PER CURIAM:

The Georgia Parties,1 Gwinnett County, Georgia, and the United States Army Corps of Engineers ("the Corps") appeal from the Middle District of Florida's grant of summary judgment in this consolidated suit. The appeal arises from more than 20 years of litigation involving the above parties as well as the States of Alabama and Florida, Alabama Power Company, the City of Apalachicola, Florida, and Southeastern Federal Power Customers, Inc. ("SeFPC"), a consortium of companies that purchase power from the federal government. All of the underlying cases2 relate to the Corps' authority to operate the Buford Dam and Lake Lanier, the reservoir it created, for local water supply. In its order, the district court found that the Corps' current operation of the Buford Project—Buford Dam and Lake Lanier collectively—had allocated more than 21% of Lake Lanier's storage space to water supply. The court determined that such an allocation exceeded the Corps' statutory authority and ordered the Corps to drastically reduce the quantity of water that it made available for water supply. The court's summary judgment order also affirmed the Corps' rejection of Georgia's 2000 request for additional water supply allocations to meet the needs of the localities through 2030. The court stayed its order for three years to give the parties time to reach a settlement or to approach Congress for additional water supply authority.

On appeal, the parties argue several jurisdictional matters. Alabama and Florida3 contend that this Court does not have appellate jurisdiction to hear the appeal of three of the four underlying cases because there is no final judgment in the cases and pendent jurisdiction is inappropriate. The Georgia Parties and the Corps argue that the district court lacked jurisdiction over these three matters because there was no final agency action, and, therefore, the Administrative Procedures Act ("APA") did not provide for judicial intervention at this juncture.

The parties also assert a number of substantive claims. The Georgia Parties argue that the district court erred by concluding that the Corps lacked authority to allocate substantial quantities of storage in Lake Lanier to water supply on the basis of the legislation that authorized the creation of the Buford Project, the 1946 Rivers and Harbors Act ("RHA"), Pub.L. No. 79-525, 60 Stat. 634 (1946). Although not in agreement with the Georgia Parties that water supply for the Atlanta area is an authorized project purpose under the RHA, the Corps does argue that the district court underestimated its authority to accommodate the water supply needs of the Atlanta area. The Georgia Parties and the Corps both assert that the district court erred by misinterpreting the scope of the Corps' authority under the 1958 Water Supply Act. The Georgia Parties and the Corps urge this Court to remand the case to the agency to make, in the first instance, a final determination of its water supply authority. Gwinnett County also individually asserts statutory, constitutional, and contractual claims relating to authority granted to it for its current withdrawals from Lake Lanier.

For the reasons explained below, we hold: First, the district court erred in finding that it had jurisdiction to hear Alabama, SeFPC, and Apalachicola because the Corps has not taken final agency action. The three cases therefore must be remanded to the Corps in order to take a final agency action. Second, the district court and the Corps erred in concluding that water supply was not an authorized purpose of the Buford Project under the RHA. The Corps' denial of Georgia's 2000 water-supply request is therefore not entitled to Chevron deference, and the request must be remanded to the Corps for reconsideration. Third, the district court erred in finding that the 1956 Act, which authorized the Corps to contract with Gwinnett County to withdraw 10 million gallons of water per day, expired after 50 years. Gwinnett County's contractual and just-compensation claims are without merit. Fourth, we also provide certain instructions to the Corps on remand. And finally, the Corps shall have one year to make a final determination of its authority to operate the Buford Project under the RHA and WSA. Our opinion is organized as follows.

    FACTS AND PROCEDURAL HISTORY

    DISCUSSION

         Part I. Jurisdictional Matters

                  A. Appellate Jurisdiction over Alabama, SeFPC, and
                        Apalachicola

                  B. Final Agency Action in Alabama, SeFPC and
                        Apalachicola

         Part II. Georgia's 2000 Request: The Corps' Water Supply Authority
                        Under the RHA

         Part III. Georgia's 2000 Request Must Be Remanded to the Corps

         Part IV. Gwinnett County's Claims Not Involving Authorization
                        Under the RHA and WSA

                  A. The Expiration of the 1956 Act

                  B. Forty MGD from the 1974 Supplemental Agreement to
                         the Corps' Contract

                  C. Just Compensation for Relocation of the Duluth Intake

         Part V. Remand Instructions to the Corps

         Part VI. Collateral Estoppel Effects on Remand Instructions

         Part VII. One-Year Time Limit on Remand

    CONCLUSION

FACTS AND PROCEDURAL HISTORY

The facts of this appeal are intertwined with the history of Buford Dam and Lake Lanier. Buford Dam sits on the Chattahoochee River, approximately forty miles upstream of Atlanta. The Chattahoochee's headwaters are in Northeastern Georgia in the Blue Ridge Mountains. The river flows southwest to Columbus and then along much of the length of the Georgia-Alabama border and into the Florida Panhandle, where it combines with the Flint River to form the Apalachicola River. The Chattahoochee, Flint, and Apalachicola Rivers together are referred to as the ACF Basin.

The Corps first began surveying the ACF Basin for suitable sites for hydroelectric facilities at the request of Congress in 1925. River and Harbor Act of 1925, Pub.L. No. 68-585, ch. 467, 43 Stat. 1186, 1194 (March 3, 1925). As a result of this survey, the Corps produced a report in 1939. See H.R. Doc. No. 76-342 (1939) [hereinafter "Park Report"]. The Park Report analyzed eleven projects at various stages of development in the ACF basin, including one at Roswell, Georgia, sixteen miles north of Atlanta. Id. ¶ 196. District Engineer Colonel R. Park, the report's author, referred to transportation, hydroelectric power, national defense, commercial value of riparian lands, recreation, and industrial and municipal water supply as "principal direct benefits" of the various projects under consideration. Park Report ¶ 243. Col. Park noted that at the time the Atlanta area had no immediate need for increased water supply, though such a future need was "not improbable." Park Report ¶ 260. He stated that a large reservoir might have value as "an assured continuous water supply" due to the "continued rapid growth of the area." Id. Though he assigned the other direct benefits a monetary value, he declined to do so for water supply, presumably because the benefit of this purpose, unlike all of the others, could only accrue in the future, rendering any valuation at that time speculative. Congress adopted the Corps' proposals in the Park Report in full in its 1945 RHA. Pub.L. No. 79-14, 59 Stat. 10, 17 (1945).

In 1946, the Corps, in its "Newman Report," recommended certain amendments and revisions to the original plan for the ACF system, including combining several of the hydroelectric sites near Atlanta into one large reservoir at Buford, Georgia to increase power generation and to better regulate flows downstream. H.R. Doc. No. 80-300, ¶ 69 (1947) [hereinafter "Newman Report"]. Division Engineer Brigadier General James B. Newman noted that the Chattahoochee River would be an excellent source of hydropower. Newman Report ¶ 7. According to Newman, a large reservoir—what would become Lake Lanier—was needed to make the locks and dams downstream more effective. The Newman Report noted that the proposed dam at Buford would be valuable for the purpose of flood control because of the frequent flooding in the basin and the severe damage that previous floods had caused. The report also explained that the various dams in the proposal would help keep flows continuous. These continuous flows would benefit navigation because they would allow barges to travel from Atlanta to Columbus and beyond, and they would assure a source of water supply for the City of Atlanta. Just as the Park Report had done before it, the Newman Report attempted to quantify the value of the benefits of the project. Only three value-calculated benefits were listed: power, navigation, and flood control. Id. ¶ 98, Table 10. It is probable that Newman, like Park, deemed there to be no immediate benefit from water supply, rendering any benefit purely prospective and any valuation of this benefit entirely speculative.

The Newman Report, at several junctures, spoke of the benefit that the dam would provide for water supply. The report concluded that the project would "greatly increase the minimum flow in the river at Atlanta," which would safeguard the city's water supply during dry periods. Id. ¶ 68. In discussing the operation of the dam, the Newman Report noted that releases of 600 cubic feet per second ("cfs") should be made during off-peak hours4 in order to ensure a continuous flow of the river at Atlanta of not less than 650 cfs, even though this flow would have a slight detrimental effect on power generation. The report noted that this "minimum release may have to be increased somewhat as the area develops." Id. ¶ 80. The Report expected that any decrease in power value would be marginal and outweighed by the benefits of an "assured" water supply for the City of Atlanta. Id. The 1946 RHA stated that the project would be "prosecuted ... in accordance with the report of the Chief of Engineers, dated May 13, 1946," Pub.L. No. 79-525, 60 Stat. 634, 635 (1946). Because that report incorporated the Newman Report in full, the Newman Report became part of the authorizing legislation for the project.

Congress continued to consider the purposes of the Buford Dam in debates about appropriations bills for the project's funding. The purposes mentioned most frequently in Congressional hearings were power, navigation, and flood control, but water supply was also discussed with some frequency. Then-mayor of Atlanta, William Hartsfield testified before a Senate subcommittee that water from the Chattahoochee was "necessary" but that Atlanta did not immediately need the water in the same manner as cities in more arid locations. Civil Functions, Dep't of the Army Appropriation Bill 1949: Hearing Before the Subcomm. of the S. Comm. on Appropriations, 80th Cong. 644 (statement of William B. Hartsfield, Mayor, Atlanta, Georgia). Congress debated whether Atlanta should be asked to contribute part of the cost of building the Buford Dam. Corps officer Colonel Potter testified that the Corps was not recommending that Atlanta be asked to pay because the services that would be provided in the field of water supply were all incidental to the purposes of hydropower and flood control and would "not cost the Federal Government 1 cent to supply." Civil Functions, Dep't of the Army Appropriations for 1952: Hearings Before the Subcomm. of the H. Comm. on Appropriations, 82d Cong. 121-122 (1951) (exchange between Rep. Gerald Ford, Member, H. Comm. on Appropriations, and Col. Potter, Corps officer). Congressman Gerald Ford presciently asked Colonel Potter whether it was foreseeable that one day in the future Atlanta would begin to request greater amounts of water from the project. Id. at 122. Col. Potter responded that the Corps would have to study the effect that such a request would have on power production. He said that the Corps would have to obtain additional water supply authorization if a request amounted to "a major diversion of water." Id. Ultimately, Atlanta was never asked to, and did not, contribute to the construction costs.

The Corps released its "Definite Project Report" for the project in 1949. The report provided a detailed discussion of the plans for the Buford Project and its operations. The report referred to flood control, hydroelectric power, navigation, and an increased water supply for Atlanta as "the primary purposes of the Buford project." U.S. Army Corps of Eng'rs: Mobile District, Definite Project Report on Buford Dam Chattahoochee River, Georgia, ¶ 48 (1949) [hereinafter "Definite Project Report"]. A later passage in the report referred to flood control, power generation, navigation, and water supply as "principle purposes of the Buford project." Id. ¶ 115. The report concluded by calculating and explaining the benefits of the various project purposes. As to water supply, it explained that the project would result in "[a] real benefit," but it did not estimate the monetary value because "definite evaluation of this benefit cannot be made at this time." Id. ¶ 124.

Buford Dam was constructed from 1950 to 1957, creating the reservoir known today as Lake Sidney Lanier. The Southeastern Power Administration ("SEPA"), the federal government agency from which SeFPC purchases the power generated at the dam, paid approximately $30 million of the $47 million of construction costs. The creation of Lake Lanier inundated the water intake structures of the Cities of Buford, Georgia and Gainesville, Georgia. As a method of compensation, the Corps signed relocation agreements with the two municipalities authorizing water withdrawals directly from the reservoir—these agreements allowed Gainesville to withdraw 8 million gallons per day ("mgd")5 and Buford 2 mgd.6 Although no storage7 was specifically allocated for water supply, the fact that the dam operated during "off-peak" hours, to the detriment of power generation, demonstrated that downstream water supply was a consideration. In accordance with the recommendations of the Newman Report, the Corps maintained the necessary minimum river flow at Atlanta by making off-peak releases of 600 cfs during these hours of the week.

During construction of the dam, Gwinnett County requested permission from the Corps to withdraw 10 mgd directly from Lake Lanier. The Corps denied the request, explaining that the approval of Congress was required for it to meet such a request. In 1955, the Corps stated to Congress that the proposed withdrawals would be in the public interest and would not have a materially adverse effect on downstream interests or power output. F.G. Turner, Ass't Chief, Eng'g Div., U.S. Army Corps of Eng'rs: Mobile District, Report on Withdrawal of Domestic Water Supply from Buford Reservoir ¶ 1 (1955). The following year, Congress passed a law that granted the Corps authority to enter into a contract with Gwinnett County for the allocation of 11,200 acre-feet of storage for regulated water supply and granted the county an easement across government property for the construction and maintenance of a pumping station and pipelines. Pub.L. No. 84-841, 70 Stat. 725 (1956).

Construction was completed in 1957. Lake Lanier covers 38,000 acres and has 692 miles of shoreline. The large size of the lake allows for a substantial benefit in the form of recreation.

Lake Lanier is divided into three tiers, or pools, divided by elevation. The first tier extends from the bottom of the lake, at an elevation of 919 feet above sea level, to an elevation of 1,035 feet. This tier holds 867,600 acre-feet of "inactive" storage. The inactive pool is generally left untouched and saved for instances of severe drought. The next tier extends from an elevation of 1,035 feet to an elevation of 1,070 feet (1,071 feet in the summer) and contains 1,049,000 acre-feet (1,087,600 acre-feet in the summer) of conservation storage. The conservation pool generally provides the water that is used for all downstream purposes. The Newman Report contemplated that conservation storage would be used primarily for hydropower and repeatedly referred to it as storage for power. The final tier extends from an elevation of 1,070 feet (1,071 feet in the summer) to an elevation of 1085 feet. This tier provides 598,800 acre-feet of flood storage. The flood pool is generally left empty so that it can accommodate excess water during flood conditions.

Buford Dam was constructed to release water from Lake Lanier through a powerhouse that generates hydropower. The powerhouse contains three turbines. Two of the turbines are large and release about 5,000 cfs when running. These two turbines operate during peak hours, when energy consumption is at its greatest. They are the most efficient source of power generation—generating 40,000 kilowatt hours ("kwh") originally and 60,000 kwh after improvements in 2004—and would be the only turbines used if a minimum off-peak flow at Atlanta were not a project concern. To accommodate this concern, the powerhouse also contains a third, smaller turbine which releases 600 cfs, generating approximately 7,000 kwh. At peak performance, the dam releases approximately 11,000 cfs of water into the river. However, when energy demand is low—so-called off-peak hours—only the small turbine is operated, allowing the dam to produce some energy while providing for a minimal continuous flow. The Corps can also release water through a small sluice gate, but this is typically done only when the small turbine is shut down for repairs or in cases of an emergency. In this manner, Buford Dam was designed to generate maximum power while also ensuring a minimum continuous flow of water downstream to accommodate water supply.

In 1958, Congress passed the Water Supply Act ("WSA"). The statute was designed to allocate some storage in multi-purpose projects like Buford to water supply. The policy underlying the statute was:

43 U.S.C. § 390b(a) (2011). To further that policy, Congress authorized the Corps to allocate storage in federal reservoirs for water supply, provided that the localities paid for the allocated storage. Id. § 390b(b). However, Congress placed the following limitation on its authorization:

Id. § 390b(d) (emphasis added). The policy of subsection (a) indicates that Congress aimed only to expand water supply allocations, not contract them by limiting previous authorizations. The articulation of the bounds of the statute's authorization makes no mention of a limit on previously granted water supply authorization. In the case of Buford, the WSA's grant of authority for water supply constitutes a supplement to any authority granted by the 1946 RHA.

In 1959, the Corps issued its Reservoir Regulation Manual for Buford Dam ("Buford Manual") as an appendix to the Corps' 1958 manual for the entire river basin. U.S. Army Corps of Eng'rs: Mobile District, Apalachicola River Basin Reservoir Regulation Manual, Appendix B (1959). The Buford Manual has not been updated and remains in effect today. The manual describes the technical features of the dam, including a description of the three tiers and their storage capacities, the size of Lake Lanier, and the general operation of the plant. It states that the project will be run to maximize releases of water during peak hours but will also utilize off-peak releases in order to maintain a minimum flow of 650 cfs at Atlanta.8 Id. at B-13. The manual makes multiple mentions of regulations that are designed to ensure this minimum flow. Id. at B-1819, B-22.

There was very little change in water supply operations at the Buford Project between 1960 and 1973. Only Gainesville and Buford withdrew water directly from Lake Lanier. Gwinnett, with which the Corps was authorized to contract, did not withdraw water directly from the reservoir. In the meantime, the City of Atlanta and DeKalb County withdrew water from the river downstream from the dam but made no recorded requests that the schedule of releases be altered to accommodate their needs. The Atlanta metropolitan area increased its water use from the Chattahoochee by 37% (from 117 mgd to 160 mgd) between 1960 and 1968. U.S. Army Corps of Eng'rs: Mobile District, Final Environmental Impact Statement: Buford Dam and Lake Sidney Lanier, Georgia (Flood Control, Navigation and Power), Statement of Findings 14 (1974). This amount was still well below the amount released by the Corps to maintain a minimum off-peak flow. Moreover, between 1956 and 1969, the number of residences within two and a quarter miles of the reservoir doubled. Id. at 15. The growing water needs of Atlanta came to the attention of the Senate, which in 1973 commissioned the Metropolitan Atlanta Area Water Resources Management Study ("MAAWRMS") to develop a plan for the long-term needs of the Atlanta area. The study was conducted by the Corps, the Atlanta Regional Commission ("ARC"), the State of Georgia, and the U.S. Environmental Protection Agency in combination.

By the 1970s, it became clear that area localities desired greater access to water in Lake Lanier. The Corps determined that it could not grant permanent water allocation rights to the localities before the completion of the MAAWRMS. While the study was being performed, the Corps entered into a number of interim contracts for water withdrawal. The first water supply contract was given to Gwinnett County and allowed for the withdrawal of up to 40 mgd directly from Lake Lanier during the course of the study. Contract of July 2, 1973, Gwinnett Record Excerpts vol. 1, ACF004024. The contract cited the WSA for authority and was based on findings of the Corps' District Engineer that the proposed withdrawals would not have significant adverse affects on the other authorized purposes of the project. No mention was made of the 1956 Act, possibly because the Act authorized only 10 mgd and thus would not have been sufficient authority for the Corps' actions. In 1975, the county informed its bond investors that the construction costs alone for its water supply facilities would be $28 million.

In 1975, the Corps concluded, and SEPA agreed, that the Buford Project could supply an annual average of 230 mgd of water for downstream withdrawal (with a maximum of 327 mgd in the summer) without significantly affecting hydropower generation. The Corps revised this number in 1979, concluding that by scheduling additional peak weekend releases it could raise the annual average to 266 mgd as an incident of power generation. In 1986, the Corps would again raise the figure for available water supply downstream incident to power generation, concluding that it could guarantee an annual average of 327 mgd by implementing a new water management system that had been proposed in the MAAWRMS.

The final report of the MAAWRMS was issued in September 1981. The report evaluated three alternative plans for dealing with Atlanta's increasing long-term water supply needs. The first alternative was to build a reregulation dam 6.3 miles below the Buford Dam. This new dam would store outflows released from Buford during peak operations and release them as needed for water supply. The study found that this alternative had the highest estimated ratio of benefits to costs. This alternative received the most support from federal and state agencies, and the study concluded it was best. The second alternative was to reallocate storage space in Lake Lanier for water supply. According to the study, in 1980, 10,512 acre-feet had been allocated to water supply, amounting to 14.6 mgd withdrawn directly from Lake Lanier. This alternative called for an increase of allocated storage space to 141,685 acre-feet by the year 2010, allowing for a total withdrawal of 53 mgd from the lake. The final alternative was to dredge the Morgan Falls Reservoir, which lies downstream from Buford, and also reallocate 48,550 acre-feet of storage space in Lake Lanier for water supply by 2010.

In 1986, Congress, in the Water Resources Development Act, authorized the construction of a reregulation dam, the MAAWRMS' favored first alternative. Pub.L. No. 99-662, § 601(a)(1), 100 Stat. 4137, 4140-41. However, the project had previously not received approval from the Office of Management and Budget, which said that state and local money should be used to construct such a project, and Congress did not appropriate any funding towards the construction of the proposed reregulation dam. Shortly thereafter, the Corps determined that the second alternative of the MAAWRMS—reallocating storage in Lake Lanier instead of constructing a new dam—would be more economical. The change was based, at least in part, on an environmental study generated by new computer models that concluded that the costs of acquiring the land flooded by the reregulation dam could rise and make the first alternative less economical than originally thought. U.S. Army Corps of Eng'rs: Mobile District, Additional Information, Lake Lanier Reregulation Dam 2 (1988).

In the late 1980s, the Corps began to prepare a Post-Authorization Change Notification Report ("PAC Report") suggesting that the authorization for the new reregulation dam be set aside in favor of the reallocation of storage alternative. A draft of the PAC Report was completed in 1989. The draft recommended that 207,000 acre-feet be allocated to water supply, allowing for 151 mgd to be withdrawn directly from Lake Lanier and 378 mgd (51 mgd more than the 327 mgd that the Corps determined was available as an incident of power supply) from the river downstream. This represented a significant increase from the 142,000 acre-feet of storage recommended by the second alternative of the MAAWRMS. The draft PAC Report also included a draft Water Control Manual that would have replaced the manual from 1958 and governed the Corps' water operations in the ACF basin. In this appeal, the Corps claims that the PAC Report's recommendations would have been made pursuant to authority from the WSA. The draft report itself noted the Corps' authority under the statute but stated that approval from Congress might be required due to the fact that the allocation exceeded 50,000 acre-feet.9 The draft report estimated that the purchased storage in Lake Lanier would cost $49,360,600. The final PAC Report was never completed due to resistance and the initiation of a lawsuit by the State of Alabama.

The Corps followed the recommendation of the MAAWRMS to make water available for water supply in the interim before a long-term solution was reached, as it had done while the study was being completed, and it entered into a temporary water supply contract with the ARC. The contract was based on the Corps' revised determination that it could provide, incidentally to power generation, 327 mgd as a year-round average with no impact on hydropower. The Corps agreed to provide releases sufficient to accommodate up to 50 mgd in withdrawals above this 327 mgd threshold, for which the ARC would pay. The contract was renewed in 1989 but expired in 1990. Since then, the ARC has continued to withdraw water from the Chattahoochee on roughly the same basis as that specified in the contract, though it has generally not needed to withdraw more than 327 mgd.

The Corps signed several other water supply contracts in the 1970s and 1980s, all of which expired in 1990, but which roughly dictate the terms under which the localities have continued to withdraw water from the Buford Project. In 1978, the Corps agreed to terms with the City of Cumming, Georgia for the paid withdrawal of 2.5 mgd. In 1985, the amount was raised to 5 mgd, and in 1987 it was raised to 10 mgd. In 1987, the Corps signed a contract with the City of Gainesville, allowing the city to withdraw up to 20 mgd, up from the 8 mgd authorized in 1953 as just compensation. The contract required Gainesville to pay for the water that it withdrew in excess of 8 mgd. In 1988, the 1973 contract with Gwinnett County was supplemented, expanding the cap on withdrawals from 40 mgd to 53 mgd. All of the contracts signed in the 1980s specifically stated that they were interim contracts to satisfy water supply needs while the Corps was studying the issue and determining a permanent plan. As a component of their interim nature, the contracts explicitly stated that they did not create any permanent rights to storage space in Lake Lanier. The Corps originally cited the Independent Offices Appropriations Act of 1952, 31 U.S.C. § 9701, as authority for these contracts, but it later deemed the statute to be ineffective in authorizing such transactions. Later contracts cited the Water Supply Act for authority.

On January 1, 1990, all of the interim contracts expired. The only remaining water supply allocations were the combined 10 mgd granted to Buford and Gainesville in the 1950s by the Corps as just compensation. However, the Corps continued to permit the localities to withdraw water from the Buford Project for water supply. Appellees refer to these water withdrawals as pursuant to "holdover" contracts.

In June, 1990, Alabama filed suit against the Corps in the Northern District of Alabama to challenge a section of the draft PAC Report and the continued withdrawal of water from the Buford Project by the Georgia Parties, which Appellees characterize as a de facto reallocation of storage. This suit is the first of the four currently on appeal. In September of 1990, Alabama and the Corps moved jointly for a stay of proceedings, which was granted, to negotiate a settlement agreement. Florida and Georgia later intervened as plaintiff and defendant, respectively. The stay order at issue in Alabama v. United States Army Corps of Engineers, 357 F.Supp.2d 1313 (N.D.Ala.2005), required the Corps not to "execute any contracts or agreements which are the subject of the complaint in this action unless expressly agreed to, in writing, by [Alabama] and Florida." Id. at 1316. The stay provided that either side could terminate it at will. It did not discuss the continued water withdrawals of the Georgia Parties.

In 1992, Alabama, Florida, Georgia, and the Corps entered into a Memorandum of Agreement ("MOA") authorizing a comprehensive study of the water supply question and requiring the Corps to withdraw the draft PAC Report along with its accompanying Water Supply Reallocation Reports and Environmental Assessments. The MOA contained a "live-and-let-live" provision that allowed the Georgia Parties to continue to withdraw water from the Buford Project at the level of their withdrawals in 1990, with reasonable increases over time. The provision made clear that it did not grant any permanent rights to the water being consumed. The MOA was originally set to last for three years but was extended several times.

In 1997, after the completion of a comprehensive study, the parties entered into the Apalachicola-Chattahoochee-Flint River Basin Compact ("ACF Compact"), which was ratified by Congress and the three states and replaced the MOA. Pub.L. No. 105-104, 111 Stat. 2219 (1997). The ACF Compact included a provision allowing continued withdrawals similar to the live and let live provision in the MOA. The Compact created an "ACF Basin Commission" composed of the governors of the three states and a non-voting representative of the federal government, to be appointed by the President. The commission was charged with establishing "an allocation formula for apportioning the surface waters of the ACF Basin among the states of Alabama, Florida, and Georgia." Id. art. VI(q)(12), 111 Stat. at 2222. Under the Compact, existing water supply contracts would be honored, and water-supply providers could increase their withdrawals "to satisfy reasonable increases in the demand" for water. Id. art. VII(c), 111 Stat. 2223-24. The Compact initially was scheduled to expire December 31, 1998, but it was extended several times; it ultimately expired on August 31, 2003, when the Commission failed to agree on a water allocation formula. The stay of the Alabama case remained in effect through the duration of the Compact. In the meantime, the Corps continued to allow the Georgia Parties to withdraw water from the Buford Project. Because there were no contracts in place, the Corps froze the rates that it charged for water and continued to proceed on the basis of the prices that were set in the interim contracts of the 1980s. This rate scheme angered hydropower customers who purchased power produced by the project directly or indirectly from SEPA.

In December 2000, SeFPC filed suit under the APA against the Corps in the United States District Court for the District of Columbia, the second of the four suits currently on appeal. SeFPC alleged that the agency had wrongfully diverted water from hydropower generation to water supply, thereby causing SeFPC's members to pay unfairly high rates for their power. SeFPC sought a judicial declaration of the Buford Project's authorized purposes as well as compensation. In March 2001, the district court referred the parties to mediation, and Georgia was joined. In January 2003, SeFPC, the Corps, and the Georgia Parties agreed to a settlement in the case, which called for an allocation of 240,858 acre-feet (estimated to be 22% of conservation storage) to water supply for once-renewable 10-year interim contracts that could be converted into permanent storage if approved by Congress (or if a court deemed Congressional approval unnecessary). In exchange, the Georgia Parties agreed to pay higher rates for water, with the income being applied as a credit against the rates charged to SeFPC's members. The D.C. district court then allowed Alabama and Florida to intervene.

In October 2003, the Alabama court enjoined the filing of the settlement agreement in the D.C. case, finding that the agreement violated the stay in its case because the approval of Alabama and Florida was not obtained. The district court in the District of Columbia approved the Agreement in February 2004, contingent upon the dissolution of the Alabama court's injunction, rejecting Alabama and Florida's argument that the Agreement exceeded the Corps' authority conferred by Congress. SeFPC v. Caldera, 301 F.Supp.2d 26, 35 (D.D.C.2004). In April 2004, a panel of this Court stayed an appeal of the Alabama court's injunction to allow the Alabama district court to decide whether to dissolve or modify the injunction in light of the D.C. district court's order approving the Agreement. In the meantime, an initial appeal of the D.C. district court's order in SeFPC by Alabama and Florida was denied by the D.C. Circuit Court of Appeals for lack of a final judgment. The Alabama district court denied a motion to dissolve the preliminary injunction, Alabama, 357 F.Supp.2d at 1320-21, but in September 2005, a panel of this Court held that the district court abused its discretion in granting the injunction and vacated the injunction. Alabama v. U.S. Army Corps of Engineers, 424 F.3d 1117, 1133 (11th Cir.2005). Once this Court dissolved the injunction over the implementation of the Agreement, the D.C. district court in March 2006 entered a final judgment in SeFPC, and Alabama and Florida again appealed to the D.C. Circuit.

In Southeastern Federal Power Customers, Inc. v. Geren, 514 F.3d 1316, 1324 (D.C.Cir.2008), the D.C. Circuit held that the settlement agreement exceeded the Corps' authority under the WSA. The parties to the settlement agreement argued that the Corps was authorized to enter into the settlement on the basis of its WSA authority alone, so the court specifically refrained from making any holdings on the basis of the RHA. Id. at 1324 n. 4. The court found that "[o]n it's face, ... reallocating more than twenty-two percent... of Lake Lanier's storage capacity to local consumption uses ... constitutes the type of major operational change referenced by the WSA." Id. at 1324. After the circuit court's remand, the Judicial Panel on Multidistrict Litigation transferred the case, along with Alabama and several others, to the Middle District of Florida.

Meanwhile, in 2000, the State of Georgia submitted a formal request to the Corps to modify its operation of the Buford Project in order to meet the Georgia Parties' water supply needs through 2030. The request was to withdraw 408 mgd from the river and 297 mgd directly from the lake, which, combined, required approximately 370,930 acre-feet of storage. In February 2001, nine months after the request was sent to the Corps and without a response from the Corps, the State of Georgia filed suit in the United States District Court for the Northern District of Georgia seeking to compel the Corps to grant its request, beginning the third of the four underlying cases. The Corps responded in April 2002 with a letter denying the request and an accompanying legal memorandum. Memorandum from Earl Stockdale, Deputy General Counsel, Department of the Army, to Acting Assistant Secretary of the Army for Civil Works: Georgia Request for Water Supply from Lake Lanier 1 (Apr. 15, 2002) [hereinafter "2002 Stockdale Memo"]. The 2002 Stockdale Memo concluded that the Corps lacked the authority to grant Georgia's request without legislative approval. The memo stated that water supply was not an authorized purpose of the Buford Project under the RHA. Further, it stated that even if water supply was authorized, the Corps would still lack the authority to make a storage allocation of the size requested because the reallocation "would involve substantial effects on project purposes and major operational changes." Id. The district court denied Florida's and SeFPC's motions to intervene in the case, but this Court reversed the denial and remanded for further proceedings. Georgia v. U.S. Army Corps of Eng'rs, 302 F.3d 1242, 1260 (11th Cir. 2002). On remand, Florida moved to dismiss or abate the proceedings; Alabama, Gwinnett County, the City of Gainesville, and the ARC moved to intervene; and Alabama moved to abate or transfer the proceedings. The district court allowed Alabama to intervene as of right and the local governments to intervene permissively, and it held that the case would be abated pending the resolution of the Alabama case. Georgia v. U.S. Army Corps of Engineers, 223 F.R.D. 691, 699 (N.D.Ga. 2004). On appeal, a panel of this Court in an unpublished decision affirmed the district court's decision to abate the case. Georgia v. U.S. Army Corps of Engineers, 144 Fed.Appx. 850 (11th Cir.2005). The case was then consolidated into the multidistrict litigation in the Middle District of Florida.

In January 2008, the City of Apalachicola sued the Corps in the federal district court for the Northern District of Florida. This is the last of the four cases being considered as part of this appeal. This case was also consolidated into the multidistrict litigation.

While the litigation was pending, the Corps began an update of its plans and operations in the ACF Basin with a focus on whether it could continue to meet the current water supply needs of the localities. In order to answer these question, specifically in light of the D.C. Circuit's Geren opinion, the Corps released a new legal memorandum by Earl Stockdale. Memorandum from Earl Stockdale, Chief Counsel, Department of the Army, to the Chief of Engineers: Authority to Reallocate Storage for Municipal & Industrial Water Supply Under the Water Supply Act of 1958, 43 U.S.C. § 390b 1 (Jan. 9, 2009) [hereinafter "2009 Stockdale Memo"]. In this memorandum, the Corps determined that the current water supply withdrawal under the "interim contracts" could be accommodated by a permanent reallocation of approximately 11.7% of the conservation storage in Lake Lanier. The Corps concluded that such a permanent reallocation would not constitute a major operational change, and it would not seriously affect any project purposes.

The Middle District of Florida, in its consideration of the multidistrict litigation, divided the trial into two phases. Phase One, which is at issue here, pertained to the Corps' authority for its operations of the project. The plaintiffs in the four underlying cases moved for summary judgment, and the Corps filed an opposition and cross-motion for summary judgement in each case. On July 17, 2009, the court granted partial summary judgment to the plaintiffs in Alabama, Apalachicola, and SeFPC and to the Corps in Georgia, and it denied summary judgment to the Georgia Parties.

The court's order concluded that the Corps had exceeded its authority in its "de facto" reallocation of storage to accommodate current water supply withdrawals. In re Tri-State Water Rights Litig., 639 F.Supp.2d 1308, 1350 (M.D.Fla.2009). The court first held that only two conclusions from of the D.C. Circuit had preclusive effect on its judgment under the principle of collateral estoppel: (1) that the WSA applied to interim reallocations of storage; and (2) that a reallocation of 22% of Lake Lanier's total conservation storage was a major operational change under the WSA. Id. at 1343. Next, the district court concluded that there was virtually no authorization for the reallocation of water supply storage in the RHA and that the Corps' sole source of authority to allocate storage for water supply was the WSA.

The court went on to hold that the 2009 Stockdale Memo was a litigation document with post hoc analysis that was not part of the administrative record. Id. at 1347. Without the memo, the district court concluded that the record contained insufficient support for the Corps' calculations of the amount of storage required for the water supply withdrawals as of 2006. The court attempted its own calculation of this figure and determined that the allocation was for 226,600 acre-feet or 21.5% of Lake Lanier's total conservation storage. Id. at 1350. In reaching its conclusion on the amount of storage necessary for the project's current operations, the court rejected several key figures that the Corps had used in making previous calculations, most notably rejecting the Corps' figure for the amount of water available for downstream withdrawal as a byproduct of hydropower operations. The district court held that the 21.5% allocation was a major operational change that exceeded the Corps' WSA authority. The court also concluded that the Corps' current operations exceeded the WSA because they seriously affected the authorized purpose of hydropower generation. Because the Georgia request represented an even larger water supply storage allocation than the current operations, the court also found that it exceeded the Corps' authority.

The district court directed the Corps to limit releases from the Buford Project to 600 cfs during off-peak hours and to discontinue all water supply withdrawals being made directly from Lake Lanier, except for the 10 mgd that Gainesville and Buford had been permitted to withdraw in their 1950s relocation agreements.10 The court stayed its order for three years, until July 17, 2012, to give the parties an opportunity to settle or to seek Congressional approval. In the meantime, the court allowed current withdrawals to continue but forbade any increases without the consent of all of the parties.

DISCUSSION

This opinion will begin by examining threshold jurisdictional questions, then will analyze the primary substantive matters involved, and finally will provide some guidance and instruction for the Corps pertaining to its analysis of its water supply authority on remand.

Part I. Jurisdictional Matters

A. Appellate Jurisdiction over Alabama, SeFPC, and Apalachicola

Alabama and Florida argue that this court lacks appellate jurisdiction over the appeals in Alabama, SeFPC, and Apalachicola. They note that the district court did not render a final judgment in the cases, as the summary judgment order on the Phase One claims did not resolve the Phase Two claims in those cases. Appellees concede that this court has jurisdiction over the appeal in the Georgia case because the district court's order did amount to a final judgment in that case. However, Alabama and Florida argue that the claims in the cases are sufficiently distinct that extending pendent jurisdiction from Georgia over the Alabama, SeFPC, and Apalachicola claims would be inappropriate.11 We disagree. Issues in the three contested cases and the Georgia case are inextricably intertwined, rendering pendent jurisdiction proper. Even if this Court did not have pendent jurisdiction over these claims, this Court would still have jurisdiction because the district court's order amounted to an injunction.

"Pendent appellate jurisdiction is present when a nonappealable decision is inextricably intertwined with the appealable decision or when review of the former decision is necessary to ensure meaningful review of the latter." King v. Cessna Aircraft Co., 562 F.3d 1374, 1379 (11th Cir. 2009) (internal quotation marks omitted). The exercise of such jurisdiction is only appropriate in "rare circumstances" so only "limited factual scenarios" will qualify. Id. at 1379-80. Thus, the critical inquiry is whether the appealable issue can be resolved without reaching the merits of the nonappealable issues. Thomas v. Blue Cross & Blue Shield Ass'n, 594 F.3d 814, 821 (11th Cir.2010).

Alabama and Florida argue that the appeal of Georgia can be resolved without addressing the issues in the other cases. They argue that the district court found that Georgia's request for a reallocation of 34% of the available storage exceeded the Corps' authorization because the argument was simply foreclosed by collateral estoppel and the preclusive effects of the D.C. Circuit's Geren decision. Thus, they argue that the merits of the Georgia appeal can be determined without considering the underlying issues in the other cases.

Alabama and Florida misread both the district court's opinion and the opinion in Geren. The district court did not, and could not, simply dismiss the Georgia case on the basis of collateral estoppel. The Corps' authority to allocate storage for water supply depends on an analysis of both the RHA and the WSA. As noted above, the district court held that only two issues were precluded, and neither of them were the authority of the Corps under the RHA. In fact, the D.C. Circuit clearly stated that the issue of water supply authority in the RHA was not before it. See Geren, 514 F.3d at 1324 n. 4 ("The court, in responding to the Corps' defense of its approval of the Agreement, has no occasion to opine whether the Corps' previous storage reallocations were unlawful."). Thus, it is clear that the holding in Geren—i.e. that a 22% reallocation of storage to water supply constitutes a "major operational change" under the WSA—cannot operate as collateral estoppel with respect to the issue of the Corps' combined authority under the RHA and the WSA.12

Thus, the district court did not apply collateral estoppel in granting summary judgment against the Georgia Parties and holding that Georgia's 2000 request exceeded the Corps' authority. Rather, the district court came to an independent conclusion on this matter and found that water supply was not an authorized purpose under the RHA. Tri-State, 639 F.Supp.2d at 1347. This finding was central to the court's holding in all four cases. Ultimately, it is impossible for this Court to rule on the merits of the appeal in Georgia without determining whether water supply was an authorized purpose of the Buford Project under the RHA. Thus, the issues raised in the various appeals are inextricably intertwined and pendent jurisdiction is proper in Alabama, SeFPC, and Apalachicola.

As an alternative basis for jurisdiction in the three cases (other than the Georgia case), this Court also has appellate jurisdiction over all of the underlying claims because the district court's order was an injunction. The order very clearly directs the parties to act, imposing a set of directives that if disobeyed could subject the parties to contempt proceedings. We are granted jurisdiction to review injunctions (or denials thereof) by 28 U.S.C. § 1292(a)(1). Alabama and Florida argue that this statute does not apply in this case because the district court's order is not an injunction. Instead, they argue, the district court merely set aside the Corps' actions because they were not in accordance with the law, and therefore in violation of the APA. See Citizens to Preserve Overton Park, Inc. v. Volpe, 401 U.S. 402, 413-14, 91 S.Ct. 814, 822, 28 L.Ed.2d 136 (1971) overruled on unrelated grounds by Califano v. Sanders, 430 U.S. 99, 105, 97 S.Ct. 980, 984, 51 L.Ed.2d 192 (1977). This argument is unpersuasive.

The district court's order is "a clearly defined and understandable directive by the court to act or to refrain from a particular action." Alabama, 424 F.3d at 1128. The district court mandated the return of operations to the levels of the mid-1970s by 2012, meaning that the Corps was required to set off-peak flows to 600 cfs and only Buford and Gainesville were allowed to withdraw any water directly from Lake Lanier (in the amounts established in their 1950s contracts). Tri-State, 639 F.Supp.2d at 1355. The court gave the Corps explicit instruction on how it was to act in the future. The Corps was also prohibited from entering into any new water supply contracts and therefore stripped of any discretion on how to allocate storage space for water supply. The district court did not refer to its order as an injunction, but the district court's intention in this regard is irrelevant. See Sierra Club v. Van Antwerp, 526 F.3d 1353, 1358-59 (11th Cir.2008) (utilizing a functional analysis to determine that the district court order was an injunction in spite of the district court's specific denial in this regard); United States v. Gila Valley Irrigation Dist., 31 F.3d 1428, 1441 (9th Cir. 1994) ("In determining whether or not an order is appealable under § 1292(a)(1), the courts do not look to the terminology of the order but to its substantial effect.") (citation omitted).

The district court's order was also sufficiently definite to be enforced via contempt proceedings. See Alabama, 424 F.3d at 1128. Federal Rule of Civil Procedure 65(d) requires that an injunctive order "state its terms specifically" and "describe in reasonable detail ... the act or acts restrained or required." Alabama and Florida do not question the definitive nature of the order in 2012. Rather, they focus on the court's directives in the interim period: "the parties may continue to operate at current water-supply withdrawal levels but should not increase those withdrawals absent the agreement of all other parties to this matter." Tri-State, 639 F.Supp.2d at 1355. Alabama and Florida argue that the order does not state with specificity the amount of water that each of the localities may withdraw. However, this is of no moment because no party would dare risk being held in contempt for violating the court's order by exploiting any of these ambiguities. The practical effect of this order makes it such that none of the localities would withdraw any more water than they did before the order was issued. Furthermore, it is irrelevant whether this injunction was defective under Rule 65(d) because this Court could still exercise jurisdiction under § 1291(a)(1) over a defective injunction. See Int'l Longshoremen's Ass'n, Local 1291 v. Phila. Marine Trade Ass'n, 389 U.S. 64, 76, 88 S.Ct. 201, 208, 19 L.Ed.2d 236 (1967). The proper remedy in such a case would be to vacate the injunction and remand the case to the district court.13 Because pendent jurisdiction is proper in this case and because the district court order is an injunction, this Court possesses appellate jurisdiction and will consider the merits of the issues raised by the parties.

B. Final Agency Action in Alabama, SeFPC, and Apalachicola

The Corps and the Georgia Parties argue that the district court did not have jurisdiction to hear the Alabama case, the SeFPC case, and the Apalachicola case because the Corps had not taken a final agency action, as required by the APA for judicial review. See 5 U.S.C. § 704. In these three cases, Appellees challenge what they have characterized as the "de facto reallocations"—the temporary water withdrawals the Corps has allowed and continues to allow. The Corps argues that it never made a formal reallocation of storage in the reservoir. Instead, it argues that it accommodated water supply under ad hoc arrangements with the localities and a series of agreements among all three States, while launching multiple, ultimately futile, attempts to reach a long-term solution to the issue. The parties all concede that the denial of Georgia's water supply request was a final agency action and that the district court possessed jurisdiction over the Georgia case. With respect to the other three cases, we conclude that there was no final agency action, and the district court therefore lacked jurisdiction to review the claims.

The APA states that "final agency action for which there is no other adequate remedy in a court [is] subject to judicial review." 5 U.S.C. § 704. The APA defines "agency action" as including "the whole or a part of an agency rule, order, license, sanction, relief, or the equivalent or denial thereof, or failure to act." 5 U.S.C. § 551(13). Because the definition of action under the APA is so broad, the critical inquiry is whether the action is final. Whitman v. Am. Trucking Ass'ns, 531 U.S. 457, 478, 121 S.Ct. 903, 915, 149 L.Ed.2d 1 (2001) ("The bite in the phrase `final action' . . . is not in the word `action,' which is meant to cover comprehensively every manner in which an agency may exercise its power. . . . It is rather in the word `final'. . . .") (citations omitted). The test for finality involves two steps:

Bennett v. Spear, 520 U.S. 154, 177-78, 117 S.Ct. 1154, 1168, 137 L.Ed.2d 281 (1997) (internal quotation marks and citations omitted).

We analyze whether the Corps' actions were final by using the two-step Bennett test. The Corps contends that it has not consummated its decisionmaking process because it has not made any final decisions on how to allocate water storage at Buford. The Corps notes that it never made any permanent water supply storage allocations and has not published any implementation guidelines. As evidence that no decisionmaking process has been consummated, the Corps notes that it has not performed the cost analyses or prepared the written reports required by the WSA, the Corps' internal guidelines, and the National Environmental Policy Act to make permanent reallocations. The Corps asserts that it attempted to start the decisionmaking process in 1989 with its draft PAC Report, which included a draft manual for operations in the ACF basin, but that the report was abandoned prior to its completion as part of the negotiations in the Alabama litigation.

The "de facto reallocations" do not meet the first prong of the Bennett test. They are based on contracts that have all expired and water withdrawals that have been extended on the basis of multi-party agreements and court orders. The various contracts that the Corps entered into with Gwinnett, the ARC, and the Cities of Gainesville and Cumming all specified their interim nature, expired in 1990, and did not purport to provide any permanent right to storage in Lake Lanier. As the Corps notes, these contracts are long expired now and are not themselves being challenged in this litigation.

What is being challenged is the continuous withdrawal of water from the Buford Project over the last forty years. Appellees assert that the Corps has utilized a practice of entering into temporary agreements in order to avoid the appearance of consummating its decisionmaking process. "[A]s a general matter, . . . an administrative agency cannot legitimately evade judicial review forever by continually postponing any consequence-laden action and then challenging federal jurisdiction on `final agency action' grounds." Nat'l Parks Conservation Ass'n v. Norton, 324 F.3d 1229, 1239 (11th Cir.2003) (citing Cobell v. Norton, 240 F.3d 1081, 1095 (D.C.Cir.2001)). The record in this case demonstrates that the Corps has not acted to avoid judicial review. Rather, the factual history of this case indicates that the Corps has made sincere efforts to effectuate permanent water supply allocations but has been thwarted by the litigation process.

The Corps has been attempting to reach a final decision on water storage allocations in the Buford Project since at least the mid-1980s, when it became aware that a permanent determination of water supply needs was vital. The agency concluded at the time that it was best to wait until the MAAWRMS was complete before making such a determination and to enter into interim contracts that would expire in 1990. Once the study was complete, the Corps embarked on the process of issuing the PAC Report and permanently reallocating certain amounts of storage to water supply—after first proposing and receiving Congressional authorization (but not funding) to build a reregulation dam. However, the Corps' plan to issue the PAC was derailed by developments in the Alabama case.

In 1992, the parties agreed to a stay and entered into a Memorandum of Agreement, which required that the Corps withdraw the PAC and prohibited it from entering into any new contracts. Memorandum of Agreement By, Between, and Among the State of Alabama, the State of Florida, the State of Georgia, and the United States Department of the Army 2 (Jan. 3, 1992). The Corps retained permission to continue to accommodate current withdrawal levels. In the memorandum, the parties agreed to conduct a Comprehensive Study, stating that "during the term of the Comprehensive Study, it is premature for the Army to commit, grant or approve any reallocation, allocation, or apportionment of water resources to service long-term future water supply." Id. Thus, during the duration of the term of the Memorandum of Agreement, the Corps was restricted from moving toward taking final agency action due to the terms agreed upon by the parties.

The same held true during the period of enforcement of the ACF Compact, the joint resolution that the three states agreed to and Congress ratified in 1997, which replaced the Memorandum of Agreement. The ACF Compact created an ACF Basin Commission charged with the power "to establish and modify an allocation formula for apportioning the surface waters of the ACF Basin among the states of Alabama, Florida and Georgia." 111 Stat. at 2222. The Compact specified that parties could continue withdrawing water but that no vested rights would be granted until the Commission adopted an allocation formula. Id. at 2223-24. The Compact provided that the Army Corps of Engineers "shall cooperate with the ACF Basin Commission in accomplishing the purposes of the Compact and fulfilling the obligations of each of the parties to the Compact regarding the allocation formula." Id. at 2225. Much like the Memorandum of Agreement before it, the ACF Compact, which remained in effect until 2003, restricted the Corps' ability to consummate a decisionmaking process on its water allocation policy. Thus, from 1992 to 2003 the Corps was operating under agreements signed by all three states that denied it the ability to make any permanent water supply allocations.

By the time the negotiations in the Alabama case fell apart in 2003, the Georgia Parties, the Corps, and SeFPC had entered into a settlement agreement in SeFPC. In pertinent part, the agreement set forth a process for entering into water supply contracts that could become permanent reallocations of storage. Evaluation of the legality of the settlement agreement in the D.C. district and circuit courts was delayed for several years by a preliminary injunction entered by the Northern District of Alabama. In March 2005, while the injunction was still in effect, the Corps filed a notice in the Alabama litigation that it intended to proceed with updating the water control plans and manuals for the ACF Basin. In response, Alabama's Congressional delegation sent a letter to the Corps stating its opposition to such actions during the pendency of the litigation, and the Corps abandoned its plans.

This Court vacated the Alabama court's injunction in late 2005 and the Corps informed the states that it considered the "relevant litigation" concluded and would proceed to update the water control manuals for the ACF Basin. However, on remand, the Alabama district court sent the parties into settlement negotiations, and the Corps again agreed to delay the update.

The SeFPC settlement was struck down in 2008 and never took effect. The Corps argues that it began the process of updating its operating manuals for the ACF basin almost immediately after the settlement agreement was invalidated. The Corps also issued a legal memorandum on its authority to allocate water storage at Buford, the 2009 Stockdale Memorandum. However, the district court deemed the 2009 Stockdale Memo to be a litigation document and not part of the administrative record. The district court issued its summary judgment order in this case on July 17, 2009, and the Corps asserts that this order once again thwarted it in its attempts to consummate the decisionmaking process. The Corps states that "every single day since 1990 the Corps was either operating under an agreement that barred it from formally taking any steps to reallocate storage, or was actively engaged in a process that could have led to a final agency action reallocating storage." The historical sequence of events supports the veracity of this claim.

This Court has accepted legal and practical barriers to administrative action as legitimate explanations for agency inaction.14 See Nat'l Parks, 324 F.3d at 1238, 1239 (holding that the National Parks Service's delay in implementing a management plan was excusable in part because a judicial order and a legislative mandate had prevented it from taking action). The courts have expressed a legitimate concern for agency avoidance of judicial review through intentional inaction. Id. at 1239. In this case, however, the lack of a definitive allocation of storage for water supply is explained by factors beyond the agency's control, rather than the Corps' inaction.

Appellees argue that the Corps has attempted to avoid judicial review of its management of the Buford Project, but they offer almost no evidence to support this contention. Alabama and Florida point out that the Corps, though required by court order not to enter into any new contracts, was not required to continue to allow the parties to withdraw water from the project. This argument misses the point. That the Corps chose to continue to permit water supply withdrawals sanctioned by the multi-party agreements does not demonstrate that the Corps effectuated a policy in regard to water supply and was attempting to avoid a judicial review of this policy. The states also note that the Corps had seven years since the time of the expiration of the agreements in 2003 to undertake a formal action. As stated above, the Corps attempted on multiple occasions after 2003 to begin the process of making final decisions on water allocations, but it was consistently thwarted by the litigation process. The Corps' two statements to the parties in 2005 that it intended to move forward with updating the water control manuals, the settlement agreement in SeFPC which was struck down in 2008, and the 2009 Stockdale Memorandum, demonstrate that the Corps intended to move forward in consummating a decisionmaking process after 2003 but could not. Appellees point to no specific final actions on the part of the Corps and there is insufficient evidence to conclude that the agency has attempted to avoid judicial review via incremental changes in operational policy. Thus, Appellees are unable to meet the first prong of the Bennett analysis.

The Corps' current operations of the Buford Project also do not meet the second prong of the Bennett test because the Corps' alleged "de facto allocations" are not actions "by which rights or obligations have been determined, or from which legal consequences will flow." 520 U.S. at 178, 117 S.Ct. at 1168 (internal quotation marks omitted). Current water supply withdrawals have taken place under the "live and let live" and other similar provisions in the stays, the Memorandum of Agreement, and the ACF Compact. These provisions, like the interim contracts that preceded them, have clearly stated the temporary nature of the allocations being made. As outlined above, the Corps has not had the opportunity to engage in a determination of rights or obligations due to the constraints imposed on it by the specific circumstances of the ongoing litigation. While it is true that access to water has been affected by the Corps' water supply allocations, as Appellees argue, that fact does not demonstrate that any future rights have been determined.15 Appellees are simply unable to produce evidence of such a determination.

Because there has been no final agency action in the Alabama, SeFPC, and Apalachicola actions, the district court lacked jurisdiction over these claims. Therefore, we vacate the district court's rulings in this regard and remand to the Corps to make final determinations pertaining to its current policy for water supply storage allocation.16

Part II. Georgia's 2000 Request: The Corps' Water Supply Authority Under the RHA

With respect to the merits, we turn first to the appeal in the Georgia case. This Court previously summarized Georgia's 2000 request of the Corps as follows:

Georgia, 302 F.3d at 1247-48.

The parties agree that the Corps' rejection of Georgia's 2000 request constituted a final agency action, of which both the district court and this Court have jurisdiction to review. Central to the Corps' rejection was the Corps' conclusion that water supply was not an authorized purpose of the Buford Project. We now hold that the Corps erred in drawing this conclusion. The text of the 1946 Rivers and Harbors Act—specifically, the Newman Report, whose language is incorporated into the statute—clearly indicates Congress' intent to include water supply as an authorized purpose in the Buford Project.

The 1945 and 1946 Rivers and Harbors Acts authorized the building of the Buford Project and serve as the baseline for the Corps' authority to operate the dam.17 The Georgia Parties contend that the district court seriously erred in its interpretation of the scope of its authority under the Act and neglected to note the specific authorization for water supply in the statute. The district court rejected this argument and held that water supply was only intended to be an incidental benefit of other operations and that the RHA did not authorize any storage for water supply in Lake Lanier.

The RHA authorized the development of the ACF Basin "in accordance with the report of the Chief of Engineers, dated May 13, 1946." 60 Stat. at 635. The Chief of Engineers Report incorporated the Division Engineer's Report—i.e. the Newman Report. Thus, the statute fully incorporated the terms of the Newman Report. The Newman Report specifically modified the recommendations of the Park Report—the foundational report for the 1945 RHA—by proposing the building of a single multi-purpose reservoir upstream of Atlanta instead of three separate reservoirs. One advantage of such a move was that the new dam could more easily accommodate water supply needs.

As the Newman Report made clear, the dam was designed with water supply specifically in mind. At times, water supply was even to be accommodated at the expense of optimal hydropower generation. The Newman Report explained:

Newman Report ¶ 80.

There are several critical provisions in this paragraph of the report. First, Congress contemplated that "the estimated present needs of the city" for water supply would be met by the initial Project by "insur[ing] at all times a flow at Atlanta not less than 650 second-feet." Id. Indeed, Congress provided in the construction of the project that, in addition to the two large turbines, there would be "a small generator to generate off-peak power as secondary energy, reserving the remaining storage for peak operation." Id. This small turbine was constructed for the sole purpose of providing off-peak releases of 600 cfs (without completely losing the value of these releases for power generation). If operated for maximum power value, the dam would be shut down during off-peak hours and would release no water, limiting water flow at Atlanta to a rate of 50 cfs. Such an operational scheme would not require a small turbine at all. Thus, the design of the project and the operational scheme were influenced by water supply concerns.

Second, the minimum flow was provided for notwithstanding the clear Congressional intent that it would be at the expense of "maximum power value." Id. Congress recognized that shutting down all water releases during off-peak hours would create an insufficient flow in the river downstream at Atlanta to meet the city's current water supply needs. Therefore, even though Congress recognized that there would be some detriment to power generation, it nevertheless provided for a minimum flow of 650 cfs at Atlanta. The Newman Report, and thus the authorizing legislation itself, explicitly stated as much. The legislation provided, in connection with the initial minimum release requirement and the contemplated increases thereof, that "the benefits to the Atlanta area from an assured water supply for the city . . . would outweigh any slight decrease in system power value." Id.

Third, Congress recognized that "[t]his minimum release may have to be increased somewhat as the area develops." Id. Indeed, in the immediately preceding paragraph, the authorizing legislation considered the growing need for water supply to the metropolitan area over a future 19-year period and spoke of increasing off-peak releases to accommodate the water supply needs of the Atlanta area in 1965. That future water supply need was estimated to require a river flow at Atlanta of 800 cfs. Id. ¶ 79. Thus, the original authorizing legislation expressly contemplated a very substantial increase in the operation of the Buford Project to satisfy the water supply needs of the Atlanta area (i.e., from 650 cfs to 800 cfs in the river flow at Atlanta).19

In light of the foregoing statutory language, and particularly Congress' intent that the Corps should have authority to accommodate the Atlanta area's water supply needs at the expense of some detriment to "system power value," we cannot conclude that Congress intended for water supply to be a mere incidental benefit. By definition, one purpose that is to be accomplished to the detriment of another cannot be incidental.20 Thus, the language of Sections 79 and 80 clearly indicates that Congress intended for water supply to be an authorized, rather than incidental, use of the water stored in Lake Lanier.21

Appellees argue that the Newman Report's references to water supply as "incidental" demonstrates that water supply was not an authorized purpose of the Buford Project. This is an attractive proposition due to its simplicity, but the context of these references undermines this claim. The language in question is as follows:

Newman Report ¶ 68 (emphasis added). We conclude that this single reference to water supply as an "incidental benefit" was an explanation for why the dam would be built above Atlanta and was not meant to confer a subordinate status.

The Corps in the Park Report proposed the construction of three dams at Cedar Creek, Lanier, and Roswell. Id. The Corps subsequently determined in the Newman Report that the system of dams in the ACF Basin would operate substantially more efficiently if one large dam was built instead. The agency decided to locate the dam at Buford, approximately 47 miles upstream of Atlanta. Paragraph 68 was an explanation for why the Corps deemed it beneficial to build the dam at this location; the explanation: water supply. An upstream location would allow the Corps to secure Atlanta's water supply as an incident of the other authorized purposes. That is to say that the aim of benefitting water supply could be accomplished without any significant detriment to hydropower, navigation, or flood control. The report stated that the revised location and size of the dam and reservoir would result in "greatly increase[d] . . . minimum flow in the river at Atlanta, thereby . . . reinforcing and safeguarding the water supply of the metropolitan area." Id. This benefit would be incidental to power generation because the water constituting the river flow at Atlanta would have generated power as it passed through the generators. There is no indication that the use of the word "incidental" in Paragraph 68 was meant to describe the importance of water supply to the project or even the importance of water supply vis-a-vis the other project purposes.

This reading is further supported by the phrase "safeguarding the water supply of the metropolitan area." Id. The fact that references to incidental benefits and the safeguarding of water supply were made in the same breath demonstrates that the Newman Report did not use the term as an indication of a subordination of the importance of water supply. Instead, the "safeguarding" language of Paragraph 68 indicates the critical nature of the water supply purpose to the project. In fact, the Newman Report went on to describe the importance of the Buford Project for the protection and assurance of Atlanta's water supply on at least four other occasions. Id. ¶¶ 73 ("would ensure an adequate water supply for the rapidly growing Atlanta metropolitan area"), 80 ("insure at all times a flow at Atlanta") ("the benefits to the Atlanta area of an assured water supply for the city"), 100 ("would ensure an adequate municipal and industrial water supply for the Atlanta area"). Congress' focus on the need to ensure the Atlanta area's water supply serves as strong evidence of the primary role given to water supply in the project.

The only other reference in the authorizing legislation to water supply as incidental appears in the Newman Report at Paragraph 100.22 There, the final sentence of the paragraph begins with the word `incidentally' and lists several project benefits that would not significantly harm other project purposes. One of the listed benefits is flood control, which Appellees concede is an authorized purpose. This use of the term `incidentally' cannot be construed to mean that water supply was intended to be a subordinate use because flood control is referred to in the same manner in the sentence. This fact is further illustrated by yet another reference to the protection of water supply in the same sentence—"would ensure an adequate municipal and industrial water supply for the Atlanta area." Id. ¶ 100. Again, as in Paragraph 68, the meaning conveyed in Paragraph 100 is a description of how the several authorized purposes could be accomplished harmoniously and the manner in which all were better served by locating the project at Buford. For these reasons, and especially because of the clear language in Paragraphs 79 and 80 of the authorizing legislation, we do not read the sparse use of the term "incidental" as indicative of the status of water supply as an authorized use vel non.

Appellees argue that the original project did not contemplate storage in Lake Lanier for water supply and that this is an indication that Congress did not intend for water supply to be an authorized purpose. We disagree. The lack of storage allocation for water supply sheds no light on the intentions of Congress. No storage allocation was specified for navigation in the Newman Report even though navigation is universally accepted as an authorized purpose of the Buford Project. See H.R. Doc. No. 80-300, Letter from Lieutenant General R.A. Wheeler, Chief of Engineers, ¶ 11(d). Furthermore, no storage was needed at the time for water supply. Almost all of the Atlanta area's water supply requirements could be met at the time as an incident to, or byproduct of, the generation of power. Thus, the lack of initially allocated storage for water supply is not at all inconsistent with the Congressional intent that water supply was an authorized purpose.

For the same reason, we believe that the fact that the localities were not asked initially to contribute to the costs of the project is of no moment in determining Congress' intent with respect to water supply authorization. Georgia, in 1946, did not require a significant amount of water beyond that which was provided by normal project operations for power generation, so a request for state contribution to the project would not have made sense. It would have meant asking the state to pay for a service that the Corps could provide essentially without cost. Moreover, at that time, Atlanta's current water supply usage required a flow of the river at Atlanta of 635 cfs.23 Even "[d]uring the extremely low-flow month of October 1941, the average flow for the month was 493 second-feet, and the minimum daily flow [was] 422 second-feet." Id. ¶ 79. In other words, before the Buford Dam was built, the river was providing the water supply needs of the Atlanta area. The requirement in the legislation that the Corps make releases "so as to [e]nsure at all times a flow at Atlanta not less than 650 second-feet," id. ¶ 80, merely provided water supply roughly commensurate to that which the river was already providing. It is not likely that the Corps or Congress would have thought it appropriate to charge Atlanta for construction costs of a project that merely replaced its currently available water supply.

The Corps could potentially have asked Georgia to pay on the basis of future water supply needs that would affect project operations, but it was not at all clear how much water would be needed in the future.24 After all, it would be almost 30 years after the Newman Report before the Corps would sign its first water supply contract—excluding the small relocation contracts made as compensation for inundating the intakes of Gainesville and Buford. Divining the value of water supply to the localities in the future would have resulted in speculative and potentially misleading results.25 Similarly, the fact that in its cost-benefit analysis, the Newman Report did not assign a particular dollar amount to water supply is not an indication that it was not authorized because the benefits of water supply were indeterminate at the time.26

One final point merits mentioning. Before the Dam was built, or even planned, the Chattahoochee provided almost all of the City of Atlanta's water supply. The building of the dam could have been a potential threat to the city's ability to withdraw water from the river because the Corps had an incentive—optimal power generation—to shut off all water flow in the river for long stretches of time. Congress responded to this concern by establishing a minimum flow requirement and noting that this requirement might have to be increased over time. Congress also clearly indicated that the Buford Project was intended to benefit the Atlanta area's needs by assuring the water supply. If water supply had been deemed a subordinate purpose by Congress, the Buford Project would have been detrimental, rather than beneficial, to the Atlanta area's water supply needs. That is to say, if the only water being supplied was to be a subordinate byproduct of power generation, then the City of Atlanta would have eventually found itself able to withdraw less water from the river than it would have been had no dam been built at all. In light of the repeated references in the authorizing legislation to safeguarding and ensuring an adequate water supply for Atlanta, Congress very clearly did not intend the dam to harm the city's water supply.

The language of the RHA clearly indicates that water supply was an authorized purpose of the Buford Project. Appellees' arguments to the contrary are unconvincing for all of the reasons mentioned above. Thus, we conclude that water supply was an authorized purpose of the RHA and that the RHA authorized the Corps to allocate storage in Lake Lanier for water supply.

Part III. Georgia's 2000 Request Must Be Remanded to the Corps

The Corps argues that its interpretation of the RHA in the 2002 Stockdale Memo, which supplied the legal reasoning for the denial of Georgia's water supply request, is entitled to deference from this Court. The RHA authorized the Corps to build the Buford Project in accordance with the Corps' plans and "with such changes therein as in the discretion of the Secretary of War and the Chief of Engineers may be advisable." H.R. Doc. No. 80-300, Letter from Lieutenant General R.A. Wheeler, Chief of Engineers ¶ 16. The Corps asserts that this gives the agency wide latitude in its interpretive authority. Additionally, the Corps argues that because it prepared the reports which comprise the language of the RHA, its "interpretation of the statute merits greater than normal weight because it was the [Corps] that drafted the legislation and steered it through Congress with little debate." Howe v. Smith, 452 U.S. 473, 485, 101 S.Ct. 2468, 2476, 69 L.Ed.2d 171 (1981). Despite the high level of respect owed to the Corps' interpretations with regard to the RHA due to its unique role in shaping the statute, we cannot defer to the Corp's interpretation of its water supply authorization in this instance. Even heightened deference cannot lead this Court to ignore the plain and express will of Congress, especially where, as here, the Corps' interpretation has not been consistent.

Under the APA, reviewing courts must set aside agency action that is "arbitrary, capricious, an abuse of discretion, or otherwise not in accordance with law." 5 U.S.C. § 706(2)(A). The denial of Georgia's request was based on a clear error of law— the Corps' misinterpretation of the RHA. Therefore, the Corps' interpretation cannot be granted deference, in spite of the agency's role in drafting the language of the legislation.

The seminal case Chevron, U.S.A., Inc. v. Natural Resources Defense Council, Inc., 467 U.S. 837, 842-44, 104 S.Ct. 2778, 2781-82, 81 L.Ed.2d 694 (1984), set up a two-step framework for evaluating whether a court must defer to an agency's construction of a statute it is charged with administering. Deference from the court is due if (1) Congress has not spoken directly on the precise question at issue and its intent is unclear, and (2) the agency's interpretation is based on a permissible construction of the statute. Id. The argument for Chevron deference in this case fails at both steps because Congress made clear its intention that water supply was an authorized purpose of the Buford Project. As discussed above, the Newman Report repeatedly stated that the Buford project would protect and assure the water supply of the Atlanta metropolitan area. Furthermore, the Report authorized the use of water for water supply at the expense of maximum hydropower generation. Congress' acknowledgment that water supply, in certain instances, was to be provided at the expense of maximum power generation necessitates the conclusion that water supply was not to be subordinate to other project purposes and was instead an authorized purpose in its own right. The Corps' interpretation that the RHA relegated water supply to incidental status cannot be reconciled with the plain language of the statute. The clear Congressional intent in the 1946 RHA was that water supply was to be an authorized purpose, and the Corps' contrary interpretation is erroneous and cannot be accepted by this Court.27

A significant fact undermining any deference to the Corps on this issue is the fact that the Corps has also been inconsistent in its statements about whether water supply was an authorized purpose. The 2002 Stockdale Memo concluded (incorrectly) that water supply was not an authorized purpose, but this is not consistent with previous Corps statements on the matter. In the Corps' 1949 Definite Project Report, the Corps referred to water supply as one of the "primary purposes" and one of "the princip[al] purposes of the Buford Project." ¶¶ 48, 115. This 1949 Report was a formal pronouncement on the issue, and the one most nearly contemporaneous to the actual enactment. In a 1987 regulation, 33 C.F.R. § 222.5, App'x E, the Corps listed water supply as a project purpose for Buford. In the Corps' comprehensive 1994 report to Congress, which listed the authorized purposes for Corps projects across the country, water supply was included as an authorized purpose of Buford under the RHA.28 U.S. Army Corps of Eng'rs: Hydrologic Engineering Center, Authorized and Operating Purposes of Corps of Engineers Reservoirs E-94 (1994). That report also defined the term "incidental benefits" and stated that incidental benefits, though they were important, were not the subject of the report and would not be listed. Id. at 3-4. The Corps and the Appellees offer no explanation for why the Corps indicated that water supply was an authorized purpose in 1949, 1987, and 1994 but took a contrary position in 2002.29

The Corps argues that even if it erred in its interpretation of the RHA, its rejection of Georgia's water supply request should be allowed to stand. The Corps suggests that it accounted for the possibility that water supply was an authorized purpose of the Buford Project and still concluded that the request exceeded its authority. In support of this contention, the Corps points to the following language in the Stockdale Memo: "Even if water supply were a specifically authorized purpose of the reservoir (and the 1958 Act did not apply), the state's request would require substantial changes in the relative sizes of project purposes. This would represent a material alteration of the project, which would require congressional action." 2002 Stockdale Memo at 11 (internal quotation marks omitted).

The Corps argues that this constitutes an alternative conclusion and that there is no reason to believe that it is incapable of considering a legal hypothetical. It argues that remanding the case for a consideration that it already gave would be duplicative. However, an administrative agency's alternative explanation for denying a state's request is "arbitrary, capricious . . . or otherwise not in accordance with law" if it is based on an impermissible reading of the authorizing statute or statutes. See Massachusetts v. Envtl. Prot. Agency, 549 U.S. 497, 532-34, 127 S.Ct. 1438, 1462-63, 167 L.Ed.2d 248 (2007) (holding that the EPA's reading of a statutory phrase in its alternative explanation for why it did not regulate greenhouse gas emissions was not in conformity with the statute, and thus remanding to the EPA). The Corps' hypothetical, which is reiterated almost verbatim in the conclusion of the memo, 2002 Stockdale Memo at 13, rests squarely on an erroneous legal proposition. In this alternative hypothetical, the Corps mistakenly assumes that the WSA would not apply to the agency's determination of its authority to grant the Georgia request if water supply were authorized by the RHA. This assumption has no foundation in law. The WSA nowhere indicates that it is superceded by, or supercedes, original authorizations for water supply. The Act was merely intended to offer greater water supply authority in federal water projects than had previously existed. For that authority to be supplemental to authority already extant in a given project is perfectly consistent with the language and purpose of the statute. To assume, as the Corps has in its alternative conclusion, that the WSA does not apply to the Buford Project merely because the authorizing statute included water supply as an authorized purpose is not supported by the language of the WSA or by its intended aim of increasing water supply authority in federal projects. The Corps' holding in the alternative must be rejected because it misinterprets the scope of the WSA.

The Corps' alternative conclusion is also undermined because, despite the Corps' contentions otherwise, its misinterpretation of the RHA was essential to its conclusion that it lacked authority to grant Georgia's request. The majority of the memorandum is devoted to the potential effects of granting the request and whether these effects would be consistent with the Corps' authority under the WSA. The discussion of the agency's authority under the WSA is predicated on the assumption that the baseline level of authorization from the RHA is zero and that no storage may be allocated to water supply pursuant to the RHA. It is only at the very end of the discussion of the Corps' authority that the agency considers arguendo the possibility that the RHA authorized water supply. Its brief discussion of this alternative is flawed in two respects. First, as noted above, the Corps erroneously assumes that if the RHA included water supply as an authorized purpose, the WSA would not be applicable at all. Second, although purporting to assume water supply was an authorized purpose of the RHA, the Corps nevertheless underestimated its RHA authority. It failed to recognize that the authorizing legislation in 1946 not only included water supply as an authorized purpose but explicitly contemplated that the Corps was authorized to increase water supply usage over time as the Atlanta area grew and that this increase would not be a change from Congressionally contemplated operations at all. Thus, the Corps never considered its authority under the RHA to substantially increase its provision of water supply and reallocate storage therefor—authority which we hold today was granted by the RHA. And the Corps never considered its WSA authority to provide water supply as an addition to (or as supplementing) its RHA authority. The failure of the Corps in these respects renders its alternative reason for denying Georgia's request arbitrary, capricious, or otherwise not in accordance with the law.30

Several other factors also indicate that the Corps' rejection of the water supply request should be remanded for further consideration. First, attached to the 2002 Memo is a "preliminary analysis of the impacts of Georgia's water supply request on authorized project purposes and operations." The Corps' analysis on the effects of the Georgia request was thus incomplete. Because the Corps' authority to grant the request may be dependant on the precise size and effect of the request, it is crucial that the Corps complete its evaluation of the request. The need for further study recommends remand to the Corps.

Second, it is also apparent that the Corps' views regarding its authority to allocate storage in Lake Lanier to water supply are evolving and that it has not come to a final, determinative decision regarding the issues underlying this authority. There are several pieces of evidence for this. In 2002, the Corps rejected the Georgia request, asserting that it did not have sufficient authority to reallocate 34% of conservation storage. However, in 2004, it agreed to settle the Geren case, in part by reallocating what the settling parties determined at that time to be 22% of the conservation storage. The Corps determined that it could make such a reallocation on the basis of its WSA authority alone. Though these decisions are not directly conflicting, the Corps never explained why it believed that the 12% storage allocation difference between the two caused it to exceed its authority. If the RHA authorizes some storage reallocation to water supply, as we hold today that it does, then the Corps should explain why this difference in allocated storage between the Georgia request and the settlement agreement pushed it beyond the boundaries of its authority. Additionally, the Corps has revised its figures for how much storage must be allocated to accommodate current levels of water supply withdrawal. In the 2002 memo, the Corps asserted that current withdrawals required a 13% reallocation of conservation storage. On appeal, the Corps claims that the current withdrawal levels are only 11.7%. It appears that the Corps may no longer conclude that Georgia's request would require an allocation as large as 34%. Any such decrease in the Corps' projection of the amount of storage it deems required for water withdrawals could also affect its determination of its authority over the Georgia request.31

Finally, because the other matters in this appeal must be remanded to the Corps, it is sensible and efficient for the agency to consider the overlapping issues that are common to Georgia and the other cases together as part of a comprehensive decision about the Corps' future water supply operations. The conclusions that the Corps reaches with respect to the questions at issue in the other cases will provide it with a more complete analysis of the issues in Georgia, as well. For example, the Corps' determinations of its authorization over current water supply withdrawals will necessitate a thorough study of the amount of storage required for water supply. Also, this appeal represents the first opportunity for a court to consider the Corps' authority under both the RHA and the WSA. Our holding—that water supply is an authorized purpose under the RHA, that the Corps does have some authority under the RHA to balance as among the authorized uses and increase the water supply purpose at the expense of the power purpose and to reallocate storage therefor, and that the Corps' authority under the WSA is in addition to its authority under the RHA—constitutes a clarification of the legal environment which will aid the Corps in its analysis on remand. For these reasons, we conclude that the Corps must reexamine the request in light of its combined authorization under the RHA and WSA.32

Part IV. Gwinnett County's Claims Not Involving Authorization Under the RHA and WSA

Gwinnett County asserts three claims that are distinct from the claims of The Georgia Parties and the Corps. First, the county asserts that a 1956 Act of Congress authorized the Corps to contract with it for 10 mgd for water supply. Second, Gwinnett asserts that the Corps contracted with it to provide permanent storage for roughly 40 mgd. Finally, the county asserts that the Buford Project rendered its intake facility at Duluth, Georgia inoperable and that it is therefore entitled to water withdrawal rights as just compensation. We find merit in the first of these claims but reject the final two.

A. The Expiration of the 1956 Act

In 1956, Congress passed an act, in part, stating the following:

Pub.L. No. 84-841, 70 Stat. 725. The district court noted in a footnote that Gwinnett had not contracted with the Corps pursuant to this authorization and held that the authorization "expired in 2006." Tri-State, 639 F.Supp.2d at 1350 n. 24. The district court has misread the plain language of the statute. The fifty-year limitation in the Act refers to the duration of any contract with Gwinnett, not to the expiration of the Act itself. The phrase "not to exceed fifty years" immediately follows the words "contract . . . upon such terms and for such period" and there is no grammatical cue that it should not be read as modifying this phrase. The district court offers no explanation for its unnatural reading of the statute and none is evident to this Court. Moreover, the Act also authorized the Corps to enter into a perpetual easement with Gwinnett, authorizing Gwinnett to build the necessary facilities to withdraw water directly from Lake Lanier on the Corps' land. 70 Stat. at 725. It would be illogical for Congress to give Gwinnett a perpetual easement to implement an authorization that would expire in fifty years. The district court's interpretation of the Act, which is espoused by Appellees in this appeal, is inconsistent with the Act's language and its grant of an easement in perpetuity.

To date there has not been a single contract between the Corps and Gwinnett predicated on the authority of the 1956 Act. Such a contract in the future would not be a reallocation of storage under the WSA or the RHA because it is directly authorized by Congress.

B. Forty mgd from the 1974 Supplemental Agreement to the Corps' Contract

Gwinnett argues that in 1974 the Corps granted the county the right to 38,100 acre-feet of permanent storage so that it could withdraw roughly 40 mgd directly from Lake Lanier. In 1973, Gwinnett and the Corps entered into an interim water contract for 40 mgd. The following year, the parties revised Article 9 of the contract to provide:

Supplemental Agreement No.1 to Contract No. DACW01-9-73-624 Between United States and Gwinnett County, Georgia for Withdrawal of Water from Lake Sidney Lanier (Apr. 29, 1974). The contract stated that this revision was being made "in order to facilitate the sale of bonds to finance [Gwinnett's] proposed water works facilities." Id. Subsequent supplemental agreements extended the life of the contract until it was finally allowed to expire in 1990.

The contract gave Gwinnett the "right to acquire" the storage space at the time of the expiration of the contract. Thus, Gwinnett possessed an option (an offer) to purchase storage space at the time of the contract's expiration, which the parties agree occurred in 1990. "If no time is prescribed for accepting an offer, it must be done within a reasonable time." Wilkins v. Butler, 187 Ga.App. 84, 369 S.E.2d 267, 268 (1988) (quotation omitted); see Home Ins. Co. v. Swann, 34 Ga.App. 19, 128 S.E. 70, 72 (1925); Restatement (Second) of Contracts § 41 (1981). Gwinnett has not demonstrated that it exercised its acceptance of the option in 1990 or at any time since then. More than twenty years have elapsed since the time that the option became available, and the right to accept the Corps' offer to acquire the 38,-100 acre-feet of storage clearly has lapsed.33

C. Just Compensation for Relocation of the Duluth Intake

On appeal, Gwinnett argues that it should have been compensated because the creation of the Buford Project led to contamination of its intake structure at Duluth, which had to be abandoned in the early 1970s. Gwinnett failed to make this argument before the district court. We generally do not consider arguments raised for the first time on appeal and need not do so here. Peek-A-Boo Lounge of Bradenton, Inc. v. Manatee Cnty., 630 F.3d 1346, 1358 (11th Cir.2011).

In any event, this argument is meritless. Gwinnett fails to discuss the rights of the federal government to make alterations to navigable waters. The federal government possesses what is known as a navigational servitude, "the privilege to appropriate without compensation which attaches to the exercise of the power of the government to control and regulate navigable waters in the interest of commerce." United States v. Va. Elec. & Power Co., 365 U.S. 624, 627, 81 S.Ct. 784, 787-88, 5 L.Ed.2d 838 (1961) (internal quotation marks omitted). The navigational servitude is a dominant servitude, trumping all competing and conflicting rights to the waterway. Id. This servitude extends to the entire river and the riverbed lying below the high-water mark. United States v. Rands, 389 U.S. 121, 123, 88 S.Ct. 265, 267, 19 L.Ed.2d 329 (1967). It is anchored in Congress' commerce clause power. "The power to regulate commerce comprehends the control for that purpose, and to the extent necessary, of all the navigable waters of the United States. For this purpose they are public property of the nation, and subject to all the requisite legislation by Congress." Id. at 122-23, 88 S.Ct. at 266-67 (alteration omitted) (internal quotation marks omitted).

The federal government does not execute a taking of riparian interests by altering rivers for navigational purposes. The government's dominant right to make use of these waterways means that its actions do not amount to an appropriation. This premise has been explicitly stated several times in the context of hydropower interests: The federal government is not required to give compensation for water power when it takes riparian lands in accordance with the navigational servitude. E.g., Va. Elec., 365 U.S. at 629, 81 S.Ct. at 788; United States v. Twin City Power Co., 350 U.S. 222, 226-27, 76 S.Ct. 259, 262, 100 L.Ed. 240 (1956); United States v. Appalachian Elec. Power Co., 311 U.S. 377, 424, 61 S.Ct. 291, 307, 85 L.Ed. 243 (1940); United States v. Chandler-Dunbar Water Power Co., 229 U.S. 53, 73-74, 33 S.Ct. 667, 676, 57 L.Ed. 1063 (1913). Gwinnett offers no explanation for why this principle should not be applied to the riparian interest in water supply.

Because of the federal government's dominant right to make alterations in the river, the effect on Gwinnett's riparian interests is not a taking. Thus, even if Gwinnett had not abandoned its claim, the claim would not be compensable.

Part V. Remand Instructions to the Corps

On remand, the Corps is to reconsider Georgia's request, as well as its authority with respect to the current provisions for water supply, in light of its authority under the RHA as well as the WSA and the 1956 Act. In particular, it should consider several important factors with respect to the Newman Report (i.e., the RHA). First, the Corps should take into consideration that water supply for the Atlanta metropolitan area was an authorized purpose of the Buford Project as well as hydroelectric power, flood control, and navigation. Second, Congress contemplated that the Corps would be authorized to calibrate operations to balance between the water supply use and the power use. Third, because Congress explicitly provided that the "estimated present needs" of the Atlanta area for water supply be satisfied at the expense of "maximum power value," Newman Report ¶ 80, we know that the water supply use is not subordinate to the power use. Fourth, from Paragraphs 79 and 80 of the Newman Report, we know that Congress contemplated that water supply may have to be increased over time as the Atlanta area grows.

However, the authorizing legislation is ambiguous with respect to the extent of the Corps' balancing authority—i.e., the extent of the Corps' authority under the RHA to provide water supply for the Atlanta area. On the one hand, the authorizing legislation recognized that the Chattahoochee River was the source of the water supply for the Atlanta area, and the legislation repeatedly referred to safeguarding or assuring the water supply of the metropolitan area. See Newman Report ¶¶ 79 and 80. It also recognized that the minimum releases initially provided by the legislation to satisfy the present water needs "may have to be increased somewhat as the area develops." Id. ¶ 80. On the other hand, the legislation also contemplates that assuring such water supply for the Atlanta area can be done with a "slight decrease in system power value." Id.34 We conclude that the Corps, the agency authorized by Congress to implement and enforce this legislation, should, in the first instance, evaluate precisely what this balance should be.35

Once the Corps has determined the extent of its authority under the RHA, it should then determine its authority pursuant to the WSA. The authority under the WSA will be in addition to the Corps' authority under the RHA and the 1956 Act.

It is apparent from the record and the evolving position of the Corps that the Corps has not arrived at a final, definitive determination of the scope of its authority to allocate storage to water supply. For example, it is not clear whether the Corps has arrived at a firm calculation of how many gallons per day can be provided for the Atlanta area's water supply needs as a mere incident to, or byproduct of, power generation. The Corps' latest figure, developed in 1986, in this regard has been 327 mgd; however, at oral argument the Corps asserted that the calculation was not definitive and deserved more study. Also, it is apparent that the Corps has not arrived at a definitive, final determination of whether, and to what extent, storage reallocation would be necessary for RHA-authorized releases from the dam primarily for water supply purposes (and how to factor in the fact that these releases will still generate some power, though not of peak value). It is also unclear whether the Corps has arrived at a final determination of the appropriate measure for determining under the RHA what the impact of increased water supply use on power is, or the appropriate measure for determining under the WSA what constitutes a "major operational change."36 Finally, the Corps has not yet articulated a policy on whether to account for return flows, and if so, how to differentiate between flows returned directly to the lake and flows returned downstream from the dam. These are some of the questions that the Corps should answer on remand, although we make no attempt to be exhaustive in that regard.

As part of the final, definitive statement of the Corps' water supply analysis, if the agency ultimately concludes that it does not have the authority to grant the Georgia request, it nevertheless should indicate the scope of the authority it thinks it does have, under the RHA, the WSA, and the 1956 Act. This way, the parties will have some further instruction, based on sophisticated analysis, of what the Corps believes to be the limitations on its power.

Part VI. Collateral Esoppel Effects on Remand Instructions

To assist the Corps in making these determinations on remand, we address here whether certain statements from this Court's decision in Alabama or the D.C. Circuit's decision in Geren carry the force of collateral estoppel. Specifically, we discuss whether either of the two claims found to have preclusive force by the district court in the instant case is binding on the Corps and whether any of Alabama and Florida's additional collateral estoppel arguments have merit. At the outset, we note that collateral estoppel applies only if (1) the issue at stake is identical to the one involved in the prior proceeding; (2) the issue was actually litigated in the prior proceeding; (3) the determination of the issue was critical and necessary to the earlier judgment; and (4) the party against whom collateral estoppel is asserted had a full and fair opportunity to litigate the issue in the prior proceeding. Christo v. Padgett, 223 F.3d 1324, 1339 (11th Cir.2000).

The district court found collateral estoppel, preclusive effect in the D.C. Circuit's conclusion that the WSA applied to interim reallocations of storage. Tri-State, 639 F.Supp.2d at 1343. We take no issue with this application of collateral estoppel. On remand, the Corps will determine the extent of its authority to supply the current water supply needs of the Atlanta area, combining its authority under the 1956 Act, the RHA, and the WSA. The Corps' authority under the WSA (as well as the statutory limits thereto) are applicable to the Corps' determination of its authority to supply current water supply needs, whether by force of collateral estoppel or clear statutory meaning or both.

The district court also found preclusive effect in the D.C. Circuit's holding that the reallocation of 22% of Lake Lanier's conservation storage is a major operational change on its face. Id. Several aspects of this holding merit discussion. First, and foremost, the Geren court considered only the Corps' authority under the WSA, not its authority under the RHA.37 Accordingly, a different issue is presented here. At the very least, this difference means that any water the Corps finds it is authorized to supply pursuant to the RHA is separate from the water it is authorized to supply pursuant to the WSA, and that this RHA-authorized water supply would not count against the Geren court's 22% limit.38

It is also possible that our reading of the authority provided by the RHA fundamentally changes the WSA analysis, given that the RHA congressionally authorizes the Corps to increase water supply in its balancing of hydropower and water supply needs, meaning that such reallocations to water supply arguably do not actually constitute a "change" of operations at all, and that the issue is therefore entirely different than the one presented to the Geren court. In other words, it is possible that the 22% holding has no preclusive force at all. However, because it is not clear that the Geren court's 22% limit will be reached in this case,39 we expressly decline to address the collateral estoppel effect of the Geren court's 22% limit.40

Second, it is clear that the question of whether percent reallocation of storage is the correct or sole measure of operational change was not actually litigated. Examination of the parties' briefs in Geren makes clear that the parties assumed, but did not put at issue, the question of whether percent reallocation of storage is the correct or sole measure of operational change. Similarly, because the parties merely assumed that percent reallocation was the appropriate measure, the Geren court made the same assumption in its opinion, without any discussion of the issue. When an issue is merely assumed, it does not meet the actual litigation requirement for collateral estoppel. See Fields v. Apfel, 234 F.3d 379, 383 (8th Cir.2000) (finding no issue preclusion with respect to whether a particular method for calculating disability benefits applied, because its applicability had merely been assumed by the court and both parties in a prior case and not placed at issue). The fact that the Geren court ruled "without thoroughly examining" the issue further undermines the preclusive effect of the ruling. A.J. Taft Coal Co. v. Connors, 829 F.2d 1577, 1581 (11th Cir.1987) (declining to apply collateral estoppel where the issue was not fully litigated, which resulted in the prior court tendering a conclusion "without thoroughly examining" the issue). Moreover, in this case, the district court did not hold that percent reallocation of storage is, as a matter of collateral estoppel, the correct or sole measure, and Appellees do not argue on appeal that we are bound by collateral estoppel to hold that percent reallocation of storage is the only appropriate measure of operational change. We conclude, for the foregoing reasons, that collateral estoppel does not bar the Corps from determining the appropriate measure of operational change on the basis of its own expertise. The Corps is free to consider on remand whether other measures, such as impact on hydropower,41 should be considered instead of or in addition to percent reallocation of storage.

Third, examination of the briefs in Geren also shows that the parties merely assumed that conservation storage was the appropriate frame of reference against which percent reallocation should be calculated, and the court likewise made this assumption. Accordingly, the actual litigation requirement is not met and the Corps is free to consider on remand whether some other portion of the dam's capacity should also be considered. For instance, it may be that the flood control storage, which sometimes contains excess water that could be released to satisfy water supply needs, should be factored into the calculation.

Alabama and Florida advance two collateral estoppel arguments in addition to those already covered above. First, they argue that the Geren court decided that, for purposes of the WSA analysis, the baseline for storage against which major operational change should be measured is zero. They further argue that the decision has the effect of collateral estoppel. We disagree. As noted above, the Geren court expressly made no decision with respect to the Corps' authority to allocate storage to water supply under the RHA.42 It addressed the issue of the appropriate baseline for the WSA analysis only in the context of rejecting the settling parties' argument that the interim reallocation level prior to the settlement was the correct baseline. A wholly different issue is presented in this appeal, in which we are required to assess the Corps' authority under the RHA to reallocate storage or otherwise provide water supply, and to factor this authority into the WSA analysis. Thus, the Geren court's decision with respect to the baseline for storage reallocation has no collateral estoppel effect in this case.

Second, Appellees argue that this Court's earlier statement in Alabama, that water supply is not an authorized purpose of the Buford Project, is preclusive. As noted earlier in this opinion, this statement does not give rise to collateral estoppel because it was not actually litigated and it was mere dicta and therefore was not critical or necessary to the judgment. See supra, note 21. In conclusion, the Corps is not bound by collateral estoppel in making the aforementioned determinations and should make its decisions on remand on the basis of its own reasoned analysis.

Part VII. One-Year Time Limit on Remand

This controversy has lasted a very long time. Since 1990, litigation related to this controversy has taken place in the Northern District of Alabama, the District Court for the District of Columbia, the Northern District of Georgia, the Northern District of Florida, the Middle District of Florida, the District of Columbia Circuit, and now five times in the Eleventh Circuit, and various attempts at compromise have been initiated and abandoned. Progress towards a determination of the Buford Dam's future operations is of the utmost importance to the millions of power customers and water users that are affected by the operations of the project. The stakes are extremely high, and all parties are entitled to a prompt resolution. Accordingly, the process for arriving at a conclusion of the bounds of the Corps' authority should be as swift as possible without sacrificing thoroughness and thoughtfulness. Given the importance of this case, the length of time it has been bouncing around the federal courts, and the amount of resources the parties and the courts have already expended, we believe that one year is sufficient for the Corps to complete its analysis of its water supply authority and release its conclusions. This panel will retain limited jurisdiction to monitor compliance with this time frame. At the end of this one-year period, we expect the Corps to have arrived at a well-reasoned, definitive, and final judgment as to its authority under the RHA and the WSA.

CONCLUSION

The Corps' did not consummate its decision-making process in the Alabama, Apalachicola, and SeFPC cases. Therefore, the district court lacked jurisdiction to hear these claims. The Corps' denial of Georgia's 2000 water supply request did constitute final agency action, and the district court's conclusion that it had jurisdiction to hear the Georgia case was proper. However, the court erred in its analysis of the Corps' rejection of the request. The decisions of the District Court and the Corps were based on a clear error of law— the determination that water supply was not an authorized purpose of the RHA. Furthermore, the Corps failed to reach a final, determinative position about its water supply authority before rejecting the state's request. Consequently, we reverse the district court's order granting the Corps summary judgment, and conclude that the Corps' decision was arbitrary and capricious or not otherwise in accordance with the law. All four cases are remanded to the district court with instructions to remand to the Corps for reconsideration. This panel will retain limited jurisdiction to monitor the one-year time limit.

Accordingly, the judgment of the district court is reversed, its findings of fact and conclusions of law in all four cases are vacated, and these cases are remanded to the district court with instructions to remand to the Corps for further proceedings not inconsistent with this opinion.

REVERSED, VACATED, AND REMANDED; LIMITED JURISDICTION RETAINED.