concurring in the denial of rehearing en banc:*
The opinion in this case speaks for itself. See United States v. Washington, 853 F.3d 946 (9th Cir. 2017). We write to respond to the views of our colleagues who dissent *1019from the decision of our court not to rehear the case en banc.
In 1854 and 1855, U.S. Superintendent of Indian Affairs and Governor of Washington Territory, Isaac I. Stevens, negotiated a series of virtually identical Treaties with the Indian Tribes that lived around Puget Sound. In return for their agreement to live on reservations, the Tribes were promised equal access to off-reservation fishing “at all usual and accustomed grounds and stations.” The Supreme Court described the importance of the promise:
During the negotiations, the vital importance of the fish to the Indians was repeatedly emphasized by both sides, and the Governor’s promises that the treaties would protect that source of food and commerce were crucial in obtaining the Indians’ assent.
Washington v. Washington State Commercial Passenger Fishing Vessel Ass’n (“Fishing Vessel”), 443 U.S. 658, 676, 99 S.Ct. 3055, 61 L.Ed.2d 823 (1979).
For more than 100 years, the State of Washington deliberately and systematically prevented the Tribes from engaging in the off-reservation fishing promised under the Treaties. The State eventually came to employ surveillance planes, high powered boats, tear gas, billy clubs and guns against tribal members engaged in off-reservation fishing. In 1970, the United States brought suit against Washington State to enforce the Treaties.
The district court held that the Treaties promised the Tribes fifty percent of the harvestable salmon in any given year. The Supreme Court affirmed, holding that the Tribes had been promised a “moderate living” from fishing, and that they were entitled to fifty percent of the harvest, up to the point where they were able to catch enough salmon to provide a moderate living. Id. at 686, 99 S.Ct. 3055. The district court entered a detailed injunction which the State strenuously resisted. The Supreme Court affirmed the injunction:
It is ... absurd to argue ... both that the state agencies may not be ordered to implement the decree and also that the District Court may not itself issue detailed remedial orders as a substitute for state supervision.
Id. at 695, 99 S.Ct. 3055.
The current proceeding is a continuation of the suit brought by the United States in 1970.
Salmon are anadromous fish—hatching in fresh water, migrating to the ocean to mature, and returning to fresh water to spawn—so access to spawning grounds is essential to their reproduction and survival. For many years, the Tribes had complained that the State had built roads across salmon-bearing streams, and that it had built culverts under the roads that allowed passage of water but not passage of salmon. The United States instituted the current proceeding in 2001 to require the State to modify its culverts to allow passage of salmon.
The State has fought the proceeding tooth and nail. The State contended, and continues to contend, that it can block every salmon-bearing stream into Puget Sound without violating the Treaties. The district court disagreed and held that the State’s affirmative act of building roads with salmon-blocking, or “barrier,” culverts violated the Treaties. The district court sought the State’s participation and assistance in drafting a remedial injunction, but the State refused to participate. Despite the State’s refusal, the district court entered an injunction that was substantially more favorable to the State than the injunction sought by the United States.
The State appealed, objecting to the district court’s holding that its affirmative acts in building roads with barrier culverts *1020violated the Treaties. Without conceding that it violated the Treaties, the State also objected to the scope of the injunction in whose formulation it had declined to participate. We affirmed.
Our dissenting colleagues object to our decision on four grounds. We respond to the objections in turn.
I. Violation of the Treaties
First, our colleagues contend that we have misread the Supreme Court’s 1979 decision in Fishing Vesssel. They contend that fifty percent of the harvestable salmon is an absolute “ceiling” on the amount of fish the Tribes have been promised. They contend that the Treaties promised only that the Tribes will get fifty percent of the harvestable salmon, and that Treaties permit the State to take affirmative acts that have the effect of diminishing the supply of salmon below the amount necessary to provide a moderate living. According to our colleagues, if the State acts affirmatively to entirely eliminate the supply of harvestable salmon, the Tribes get fifty percent of nothing.
Our colleagues misread Fishing Vessel. The Court recognized that the Treaties promised that the Tribes would have enough salmon to feed themselves. In the words of the Court, the Treaties promised that the Tribes would have enough har-vestable salmon to provide a “moderate living.” Fishing Vessel, 443 U.S. at 686, 99 S.Ct. 3065. The Tribes get only fifty percent of the catch even if the supply of salmon is insufficient to provide a moderate living. However, there is nothing in the Court’s opinion that authorizes the State to diminish or eliminate the supply of salmon available for harvest.
It is undisputed that at the present time fifty percent of the harvestable salmon in Puget Sound does not provide a moderate living to the Tribes. It is also undisputed that the State has acted affirmatively to build roads with barrier culverts that block the passage of salmon, with the consequence of substantially diminishing the supply of harvestable salmon. Evidence at trial showed that remediation of the State’s barrier culverts will increase the yearly supply of salmon by several hundred thousand adult salmon. Half of the newly produced harvestable salmon will be available to the Tribes. The other half will be available to non-Indians.
Our opinion does not hold that the Tribes are entitled to enough salmon to provide a moderate living, irrespective of the circumstances. We do not hold that the Treaties’ promise of a moderate living is valid against acts of God (such as an eruption of Mount Rainier) that would diminish the supply of salmon. Nor do we hold that the promise is valid against all human-caused diminutions, or even against all State-caused diminutions. We hold only that the State violated the Treaties when it acted affirmatively to build roads across salmon bearing streams, with culverts that allowed passage of water but not passage of salmon.
II. Effect and Scope of the Holding
Second, our colleagues contend that our decision may open the door to “a whole host of future suits,” and that we do “nothing to cabin [our] opinion.” We are not sure what the hypothesized future suits would be. But we are sure that we have not opened the floodgates to a host of future suits.
Because of the Eleventh Amendment, a further suit against Washington State seeking enforcement of the Treaties cannot be brought by the Tribes. Nor can it be brought by non-Indians who would benefit from an increase in harvestable salmon (recall that 50% of any increased salmon harvest will go to non-Indians). Nor can it be brought by environmental groups. The *1021only possible plaintiff is the United States. The United States is a responsible litigant and is not likely to burden the States ■without justification. The history of this litigation demonstrates that it was no easy thing for the Tribes to persuade the United States to institute proceedings against the state of Washington to seek remediation of the State’s barrier culverts, and will be no easy thing for other Northwest tribes to persuade the United States to bring comparable suits: against other States.
Our opinion describes the facts of this litigation carefully and in detail, as required by our decision in United States v. State of Washington, 759 F.2d 1353, 1357 (9th Cir. 1985) (en banc) (“[T]he measure of the State’s [Treaty] obligation will depend for its precise legal formulation on all of the facts presented by a particular dispute.”). Cabining our opinion by means other than a careful, detailed description of the facts presented would have entailed positing hypothetical facts 'in cases not before us and giving an improper advisory opinion. On the facts presented to us, we held that the State violated the Treaties when it acted affirmatively to block salmon-bearing streams by building roads with culverts that protected the State’s roads but killed the Tribes’ salmon. Other cases with different facts might come out differently, but we did not decide—and should not have decided—such cases.
III. Laches
' Third, our colleagues contend that the United States’ suit on behalf of the Tribes is barred by laches. There is an established line of cases holding that the United States cannot, based on laches or estoppel, render unenforceable otherwise valid Indian treaty rights. Our colleagues contend that these cases have been overruled by City of Sherrill v. Oneida Indian Nation of N.Y., 544 U.S. 197, 125 S.Ct. 1478, 161 L.Ed.2d 386 (2005), and that laches applies here.
This contention is belied by Sherrill itself. In 1788, the Oneida Indian Nation (“OIN”), located in New York State, had a reservation of 300,000 acres. By 1920, the OIN had sold off all but 32 acres. In 1985, the Supreme Court held that the sale of OIN lands had been illegal, and that the OIN was entitled to monetary compensation for the sales. County of Oneida v. Oneida Indian Nation of N.Y., 470 U.S. 226, 105 S.Ct. 1245, 84 L.Ed.2d 169 (1985). The OIN subsequently bought two' parcels of land -within the boundaries of its ancestral reservation. The parcels had been sold to a non-Indian in 1807. The OIN asserted that the repurchased parcels were sovereign tribal property and therefore free from local taxation. The Supreme Court disagreed. It wrote, “[T]he'Tribe cannot unilaterally revive its ancient sovereignty ... over the parcels at issue. The Oneidas long ago relinquished the reins of government and cannot regain them through open market purchases from current titleholders.” Sherrill, 544 U.S. at 203, 125 S.Ct. 1478.
.The case before us is different from Sherrill. The question in our ease is not whether, as in Sherrill, a tribe can reassert sovereignty, over land within the boundaries of an abandoned reservation. The Tribes have not abandoned their reservations. Nor is the question whether, as in Sherrill, the Tribes have acted to relinquish their rights under the Treaties. The ■Tribes have done nothing to authorize the State to construct and maintain barrier culverts. Nor, finally, is the question whether,- as in Sherrill, to allow the revival of disputes or claims that have long been dormant. Washington and. the Tribes have been in a continuous state of conflict over treaty-based fishing rights for well over one hundred years.
*1022IV. Breadth of the Injunction
Fourth, our colleagues contend that the injunction is overbroad. The United States requested an injunction that would have required the remediation, of all of the State’s barrier culverts within five years. The district court declined that request. Instead, it issued a nuanced injunction requiring the remediation of some, but not all, of the barrier culverts within seventeen years.
Briefly stated, the injunction provides as follows. The only seriously debated culverts are those under the control of the Washington State Department of Transportation (“WSDOT”). The court ordered the State to prepare a list of all of WSDOT barrier culverts within the area covered by the Treaties. In Paragraph 6 of the injunction, the court ordered WSDOT to provide, within seventeen years, fish passage for each barrier culvert with more than 200 linear meters of accessible salmon habitat upstream to the first natural passage barrier. In Paragraph 7, the court ordered WSDOT to replace existing barrier culverts above which there was less than 200 linear meters of upstream accessible salmon habitat only at the “end of the useful life” of the culverts, or sooner “as part of a highway project.” In Paragraph 8, the court allowed WSDOT to defer correction of some of the culverts described in Paragraph 6. Deferred culverts can account for up to ten percent of the total accessible upstream habitat from the culverts described in Paragraph 6. WSDOT can choose which culverts to defer, after consulting with the United States and the Tribes. Culverts deferred under Paragraph 8 need only be replaced on the more lenient schedule specified in Paragraph 7.
The injunction thus divided WSDOT barrier culverts into two categories. High priority category culverts must be remedi-ated within seventeen years. Low priority category culverts must be remediated only at the end of the natural life of the existing culvert, or in connection with a highway project that would otherwise require replacement of the culvert. Deferred culverts in the high priority category (culverts blocking a total of ten percent of the accessible upstream habitat above all the high priority culverts) can be remediated on the schedule of low priority culverts.
In identifying the State’s barrier culverts and sorting them into the two categories, the district court focused on the amount of available upstream spawning habitat before encountering a natural barrier. Culverts with more than 200 linear meters of accessible upstream habitat are in the high category; culverts with less than 200 meters are in the low category. The court ignored the existence of man-made barriers, including those downstream of the State’s barrier culverts. In so doing, the court followed the methodology of the State in identifying and prioritizing culverts that should be remediated. The State could have objected to the court’s reliance on its own methodology, but it did not do so.
There were good reasons for the district court to ignore, for purposes of its injunction, the existence of downstream barriers. The most obvious reason is the following: The State identified a total of 817 state-owned barrier culverts, including both high and low priority culverts. On streams yvhere there are both state and non-state barrier culverts, there are 1,590 non-state culverts. Of those, 1,370 are upstream of the state culverts; only 220 are downstream. Of those 220 downstream culverts, 152 allow partial passage of salmon; only 68 entirely block passage.
Even if we were to make the assumption that all 817 of the identified barrier culverts are high priority culverts (which they clearly are not), state-provided documents introduced at trial showed that roughly 230 of them—more than all of the 220 non-*1023state downstream culverts combined— need not be remediated within seventeen years. They may be deferred and need be remediated only at the end of their natural life or in connection with an independently undertaken highway project. Further, Washington law already imposes some obligation on the part of owners of non-state barrier culverts to repair or replace them, at their own expense, to allow fish passage.
Our dissenting colleagues emphasize the high cost of complying with the injunction. Our colleagues, like the State, exaggerate the cost. The State claimed in its brief to us that compliance with the injunction will cost a total of $1.88 billion. Our colleagues highlight that figure at the beginning of their dissent. There is no plausible basis for the State’s claim of $1.88 billion. We analyze the evidence in detail in our opinion, to which we refer the reader. For present purposes, it is sufficient to note, as we point out in our opinion, that ‘Washington’s cost estimates are not supported by the evidence.” United States v. Washington, 858 F.3d at 976.
* * *
In sum, the district court properly found that Washington State violated the Treaties by acting affirmatively to build state-owned roads, and to build and maintain salmon-blocking culverts under those roads. By allowing passage of water, the culverts protect the State’s roads. But by not allowing passage of fish, the culverts kill the Tribes’ salmon. There is ample evidence in the record that remediation of the State’s barrier culverts will have a substantial beneficial effect on salmon populations, resulting in more harvestable salmon for the Tribes. As an incidental result, there will also be more harvestable salmon for non-Indians. The United States requested an injunction requiring remediation of all of the State’s barrier culverts within five years. The district court crafted a careful, nuanced injunction, giving the United States much less than it requested. We unanimously concluded that the district court properly found a violation of the Treaties by the State, and that it acted within its discretion in formulating its remedial injunction.
O’SCANNLAIN, Circuit Judge,* with whom KOZINSKI, TALLMAN, CALLAHAN, BEA, IKUTA, and N.R. SMITH, Circuit Judges, join, and with whom BYBEE and M. SMITH, Circuit Judges, join as to all but Part IV, respecting the denial of rehearing en banc:Fashioning itself as a twenty-first century environmental regulator, our court has discovered a heretofore unknown duty in the Stevens Indian Treaties of 1854 and 1855. The panel opinion in this case enables the United States, as a Treaty signatory, to compel a State government to spend $1.88 billion1 to create additional salmon habitat by removing or replacing culverts2 under state-maintained highways *1024and roads, wherever found. Pacific Northwest salmon litigation has been -ongoing for almost fifty years,3 has been before our court multiple times, and has been up to and down from the Supreme Court. Nonetheless, it apparently just occurred to the Tribes, the United States, and our court that in order to fulfill nineteenth century federal treaty obligations, the State of Washington must now be required to remove physical barriers which might impede the passage of salmon. See Washington V, 853 F.3d at 966.
Given the significance of this case—both in terms of dollars and potential prece-dential effect—it seemed the ideal candidate for en banc review and, hopefully, correction on the merits. But rather than reining in a runaway decision, our court has chosen to do nothing—tacitly affirming the panel opinion’s erroneous reasoning.
With utmost respect, I believe our court has made a regrettable choice. .
I
In reaching its conclusion, the panel opinion makes four critical errors.
First, it misreads Washington v. Washington State Commercial Passenger Fishing Vessel Association (“Fishing Vessel”), 443 U.S. 658, 99 S.Ct. 3055, 61 L.Ed.2d 823 (1979), as requiring Washington to ensure that there are a certain “number of fish” available for the Tribes, “sufficient to provide a ‘moderate living;”’ Washington V, 853 F.3d at 965 (quoting Fishing Vessel, 443 U.S. at 686, 99 S.Ct. 3055).
Second, by holding that' culverts need to be removed because they negatively impact the fish population, the panel opinion sets up precedent that could be used to challenge activities that affect wildlife habitat in other western states, which led Idaho and Montaná to join Washington in requesting rehearing. The panel opinion fails to articulate a limiting legal principle that will prevent its holding from being used to attack a variety of development, construction, and farming practices, not just in Washington but throughout the Pacific Northwest.
Third, the panel opinion contravenes City of Sherrill v. Oneida Indian Nation of New York, 544 U.S. 197, 125 S.Ct. 1478, 161 L.Ed.2d 386 (2005), by refusing to apply the doctrine of laches to the United States,
Fourth, the panel opinion upholds an injunction that is overbroad—requiring the State to spend millions of dollars on repairs that will have no immediate effect on salmon habitat.
II
The Stevens Treaties4 provide that “(t]he right of taking fish, at all usual and accustomed grounds and stations, is fur*1025ther secured to said Indians, in common with all citizens of the Territory.” Fishing Vessel, 443 U.S. at 674, 99 S.Ct. 3066. The precise contours of this guarantee remain hotly contested but were most fully addressed by the Supreme Court’s opinion in Fishing Vessel.
A
The panel opinion reads language in Fishing Vessel as requiring that there be enough fish to provide a “moderate living” for the Tribes. See Washington V, 853 F.3d at 965-66. It is true that the Court stated that “Indian treaty rights to a natural resource [i.e. fish] ... secures so much as, but no more than, is necessary to provide the Indians with a livelihood—that is to say, a moderate living.” Fishing Vessel, 443 U.S. at 686, 99 S.Ct. 3055. In isolation, this statement might be read as guaranteeing the Tribes a certain number of fish, hut only if one ignores the rest of the opinion. In Fishing Vessel, the Supreme Court adopted the United States’ position that the Treaties entitled the Tribes “either to a 60% share of the ‘harvestable’ fish” passing through their fishing grounds “or to their needs, whichever was less.” Id. at 670, 99 S.Ct. 3055 (emphasis added); see also id, at 685-86, 99 S.Ct. 3055.
Thus, notwithstanding the significance of fish to the Tribes, the Court recognized' that “some ceiling should be placed on the Indians’ apportionment to prevent- their needs from exhausting the entire resource and thereby frustrating the treaty right of ‘all [other] citizens of the Territory.’ ” Id. at 686, 99 S.Ct. 3055. The Court ruled that 50%.of the available fish was the appropriate limit. See id. (“[T]he 50% figure imposes a maximum ... allocation.”) (“[T]he maximum possible allocation to the Indians is fixed at 50%.”); id. at 686, 99 S.Ct. 3055 n.27 (“Because the 50% figure is only a ceiling, it is not correct to characterize our holding as ‘guaranteeing the Indians a specified percentage’ of the fish.”).
Such ceiling makes intuitive sense. With or without pre-existing barriers, the population of fish varies dramatically from year to year and season to season. In a year with a low run of fish, absent a ceiling, the Tribes’ needs could easily predominate, leaving few fish for other citizens. Thu.s, to protect the rights of all parties, to the Treaties, the Court imposed a 60% ceiling.
Since the fish population varies, however, the presence of the ceiling necessarily entails that the Tribes may not always receive enough fish to provide a “moderate living,” Indeed, the Court emphasized that the Treaties secured to the Tribes “a fair share of the available fish,” rather than a certain number of fish. Id. at 685, 99 S.Ct. 3055 (emphasis added). The total number of fish that the Tribes receive indubitably will vary with the run of fish. See id. at 679, 99 S.Ct. 3055 (observing that the Treaties “secure the Indians’ right to take a share of each run of fish that passes through tribal fishing areas” (emphasis added)); id. at 687, 99 S.Ct. 3055 (discussing the “50% allocation of an entire run that passes through ... customary fishing grounds”).
Thus, by imposing a percentage ceiling tied to the relevant run rather than a fixed numerical floor, the Court rejected the proposition that the Tribes were entitled to a certain number of fish. Indeed, “while the maximum possible allocation to the Indians is fixed at 50%, the minimum is not; the latter will, upon proper submissions to the District Court, be modified in response to changing circumstances.”5 Id. *1026at 686-87, 99 S.Ct. 3055. Our court has confirmed this holding multiple times.
In United States v. Washington, 759 F.2d 1353, 1359 (9th Cir. 1985) (“Washington III”), our en banc court explained:
[T]he Supreme Court in Fishing Vessel did not hold that the Tribes were entitled to any particular minimum allocation of fish. Instead, Fishing Vessel mandates an allocation of 50 percent of the fish to the Indians, subject to downward revision if moderate living needs can be met with less. The Tribes have a right to at most one-half of the harvesta-ble fish in the case area.
Id. (emphasis added). Likewise in Midwater Trawlers Co-operative v. Department of Commerce, 282 F.3d 710, 719 (9th Cir. 2002), we observed that under Fishing Vessel, the Makah Tribe was entitled “to one-half the harvestable surplus of Pacific whiting that passes through its usual and accustomed fishing grounds, or that much of the harvestable surplus as is necessary for tribal subsistence, whichever is less.” Id. (emphasis added). Most recently in Skokomish Indian Tribe v. United States, 410 F.3d 506, 513 (9th Cir. 2005), our en banc court again described Fishing Vessel as holding that the Tribes were “entitled to an equal measure of the harvestable portion of each run that passed through a ‘usual and accustomed’ tribal fishing ground, adjusted downward if tribal needs could be satisfied by a lesser amount.” Id. (emphasis added) (quoting Fishing Vessel, 443 U.S. at 685-89, 99 S.Ct. 3055).
By holding that the Treaties guarantee “that the number of fish would always be sufficient to provide a ‘moderate living’ to the Tribes,” Washington V, 853 F.3d at 965 (emphasis added), the panel opinion turns Fishing Vessel on its head. It imposes an affirmative duty upon the State to provide a certain quantity of fish, which reads out the 50% ceiling entirely.
Instead, the panel opinion ignores the 50% ceiling, effectively adopting the position urged by the Tribes in Fishing Vessel that “the treaties had reserved a pre-exist-ing right to as many fish as their commercial and subsistence needs dictated.” 443 U.S. at 670, 99 S.Ct. 3055. Yet, as explained, the Supreme Court has already rejected this approach, following instead the United States’ position that the Tribes were guaranteed the lesser of then- needs or 50% of the available run. See id. at 670, 685, 99 S.Ct. 3055. Likewise, our court has rejected interpretations of Fishing Vessel that would entitle the Tribes to a “particular minimum allocation of fish.” Washington III, 759 F.2d at 1359. The panel opinion’s holding misconstrues not only the Supreme Court’s decision in Fishing Vessel but also our decisions in Washington III, Midwater Trawlers, and Skokomish Indian Tribe.
B
To reach its conclusion, the panel points to various statements allegedly made by Governor Stevens to the Tribes at the time the Treaties were negotiated in the 1850s. Washington V, 853 F.3d at 964-65. As the Supreme Court observed in Fishing Vessel, however, “[bjecause of the great abundance of fish and the limited population of the area, it simply was not contemplated that either party would interfere with the other’s fishing rights.” 443 U.S. at 668, 99 S.Ct. 3055. Indeed, the Supreme Court considered the very same statements in Fishing Vessel yet still chose to impose a 50% cap on the Tribes’ share of available fish. See id. at 666-68 & nn. 9 & 11, 99 S.Ct. 3055.6 Such cap necessarily means that the Tribes are not always guaranteed *1027enough fish to meet their needs. If the Supreme Court considered Stevens’ statements and declined to find that the Tribes were entitled to a certain minimum quantity of fish, it eludes me how a panel of our court can reach the opposite conclusion by relying on these statements now. The panel opinion utterly fails to grapple with the 50% cap imposed by Fishing Vessel.
The panel opinion further cites to the Supreme Court’s opinion in Winters v. United States, 207 U.S. 564, 576-77, 28 S.Ct. 207, 52 L.Ed. 840 (1908), and our opinion in United States v. Adair, 723 F.2d 1394, 1409, 1411 (9th Cir. 1983), as supporting its conclusion that the Stevens Treaties guarantee the Tribes a specific quantity of fish. Yet, neither Winters nor Adair is factually relevant. Each involved the question of whether certain tribes were entitled to various water rights on their reservations under the treaties creating the reservations.
In Winters, the Supreme Court held that the lands ceded to create the Fort Belknap Indian Reservation necessarily included the water rights accompanying such lands. See 207 U.S. at 565, 576-77, 28 S.Ct. 207. Likewise in Adair, we held “that at the time the Klamath Reservation was established, the [United States] and the Tribe intended to reserve a quantity of the water flowing through the reservation.” 723 F.2d at 1410. Thus, both cases stand for the somewhat unremarkable proposition that in the context of Native American reservations, water rights accompany land rights.
It is true that both cases found water rights that were not explicitly detailed in the text of the treaties. Nonetheless, if we read these cases broadly to mean that we can and should infer a whole host of rights not contained in the four corners of tribal treaties, the possibilities are endless. Since the Supreme Court made it plain in Fishing Vessel that the Tribes are not entitled to a certain numerical amount of fish, we certainly should not rely on Winters and Adair to hold otherwise.
Ill
Even if one agrees with the panel opinion that the Tribes are entitled to a specific quantity of fish, however, it does not necessarily mean that the installation and maintenance of culverts run afoul of the Treaties. But assuming that they do, it is far from clear that the drastic remedy of removal or repair should be required.
A
Before reaching its conclusion that the State violated the Treaties, the panel opinion devotes minimal treatment to showing (1) that tribal members would engage in more fishing if there were more salmon and (2) that removing culverts would increase this salmon population. See Washington V, 853 F.3d at 966 (devoting three paragraphs to these issues).7 The panel opinion acknowledges that the State of Washington was not intentionally trying to impact the fish population when it installed culverts under state highways and other roads.8 Id. Nonetheless, the panel opinion *1028concludes that because there was evidence .that culverts affect fish population, and because the fish population is low, the State violated the Treaties by building and maintaining its culverts. See id.
This overly broad reasoning lacks legal foundation. There are many factors that affect fish population and multiple fish populations that are low.9 Is any surface physical activity, wherever found, that negatively affects fish habitat an automatic Treaty violation? If so, the panel’s opinion could open the door to a whole host of future suits.
While such speculation may sound farfetched, in actuality, it is already occurring. Legal commentators have noted that plaintiffs could use the panel’s decision to demand the removal of dams and attack a host oí other practices that can degrade fish habitat (such as logging, grazing, and construction).10 The panel does nothing to cabin its opinion. Nor does it provide any detail for how to determine if a fish population has reached an appropriate size, making further remedial efforts unnecessary.
B
Furthermore, the future reach of this decision extends far beyond the State of Washington. As the. amici observe, the same fishing rights are reserved to tribes in Idaho, Montana, and Oregon. Further, the Stevens Treaties also guarantee the Tribes the privilege of hunting. See Fishing Vessel, 443 U.S. at 674, 99 S.Ct. 3055. There seems little doubt that future litigants will argue that the population of various birds, deer, elk, bears, and similar animals, which were traditionally hunted by the Tribes, have been impacted by Western development. If a court subsequently concludes that hunting populations are covered by the reasoning of this decision, the potential impact of this case is virtually limitless.
C
Yet, our court has already held that the Stevens Treaties cannot be used to attach broad “environmental servitudes” to the land. See United States v. Washington, 694 F.2d 1374, 1381 (9th Cir. 1982) (coining the term “environmental servitude”), vacated on reh’g, Washington III, 759 F.2d at 1354-55 (but reaching similar result). Thus, in Washington III, our en banc court vacated a declaratory judgment from the district court which held “that the treaties impose upon the State a corresponding duty to refrain from degrading or authorizing the degradation of the fish habitat to an extent that would deprive the treaty Indians of their moderate living needs.” 759 F.2d at 1355, vacating United States v. Washington, 506 F.Supp. 187, 208 (W.D. Wash. 1980) (“Washington IF), While the panel’s opinion here deals with .the specific issue of culverts, its reasoning is .not so confined; it effectively imposes the same boundless standard ■ upon the State—preventing habitat- degradation— that we rejected in Washington III. .
D -
Once a court has decided that there has been a violation, it must address the remedy. The panel opinion acknowledges “that *1029correction of barrier culverts is only one of a number of measures that can usefully be taken to increase, salmon -production.”11 Washington V, 853 F.3d at 974. And, the panel opinion further concedes “that the benefits of culvert correction differ depending on the culvert in question.” Id. Yet, if culverts are only one “measure” that could affect the salmon population, what about the other, measures? Why is it appropriate to require the State to correct culverts rather than something else? Since, at some level, almost all urban growth can impact fish populations, should the State be required to reverse decades of development in an effort to increase the number of fish? Is the answer that any activity that amounts to a Treaty violation must be halted or removed? The panel opinion offers no cost-benefit analysis, or any other framework, to guide future courts on what is an appropriate remedial measure (and what is not).12
In effect, the panel’s decision opens a backdoor to a whole host of potential federal environmental regulation-making. And, it invites courts, who have limited expertise in this area, to serve as policymakers.
But the issues at the heart of this suit— development versus wildlife habitat, removal versus accommodation—are properly left to the political process. Judges are ill-equipped to evaluate these questions. We deal in closed records and have difficulty obtaining and evaluating on-the-ground information—for example, which culverts it would be most cost-effective to remove over the next seventeen years.
Here, the State recognizes that “[s]alm-oñ are vital to Washington’s economy, culture, and diet.” Prior to the injunction, the State was already working to address problematic culverts, and the State has spent “hundreds of millions of dollars” on programs designed “to preserve and restore salmon runs.” There is no justification for interfering with the State’s existing programs.
IV
Notably, the panel opinion does not prohibit the State from installing future culverts. Instead, it orders the State to correct existing culverts. See Washington V, 853 F.3d at 979-80. Yet, according to the State,-it was the federal .government,, now bringing suit in its capacity as trustee for the Tribes, which “specified the design for virtually all of the culverts at issue.” Further, these culverts have been in place for many decades. According to the State, “Washington’s state highway system has been essentially the same size since the 1960’s,” and thus presumably many culverts predated this litigation, which has been ongoing for almost fifty years. Apparently, however, no one thought that the culverts might be a problem until 2001 when the Tribes filed a request for determination that such pre-existing barriers were infringing the Treaties. See Washington V, 853 F.3d at 954.
*1030Given the United States’ involvement in designing the culverts and its long acquiescence in their existence, one might suppose that an equitable doctrine such as laches would bar suit by the United States. Indeed, “[i]t is well established that laches, a doctrine focused on one side’s inaction and the other’s legitimate reliance, may bar long-dormant claims for equitable relief.” City of Sherrill, 544 U.S. at 217, 125 S.Ct. 1478.
According to the panel opinion, however, “[t]he United States cannot, based on lach-es or estoppel, diminish or render unenforceable otherwise valid Indian treaty rights.” Washington V, 853 F.3d at 967. The panel opinion cites several cases for this proposition, including the 1923 opinion of Cramer v. United States, 261 U.S. 219, 234, 43 S.Ct. 342, 67 L.Ed. 622 (1923) (holding that a government agent’s unauthorized acceptance of leases of tribal land could not bind the government or tribe), and United States v. Washington, 157 F.3d 630, 649 (9th Cir. 1998) (“Washington IV’) (“[Ljaches or estoppel is not available to defeat Indian treaty rights.”). See Washington V, 853 F.3d at 967.
Yet, the panel opinion’s rejection of lach-es contravenes the Supreme Court’s subsequent 2005 decision in City of Sherrill, 544 U.S. at 221, 125 S.Ct. 1478. That case involved an attempt by the Oneida Indian Nation to reassert sovereignty over newly-purchased land that had once belonged to the Nation but had been sold in contravention of federal law (although with the apparent acquiescence of federal agents) approximately two hundred years before. Id. at 203-05, 211, 125 S.Ct. 1478. In particular, the Nation sought to avoid local regulatory control and taxation of its newly-purchased parcels. Id. at 211, 125 S.Ct. 1478.
The Supreme Court analogized the situation to a dispute between states, explaining that “long acquiescence may have con-
trolling effect on the exercise of dominion and sovereignty over territory.” Id. at 218, 125 S.Ct. 1478. The Court further “recognized the impracticability of returning to Indian control land that generations earlier passed into numerous private hands.” Id. at 219, 125 S.Ct. 1478. Therefore, the Court concluded, “the Oneidas’ long delay in seeking equitable relief against New York or its local units, and developments in the city of Sherrill spanning several generations, evoke the doctrines of laches, acquiescence, and impossibility, and render inequitable the piecemeal shift in governance this suit seeks unilaterally to initiate.” Id. at 221, 125 S.Ct. 1478.
Thus, Sherrill indicates that our court’s previous holding in Washington IV, 157 F.3d at 649, that laches cannot be used “to defeat Indian treaty rights” is wrong and impliedly overruled. Cf. Miller v. Gammie, 335 F.3d 889, 900 (9th Cir. 2003). The Second Circuit has recognized as much, observing that Sherrill “dramatically altered the legal landscape” by permitting “equitable doctrines, such as laches, acquiescence, and impossibility” to “be applied to Indian land claims.” Cayuga Indian Nation v. Pataki, 413 F.3d 266, 273 (2d Cir. 2005).
Yet, the panel opinion blindly cites Washington IV and sidesteps the central tenet of Sherrill by attempting to distinguish it on its facts. See Washington V, 853 F.3d at 967-68. The panel opinion tries to draw three distinctions: (1) this case does not involve the question of whether the Tribes can regain sovereignty over abandoned land; (2) the Tribes never authorized the design or construction of the culverts; and (3) the Tribes are not trying to revive claims that have lain dormant. Id. at 968.
The first distinction is irrelevant; since Sherrill made clear that laches can apply to Indian treaty rights, it should not mat*1031ter whether a party is seeking to apply laches in the context of sovereignty over land or the enforcement of rights appurtenant to land (the ability to fish).
Second, as Montana and Idaho observe, it does not matter that the Tribes never authorized the design or construction of the culverts because Washington is seeking to impose the doctrine of laches against the United States, not the Tribes. And, as the Second Circuit has made plain, the logic of Sherrill applies to the United States when it is acting as trustee for the Tribes. See Oneida Indian Nation v. Cty. of Oneida, 617 F.3d 114, 129 (2d Cir. 2010).
Notably, only the United States could bring suit against Washington for alleged culvert violations because Washington is protected by sovereign immunity against suit from the Tribes. See Idaho v. Coeur d’Alene Tribe of Idaho, 521 U.S. 261, 268, 117 S.Ct. 2028, 138 L.Ed.2d 438 (1997). The panel opinion asserts that the United States cannot waive treaty rights, and this may be true as a general matter. Washington V, 853 F.3d at 967. Nonetheless, in the context of specific litigation, since the United States acts as the Tribes’ trustee, such representation necessarily entails the ability to waive certain litigation rights (failing to bring a claim within the statute of limitations for example). Thus, the fact that the Tribes did not authorize the culverts is irrelevant; the United States did, and it further failed to object to the culverts for many years.
Finally, I disagree with the panel opinion’s assertion that the United States is not trying to revive claims that have lain dormant. Presumably, the State’s alleged violation of the Treaties was complete when it constructed the culverts (and relevant highways) in the 1960s. The United States first brought suit to enforce the Tribes’ fishing rights in 1970. Washington V, 853 F.3d at 958. Yet, the United States found no problem with the culverts until 2001. While the claims did not lie dormant for 200 years as in Sherrill, they were dormant for over 30 years. And as in Sher-rill, there are significant practical issues involved with asserting the claims now such as the time, expense, and efficacy of removing the culverts. See 544 U.S. at 219, 125 S.Ct. 1478.
Thus, while Sherrill may be factually distinct, it is also directly on point. The panel opinion errs by ignoring its central teaching. There is good reason to contend that the United States is barred from bringing this suit by the doctrine of laches. And, if the United States is barred from suit, the entire suit is prohibited, since the Tribes cannot puncture the State’s defense of sovereign immunity on their own. See Coeur d’Alene Tribe, 521 U.S. at 268, 117 S.Ct. 2028.
Rather than taking the opportunity to harmonize our precedent, the panel opinion ignores the changes wrought by Sher-rill, defying the Supreme Court’s direction.
V
Even if one concludes (1) that the Treaties guarantee the Tribes enough fish to sustain a “moderate living,” (2) that violation of such guarantee can and should be remedied by removing culverts, and (3) that the suit is not barred by the doctrine of laches, there is still good reason to reject the injunction itself as overbroad. As the State explains, the injunction requires it to replace or repair all 817 culverts located in the area covered by the Treaties without regard to whether replacement of a particular culvert actually will increase the available salmon habitat.
In addition to state-owned culverts, there are a number of other privately-owned culverts and barriers on the streams in question which are not covered by the injunction. Where there are non-*1032state-owned culverts blocking fish passage downstream or immediately upstream from state-owned culverts, replacement of the State’s culverts will make little or no difference on available, salmon habitat. Indeed, the State observes that
(1) roughly 90% of state barrier culverts are upstream or downstream of other barriers ... (2) state-owned culverts are less than 25% of known barrier culverts .,. and (3) in many watersheds, non-state barrier culverts drastically exceed state-owned culverts, by up to a factor of 36 to 1[J
The panel attempted to address this issue in its revised opinion. First, the opinion quotes testimony from a former State employee stating that Washington itself does not take into account the presence of non-state-owned barriers when calculating the priority index for which culverts to address. Washington V, 853 F.3d at 973. What the opinion does not reveal, however, is that this same expert also testified that correcting ■ state-owned culverts that are downstream from non-state barriers “generally” will not have an immediate impact or benefit on salmon habitat. And, according to the State of Washington, the. priority index, notwithstanding its name, typically does not dictate which barriers the State addresses first; instead the State focuses on culverts in streams without barriers.
Next, the panel opinion points out that Washington law requires dams or other stream obstructions to include a fishway and observes that the State may take corrective action against private owners who fail to comply with this obligation. Washington V, 853 F.3d at 973 (quoting Wash. Rev, Code Ann. § 77.57.030(1)-(2)). Yet, what the panel opinion fails to disclose is that this law only went into effect in 2003 and specifically “grandfathered in” various obstructions that were installed before May 20, 2003. Wash. Rev. Code Ann. § 77.57.030(3). Presumably, some of the non-state barriers would fall under this exception.
Finally, the panel opinion observes that [I]n 2009, on streams where there were both state and non-state barriers, 1,370 of the 1,590 non-state barriers, or almost ninety percent, were upstream of the state barrier- culverts. Sixty nine percent of the- 220 downstream non-state barriers allowed partial passage of fish. Of the. 152 that allowed partial passage, “passability” was 67% for 80 of the barriers and 33% for 72 of them. - . ■■■■
Washington V, 853 F.3d at 973.
Given the significant cost of replacing barriers, however, being forced to replace even a single barrier that will have no tangible impact on thé salmon population is an unjustified burden. Even using the most conservative estimates found by the district court, the average cost of replacing a single culvert is between $658,639 and $1,827,168. Washington V, 853 F.3d at 976.13 We do not know the precise number of state-owned culverts that are located above non-state-owned culverts which prevent all fish passage. Yet, considering that there are at least sixty-eight ■ non-state-owned barriers blocking all passage downstream from state-owned culverts,14 there are almost certainly more than one or two culverts whose replacement would have no impact whatsoever on salmon habitat. The *1033panel’s opinion utterly fails to explain why the State should waste millions' of dollars on such culverts in particular.
Further, even if the majority of non-state barriers are upstream, the court should still take into account the location of these barriers. As noted, if a non-state upstream barrier is close to or immediately above a state barrier, replacing the state barrier will have little effect on the size of salmon habitat, but it will come at a significant cost to the State.
The panel opinion observes that the injunction offers the State a longer schedule for replacing barriers that will open up less habitat. See Washington V, 853 F.3d at 974-75. It may be advantageous to the State to have the cost spread out over a longer time period, but whether it occurs five years or twenty-five years from now, the panel opinion fails to explain why taxpayers should be required to replace barriers that will not change the available salmon habitat.15
Thus, significant overbreadth problems remain. There is no doubt that the record in this ease is voluminous and pinpointing the specific culverts whose removal might actually impact the available salmon habitat is an arduous task. Both the panel and district court made a valiant effort to wade through the many pages of maps and statistics.16 As it currently stands, however, the injunction is unsupportable.
VI
In sum, there were many reasons. to rehear this case en banc. The panel opinion’s reasoning ignores the Court’s holding in Fishing Vessel and our own cases, is incredibly broad, and if left unchecked, could significantly affect natural -resource management throughout the Pacific Northwest, inviting judges to become environmental regulators. By refusing to consider the doctrine of laches, the panel, opinion further disregards the Supreme ■Court’s decision in Sherrill, relying instead on outdated and impliedly overruled precedent from our court. Finally, the panel opinion imposes a poorly-tailored injunction which will needlessly cost the State hundreds of millions of dollars.
Rather than correcting these errors, our court has chosen the path of least resistance. We should have reheard this case en banc.
Separate Statement of HURWITZ, Circuit Judge:
The dissent from the denial of rehearing en banc unfortunately., perpetuates the false notion that the full court’s refusal to exercise its discretion under Federal Rule of Appellate Procedure 35(a) is tantamount to the court “tacitly affirming the panel opinion’s erroneous reasoning.” This effectively rewrites Rule 35(a). The Rule is entirely discretionary, providing that the court “may order” rehearing en banc, and cautioning that such an order “is not favored” and is reserved for “a question of exceptional importance” or “to secure or maintain uniformity' of the court’s decisions.”
Like the denial of certiorari by the Supreme Court, the denial of rehearing en banc simply leaves a panel decision undis*1034turbed. There are at least as many valid reasons for a circuit judge to decide not to vote to rehear a case en banc as there are for a Supreme Court justice to decide not to vote to grant certiorari. Indeed, there is at least one additional reason—Supreme Court review remains available to the losing litigant in our court, so it is not necessary that each of us have the last word on every case. No one would suggest that when the Supreme Court exercises its discretion not to grant certiorari, it is “tacitly affirming” the decision below. No different legal or factual conclusion can be made here.
Judges on our court—even those who cannot participate in the voting—are entirely free to criticize the court’s failure to grant rehearing en banc and express their own views as to why a panel decision is incorrect. But it is not correct to impute hidden meanings to the discretionary decisions of others. When a judge chooses not to indicate views on the merits of a controversy, colleagues should not invent them.
District Judge Ezra was a member of the three-judge panel that decided this case. Because Judge Ezra is not a member of the Ninth Circuit, he does not have the authority to vote on a petition for rehearing en banc.
As a judge of this court in senior status, I no longer have the power to vote on calls for rehearing cases en banc or formally to join a dissent from failure to rehear en banc. See 28 U.S.C. § 46(c); Fed. R. App. P. 35(a). Following our court’s general orders, however, I may participate in discussions of en banc proceedings. See Ninth Circuit General Order 5.5(a).
. According to the State's estimate. There is a dispute about the actual cost of the injunction, but even using the more conservative estimates on which the district court relied, the cost of replacing all 817 culverts ranges from $538 million to $1.5 billion (the average cost of replacing a culvert was $658,639 to $1,827,168). See United States v. Washington, 853 F.3d 946, 976 (9th Cir. 2017) ("Washington V”).
. A culvert is "[a] tunnel carrying a stream or open drain under a road or railway.” Culvert, OxfordDictionaries. com, https ;//en. oxford dictionaries.com/definition/culvert (last visited April 29, 2017).
. Five iterations of the United States v. Washington litigation, including this case, which is referred to as Washington V, are mentioned herein and are referred to as Washington I, Washington II, etc.
. The Treaties are a series of Senate-ratified agreements between the United States and various Indian tribes that were negotiated in the 1850s by Isaac Stevens, then-federal Governor and Superintendent of Indian Affairs of the Washington Territory (pre-statehood), under which the Tribes agreed to give up land in exchange for monetary payments. Fishing Vessel, 443 U.S. at 661-62, 666, 99 S.Ct. 3055, The Treaties contained clauses reserving the Tribes’ right to fish on ceded land. See, e.g., Treaty of Medicine Creek, 10 Stat. 1132 (1854). Beginning with U.S. District Court Judge George Boldt’s decision in 1974, United States v. State of Washington, 384 F.Supp. 312 (W.D. Wash. 1974) ("Washington /”), the contours of these fishing rights have been the subject of extensive litigation before the district court, our court, and the Supreme Court and tumultuous protests by the people impacted by these decisions,
. Such changing circumstances include the Tribes finding "other sources of support that lead it to abandon its fisheries.” Id. at 687, 99 S.Ct. 3055. Washington does not present this contention, but arguably the tribal economy has changed dramatically since the enactment of the Stevens. Treaties, leading the Tribes to rely less on fish for their subsistence.
. In fact, the panel opinion quotes Fishing Vessel for some of these statements. See Washington V, 853 F.3d at 964-65.
. The panel opinion provides more factual support for the proposition that culverts adversely affect the population of salmon in considering the injunction, see Washington V, 853 F.3d at 972-75, but at that point it had already found that the Treaties were violated.
. The concurrence makes the extravagant assertion that I maintain that the Treaties allow the State to act "affirmatively to entirely eliminate the supply of harvestable salmon.” What utter nonsense! I said no such thing! In building and maintaining the culverts, the State was not acting affirmatively to destroy the salmon population—any negative effects were incidental—as the panel opinion acknowledged. See Washington V, 853 F.3d at 966. Far from seeking to eliminate the salmon population, the State recognizes that it is a treasured resource and has spent millions of *1028dollars on programs specifically designed to preserve, to protect, and to enhance the salmon population.
. See, e.g., Washington Department of Fish & Wildlife, Washington’s Native Char, http:// wdfw.wa.gov/fishing/char/ (noting that the bull trout population is "low and in some cases declining”).
. See, e.g., Michael C. Blumm, Indian Treaty Fishing Rights and the Environment: Affirming the Right to Habitat Protection and Restoration, 92 Wash. L. Rev. 1, 29-31 (2017).
. Indeed, the State, argues that while the culverts have been in place, the fish harvest has fluctuated dramatically from "nearly 11 million fish in 1985” to "900,000 fish” in 1999, and then báck to "over 4 million fish by 2003.” Such. evidence tends to suggest that culverts are not a primary driver of fish population.
. It seems highly - likely that if the panel opinion had engaged in such cost-benefit analysis, there would be more cost-effective ways to remedy the alleged Treaties violation. For example, a 1997 state report estimated that if the State replaced the culverts maintained by the Washington State Department of Transportation (which controls a majority of culverts), it would result in an annual increase of 200,000 salmon. Washington V, 853 F.3d at 970. It might be cheaper to stock an additional 200,000 salmon into Washington’s streams each year.
. Contrary to the curious claim in the concurrence that the costs are exaggerated, these figures were relied upon in the panel's own opinion!
. Sixty-eight equals thirty-one percent of 220. See Washington V, 853 F.3d at 973 (explaining that "[sjixty nine percent of the 220 downstream non-state barriers [i.e. 152 culverts] allowed partial passage of fish,” and thus by implication, thirty-one percent (i.e. 68 culverts) blocked all passage).
. In addition to the obvious financial cost to the State, there is also a broader cost to residents. Shortly after the panel’s opinion was issued, various news stories informed residents of highway closings resulting from the repair of culverts associated with the injunc-tiom.See, e.g., KIRO7, S[R] 167 to be closed all weekend -from Sumner to Auburn (Aug. .19, 2016), http://www.kiro7.com/news/local/sb-167-to-be-closed-all-weekend-from-sumner-to-auburn/426411799.
. Indeed, the difficulties of crafting an appropriate injunction illustrate why it is an undertaking best left to-the State.