4.5 mbfiacre provide approximately $112/ mbf, compared to only $38/mbf for 1.5 mbfiacre (Lamdram, pers comm).
The size of tree made available for harvest has a significant influence on sale volume per acre averages and thus, per unit bid values. Assuming typical heights, the board foot volume for a 12-inch dbh tree is 39, compared to 317 for a 20 inch tree and 710 for a 24 inch tree. Using these assumptions, 77 twelve-inch dbh tress would be needed to reach the minimum economically feasible sale volume (estimated at 3 mbflacre). This compares to 9 trees of 20-inch dbh and 4 trees of 24-inch dbh. In summary, including only a few medium-sized trees can make an impact on the economic viability of a given project.
A number of options are available for deriving eommercially-valuable wood products from fuels treatments. Where wood-fired electrical generation facilities exist and sufficient sawtimber value is present, small trees, e.g. biomass, can be removed. Bids in excess of required collections may also be made available for fuel reduction treatments within the sale area boundary. These may include:
1) Shredding of ladder fuels, i.e. small trees, woody shrubs, and surface fuel,
2) Prescribed fire treatment following timber harvest, or
3) Fuel reduction treatment outside timber sale units (within the time sale area boundary).
Alternatively, a stewardship contract package (a service contract, not a timber sale contract), that includes commereially-valuable sawtimber, may provide for cost-effective implementation of multiple fuels reduction projects within the contracted area.
In amplification, the USFS replied to the following public comment:
*1233 9.2.1. Public Concern: The Final SEIS should not claim that increased logging levels wiU increase forest protection, or it should scientifically justify that assertion.
Response: Alternative S2 in the SEIS was developed to provide opportunities for increasing available funds for fuels reduction work on the national forests. This alternative increases revenues by permitting the removal of some medium-sized trees from some areas. The SEIS does not suggest that removing these trees will alter stand structure in ways that significantly enhance fire protection. It is the increase in available funds from logging that can be used to increase fuels reduction work. But the work would be done on other lands. See the discussion on fuels treatment economics in the SEIS (Chapter 4, Economics of Fuels Treatments) for more information about treatment costs and the value of additional timber harvest to fuels reduction work. The Final SEIS (Chapter 4, Fire and Fuels Management) has an expanded discussion regarding the economics of fuels treatments.
Sell trees to loggers. Use the money to clear areas of what is potential fuel for fire. The solution has a secondary benefit: what the loggers cut can, at least in part, be timber that was potential for fire. In one sale, a fire hazard can be removed and the USFS paid so that it can remove the fuel of future fires.
Two for one always has an attractive ring. But are there no alternative ways of getting money to do the clearing that is imperative? Obviously, there may be. First of all, there is the USFS’s own budget. Does that budget contain any funds that could be devoted to fuel removal? Is every one of its activities so necessary and so tightly allocated that no money could be shifted? We do not know the answer because this alternative has not been explored.
Suppose that the USFS and its parent, the Department of Agriculture, cannot spare a dime. What then? Appropriate appropriations come from Congress. The work of fire prevention is work of the first importance. If the USFS does not have enough, why should not Congress be asked to give it more? Surely the avoidance of catastrophic fire in the national forests must rate a high priority among the needs of the nation.
Alternatives considered in the 2001 FEIS address the critical problem of fuel reduction. Several of them (F3, F4, F6, and F7) are projected as achieving an acreage reduction of over 30% in the first five decades as opposed to a 22% reduction that is projected in the adopted Alternative S2. These alternatives do not appear to have been reexamined in the light of the new urgency of fire prevention.
The Attorney General of California raised several alternative methods to fund USFS’s fire reduction objectives, including requesting a special appropriation from Congress, re-prioritizing other funding, and altering its fuel treatment program. USFS failed to consider these alternatives in its implementation of the 2004 SEIS. So long as all these alternatives remain unexamined or unreexamined, so long does the SEIS fail to conform to the law. The district court abused its discretion in concluding that USFS complied with NEPA’s requirement to “[rjigorously explore and objectively evaluate all reasonable alternatives.” 40 C.F.R. § 1502.14(a) (2000).
Balancing of Equities
The legal merits of the Sierra Forest’s case, at this stage of the litigation, are strong. To justify a preliminary halt to the projects the real possibility of irreparable harm is still required. It is not necessary to canvass all the species that *1234may be affected and all the environmental harm that might ensue. It suffices in this case to take account of the status of the spotted owl whose range relates to the affected forests. True, the species exists in southern California as well as in the northwest; but the species as a whole has been classified as “sensitive” by the Forest Service. The proposed logging will not destroy the species. What it will do is reduce its established habitat. The possibility that this reduction in its range will irreparably damage the sensitive species cannot be dismissed.
Postponement of the Forest Service plans may increase the danger posed by fires; but the Forest Service and Congress do not appear helpless to find the funds to decrease the dangers. The question we address here is whether USFS’s choice of funding for fire reduction — rather than fire reduction itself — outweighs California’s preservation interests. We conclude that it does not, given that “special solicitude” should be afforded California’s stake in its natural resources and that the Forest Service did not consider alternatives to its choice of funding. Massachusetts v. Envtl. Prot. Agency, — U.S. -, 127 S.Ct. 1438, 1454-55, 167 L.Ed.2d 248 (2007).
Public interests are further implicated: the importance of preserving the environment and of enforcing the law intended to preserve it. See Amoco Prod. Co. v. Vill. of Gambell, 480 U.S. 531, 545, 107 S.Ct. 1396, 94 L.Ed.2d 542 (1987).
For the reasons stated, we REVERSE the district court’s denial of a preliminary injunction and REMAND with instructions to grant immediately a preliminary injunction on the three proposed projects to the extent that they are inconsistent with the 2001 FEIS.