delivered the opinion of the Court.
In this action invoking the Court’s original jurisdiction, the State of Idaho seeks an equitable apportionment against the States of Oregon and Washington of the anadromous fish that migrate between the Pacific Ocean and spawning grounds in Idaho. The Special Master has filed his final report on the merits and recommends that the action be dismissed without prejudice. We have before us Idaho’s exceptions to that report.
I
Although somewhat repetitive of the Court’s prior writings in this litigation, 444 U. S. 380 (1980), we feel it worthwhile to outline once again the facts of the case and the Court’s prior rulings. The dispute concerns fish, one of the valuable *1019natural resources of the Columbia-Snake River system in the Pacific Northwest. That system covers portions of Wyoming, Idaho, Washington, Oregon, and British Columbia. From its origin in northwest Wyoming, the Snake River flows westerly across southern Idaho until it reaches the Idaho and Oregon border. At that point, the river winds northward to form the border between those States for approximately 165 miles, and then the border between Washington and Idaho for another 30 miles. Next, it turns abruptly westward and flows through eastern Washington for approximately 100 miles, finally joining the Columbia River. The Columbia, before this rendezvous, flows southward from British Columbia through eastern Washington. After it is supplemented by the Snake, the Columbia continues westward 270 miles to the Pacific Ocean. For most of the distance, it forms the boundary between Washington and Oregon.
A
Among the various species of fish that thrive in the Columbia-Snake River system, anadromous fish — in this case, chi-nook salmon and steelhead trout — lead remarkable and not completely understood lives. These fish begin life in the upstream gravel bars of the Columbia and Snake and their respective tributaries. Shortly after hatching, the fish emerge from the bars as fry and begin to forage around their hatch areas for food. They grow into fingerlings and then into smolt; the latter generally are at least six inches long and weigh no more than a tenth of a pound. The period the young fish spend in the hatching areas varies with the species and can last from six months to well over a year.
At the end of this period, the smolts swim down river toward the Pacific.1 In the estuary of the Columbia, the *1020young fish linger for a time in order to grow accustomed to the chemical cues of the water. A. Netboy, The Columbia River Salmon and Steelhead Trout 44 (1980). It is believed that they pick up the river’s scent so that in their twilight years they can return to their original home. Tr. of Oral Arg. 19. Even under the best of conditions, only a small fraction of the smolts that set out from the gravel bars ever reach the ocean.
Once in the ocean, the smolts grow into adults, averaging between 12 and 17 pounds. They spend several years traveling on precise, and possibly genetically predetermined, routes. See A. Netboy, supra, at 46-49. At the end of their ocean ventures, the mature fish ascend the river. They travel in groups called runs, distinguishable both by species and by the time of year. All the fish return to their original hatching area, where they spawn and then die. At issue in this case are the runs of spring chinook between February and May, the runs of summer chinook in June and July, and the runs of summer steelhead trout in August and September.
B
Since 1938, the already arduous voyages of these fish have been complicated by the construction of eight dams on the Columbia and Snake Rivers.2 First, interdicting the flow of the Snake River in Washington are the Lower Granite (constructed in 1969), the Little Goose (1968), and the Lower Monumental (1967) Dams. The Ice Harbor Dam (1961) sits astride the Snake just above its confluence with the Columbia. Four more dams interrupt the Columbia on its way to the Pacific: the McNary (1963), the John Day (1968), the Dalles (1957), and the original dam, the Bonneville (1938).
*1021In order to produce electrical power, these dams divert a flow of water through large turbines that have devastating effect on young smolts descending to the Pacific. Spillways have been constructed to permit the smolts to detour around the turbines.3 The dams also present great obstacles to the adults. Fish ladders — water-covered steps — enable the returning adults to climb over the dams; in addition, the ladders provide an opportunity for compiling statistics.4 Varying water conditions and the demand for power can increase the mortality of both descending smolts and ascending adults. The mortality rate for oceanbound smolts averages approximately 95%. Report of Special Master 7. Their adult counterparts die at a rate of 15% at each dam. Only 25% to 30% of the adults passing over the first dam, the Bonneville, succeed in running the gauntlet to traverse the Lower Granite Dam and enter Idaho, /bid.5
*1022Another factor depleting the anadromous fish population is fishing, sometimes referred to as “harvesting.” In 1918, Oregon and Washington, with the consent of Congress, Act of Apr. 8, 1918, ch. 47, 40 Stat. 515, formed the Oregon-Washington Columbia River Fish Compact to ensure uniformity in state regulation of Columbia River anadromous fish. Idaho has sought entry into the Compact on several occasions, but has been rebuffed. Under the Compact, Oregon and Washington have divided the lower Columbia into six commercial fishery zones: zones one through five cover the Columbia from its mouth to the Bonneville Dam; zone six stretches from the Bonneville Dam to the McNary Dam below the confluence with the Snake. Each year, authorities from both States estimate the size of the runs to determine the length of a fishing season the runs can support. The States do not permit commercial harvests of chinook salmon or steelhead trout in any of their Columbia River tributaries; they do, however, permit sport fishing in most locations.
Pursuant to treaties ratified in 1859, several Indian Tribes have “the right of taking fish at all usual and accustomed places.” Sohappy v. Smith, 302 F. Supp. 899, 904 (Ore. 1969). In 1977, after lengthy litigation over Indian treaty rights,6 Oregon and Washington agreed with the Indians to preserve zone six solely for Indian fishing. They also agreed *1023to limit commercial harvests in zones one through five to an amount that permits sufficient numbers of fish to pass over the Bonneville Dam to provide an equitable share for the Indians and to leave enough fish to replenish the runs. Under the plan, escapement goals — the number of fish passing the Bonneville — are set for each run. When the estimated size of the run exceeds the escapement goal by a specified amount, the surplus is allocated between non-Indian fishers below the Bonneville and Indian fishers above that dam. Two Indian Tribes recently have withdrawn from the agreement, however, casting its future effectiveness into doubt.
Although the parties disagree as to the causes, runs of all the relevant species since 1973 have been significantly lower. See Report of Special Master 46-51 (tables). Since that year, Oregon and Washington have not permitted commercial harvests of summer chinook; in both States, steelhead trout are now designated game fish and may not be harvested commercially. Harvests of spring chinook have been permitted only in 1974 and 1977. In the years since 1973, there has been some sport fishing of all three runs.
C
In 1976, the Court granted Idaho leave to file its complaint requesting an equitable apportionment of anadromous fish in the Columbia-Snake River system. 429 U. S. 163. The matter was referred to a Special Master, the Honorable Jean S. Breitenstein, Senior Judge for the United States Court of Appeals for the Tenth Circuit. See 431 U. S. 952 (1977). The Special Master initially recommended that the suit be dismissed without prejudice for failure to join an indispensable party, the United States. That recommendation was not accepted, and the case was remanded for trial. 444 U. S. 380 (1980). The Court stated that Idaho “must shoulder the burden of proving that the [non-Indian] fisheries in [Oregon and Washington] have adversely and unfairly affected the number of fish arriving in Idaho.” Id., at 392.
*1024After trial and oral argument, the Special Master issued his final report on the merits. He has recommended that the action be dismissed without prejudice, apparently for two distinct reasons. First, he found that Idaho has not demonstrated that it has suffered any injury at the hands of Oregon and Washington. Second, even assuming that it has suffered such an injury, he found it impossible to fashion a decree to apportion the fish fairly among the parties. Idaho has filed exceptions to the report.7
B
>
As an initial matter, the Special Master correctly concluded that the doctrine of equitable apportionment is applicable to this dispute. Although that doctrine has its roots in water rights litigation, see Kansas v. Colorado, 206 U. S. 46, 98 (1907), the natural resource of anadromous fish is sufficiently similar to make equitable apportionment an appropriate mechanism for resolving allocative disputes.8 The anad-romous fish at issue travel through several States during their lifetime. Much as in a water dispute, a State that over-fishes a run downstream deprives an upstream State of the fish it otherwise would receive. A dispute over the water flowing through the Columbia-Snake River system would be resolved by the equitable apportionment doctrine; we see no reason to accord different treatment to a controversy over a similar natural resource of that system. *1025The doctrine of equitable apportionment is neither dependent on nor bound by existing legal rights to the resource being apportioned. The fact that no State has a pre-existing legal right of ownership in the fish, Hughes v. Oklahoma, 441 U. S. 322, 329-336 (1979), does not prevent an equitable apportionment. Conversely, although existing legal entitlements are important factors in formulating an equitable decree, such legal rights must give way in some circumstances to broader equitable considerations. See Colorado v. New Mexico, 459 U. S. 176, 184 (1982); id., at 195 (O’CONNOR, J., concurring); Nebraska v. Wyoming, 325 U. S. 589, 618 (1945); Connecticut v. Massachusetts, 282 U. S. 660, 670-671 (1931).
At the root of the doctrine is the same principle that animates many of the Court’s Commerce Clause cases: a State may not preserve solely for its own inhabitants natural resources located within its borders. See Philadelphia v. New Jersey, 437 U. S. 617, 627 (1978); see also New England Power Co. v. New Hampshire, 455 U. S. 331, 338 (1982); Hughes v. Oklahoma, 441 U. S., at 330. Consistent with this principle, States have an affirmative duty under the doctrine of equitable apportionment to take reasonable steps to conserve and even to augment the natural resources within their borders for the benefit of other States. Colorado v. New Mexico, 459 U. S., at 185; Wyoming v. Colorado, 259 U. S. 419, 484 (1922). Even though Idaho has no legal right to the anadromous fish hatched in its waters, it has an equitable right to a fair distribution of this important resource.
B
Because apportionment is based on broad and flexible equitable concerns rather than on precise legal entitlements, see Colorado v. New Mexico, 459 U. S., at 183; Nebraska v. Wyoming, 325 U. S., at 618, a decree is not intended to compensate for prior legal wrongs. Rather, a decree prospectively ensures that a State obtains its equitable share of a re*1026source. A decree may not always be mathematically precise or based on definite present and future conditions. Uncertainties about the future, however, do not provide a basis for declining to fashion a decree. Reliance on reasonable predictions of future conditions is necessary to protect the equitable rights of a State.
To the extent that the Special Master found that the formulation of a workable decree is impossible, we must disagree. See Washington v. Fishing Vessel Assn., 443 U. S. 658, 663 (1979) (regular habits of anadromous fish make it possible to forecast size of runs). Idaho’s proposed formula for apportioning the fish is one possible basis for a decree.9 It relies on the number of jackfish — reproductively precocious male fish, which return a year ahead of other members of their age group — passing over the Bonneville and the lee Harbor Dams to predict the size of the run the following year and the percentage of fish in the run that originate in Idaho.10 *1027Although the computation is complicated and somewhat technical, that fact does not prevent the issuance of an equitable decree. See 444 U. S., at 390; Nebraska v. Wyoming, 325 U. S., at 616-617. Nothing in the record undermines the assumption supporting Idaho’s formula that there is a definite relationship between the number of jackfish and the total number of fish in a particular run the following year. Thus, if Idaho suffers from the injury it alleges, we see no reason why that injury could not be remedied by an equitable decree.
C
The Special Master also found, however, that Idaho has not demonstrated sufficient injury to justify an equitable decree. A State seeking equitable apportionment under our original jurisdiction must prove by clear and convincing evidence some real and substantial injury or damage. Colorado v. New Mexico, 459 U. S., at 187-188, n. 13; Connecticut v. Massachusetts, 282 U. S., at 672; see New Jersey v. New York, 283 U. S. 336, 344-345 (1931). In reaching his conclusion, the Special Master stated that the determination should be based on present conditions. Report of Special Master 25-26. He therefore focused on the most recent time period, 1975 through 1980, during which all the dams and various conservation programs were in operation.
We approve this approach. The Special Master found that, due to the operation of the dams, the fish runs have been depressed since 1970. Id., at 26, 34. It is highly unlikely that the dams will be removed or the number of deadly turbines reduced; all parties must live with these conditions in the determinable future.11 Although Oregon and Wash*1028ington may have harvested a disproportionate share of anad-romous fish over the long run,12 Idaho took 58.72% of the total harvest in the period from 1975 through 1980. Id., at 44. Equitable apportionment is directed at ameliorating present harm and preventing future injuries to the complaining State, not at compensating that State for prior injury. We agree with the Special Master that these figures do not demonstrate that Oregon and Washington are now injuring Idaho by overfishing the Columbia or that they will do so in the future.
Moreover, Idaho has not proved that Oregon and Washington have mismanaged the resource and will continue to mismanage. The two States in 1974 did permit some over*1029fishing of the Columbia.13 Idaho, however, has produced no concrete evidence of other mismanagement, and the Special Master concluded that “[t]he record shows no repetition or threatened repetition of [prior mismanagement].”14 Id., at 32. Although it is possible that Washington and Oregon will mismanage this resource in the future, Idaho has not carried its burden of demonstrating a substantial likelihood of injury.
HH l-H
For the foregoing reasons, we adopt the Special Master’s recommendation and dismiss the action without prejudice to the right of Idaho to bring new proceedings whenever it shall appear that it is being deprived of its equitable share of anad-romous fish.
It is so ordered.
The smolts, apparently, prefer not to swim. They face upstream, open their mouths, and permit the current to carry them downstream. Should they come upon a quiet spot, they turn around and swim. A. Netboy, The Columbia River Salmon and Steelhead Trout 44 (1980).
Three dams in Idaho — the Brownlee (constructed in 1968), the Oxbow (1961), and the Hells Canyon (1967) Dams — have closed off the upper Snake River entirely to this piscean traffic. This renders unusable much good spawning area.
Most dams are also equipped with screens that divert the smolts away from the turbines and into the spillways. Since 1969, however, the number of turbines operating on the dams has increased from 8 to 24, causing more water to be directed through turbines and reducing the water flow down the spillways. This has increased smolt mortality dramatically. There is an experimental plan to place smolts in tanks and “bus” them around all the dams for release below the Bonneville Dam. See Tr. of Oral Arg. 15; Idaho’s Exceptions to Master’s Final Report on Merits 102-103 (Idaho’s Exceptions).
At each fish ladder, the Army Corps of Engineers has constructed observation windows from which it counts and records the number of ascending fish and notes their variety. This count must be adjusted for the phenomenon of “fall back”: often adult fish that have been counted are swept back over the dam or down the ladder by strong currents. In addition to the effect this phenomenon has on the complexity of the count, the fall over the dam causes nitrogen supersaturation, making the fish slightly giddy and disoriented, and serving to increase adult mortality.
Apparently, the John Day Dam, constructed in 1968, is “the big killer” of ascending adults. See Tr. of Oral Arg. 16. To mitigate the effects of the high mortality rate caused by all the dams, hatchery programs hatch and nurture millions of smolts and release them into the Snake River. The Idaho Power Company finances several Idaho hatcheries, pursuant to a condition imposed by the Federal Energy Regulatory Commission in *1022granting the company’s application for a license to construct dams along the upper portions of the Snake. Report of Special Master 9; see n. 3, supra. In addition, the parties have agreed to construct 10 hatcheries, 6 in Idaho, to compensate for losses caused by the four dams on the lower Snake River.
The Sohappy District Court in 1974 held that the Indians were entitled to 50% of the fish destined to pass over the Bonneville. See Sohappy v. Smith, 529 F. 2d 570, 572 (CA9 1976); cf. Washington v. Fishing Vessel Assn., 443 U. S. 658, 685-689 (1979) (approving similar 50% allocation to Indians). The Court of Appeals for the Ninth Circuit vacated the order and remanded the case to the District Court for consideration of other factors bearing on the apportionment. 529 F. 2d, at 573-574. The parties reached the agreement described in the text before any further District Court action.
Washington filed no exceptions of its own, but has responded to those of Idaho. Oregon did not participate in our review of the Special Master’s report.
The Court in Kansas v. Colorado said:
“[W]henever... the action of one State reaches through the agency of natural laws into the territory of another State, the question of the extent and the limitations of the rights of the two States becomes a matter of justi-ciable dispute between them, and this court is called upon to settle that dispute in such a way as will recognize the equal rights of both and at the same time establish justice between them.” 206 U. S., at 97-98.
Oregon and Washington authorities employ a similar formula in estimating the size of runs and in setting Bonneville Dam escapement goals pursuant to the Indian treaty rights settlement agreement. In addition to the apportionment formula, Idaho’s plan would require Oregon and Washington (1) to continue the same primary management techniques they have been using; (2) to estimate the size of future runs and dam mortality rates; (3) to meet the escapement requirements they have set for the last five years; (4) to determine the number of fish in each run that originated in Idaho; (5) to determine the harvestable surplus of Idaho-origin fish; (6) to allot to Idaho a share of that surplus (after subtracting Indian fisheries) equal to the percentage that Idaho-origin fish are of the total Columbia River run; and (7) to make up any shortfall in Idaho’s allocated harvest out of the next year’s harvest.
The latter prediction is possible because most fish that surmount the Ice Harbor Dam are headed for spawning grounds in Idaho. We express no view on the appropriateness of Idaho’s proposed formula. We note that it apportions fish solely on the basis of their origin. Flexibility is the linchpin in equitable apportionment cases, and, in our prior decisions, we have based apportionment on the consideration of many factors to ensure a fair and equitable allocation. See Colorado v. New Mexico, 459 U. S. 176, 183 (1982).
Idaho accepts, as it must, see 444 U. S., at 388, the continued operation of the dams and their adverse impact on the runs. See Idaho’s Exceptions 46, 87. Its argument that the parties must share that adverse impact equally, id., at 87, is relevant to the fashioning of an equitable decree, but not to the existence of a cognizable injury.
Idaho claims that from 1962 through 1980, when spring chinook that originated in Idaho constituted 50% of the total runs, Oregon and Washington took 83% of the Idaho spring chinook. According to Idaho, they also harvested 75% of the Idaho-origin summer chinook, which during the period constituted 40% of all summer chinook runs. As to steelhead trout, Idaho asserts that Oregon and Washington took 58% of the harvest of Idaho-origin fish, which was 48% of the total steelhead runs. Id., at 49-50.
Of course, these figures presume, as does Idaho’s entire argument, that Idaho is entitled to those fish that originate in its waters. After Hughes v. Oklahoma, 441 U. S. 322 (1979), however, Idaho cannot claim legal ownership of the fish. While the origin of the fish may be a factor in the fashioning of an equitable decree, it cannot by itself establish the need for a decree. Instead, the Court must look to factors such as disproportionate reductions in Idaho’s normal harvest, or reductions in the total fish in the runs caused by mismanagement or overfishing by Washington and Oregon. As a historical matter, Idaho’s own tables demonstrate that its proportion of the harvest of Idaho-origin spring chinook increased from 13.5% in 1962 through 1967 to 45.5% in 1975 through 1980, and its percentage of the harvest of Idaho-origin steelhead trout increased in the same period from 35.1% to 90.7%. Idaho’s harvest percentage of Idaho-origin summer chi-nook did decrease between the two periods, but only 192 fish from that run were caught in the latter period, a de minimis number. Idaho’s Exceptions 53-54 (tables 6, 7, and 8). Although we reject the assumption of entitlement underlying Idaho’s comparisons, even under that assumption, Idaho’s portion of the harvest has been increasing.
The Special Master found that the last incident of mismanagement occurred in 1974 when, despite the recommendation of experts, Oregon and Washington permitted a limited harvest. They overestimated the Bonneville count by failing to consider the fall back phenomenon, and underestimated the Indian fishery for the year. The overfishing reduced the number of fish returning to spawn. Report of Special Master 32.
Moreover, despite Idaho’s claim that Oregon and Washington managed only for minimum escapements over the Bonneville, the Special Master found that Idaho had never requested those States to increase the escapement goal. Id., at 31. In fact, Idaho seems quite content with the current escapement goals; its plan requires that Oregon and Washington “manage to meet the same spawning escapements they have been managing for over the last five years.” Idaho’s Exceptions 82.