EASTERBROOK, Circuit Judge.

So much of the Chicago metropolitan area is covered with concrete or other impermeable surfaces that the remaining ground cannot absorb the water from heavy rain. The excess goes into a combined stormwater and sewer system, which can overflow and escape through outfalls located on the banks of canals and rivers. In 1975 the Metropolitan Water Reclamation District, which manages sewage control (including purification plants), began construction on an ambitious project to impound water until it can be cleaned up and released safely: the Tunnel and Reservoir Plan, sometimes called TARP and commonly known as Deep Tunnel.

About 110 miles of large-diameter tunnels, as much as 350 feet underground (hence "deep" tunnel), collect runoff water and sewage during rainfall. But these tunnels, large and extensive as they are, can hold "only" 2.3 billion gallons of water, and heavy or extended rain may exceed that capacity. The plan therefore includes reservoirs, to which the tunnels direct their contents during high-inflow conditions. Two reservoirs, which between them can hold 3.4 billion gallons, are operational today. One of these is scheduled to be replaced later this year by the Thornton Composite Reservoir, which can accommodate 4.8 billion gallons from TARP. (This reservoir can hold 7.9 billion gallons, but 3.1 billion gallons of that capacity is for overflow from Thorn Creek and is not counted as part of the Deep Tunnel system.) The final piece of the system, the McCook Reservoir, is scheduled for completion in 2029 with a capacity of 10 billion gallons (and an interim capacity of 3.5 billion gallons by 2017). Deep Tunnel's final capacity will be 17.5 billion gallons.

The Thornton and McCook reservoirs have taken a long time to build because both will occupy worked-out limestone quarries. The demand for limestone, which has declined in recent years, affects the date of completion. (Paying to have limestone dug up in advance of demand from building and roadwork projects not only would be expensive but also would require the acquisition of land on which to dump huge piles of limestone, which would be unsightly and also create environmental problems as minerals leached from the limestone during rain.) But the delay in finishing the reservoirs, plus some problems in the design of the tunnels that can require the District to hold them below capacity, have hindered the system's ability to prevent stormwater and sewage releases. The record does not show how many untreated releases occur annually at the approximately 375 locations along rivers and canals where this is possible. It does show that flows from all sources (including direct rain runoff) are so substantial approximately once a year that the rivers and canals themselves would overflow unless permitted to discharge into Lake Michigan (normally locks prevent this, but they are opened when necessary to keep the rivers' levels under control).

In 2011 the United States and the State of Illinois jointly filed this suit, under sections 301 and 309 of the Clean Water Act, 33 U.S.C. §§ 1311, 1319, seeking an order that the District improve the TARP's performance, accelerate its completion date, and do more to contain and mitigate overflows in the interim. The district court permitted the Alliance for the Great Lakes and four other environmental organizations (collectively the Alliance) to intervene under 33 U.S.C. § 1365(b)(1)(B). See 2012 U.S. Dist. LEXIS 111223 (N.D.Ill. Aug. 7, 2012).

The complaint was accompanied by a proposed consent decree, which the pollution-control agencies had been negotiating with the District for four years. The settlement requires the District to complete the Deep Tunnel project, meet operational criteria when construction has been completed, monitor the system's performance, develop additional measures if needed to attain compliance with the Act and applicable permits in the interim, and maintain the decree in force until the district court concludes that compliance with the Act has been achieved. The Alliance opposed this proposal, contending that it requires the District to do too little and takes too long even for what it accomplishes. In a lengthy opinion, the district court rejected the Alliance's protest and entered the proposed decree. 2014 U.S. Dist. LEXIS 2049 (N.D.Ill. Jan. 6, 2014). The opinion contains many details about the settlement that we do not need to recapitulate, though we describe a few details later.

The district judge also concluded that the settlement binds the Alliance, and we start with this decision. A consent decree is at base a contract, see United States v. ITT Continental Baking Co., 420 U.S. 223, 95 S.Ct. 926, 43 L.Ed.2d 148 (1975), and the Alliance asks how it can be bound by a contract to which it did not agree. The district court's answer was "res judicata" (the combination of issue and claim preclusion), but that doctrine prevents one party from litigating the same claim or issue in multiple suits. See Robinson v. Harvey, 617 F.3d 915, 916 (7th Cir.2010). The Alliance is not the United States or the State of Illinois, and there is only one lawsuit. The district court relied on Friends of Milwaukee's Rivers v. Milwaukee Metropolitan Sewerage District, 382 F.3d 743, 757-65 (7th Cir.2004), and United States v. Metropolitan St. Louis Sewer District, 952 F.2d 1040 (8th Cir.1992), but each of these decisions dealt with multiple suits (filed only a few hours apart in Friends of Milwaukee's Rivers).

Our situation is governed not by principles of res judicata but by the fact that legislation may allocate litigation rights to an agency that represents the whole public. See Taylor v. Sturgell, 553 U.S. 880, 895, 128 S.Ct. 2161, 171 L.Ed.2d 155 (2008) (discussing the sixth exception to the norm that principles of virtual representation do not generally bind one person to the results of another's suit); Restatement (Second) of Judgments § 41(1)(d). The statute modifying the common law is § 1365(b), which provides:

No action may be commenced —

Section 1365(b)(1)(B) tells us that no private litigation may be "commenced" if the EPA or a state "has commenced and is diligently prosecuting a civil ... action" about the same matter the private litigant wants to raise.

We held in Friends of Milwaukee's Rivers that this language also means that the resolution (including a settlement) of a federal or state suit is binding on a private litigant whose suit was filed after the state or federal government's, if the state or federal action was diligently prosecuted. Accord, Louisiana Environmental Action Network v. Baton Rouge, 677 F.3d 737, 749-50 (5th Cir.2012); Piney Run Preservation Association v. Carroll County, 523 F.3d 453, 459-60 (4th Cir.2008). The United States and Illinois contend that the outcome of a governmental suit is equally conclusive for private claims asserted by intervenors.

This leads the Alliance to ask what the point of intervention might be — for intervention, no less than a ban on standalone private litigation, is part of this scheme. The answer, we think, is that intervention carries four rights: to introduce evidence if the case goes to trial; to object to a proposed settlement (a right the Alliance has exercised); to appeal if the intervenor thinks that the government has accomplished too little (another right the Alliance has exercised); and to enforce any judgment, just as the United States and Illinois can do. This decree orders the District to come into compliance with the Act and its permits, providing details about how and when. If the District falls short, either in implementing the interim measures or achieving compliance by 2029, then the Alliance can ask the district court for relief. And if the District asks the district court to dissolve the decree in 2030, the Alliance, as a party, can protest that too and appeal from an adverse decision.

That is a more modest role than a full-fledged independent litigator would have, but § 1365(b)(1)(B) tells us that a private party is not supposed to be a full-fledged independent litigator, if the state or federal government diligently prosecutes a suit. This is also the implication of the Supreme Court's observation that private intervenors are supposed to "supplement rather than to supplant" public litigation. Gwaltney of Smithfield, Ltd. v. Chesapeake Bay Foundation, Inc., 484 U.S. 49, 60, 108 S.Ct. 376, 98 L.Ed.2d 306 (1987). If the Alliance could carry on just as if it were the plaintiff in a separate suit, that would "supplant" the governmental case. Why would anyone settle with the EPA or a state, if the settlement did not buy peace? The District made costly promises, but if the Alliance is right then it got nothing in return, for the Alliance can carry on with the suit. And if the Alliance also settled, then another person could intervene to demand still more relief. Depriving the original parties of their ability to settle disputes is not consistent with the observation in Gwaltney, the holding of Friends of Milwaukee's Rivers, or the structure of § 1365(b).

Section 1365(b)(1)(B) blocks stand-alone private litigation only if the public agencies "diligently" prosecute their suits. The Alliance maintains that if there is a chance that the consent decree will fail to achieve full compliance with the Act and all permits, then the government has not prosecuted the claims diligently. That can't be right. Even the most diligent litigator may conclude that settlement is the best option — if only because it frees up enforcement resources for use else-where — and to achieve a settlement a litigant must accept something less than the most favorable outcome.

In Friends of Milwaukee's Rivers we concluded that a settlement likely to achieve the principal enforcement goals demonstrates diligent prosecution. 382 F.3d at 759-60. For a settled case, this pretty much turns the diligence issue into the question whether the proposed consent decree is reasonable. We remanded in Friends of Milwaukee's Rivers because the district court had approved that settlement without inquiring how likely it was to be an adequate solution. Here, by contrast, the district court considered the proposal's adequacy at length and evaluated the Alliance's objections. This enables us to merge the inquiries: if the consent decree is a reasonable settlement likely to bring about compliance with the Act, it also demonstrates diligent prosecution.

The Clean Water Act forbids "the discharge of any pollutant" (a defined phrase, see South Florida Water Management District v. Miccosukee Tribe, 541 U.S. 95, 124 S.Ct. 1537, 158 L.Ed.2d 264 (2004)), except when authorized by permits. The District holds multiple permits that govern releases from its system. These permits recognize the inevitability of discharges during storms but impose three kinds of conditions: the discharged water must have minimum oxygen levels, must keep solid matter ("floatables") under a specified level, and must provide for some rudimentary treatment (what the permits call "primary treatment ... with adequate retention time").

The federal and state complaint alleges that some overflows do not meet one or more of these three conditions. If the Deep Tunnel plan, augmented by the provisions of the consent decree, succeeds, then by 2029 (if not earlier) it will be possible to retain the water long enough to route it through the District's seven treatment plants at a rate that will allow full treatment. But the Alliance predicts that it will not succeed. The district court rejected the Alliance's protests in part because the alternatives it proposed would be costly (more than $1 billion), much of which would turn out to be wasted if the state and federal government are right about the plan's adequacy, and because even if the Alliance's predictions are right the rate of backflows from rivers and canals into Lake Michigan will drop from once a year to once a decade when the system is finished.

The Alliance's main argument on appeal can be summed up as: "It just won't work." The argument has three components. First, a study completed in 1994, and the testimony of one of the District's consultants before the Illinois Pollution Control Board, state that heavy rainfall will overwhelm the Deep Tunnel system. But the EPA, the Pollution Control Board, and the District are more optimistic about the system's likely functioning when completed. Dissatisfied with the 1994 study, which was based on only one modeled rainfall event and did not analyze the system's current operational plan, the EPA conducted its own. This new study, released in 2009, concluded that the finished Deep Tunnel system will work, at least during average and above-average rainfall years. (The study did not consider once-in-100-year events and other huge storms.) Recent data covering 2005 through 2011 in the Upper Des Plaines segment of the TARP system, which is finished (including its full reservoir capacity), show an average of fewer than two overflow events a year, and this period includes 2008, which had the largest recorded rainfall in the Chicago area's history. The district court was entitled to think a wait-and-see approach reasonable. The best way to decide between competing predictions is to see what happens.

But the Alliance maintains that the consent decree itself concedes that the system won't work. That is because the decree authorizes untreated discharges when (a) the tunnels are full, or (b) a "transient event" occurs — roughly, when the rate of inflow is so great that allowing it to continue would damage the tunnel system, and the District must close one or more gates to protect it. The Alliance says that this will occur whenever there is a "heavy rainstorm" but does not quantify that term (how heavy? how often? how much untreated water will be discharged?). More important, however, is the structure of the Clean Water Act: discharges are forbidden except when authorized. If the EPA and the Pollution Control Board are willing to authorize untreated discharges when there is no alternative (and that's what conditions (a) and (b) amount to), then there's no violation of the Act.

It's not as if the Alliance were contending that the tunnel system itself should be enlarged or its design changed, long after its construction, to avoid damage from particularly large inflows. Water weighs 62.4 pounds per cubic foot, or 8 pounds a gallon, and can drop a long distance in the tunnel system — greater than the height of Niagara Falls, where dropping water generates vast quantities of energy. This shows the power of gravity and the need to protect even tunnels with thick concrete walls. The consent decree allows the District to protect this costly pollution-control asset.

In the district court, the Alliance proposed that the District build treatment plants at 105 outfalls, increasing the number of plants from 7 large ones to 112 mostly small ones — just in case. But the district judge sensibly concluded that it would be imprudent to spend $1 billion on that project if it turns out that the system will be largely effective when the reservoirs are completed, for by the time the 105 new plants were operational they might be unnecessary. The Alliance's appellate brief drops this proposal, leaving no alternative to the decree's toleration of the inevitable.

The Alliance's final argument that the decree won't work is that it permits the release of floatables in excess of the quantity allowed by the "Combined Sewer Overflow Control Policy" adopted by the EPA in 1994. See 59 Fed.Reg. 18,688 (Apr. 19, 1994). The district court found otherwise, concluding that after the Deep Tunnel system is complete the District will be in compliance with the floatables policy. That finding is not clearly erroneous. If the Chicago area had a different kind of sewer system, or if Deep Tunnel had been designed differently, then a consent decree might have been able to achieve full compliance faster. That the settlement of a given case takes the world as it exists, however, does not show a lack of diligent prosecution or a substantively unreasonable outcome.

The EPA and the State sought to deal with the limitations imposed by the design of the sewer system and the Deep Tunnel project by using realistically available options. The District already has two pontoon boats that deal with floatables; the decree requires it to add two specially designed skimmer boats to keep floatables under control (even if their source is something other than the District's system) and to put a boom around one outfall that has experienced frequent discharges. (Booms elsewhere might interfere with navigation.) Another part of the decree requires the District to adopt a "green infrastructure" program that will reduce the amount of water flowing into the system during rains. The District must supply rain barrels to catch runoff from buildings and increase the permeability of the surface so that the ground can retain more water. The District must complete the reservoirs on schedule and pay as much as $5,000 a day for failure to do so; a desire to avoid these fines may lead the District to pay the limestone miners to remove the rock faster. Once each of the new reservoirs is completed, the District must provide enough on-site water treatment capacity to cover "the maximum flow accounting for all hydraulic and hydrologic factors that can pass through" the system. It must provide extra pumps if needed to move water faster from reservoirs to treatment plants, so that storage capacity is available in the reservoirs for incoming water. And if monitoring reveals that these steps don't achieve their goal, then the District must come up with and implement a new plan that will. These are costly promises by the District, and the district court did not abuse its discretion (the applicable standard, see United States v. George A. Whiting Paper Co., 644 F.3d 368, 372 (7th Cir.2011)) in concluding that the decree carries a reasonable prospect of success.

As the Alliance sees things, some of the District's promises are worthless (or at least incomplete) because the decree does not spell out "which pollutants will be monitored, at which locations, how frequently, or by what method." And it does not specify what, precisely, will be done if problems remain in 2029. Some other consent decrees that the EPA has negotiated do include these things. Yet the District is so large, and the locations of potential outfalls so numerous, that it's just not practical to try to cover all details in one document. The EPA anticipates working out details as time passes and additional reservoir capacity becomes available (which influences what needs to be looked at and tested for), and if the District does not cooperate the court can afford supplemental relief.

As for what happens in 2029 or later if untreated discharges continue at an unacceptable rate, the next steps ought to depend on what's not then working well. If the EPA (or a court) could be sure in 2014 what the exact nature of the problem (if any) would be in 2029, then it would be sensible to start planning and building the remedy today; but if either there won't be a serious problem in 2029, or the problem is something not now foreseen, then relying on a 2014 decree for the solution would be foolish. Yogi Berra observed that it is hard to make predictions, especially about the future. State and federal agencies are entitled to rely more on experience and less on predictions.

The consent decree that the district court has approved is reasonable in light of the current infrastructure, the costs of doing things differently (no one proposes to build a new sewer system or redo the Deep Tunnel project), and the limits of knowledge about what will happen when the system is completed. The decree is the outcome of diligent prosecution and therefore binds would-be private litigants such as the Alliance.

AFFIRMED