with whom Justice Stevens joins as to Part I, concurring in part and dissenting in part.
As of December 2006, the United States Navy planned to engage in a series of 14 antisubmarine warfare training exercises off the southern California coast. The Natural Resources Defense Council, Inc., and others (NRDC) brought this case in Federal District Court claiming that the National Environmental Policy Act of 1969 (NEPA) requires the Navy to prepare an environmental impact statement (EIS) (assessing the impact of the exercises on marine mammals) prior to its engaging in the exercise's. As the case reaches us, the District Court has found that the NRDC will likely prevail on its demand for an EIS; the Navy has agreed to prepare an EIS; the District Court has forbidden the Navy to proceed with the exercises unless it adopts six mitigating measures; and the Navy has agreed to adopt all but two of those measures.
The controversy between the parties now concerns the two measures that the Navy is unwilling to adopt. The first concerns the “shutdown zone,” a circle with a ship at the center within which the Navy must try to spot marine mammals and shut down its sonar if one is found. The controverted condition would enlarge the radius of that circle from about one-tenth of a mile (200 yards) to one and one-quarter miles (2,200 yards). The second, concerns special ocean conditions called “surface ducting conditions.” The controverted condition would require the Navy, when it encounters any such condition, to diminish the sonar’s power by 75%. The Court of Appeals affirmed the District Court order that contained these two conditions. 518 F. 3d 658, 703 (CA9 2008).
I
We must now decide whether the District Court was legally correct in forbidding the training exercises unless the Navy implemented the two controverted conditions. In *35doing so, I assume, like the Court, that the NRDC will prevail on its demand for an EIS. (Indeed, the Navy is in the process of preparing one.) And, I would ask whether, in imposing these conditions, the District Court properly “balance[d the] harms.” See, e.g., Amoco Production Co. v. Gambell, 480 U. S. 531, 545 (1987).
Respondents’ (the plaintiffs) argument favoring the District Court injunction is a strong one. As Justice Ginsburg well points out, see post, at 47-48 (dissenting opinion), the very point of NEPA’s insistence upon the writing of an EIS is to force an agency “carefully” to “consider ... detailed information concerning significant environmental impacts,” while “giv[ing] the public the assurance that the agency ‘has indeed considered- environmental concerns in its decision-making process.’” Robertson v. Methow Valley Citizens Council, 490 U. S. 332, 349 (1989). NEPA seeks to assure that when Government officials consider taking action that may affect the environment, they do so fully aware of the relevant environmental considerations. An EIS does not force them to make any particular decision, but it does lead them to take environmental considerations into account when they decide whether, or how, to act. Id., at 354. Thus, when a decision to which EIS obligations attach is made without the informed environmental consideration that NEPA requires, much of the harm that NEPA seeks to prevent has already taken place. In this case, for example, the absence of an injunction means that the Navy will proceed with its exercises in the absence of the fuller consideration of environmental effects that an EIS is intended to bring. The absence of an injunction thereby threatens to cause the very environmental harm that a full preaction EIS might have led the Navy to avoid (say, by adopting the two additional mitigation measures that the NRDC proposes). Consequently, if the exercises are to continue, conditions designed to mitigate interim environmental harm may well be appropriate.
*36On the other hand, several features of this case lead me to conclude that the record, as now before us, lacks adequate support for an injunction imposing the two controverted requirements. First, the evidence of need for the two special conditions is weak or uncertain. The record does show that the exercises as the Navy originally proposed them could harm marine mammals. The District Court found (based on the Navy’s study of the matter) that the exercises might cause 466 instances of Level A harm and 170,000 instances of Level B harm. App. to Pet. for Cert. 196a-197a. (The environmental assessment actually predicted 564 instances of Level A harm. See App. 223-224.) The study defines Level A injury as “any act that injures or has the significant potential to injure a marine mammal or marine mammal stock in the wild” through “destruction or loss of biological tissue,” whether “slight to severe.” Id., at 160. It defines Level B harm as “ ‘any act that disturbs or is likely to disturb a marine mammal... by causing disruption of natural behavioral patterns including, but not limited to, migration, surfacing, nursing, breeding, feeding, or sheltering to a point where such behaviors are abandoned or significantly altered’ ” and describes it as a “short term” and “temporary” “disturbance.” Id., at 161, 175.
The raw numbers seem large. But the parties argue about the extent to which they mean likely harm. The Navy says the classifications and estimates err on the side of caution. (When in doubt about the amount of harm to a mammal, the study assumed the harm would qualify as Level A harassment. Id., at 200.) The Navy also points out that, by definition, mammals recover from Level B injuries, often very quickly. It notes that, despite 40 years of naval exercises off the southern California coast, no injured marine mammal has ever been found. App. to Pet. for Cert. 274a-275a. (It adds that dolphins often swim alongside the ships. Id., at 290a, 346a.) At the same time, plaintiffs point to instances where whales have been found stranded. They add *37that scientific studies have found a connection between those beachings and the Navy’s use of sonar, see, e. g., App. 600-602, and the Navy has acknowledged one stranding where “U. S. Navy mid-frequency sonar has been identified as the most plausible contributory source to the stranding event,” id., at 168.
Given the uncertainty the figures create in respect to the harm caused by the Navy’s original training plans, it would seem important to have before us at least some estimate of the harm likely avoided by the Navy’s decision not to contest here four of the six mitigating conditions that the District Court ordered. Without such evidence, it is difficult to assess the relevant harm — that is, the environmental harm likely caused by the Navy’s exercises with the four uncontested mitigation-measures (but without the two contested mitigation measures) in place.
Second, the Navy has filed multiple affidavits from Navy officials explaining in detail the seriousness of the harm that the delay associated with completion of this EIS (approximately one year) would create in respect to the Navy’s ability to maintain an adequate national defense. See generally App. to Pet. for Cert. 260a-357a. Taken by themselves, those affidavits make a strong case for the proposition that insistence upon the two additional mitigating conditions would seriously interfere with necessary defense training.
The affidavits explain the importance of training in antisubmarine warfare, id., at 263a; the need to use active sonar to detect enemy submarines, id., at 266a-267a, App. 566; the complexity of a training exercise involving sonar, App. to Pet. for Cert. 343a; the need for realistic conditions when training exercises take place, id., at 299a-300a, App. 566; the “cascading” negative “effect” that delay in one important aspect of a set of coordinated training exercises has upon the Navy’s ability “to provide combat ready forces,” App. to Pet. for Cert. 343a; the cost and disruption that would accompany the adoption of the two additional mitigating conditions that *38the NRDC seeks, ibid.; the Navy’s resulting inability adequately to train personnel, id., at 278a; the effectiveness of the mammal-protecting measures that the Navy has taken in the past, id., at 285a-298a; and the reasonable likelihood that the mitigating conditions to which it has agreed will prove adequate, id., at 296a.
Third, and particularly important in my view, the District Court did not explain why it rejected the Navy’s affidavit-supported contentions. In its first opinion enjoining the use of sonar, the District Court simply stated:
“The Court is ... satisfied that the balance of hardships tips in favor of granting an injunction, as the harm to the environment, Plaintiffs, and public interest outweighs the harm that Defendants would incur if prevented from using [mid-frequency active (MFA)] sonar, absent the use of effective mitigation measures, during a subset of their regular activities in one part of one state for a limited period.” Id., at 217a-218a.
Following remand from the Court of Appeals, the District Court simply repeated, word for word, this same statement. It said:
“The Court is ... satisfied that the balance of hardships tips in favor of granting an injunction, as the harm to the environment, Plaintiffs, and public interest outweighs the harm that Defendants would incur (or the public interest would suffer) if Defendants were prevented from using MFA sonar, absent the use of effective mitigation measures, during a subset of their regular activities in one part of one state for a limited period.” 530 F. Supp. 2d 1110, 1118 (CD Cal. 2008).
With respect to the imposition of the 2,200-yard shutdown zone, the District Court noted evidence of the harm that MFA sonar poses to marine mammals, and then concluded that “[t]he Court therefore is persuaded that while the 2200 yard shutdown requirement may protect marine mammals *39from the harshest of sonar-related consequences, it represents a minimal imposition [on] the Navy’s training exercises.” Id., at 1119. The District Court did not there explain the basis for that conclusion. With respect to the imposition of the surface ducting condition, the District Court said nothing about the Navy’s interests at all. Id., at 1120-1121.
While a district court is often free simply to state its conclusion in summary fashion, in this instance neither that conclusion, nor anything else I have found in the District Court’s opinion, answers the Navy’s documented claims that the two extra conditions the District Court imposed will, in effect, seriously interfere with its ability to carry out necessary training exercises.
The first condition requires the Navy to reduce the power of its sonar equipment by 75% when the ship encounters a condition called “surface ducting” that occurs when the presence of layers of water of different temperature make it unusually difficult for sonar operators to determine whether a diesel submarine is hiding below. Rear Admiral John Bird, an expert in submarine warfare, made clear that the 75% power-reduction requirement was equivalent to forbidding any related training. App. to Pet. for Cert. 297a. But he says in paragraph 52 of his declaration: “Training in surface ducting conditions is critical to effective training because sonar operators need to learn how sonar transmissions are altered due to surface ducting and how submarines may take advantage of them.” Id., at 299a-300a. The District Court, as far as I can tell, did not even acknowledge in its opinion the Navy’s asserted interest in being able to train under these conditions. 530 F. Supp. 2d, at 1120-1121.
The second condition requires the Navy to expand the sonar “shutdown” area surrounding a ship (i e., turn off the sonar if a mammal is spotted in the area) from a circle with a radius of about one-tenth of a mile to a circle with a radius of about one mile and a quarter. Both sides agree that this *40requirement will lead to more shutdowns. Admiral Gary Roughead, Chief of Naval Operations, states in paragraph 12 of his declaration that this expanded zone requirement “will result in increased interruptions to training exercises, . . . vastly increasing] the risk of negating training effectiveness, preventing strike group certification, and disrupting carefully orchestrated deployment plans to meet world-wide operational commitments.” App. to Pet. for Cert. 344a. Again, I can find nothing in the District Court’s opinion that specifically explains why this is not so. 530 F. Supp. 2d, at 1119-1120.
Fourth, the Court of Appeals sought, through its own thorough examination of the record, to supply the missing explanations. But those explanations are not sufficient. In respect to the surface ducting conditions, the Court of Appeals rejected the Navy’s contentions on the ground that those conditions are “rar[e],” and the Navy has certified trainings that did not involve any encounter with those conditions. 518 F. 3d, at 701-702. I am not certain, however, why the rarity of the condition supports the District Court’s conclusion. Rarity argues as strongly for training when the condition is encountered as it argues for the contrary.
In respect to the expansion of the “shutdown” area, the Court of Appeals noted that (1) the Navy in earlier exercises had shut down its sonar when marine mammals were sited within about one-half a mile, (2) the Navy has used a larger shutdown area when engaged in exercises with lower frequency sonar equipment, and (3) foreign navies have used larger shutdown areas. Id., at 699-701, and nn. 63, 67. But the Navy’s affidavits state that (1) earlier shutdowns when marine mammals were spotted at farther distances “likely occurred during tactically insignificant times,” App. to Pet. for Cert. 356a, (2) ships with low frequency sonar (unlike the sonar here at issue) have equipment that makes it easier to monitor the larger area, particularly by significantly reducing the number of monitoring personnel necessarily involved, and (3) foreign navy experience is not relevant given the *41potentially different military demands upon those navies, App. 508-509.
Finally, the Court of Appeals, mirroring a similar District Court suggestion in the language I have quoted, says that “the exercises in southern California are only a subset of the Navy’s training activities involving active sonar.” 518 F. 3d, at 702. It adds that the Navy’s study “shows the Navy is still able to conduct its exercises in alternative locations, in reduced number, or through simulation.” Ibid., n. 69. The Court of Appeals, however, also concluded that the study “provides reasonably detailed justifications for why the Southern California Operating Area is uniquely suited to these exercises, and demonstrates that the Navy would suffer a certain hardship if the considered alternatives were employed instead.” Ibid.
Fifth, when the Court of Appeals first heard this case following the District Court’s imposition of a broad, absolute injunction, it held that any injunction must be crafted so that the Navy could continue its training exercises. Noting that the Navy had, in the past, been able to use mitigation measures to “reduce the harmful effects of its active sonar,” it “vacate[d] the stay and remand[ed] this matter to the district court to narrow its injunction so as to provide mitigation conditions under which the Navy may conduct its training exercises.” 508 F. 3d 885, 887 (CA9 2007) (emphasis added). For the reasons just stated, neither the District Court nor the Court of Appeals has explained why we should reject the Navy’s assertions that it cannot effectively conduct its training exercises under the mitigation conditions imposed by the District Court.
I would thus vacate the preliminary injunction imposed by the District Court to the extent it has been challenged by the Navy. Neither the District Court nor the Court of Appeals has adequately explained its conclusion that the balance of the equities tips in favor of plaintiffs. Nor do those parts of the record to which the parties have pointed supply the missing explanation.
*42II
Nonetheless, as the Court of Appeals held when it first considered this case, the Navy’s past use of mitigation conditions makes clear that the Navy can effectively train under some mitigation conditions. In the ordinary course, I would remand so the District Court could, pursuant to the Court of Appeals’ direction, set forth mitigation conditions that will protect the marine wildlife while also enabling the Navy to carry out its exercises. But, at this point, the Navy has informed us that this set of exercises will be complete by January, at the latest, and an EIS will likely be complete at that point, as well. Thus, by the time the District Court would have an opportunity to impose new conditions, the case could very well be moot.
In February of this year, the Court of Appeals stayed the injunction imposed by the District Court — but only pending this Court’s resolution of the case. The Court of Appeals concluded that “[i]n light of the short time before the Navy is to commence its next exercise, the importance of the Navy’s mission to provide for the national defense and the representation by the Chief of Naval Operations that the district court’s preliminary injunction in its current form will ‘unacceptably risk’ effective training and strike group certification and thereby interfere with his statutory responsibility ... to ‘organiz[e], train[], and equip[] the Navy,’” interim relief was appropriate, and the court then modified the two mitigation conditions at issue. 518 F. 3d 704, 705 (CA9 2008).
With respect to the 2,200-yard shutdown zone, it required the Navy to suspend its use of the sonar if a marine mammal is detected within 2,200 yards, except when sonar is being used at a “critical point in the exercise,” in which case the amount by which the Navy must power down is proportional to the mammal’s proximity to the sonar. Id., at 705-706 (internal quotation marks omitted). With respect to surface ducting, the Navy is only required to shut down sonar alto*43gether when a marine mammal is detected within 500 meters and the amount by which it is otherwise required to power down is again proportional to the mammal’s proximity to the sonar source. Ibid. The court believed these conditions would permit the Navy to go forward with its imminently planned exercises while at the same time minimizing the harm to marine wildlife.
In my view, the modified conditions imposed by the Court of Appeals in its February stay order reflect the best equitable conditions that can be created in the short time available before the exercises are complete and the EIS is ready. The Navy has been training under these conditions since February, so allowing them to remain in place will, in effect, maintain what has become the status quo. Therefore, I would modify the Court of Appeals’ February 29, 2008, order so that the provisional conditions it contains remain in place until the Navy’s completion of an acceptable EIS.