In the Supreme Court of the United States
CENTRAL GREEN CO., PETITIONER
UNITED STATES OF AMERICA
ON WRIT OF CERTIORARI
TO THE UNITED STATES COURT OF APPEALS
FOR THE NINTH CIRCUIT
BRIEF FOR THE UNITED STATES
SETH P. WAXMAN
Counsel of Record
DAVID W. OGDEN
Assistant Attorney General
BARBARA D. UNDERWOOD
Deputy Solicitor General
DAVID C. FREDERICK
Assistant to the Solicitor General
ROBERT S. GREENSPAN
IRENE M. SOLET
Department of Justice
Washington, D.C. 20530-0001
Whether 33 U.S.C. 702c, which provides that "[n]o liability of any kind shall attach to or rest upon the United States for any damage from or by floods or flood waters at any place," bars petitioner's tort action arising from property damage sustained as a result of allegedly negligent construction and maintenance of an irrigation canal that is part of a multi-purpose federal flood control project.
In the Supreme Court of the United States
CENTRAL GREEN CO., PETITIONER
UNITED STATES OF AMERICA
ON WRIT OF CERTIORARI
TO THE UNITED STATES COURT OF APPEALS
FOR THE NINTH CIRCUIT
BRIEF FOR THE UNITED STATES
The opinion of the court of appeals (Pet. App. 1-9) is reported at 177 F.3d 834. The opinion of the district court (Pet. App. 10-20) is unreported.
The judgment of the court of appeals was entered on May 20, 1999. A petition for rehearing was denied on September 7, 1999 (Pet. App. 21). The petition for a writ of certiorari was filed on November 19, 1999, and was granted on March 20, 2000. The jurisdiction of this Court rests on 28 U.S.C. 1254(1).
33 U.S.C. 702c, in pertinent part, is in App., infra, 1a.
The Flood Control Act of 1928, 33 U.S.C. 702c, provides that "[n]o liability of any kind shall attach to or rest upon the United States for any damage from or by floods or flood waters at any place." Petitioner claims that the waters that damaged its property-allegedly the result of a leak from the Madera Canal in the Friant Division of California's Central Valley Project (CVP)-were irrigation waters and not flood waters, and that the court of appeals used the wrong test to determine that they were flood waters. To place this issue in context, it is necessary to understand the interrelated history of irrigation and flood control in California's Central Valley, the purposes and operations of the CVP and its Friant Division, and the relationship of the Madera Canal to both irrigation and flood control.
1. History of Flooding in the Central Valley. The Central Valley consists of the basins of the Sacramento River in the north and the San Joaquin River in the south. Bordered on the east by the Sierra Nevada Mountains and on the west by the Coast Range, the valley extends almost 500 miles and includes more than one third of the State. The two rivers flow toward each other and join in the Sacramento-San Joaquin delta, eventually emptying into San Francisco Bay and, from there, into the Pacific Ocean. S. Doc. No. 113, 81st Cong., 1st Sess. 83 (1949) (S. Doc. No. 113).
Efforts to address the problem of flooding in the Central Valley began as early as 1880. See Report of the State Engineer To The Legislature of California-Session of 1880, Pt. II, at 5 (1880). California's State Engineer reported that the low lands of the Sacramento and San Joaquin valleys "have been and still in a great measure are subject to annual inundation." Ibid. According to the State Engineer, the Sacramento River "brings down a formidable flood volume almost every year, inundates a large area of country, and seriously threatens with devastation several hundred thousands of acres." Ibid. The San Joaquin was "in correspondingly high flood" approximately once every four years. Ibid. Periodic flooding in the winter contrasted with generally inadequate sustained rainfall during the summer, making "the use of irrigation imperative for production of most agricultural crops." S. Doc. No. 113, at 85.
Beginning early in the 20th century, California actively pursued a coordinated approach to flooding and water shortages. In 1915, the state legislature authorized the convening of a conference "[f]or the purpose of considering and recommending a unified state policy with reference to irrigation, reclamation, water storage, flood control, municipalities, and drainage, with due regard to the needs of water power, mining, and navigation." H.R. Doc. No. 416, 84th Cong., 2d Sess. Pt. 1, at 109 (1956) (H.R. Doc. 416).1 The following year, the State Water Problems Conference issued its report. Even at that early planning stage, the Conference recognized that the use of dams and their resulting reservoirs for both flood control and irrigation involved a careful balancing of "antagonistic interests." Id. at 128. The report explained that "[a] reservoir for highest economic efficiency in flood control should be kept empty until actual flood" and then gradually emptied "to give storage for another flood. That same reservoir, if used for power or irrigation would, on the contrary, be filled as soon as possible-before actual flood if conditions permitted-and kept full, lest there should not be subsequent flood to fill it." Id. at 129.
2. Early Federal Involvement in Central Valley Flood Control. In 1893, Congress established the California Debris Commission, a panel of three Army engineers appointed by the President. Congress directed the Commission to adopt plans to address flooding and navigation problems caused by tailings from hydraulic mining that had clogged the channels of the Sacramento and San Joaquin Rivers. Caminetti Act, ch. 183, 27 Stat. 507; see also H.R. Doc. 416, at 50.2 In 1917, Congress authorized the Secretary of War to carry out the Commission's plans to control floods on the Sacramento River. Act of Mar. 1, 1917, ch. 144, § 2, 39 Stat. 949. That Act authorized expenditure of $45 million for flood control work on the Mississippi River, conditioned upon contributions of at least half that amount from "local interests protected thereby." § 1, 39 Stat. 948. Congress continued to link flood control in the Central Valley and along the Mississippi in the Flood Control Act of 1928, Act of May 15, 1928, ch. 569, § 1, 45 Stat. 534-535, which contains the immunity provision at issue in this case (§ 3, 45 Stat. 535-536). That Act increased the authorized federal expenditures for flood control on the Sacramento River to a total of $17.6 million (§ 13, 45 Stat. 539).
3. Construction of the Central Valley Project and Friant Dam. As the federal government was making its first investments in flood control on the Sacramento River, California was investigating a coordinated approach to water problems, including flood control, throughout the Central Valley. See H.R. Doc. 416, at 139-256. Those investigations culminated in 1931 in a State Water Plan submitted to the California legislature by the State's Director of Public Works. Id. at 257. That plan addressed California's "twofold" water problem, "involving first the conservation and utilization of its water resources, and second, the control of floods." Id. at 258-259. The plan called for a system of dams, reservoirs, and canals in the Central Valley that would, among other functions, pump water from the Sacramento-San Joaquin delta into the San Joaquin basin, at Mendota, for irrigation. That system would permit the waters of the San Joaquin to be diverted from farther upstream, at Friant, to irrigate previously arid upstream lands. Id. at 261.
The 1931 State Water Plan specifically called for construction of Friant reservoir, as well as the Madera and Friant-Kern Canals to be fed by that reservoir. H.R. Doc. 416, at 271-272. Moreover, the plan proposed "[t]he reservation of space and its operation for flood control * * * in each of the major reservoirs on the more important streams." Id. at 295. The plan further specified the quantity of storage space to be reserved for flood control in each of 14 proposed reservoirs, including Friant, and explained that "operation of these reservoirs for flood control would not materially impair their value for conservation purposes, nor materially decrease the amount of value of the electric energy generated by water released from them." Id. at 295. Operating the reservoirs "specifically for flood control * * * would result in a substantial reduction of floods and in an increased degree of protection to the areas subject to overflow." Id. at 297.3
In 1933, the California legislature authorized construction of a coordinated water project along the lines of the State Water Plan. That planned "Central Valley Project," which included both Friant Dam and the Madera Canal, was to be financed by the sale of revenue bonds. H.R. Doc. 416, at 412, 423. Almost from the beginning, however, state financing appeared infeasible and federal involvement became essential to the project's realization. Id. at 522. In 1935, the State legislature acknowledged those facts and authorized the State's Department of Public Works "to prosecute efforts to secure Federal aid and assistance in financing the construction of the Central Valley project." Id. at 555.
In the ensuing years, the federal government acted to advance construction of the CVP and, ultimately, to make it an entirely federal project. In 1935, Congress authorized the Army Corps of Engineers to undertake construction of Kennett Dam (now called Shasta Dam) on the Sacramento River, one of the major components of the coordinated project that was initially planned and authorized by the State. Rivers and Harbors Act of 1935, ch. 831, 49 Stat. 1038; see also H.R. Doc. 416, at 544-555. Also in 1935, the President approved a report by the Secretary of the Interior on the feasibility of the entire Central Valley Project as a reclamation project pursuant to federal reclamation laws. H.R. Doc. 416, at 562-567.4 As described in that report, the CVP included among its principal features Kennett Reservoir on the Sacramento River, pumping plants and canals to deliver water from the Sacramento-San Joaquin delta to Mendota, Friant Reservoir on the San Joaquin River, and the Madera and Friant-Kern Canals to be fed by the Friant reservoir. Id. at 564-565. In 1936, Congress appropriated funds specifically for "construction of Friant Reservoir and irrigation facilities therefrom." Act of June 22, 1936, ch. 689, 49 Stat. 1622.
Finally, in 1937, Congress transferred responsibility for construction of the CVP from the Corps of Engineers to the Secretary of the Interior. Rivers and Harbors Act of 1937, ch. 832, § 2, 50 Stat. 850. The same statute provided that:
the entire Central Valley project * * * is hereby reauthorized and declared to be for the purposes of improving navigation, regulating the flow of the San Joaquin River and the Sacramento River, controlling floods, providing for storage and for the delivery of the stored waters thereof, for the reclamation of arid and semiarid lands and lands of Indian reservations, and other beneficial uses, and for the generation and sale of electric energy as a means of financially aiding and assisting such undertakings and in order to permit the full utilization of the works constructed to accomplish the aforesaid purposes.
Ibid. Congress also specified that the project's dams and reservoirs "shall be used, first, for river regulation, improvement of navigation, and flood control; second, for irrigation and domestic uses; and, third, for power." Ibid.
Construction of Friant Dam began in October 1939 and the dam was in partial operation by November 1942. U.S. Army Corps of Engineers, Report on Reservoir Regulation for Flood Control, Friant Dam and Millerton Lake, San Joaquin River, California 2 (Dec. 1965, rev. Aug. 1980) (Friant Flood Report) (lodged with the Clerk). Much of the remaining construction work on the dam and related features of the project was deferred during the Second World War (ibid.), but by February 1947, the dam itself was "largely constructed," the Madera Canal was "essentially complete," and construction of the Friant-Kern Canal was underway. H.R. Doc. No. 146, 80th Cong., 1st Sess. 2 (1947). Friant Dam and its reservoir, named Millerton Lake, were completed in 1949. U.S. Army Corps of Engineers, Post-Flood Assessment for 1983, 1986, 1995, and 1997, Central Valley, California 3-32 (1999) (Post-Flood Assessment).5
Congress confirmed the flood control role of Friant Dam while it was still under construction. In the Flood Control Act of 1944, ch. 665, 58 Stat. 887, Congress centralized responsibility for flood control operations at federal reservoirs by directing the Secretary of War to prescribe regulations governing such operations at "all reservoirs constructed wholly or in part with Federal funds provided on the basis of [flood control or navigation] purposes." § 7, 58 Stat. 890. The House report on the 1944 Act described the "widespread and damaging floods [that had] occurred in several of the major river basins of the Nation" in 1942 and 1943, with particular emphasis on floods in "the lower San Joaquin Valley." H.R. Rep. No. 1309, 78th Cong., 2d Sess. 1, 2 (1944) (H.R. Rep. No. 1309). That report further stated that recent experience "in the operation of multiple-purpose reservoirs during major floods has demonstrated that * * * reservoirs constructed wholly or in part with federal funds * * * should have their flood-control features operated under the supervision of the Chief of Engineers in accordance with regulations prescribed by the Secretary of War." Id. at 7. In addition, however, the 1944 Act expressly approved the plan "for flood control and other purposes on the Lower San Joaquin River and tributaries * * * in accordance with the recommendations of the Chief of Engineers" and authorized $8 million for "initiation and partial accomplishment of the plan." § 10, 58 Stat. 901.
4. Operation of Friant Dam and the Madera Canal. Friant Dam, located 25 miles northeast of Fresno, is a concrete structure that can store 520,500 acre-feet of water. Friant Flood Report 11. Water is released from the dam into the Madera and Friant-Kern Canals as well as down the river either through outlets in the dam or over its spillway. Id. at 15. The 36-mile long Madera Canal extends northward from the dam, crossing the Fresno River and emptying into the Chowchilla River. Id. at 11; see App., infra, at 2a (diagram depicting Central Valley Project). Approximately the first 7.7 miles of the canal, including the portion that traverses petitioner's property, is lined by concrete, with the remainder unlined and formed by dug-out clay and rock. The Friant-Kern Canal extends 152 miles southward from the dam and connects to the Kern River. Friant Flood Report 11; Post-Flood Assessment 3-32.
Friant Dam is owned by the United States and operated by the Department of the Interior as part of the CVP. Post- Flood Assessment 3-32. The Madera Canal, likewise part of the CVP (see S. Doc. No. 113, at 130), is owned by the United States (through the Department of the Interior) and operated by the Madera Irrigation District (and its legal successor) under contract with and supervision of the Department. J.A. 8, 19-47. The Department's Bureau of Reclamation controls irrigation operations in all parts of the CVP, including the Friant Division. Friant Flood Report App. A.
Pursuant to the Flood Control Act of 1944, however, the Friant Division's flood control operations are regulated by the Corps of Engineers. 33 C.F.R. 208.11. In 1955, the Corps first promulgated regulations governing "the use and operation of Friant Dam and Reservoir on San Joaquin River, California, for flood control purposes." 20 Fed. Reg. 9181. In 1976, the Corps replaced the regulations specifically addressing Friant Dam, and similar regulations for other individual reservoirs, with new regulations governing the operation of all projects covered by the 1944 Act or otherwise subject to the Corps' authority to direct flood control operations, other than those owned and operated by the Corps itself. 41 Fed. Reg. 20,400 (1976).
Under those regulations (which list covered projects, including "Friant Dam and Millerton Lake," by name), each project's flood control operations are governed by a "water control plan and manual" prepared by the Corps of Engineers specifically for that project. 33 C.F.R. 208.11(d)(4). The project owner commits to the terms of the plan and manual in a "letter of understanding" or a "field working agreement" with the Corps. 33 C.F.R. 208.11(c)(3). See Friant Flood Report App. A. The Corps' plan for the Friant Division is designed to address two kinds of floods that occur in the Central Valley-floods caused by rain and floods caused by melting snow. Rainfloods, which typically occur in late fall and winter, are characterized by high peak discharges lasting only a few days at a time. Notwithstanding their short duration, such rains historically have flooded natural river channels unimproved by man-made levees and dams. In contrast, snow-melt floods, which occur during the late spring and summer, produce sustained, moderate flows over a period of two to three months and result in a much larger volume of total runoff. Id. at 9-10. Those runoffs require extensive monitoring to ensure against flooding.
The plans in effect pursuant to 33 C.F.R. 208.11 for flood control operations of Friant Dam and its related canals address both rain and snow-melt floods. Regulations require each project to have a "Flood Control Diagram" describing the multiple uses of storage space behind the dam over the course of the year. 33 C.F.R. 208.11(d)(5)(i). The Friant diagram shows that for ten months of the year (October through July), the reservoir stores water both to serve irrigation needs and to control floods. Friant Flood Report App. A, Chart A-11. "The flood control operation is determined daily as described on chart A-11." Id. at A-2. The Flood Control Diagram also shows that during some portions of the year, depending on the weather, the total quantity of stored water necessary to satisfy irrigation needs and to control floods may exceed the storage capacity of the dam. Id. Chart A-11. At those times, "supplemental releases"-releases solely for flood control-are mandatory. Ibid. Some of those flood releases are made through the Madera and Friant-Kern Canals. Post-Flood Assessment 3-32 to 3-33.6 The Corps of Engineers has estimated that operation of Friant Dam reduced flood damage by $23,690,000 in 1983, $33,190,000 in 1986, $54,310,000 in 1995, and $3,320,000 in 1997. Id. at 5-15, 5-27, 5-38, 5-48.
5. This Litigation. Petitioner Central Green Company, the owner of pistachio orchards in Madera County, California, brought this action against the United States under the Federal Tort Claims Act, 28 U.S.C. 2671 et seq., alleging that negligent design, construction, or maintenance of the Madera Canal by the federal government resulted in property-damaging leakage of water from the canal. Pet. App. 2; J.A. 8-9.7 The district court granted the government's motion for judgment on the pleadings. Pet. App. 20. Rejecting petitioner's contention that the Flood Control Act of 1928, 33 U.S.C. 702c et seq., did not apply because the waters carried through the Madera Canal are used for irrigation purposes and not for flood control, the district court observed that "[t]he legislative history [of Section 702c] * * * is very clear that the scope of the immunity is very broad." Pet. App. 19. The district court concluded that petitioner had "cited nothing that allows the undermining of the scope of that immunity when a multi-purpose project is involved." Id. at 19- 20.
The court of appeals affirmed. Pet. App. 1-9. The court recognized that Section 702c "confers broad immunity for claims arising from the design, operation, or management of federally authorized flood control projects." Id. at 2. Petitioner's "sole argument" against application of the immunity here, according to the court, was that the water that damaged its property was not "flood water" within the meaning of Section 702c because it was held "for irrigation purposes rather than flood control." Id. at 3. The court concluded that its own decisions, both before and after this Court's decision in United States v. James, 478 U.S. 597 (1986), compelled rejection of that argument because the Madera Canal is part of a federal flood control project, the CVP. Pet. App. 5-8.
SUMMARY OF ARGUMENT
I. In United States v. James, 478 U.S. 597 (1986), this Court construed the Flood Control Act of 1928 to confer broad immunity from suit against the United States for damages by flood waters, which the Court defined as "all waters contained in or carried through a federal flood control project for purposes of or related to flood control, as well as to waters that such projects cannot control." Id. at 605. That definition is satisfied in this case.
II. In several statutes, Congress expressly provided that the entire Central Valley Project (CVP) is a flood control project. Both as it was originally conceived by Congress and as it operates in practice, the CVP's flood control functions depend on the integrated operations of the project's numerous subdivisions, which include Friant Dam and the Madera Canal. The timing and amount of water releases at each place in the CVP have consequences for flood control. Moreover, Congress also specified in the Flood Control Act of 1944 that the Friant Division (of which the Madera Canal is a part) is a flood control project by incorporating into the legislation a set of Corps of Engineers' recommendations that clarified the important flood control purpose of the Friant Dam and the use of the Madera Canal for releases of flood waters.
The 1944 Act also conferred legislative rulemaking authority on the Corps of Engineers to promulgate flood control rules for multiple-use projects. In 1955, the Corps promulgated such regulations for the Friant Division. That authoritative determination, which has been maintained consistently by the Corps since that time, is entitled to deference. United States v. Gerlach Live Stock Co., 339 U.S. 725 (1950), is not to the contrary. The issue in that case was whether the reclamation laws required the government to compensate downstream water rights owners. In upholding the claims of the owners, the Court did not foreclose the Friant Division from being deemed a flood control project under Section 702c.
III. Not only is the Friant Division (and the Madera Canal) a flood control project, but also the Madera Canal carries flood waters within the meaning of James. For ten months of the year, Millerton Lake stores water to control floods and to serve irrigation needs. Petitioner's approach rests on an untenable theory of legal distillation of those waters: that at some point in the process of capturing, storing, and releasing those waters it is possible to identify some as "flood waters" and some as "irrigation waters." That theory is belied by (A) the regulatory requirements governing flood releases from Friant Dam, which may be satisfied only by channeling some of those releases into the Madera Canal; (B) the contract with the Madera Irrigation District, which expressly provides for flood waters to be released into the Madera Canal; and (C) the historical record of water flows in the canal, which establish empirically that more water has been released into the Madera Canal for flood control than for principally irrigation purposes over the past twenty years. There thus can be no question that the Madera Canal is a flood control project and the waters that allegedly leaked from the canal were "for purposes of or related to flood control." James, 478 U.S. at 605. Moreover, petitioner's allegation that water leaked from the canal is sufficient by itself to satisfy the alternative definition of "flood waters" in James as "waters that such [flood control] projects cannot control." Ibid.
IV. In this Court petitioner advances for the first time a novel construction of the Flood Control Act not heretofore adopted by a single court: that the immunity applies only where the waters causing damage are "primarily" for the purpose of flood control. That approach conflicts with the broad immunity recognized by the James Court as compelled by the plain language of Section 702c. Petitioner's approach is unworkable since it requires courts to engage in an analysis of a project's "primary" purpose that Congress itself does not specify when enacting multiple-use water projects. Petitioner's approach also produces the result- plainly inconsistent with congressional intent-that scores of multiple-use projects recognized by the Corps under its delegated rulemaking authority as flood control projects would be ineligible for Section 702c immunity.
I. UNITED STATES v. JAMES ESTABLISHES THAT THE IMMUNITY PROVIDED BY THE FLOOD CONTROL ACT MUST BE BROADLY CONSTRUED TO INCLUDE ALL DAMAGES CAUSED BY FLOOD WATERS IN A FLOOD CONTROL PROJECT
In the Flood Control Act of 1928, Congress embarked upon a multi-decade program of unprecedented scope to construct dams and other structures for flood control. See United States v. Sponenbarger, 308 U.S. 256, 262 (1939). Congress limited the government's exposure to financial liability for damages resulting from flood control activities by including in the 1928 legislation a provision that "[n]o liability of any kind shall attach to or rest upon the United States for any damage from or by floods or flood waters at any place." 33 U.S.C. 702c.
A. The James Test Confers Immunity From Liability For Damage From Waters Carried In Flood Control Projects That Either Are Related To Flood Control Or Cannot Be Controlled In The Project
In United States v. James, 478 U.S. 597 (1986), this Court rejected a narrow construction of Section 702c immunity, construing it to extend to both property damage and personal injury. Id. at 605. Noting that it "is difficult to imagine broader language" than Section 702c, the Court held that the "sweeping language" of the text and the "equally broad and emphatic language" of its legislative history support recognition of a broad immunity for the United States. Id. at 604, 608. The Court added that "Congress' choice of the language 'any damage' and 'liability of any kind' further undercuts a narrow construction." Id. at 605 (quoting 33 U.S.C. 702c).8
Justice Powell's opinion for the Court in James emphasized that a broad immunity is consistent with the purpose of Section 702c. The Court explained that "the legislative history of § 702c shows a consistent concern for limiting the Federal Government's financial liability to expenditures directly necessary for the construction and operation of the various projects." 478 U.S. at 607. The Court also stressed that "[n]umerous statements concerning the immunity provision confirm that it was intended to reaffirm sovereign immunity in such a dangerous and extensive project," including this statement from the Chair of the House Rules Committee:
I want this bill so drafted that it will contain all the safeguards necessary for the Federal Government. If we go down there and furnish protection to these people- and I assume it is a national responsibility-I do not want to have anything left out of the bill that would protect us now and for all time to come. I for one do not want to open up a situation that will cause thousands of lawsuits for damages against the Federal Government in the next 10, 20, or 50 years. 69 Cong. Rec. 6641 (1928) (remarks of Rep. Snell).
Ibid. Accordingly, Section 702c "safeguarded the United States against liability of any kind for damage from or by floods or flood waters in the broadest and most emphatic language." Id. at 608 (emphasis added) (quoting National Mfg. Co. v. United States, 210 F.2d 263, 270 (8th Cir.), cert. denied, 347 U.S. 967 (1954)). The language the Court chose to define the scope of that immunity was correspondingly broad: "The Act concerns flood control projects designed to carry floodwaters. It is thus clear from § 702c's plain language that the terms 'flood' and 'flood waters' apply to all waters contained in or carried through a federal flood control project for purposes of or related to flood control, as well as to waters that such projects cannot control." 478 U.S. at 605 (emphasis added).9 Under James, therefore, the government is entitled to immunity in this case because, as we show below, it is a matter of public record that: (1) the Madera Canal is part of a federal flood control project and (2) the waters that caused the flood either (a) are carried in the project "for purposes of or related to flood control" or (b) are "waters that such projects cannot control." Ibid.
B. The Court Of Appeals Properly Upheld The Government's Immunity Under James
The courts below properly dismissed petitioner's claims on the ground that petitioner was seeking compensation for "damage from or by floods or flood waters," within the meaning of the immunity conferred by 33 U.S.C. 702c. The court of appeals correctly held that the James test was satisfied because the Madera Canal is part of the Central Valley Project, which is indisputably a flood control project, and thus all waters in the Project are contained in or carried through it for purposes that are at least related to flood control. Pet. App. 3-8. Petitioner has questioned that determination, drawing support from the court of appeals' statement that the Madera Canal itself "is not a flood control project and serves no flood control purpose." Id. at 9. That observation, which was made without citation to any sources, is both inaccurate and legally irrelevant.
As we demonstrate in Part II, infra, numerous statutory and regulatory authorities establish not only that the CVP is a flood control project, but also that the Friant Division of the CVP-of which the Madera Canal is a critical component -is a flood control project. In addition, as we demonstrate in Part III, infra, judicially-noticeable materials establish that the Madera Canal itself serves important flood control pur-poses.
Moreover, regardless of the purposes for which the water is contained in or carried through the Madera Canal, the allegations in petitioner's complaint establish the second prong of the James test: that the property was damaged by waters that a federal flood control "project cannot control." 478 U.S. at 605. The complaint alleges that the "Madera Canal was constructed or has been maintained in such a fashion that substantial amounts of water leak and have leaked from [respondent's] canal," and that petitioner's property was damaged by those alleged leaks. J.A. 9. Thus, if the Court concludes that the Madera Canal is part of a flood control project as set forth in Part II, the legal requirements for immunity under both prongs of the James test have been satisfied. That alternative ground for affirmance (the second prong of the James test) is consistent with the analysis of the court of appeals in Morici Corp. v. United States, 681 F.2d 645 (9th Cir. 1982), on which the court below relied. Pet. App. 6. In Morici, the Ninth Circuit had held that the government was immune from suit for flooding caused by waters that could not be controlled in the CVP because of water seepage from the Sacramento River. 681 F.2d at 646-648.
II. CONGRESS UNAMBIGUOUSLY ESTABLISHED THE FRIANT DIVISION OF THE CENTRAL VALLEY PROJECT AS A FLOOD CONTROL PROJECT
A. The Entire Central Valley Project Is A Federal Flood Control Project
In the Rivers and Harbors Act of 1937, ch. 832, § 2, 50 Stat. 850, Congress authorized all of the Central Valley Project, including its irrigation canals, as a federal flood control project. In particular, that statute provides: "[T]he entire Central Valley project, California, * * * is hereby reauthorized and declared to be for" purposes including "controlling floods" (emphasis added).10 Petitioner is mistaken that the 1937 Act's authorization of the CVP for flood control purposes extended only to its "'dam[s] and reservoirs' but not its irrigation canals." Br. 27 (quoting text of statute).11 To the contrary, Friant Dam and the Madera and Friant-Kern Canals were expressly included in plans for the CVP. H.R. Doc. 416, at 564-565.
As originally conceived and as operated for more than 50 years, the CVP's flood control operations have been integrated among the various subdivisions of the project. Because the Sacramento and San Joaquin Rivers eventually merge at a delta near San Francisco Bay, waters released from any dam on either river (or their tributaries) will affect the volume of water in the delta and such releases throughout the project must be coordinated carefully to avoid floods. Thus, courts in the Ninth Circuit have uniformly held that, notwithstanding its multiple purposes, the entire CVP is a federal flood control project for which the flood immunity applies. See, e.g., Morici Corp., 681 F.2d at 648 (government immune from suit for crop damages caused by seepage from Sacramento River); Islands, Inc. v. United States Bureau of Reclamation, 64 F. Supp. 2d 966, 969 (E.D. Cal. 1999); United States v. Iron Mountain Mines, Inc., 881 F. Supp. 1432, 1439 (E.D. Cal. 1995).
A few examples using projects in the San Joaquin River basin illustrate why the CVP is properly viewed as an integrated flood control project. See App., infra, 2a (map with water release capacities). Three dams regulate flows into the San Joaquin upstream of the Mariposa Bypass: Friant Dam (operated by the Bureau of Reclamation) on the San Joaquin River; Hidden Dam (operated by the Corps of Engineers) on the Fresno River, which flows into the San Joaquin north of Mendota; and Buchanan Dam (operated by the Corps) on the Chowchilla River, which flows into the San Joaquin River (roughly parallel to but north of the Fresno River). The Madera Canal starts at the San Joaquin, traverses the Fresno (with one bypass enabling water to be released into the Fresno and another to send water underneath the river), and empties into the Chowchilla. The Corps has calculated flood channel flows-the maximum amount of water that will not cause flooding-for those rivers in cubic feet per second (cfs). The flood channel flow of the San Joaquin immediately below Friant Dam is 8000 cfs; of the Fresno River below Hidden Dam, 5000 cfs; and of the Chowchilla River below Buchanan Dam, 5000 cfs. Thus, if each dam releases the maximum that can be accommodated by its immediate downstream channel, the releases would total 18,000 cfs. But the flood channel flow for the San Joaquin River downstream of where the Fresno and Chowchilla Rivers enter it is only 16,500 cfs. Accordingly, the three dams must coordinate their releases so that not all three are releasing their maximum at the same time, which would exceed channel capacity by 1500 cfs.
Likewise, when releases from Friant Dam must exceed 8000 cfs to avoid surpassing the reservoir's capacity, the Bureau releases water into the Madera and Friant-Kern Canals to avoid flooding on the San Joaquin below Friant Dam.12 Excess water from Friant may travel through the Madera Canal to be discharged into the Fresno and/or Chowchilla Rivers to avoid flooding a narrow stretch of the San Joaquin at Skaggs Branch before the Fresno flows into the San Joaquin River. But such releases must be coordinated carefully with the Corps of Engineers, which may be making releases of water from its Hidden and Buchanan Dams. See Post-Flood Assessment 3-34 to 3-35 (noting that "[d]uring flood management operations, floodwaters can be routed to the Fresno River via the Madera Canal" and "flows from the Madera Canal can be directed down Ash (5000 cfs) and Berenda (2000 cfs) Sloughs, about 10 miles downstream from Buchanan Dam" on the Chowchilla River). These examples of routine practice illustrate what is true of the entire Central Valley Project: that the release of water at one place necessarily affects decisions by Corps and Bureau personnel as to whether water should be released at another place. The court below thus properly held that the United States is immune from petitioner's suit because the CVP, of which the Madera Canal is a part, is a flood control project. See Pet. App. 8.
B. Congress Specified That The Friant Division Of The CVP Is A Flood Control Project
Even if Congress's intent had been unclear in designating the entire Central Valley Project as a federal flood control project warranting application of Section 702c immunity, any such ambiguity would be irrelevant here because Congress has provided that the Friant Division is itself a federal flood control project. Petitioner correctly identifies the reclamation purposes of the Friant Division (Br. 27-31), but overlooks key enactments establishing Congress' intent that the project also perform flood control functions.
In 1936, Congress "recognized that destructive floods upon the rivers of the United States * * * constitute a menace to national welfare." Flood Control Act of 1936, ch. 688, § 1, 49 Stat. 1570. Congress directed the Corps of Engineers to embark on "Federal investigations and improvements of rivers and other waterways for flood control and allied purposes," further providing that such investigations "not interfere with investigations and river improvements incident to reclamation projects * * * undertaken by the Bureau of Reclamation." § 2, 49 Stat. 1570-1571. Congress specifically instructed the Corps "to cause preliminary examinations and surveys for flood control at the following-named localities," specifying further that "the Government shall not be deemed to have entered upon any project for the improvement of any waterway mentioned in this Act until the project for the proposed work shall have been adopted by law: * * * San Joaquin River from Herndon to Antioch and its main east side tributaries." § 6, 49 Stat. 1592, 1595. The 1936 Flood Control Act required the Corps to study scores of rivers and to implement flood control improvements in dozens of others.
In 1938, Congress authorized an appropriation of $375 million for flood control works nationwide over the five-year period ending June 30, 1944. See Act of June 28, 1938, ch. 795, §9, 52 Stat. 1226. The following year, a House report advocated "annual authorizations for flood-control projects to protect lives and property in the orderly development of the national policy of flood control." H.R. Misc. Rep. No. 799, 76th Cong., 1st Sess. 2 (1939). By 1941, that general recognition of the widespread need for a comprehensive flood control policy had expanded to multiple-use projects. "Plans for control of destructive floodwaters should to the extent practicable provide for their conservation and use for the benefit of all the people * * * to meet urgent needs for flood control, irrigation, and navigation and to satisfy the growing demands for electric power." H.R. Misc. Rep. No. 759, 77th Cong., 1st Sess. 3, 5 (1941).
The Nation's entry into World War II delayed implementation of that policy, but by 1944 Congress again turned its attention to flood control. The Flood Control Act of 1944 confirms that Congress intended the Friant Division (as well as many other multiple-use projects) to be a federal flood control project. Section 1 expressed Congress's intent to "preserve and protect to the fullest possible extent established and potential uses, for all purposes, of the waters of the Nation's rivers" and "to facilitate the consideration of projects on a basis of comprehensive and coordinated development." 58 Stat. 888. The Act mandated coordination between and among the Corps of Engineers, the Bureau of Reclamation, and the Secretary of Agriculture (who had responsibility for assessing flood effects on soil erosion and watersheds) in the investigation, planning, and operation of projects. § 1, 58 Stat. 888-889.
In Section 7 of the 1944 Act, Congress provided specifically that multiple-use projects operated by agencies other than the Corps be deemed federal flood control projects, and that they would be operated for that purpose in a manner prescribed by the Corps, pursuant to expressly delegated rulemaking authority:
Hereafter, it shall be the duty of the Secretary of War to prescribe regulations for the use of storage allocated for flood control or navigation at all reservoirs constructed wholly or in part with Federal funds provided on the basis of such purposes, and the operation of any such project shall be in accordance with such regulations.
58 Stat. 890. See also H.R. Rep. No. 1309, at 7 (noting that such legislation was necessary because "recent experiences in the operation of multiple-purpose reservoirs during major floods has demonstrated that to assure the expected flood-control benefits" such projects should adhere to Corps of Engineers flood-control operational rules). The Corps has exercised that rulemaking expressly with respect to the flood control operations of the Friant Division. See pp. 27-30, infra.
The 1944 Act also specifically recognized the Friant Division as a federal flood control project. As was its normal practice in those years, Congress adopted the report and recommendations of the Corps of Engineers directly in the text of the enacted legislation, thereby putting a con-gressional imprimatur on the specific findings and policies proposed by the Corps. Thus, in addition to the many other multiple-use projects authorized in the Flood Control Act of 1944, Section 10 provided that: "The plan of improvement for flood control and other purposes on the Lower San Joaquin River and tributaries, including Tuolumne and Stanislaus Rivers, in accordance with the recommendations of the Chief of Engineers in Flood Control Committee Document Numbered 2, Seventy-eighth Congress, second session, is approved, and there is hereby authorized $8,000,000 for initiation and partial accomplishment of the plan." 58 Stat. 901.
In Flood Control Document No. 2, in turn, the Corps recommended operating the Friant Division as a flood control project. See H.R. Comm. on Flood Control Doc. No. 2, 78th Cong., 2d Sess., San Joaquin River and Tributaries (Comm. Print 1944) (Flood Control Doc. No. 2). The Corps noted that "[t]he floods of 1937 and 1938 resulted in widespread damage on the San Joaquin River and its majortributaries and made evident the necessity for a comprehensive plan of control embracing the entire San Joaquin River system." Id. at 23. It further tabulated data indicating that the 1937 floods had produced peak natural flow from rain floods on the San Joaquin River at Friant of 77,000 cubic feet per second (cfs). Id. at 26, Tab. 15. To place that huge volume in proper context, the Corps recommended that Friant Dam be operated so that no more than 7000 cfs would be released into the San Joaquin above Skaggs Branch. Id. at 39, Tab. 22. Notwithstanding the paucity of economic development along the San Joaquin in that era, the Corps estimated that the 1937 floods caused flood damage of $1.46 million. Id. at 32, Tab. 20.
The massive flooding of 1937-1938 initiated a major re-orientation in the thinking of the Corps and Congress. The Corps formally recommended a "plan of improvement" for the Friant Reservoir [Millerton Lake] "to obtain the maximum flood control possible without impairment of irrigation yield, i.e., operation in accordance with criteria given in paragraph 82,(a)." Flood Control Doc. No. 2, at 35. It is apparent that the Corps factored in the use of the Madera and Friant-Kern Canals for flood control in its calculations: "Present plans provide for gross storage of 520,000 acre-feet, inactive storage (below canal outlets) of 130,000 acre-feet, and flood outlet capacity exceeding 12,000 cubic feet per second before any of the active storage space is occupied." Id. at 35. In calculating the 12,000 cfs flood release from the Friant Dam, the Corps expressly noted that only 7000 cfs could be released from Friant Dam into the San Joaquin River without causing a flood between Skaggs Branch and Mendota. See id. at 39, Tab. 22. The capacities of the Madera Canal of approximately 1000 cfs and the Friant-Kern Canal of approximately 4000 cfs, added to the 7000 cfs capacity that the Corps calculated could be safely released from the dam into the river, thus accounted for the Corps' conclusion that the project had a "flood outlet capacity exceeding 12,000 cubic feet per second." Ibid.
In forwarding the Corps' recommendations to Congress, which the 1944 Act subsequently enacted, the Chief of Engineers stressed that "[t]he plan also contemplates that Friant Reservoir, now under construction as an element of the Central Valley project, will be so operated as to afford the maximum degree of flood control compatible with its primary use for irrigation." Flood Control Doc. No. 2, at 3. The Corps recommended specific design features for the project so that the flood control value of the total project costs estimated at $3.5 million could be realized. Id. at 45. Operated according to plan, "about 82 percent of the total average annual flood damage for the period of record would be prevented." Ibid. The 1944 Act and its legislative history thus conclusively refute petitioner's assertion (Br. 40) that "there simply is no basis to conclude that Congress con-ceived of [the Friant Division] as a 'flood control project.'"
C. The Corps of Engineers' Longstanding And Consistent Determination That The Friant Division Is A Flood Control Project Is Entitled To Deference
1. In 1955, the Corps of Engineers first promulgated regulations pursuant to the delegated rulemaking authority of Section 7 of the 1944 Act that governed "the use and operation of Friant Dam and Reservoir on San Joaquin River, California, for flood control purposes." 20 Fed. Reg. 9181. Those rules, which have since been augmented, establish three required operational conditions that can only be achieved if the Madera and Friant-Kern Canals are used for release of flood waters: (1) releases from Friant Dam "shall be restricted to quantities which will not cause downstream flows to exceed, insofar as possible, * * * 8,000 cubic feet per second between Friant and Skaggs Bridge";13 (2) "[s]torage space in Friant Reservoir [Millerton Lake] shall be kept available for flood control purposes in accordance with the Flood Control Storage Reservation Diagram currently in force," ibid.; and (3) "[i]n the event that the reservoir level rises above elevation 578 at the dam (top of spillway gates), subsequent operation of the dam shall be such as to cause downstream flows and stages to exceed as little as possible the criteria prescribed in paragraph (a) of this section," ibid. While the Corps' criteria do not expressly specify use of the Madera and Friant-Kern Canals, they must be used for the release of water in excess of 8000 cfs to comply with those rules. Since the issuance of those formal rules in 1955, the Corps has included the Friant Division in its list of projects operated by other agencies under Corps-directed flood control rules pursuant to Section 7 of the 1944 Act. See 33 C.F.R. 208.11(e).14
2. Plans for the Friant Division adopted pursuant to those regulations address the operation of the dam and its related canals. The Flood Control Storage Reservation Diagram referenced in 20 Fed. Reg. at 9181 is Chart A-11 to the Friant Flood Report. That Diagram shows generally what space should be made available in Millerton Lake for various purposes at different times during the average water year. From approximately October 1 through August 1, the Corps requires the Bureau to maintain space in Millerton Lake to handle either rain floods (October 1 to March 15) or snow-melt runoff (March 15 to July 31). According to the Flood Control Diagram, "[w]ater stored in rainflood space shall be released as rapidly as possible without exceeding 8,000 cfs below Little Dry Cr[eek] or 6,500 cfs at the near Mendota gage." Friant Flood Report Chart A-11. Similarly, during the period of snow-melt runoff, implementation of the Diagram requires that space referred to as "conditional space" be reserved to meet the dual needs of accommodating flood water and providing water for irrigation. "When water is stored in the portion of conditional space required for flood control, a supplemental release in addition to irrigation demand must be made." Ibid. The Corps calculates the current maximum initial capacity of the Madera Canal as 1200 cfs, declining to 625 cfs at the canal's terminus at the Chowchilla River, and of the Friant-Kern Canal as 4500 cfs declining gradually to 2000 at its terminus. Id. at 11. As a general matter, flood releases into the Madera and Friant-Kern Canals because of rain floods occur from December until April, when irrigation demand is low. Thereafter, melting snow is the primary source of flood releases into the canals. Such flood releases may coincide with or far exceed irrigation needs. See pp. 37-40, infra.
The Corps' determination, through rulemaking, that the Friant Division operates as a flood control project is entitled to this Court's deference. Chevron U.S.A. Inc. v. Natural Resources Defense Council, Inc., 467 U.S. 837, 842-844 (1984); accord Christensen v. Harris County, 120 S. Ct. 1655, 1662 (2000).15
D. Petitioner's Reliance On Gerlach Is Misplaced
Petitioner erroneously contends (Br. 35-40) that this Court's decision in United States v. Gerlach Live Stock Co., 339 U.S. 725 (1950) (Gerlach), establishes that Section 702c does not apply here because the Friant Division is a reclamation project and not a flood control project. In Gerlach, property owners downstream from Friant Dam who had rights to the waters of the San Joaquin River had obtained awards from the court of claims for takings caused by the dam's restriction of the river's flow. Challenging those awards, the United States argued in this Court that it was not liable for any such takings because the authorized purposes of the CVP as a whole included flood control and navigation. The theory behind the government's reading of the applicable statutes was that those purposes supported a "superior navigation easement" that superseded the interests of the water rights owners in downstream riparian rights. 339 U.S. at 736.
In disagreeing with that position, this Court held that Congress intended to compensate owners of water rights. The Court explained that Congress had made the Friant Division subject to federal reclamation law, which requires recognition of all vested rights under state law, including water rights. 339 U.S. at 734-736. The Court also noted that, prior to that litigation, the Bureau of Reclamation had consistently respected the water rights of land owners affected by reclamation projects by, inter alia, requesting appropriations to purchase such rights. Id. at 735 & n.9.
The Court did not decide in Gerlach that because the Friant Division is a reclamation project, it is therefore not a flood control project. Indeed, the Court noted that under a formula "approved by the President, * * * multiple purpose dams [in the CVP] are the responsibility of the Bureau of Reclamation," whereas "dams and other works only for flood control are exclusively the responsibility of the Army Engineers." 339 U.S. at 733-734 (emphasis added). Thus, by concluding that the Friant Division is a reclamation project subject to the reclamation laws, the Court simply acknowledged that flood control is not its sole purpose. Gerlach does not hold, or even suggest, that the Friant Division is not a "flood control project" as this Court used those words in James.16 In any event, even if Gerlach could be read as petitioner urges, the Corps of Engineers determined five years later that the CVP's Friant Division is a federal flood control project pursuant to its delegated rulemaking authority under the Flood Control Act of 1944. See 20 Fed. Reg. at 9181.
III. THE MADERA CANAL CARRIES WATERS RELATED TO FLOOD CONTROL WITHIN THE MEANING OF JAMES
Petitioner argues (Br. 23) that "flood waters" are distinct from "irrigation water," and that only the latter flowed through the Madera Canal, allegedly damaging petitioner's property. The fact that water may be put to beneficial uses -such as irrigation-does not preclude it from being "flood water" within the meaning of Section 702c. In James, this Court provided a simple, functional definition for that term. Water in a federal flood control project is "flood water" if it is carried through the project "for purposes of or related to flood control" or if it is "water that such [a] project cannot control." 478 U.S. at 605. The water in the Madera Canal meets both alternative facets of that definition.17
A. Waters Released Into Madera Canal Cannot Be Segregated Into "Flood" And "Irrigation" Waters
The Flood Control Diagram for Friant Dam shows that, during most of the year (October through July), Millerton Lake stores water both to control floods and to meet irrigation needs. Friant Flood Report App. A, Chart A-11. Water can be released from the reservoir in three ways: (1) through the canals, (2) through outlets in the dam that send water down the river, and (3) over the dam's spillway, which likewise sends water down the river. Id. at 15. Discharges to serve the respective demands of flood control and irrigation are made from a single, undifferentiated body of water behind the dam-Millerton Lake. As an empirical matter, it is not true that the quantity of water released down the river always at least equaled the total quantity that the Flood Control Diagram requires to be released to create the necessary storage space for flood control. See pp. 37-40, infra. Even if that were not so, it would be unreasonable to regard every drop of "flood water" as going down the river (see Pet. Br. 25) and, concomitantly, all water flowing into the Madera Canal as consisting purely of "irrigation water." Petitioner's theory misperceives the Friant Division's operations in practice and the historical record of releases of water into the irrigation canals.
Between 1873 and 1979, the annual flows below Friant varied widely, with a low of 361,500 acre-feet (representing approximately 60 percent of Millerton Lake's storage capacity) in 1977 to a high of 4,367,800 acre-feet in 1906 (meaning that the reservoir would have been filled and emptied more than eight times during the course of the year). The mean annual unimpaired flow for that period is 1,790,700 acre-feet, which is three-and-one-half times the storage capacity of the reservoir. See Friant Flood Report Chart 5. Thus, except in the most severe drought years, the waters stored to prevent floods behind Friant Dam are transferred to the canals (during the course of the year) for both flood control and irrigation purposes in a manner that is impossible to segregate. Releases for irrigation in high flow years also serve a flood control purpose, and water released for flood control may be stored by irrigators for later irrigation use or to recharge ground water. That type of multiple-use was precisely what Congress intended for the project. See pp. 7, 24-25, supra.
To characterize the water carried through the Madera Canal as "related to flood control" (James, 478 U.S. at 605) is thus hardly "metaphysical," as petitioner asserts (see Br. 12). Rather, that label better describes petitioner's conception that dammed water in a reservoir can somehow be legally distilled into "irrigation" molecules and "flood control" molecules. Water is stored and released when and where it can best achieve the multiple purposes of the project. If a flood can be avoided by releasing water into the canals, that water does not lose its character as flood waters, and the government does not lose its immunity, merely because the government has arranged to provide some of that water to irrigators.18 Moreover, petitioner's insistence that the beneficial use of stored water-here for irrigation- deprives it of legal status as "flood water" is inconsistent with James. The Court held there that Section 702c applied to damages to recreational users of federal flood control lakes who were dragged through dam gates when water was being released for flood control purposes. According to petitioner, "James indicates that [Section 702c] encompasses waters originally stored and later diverted for purposes of flood control notwithstanding that in the interim they may have provided incidental benefits, including recreational activities such as boating." Br. 19. That reasoning is unpersuasive. First, it erroneously assumes that a singular purpose can be ascribed to water in a multiple-use project. Second, there is no basis for petitioner's apparent position that water in a federal flood control lake retains its character as "flood water" if its beneficial use precedes its release from the dam, but not when its beneficial use follows such release. Moreover, a single release may accomplish two purposes concurrently; the flood control benefit of a release for irrigation may rest in increasing storage capacity for an anticipated flood. Finally, the allegations in petitioner's complaint in any event do not require the Court to make such a fine distinction, because the leaks resulting from long-term, structural damage asserted by petitioner cannot be ascribed to a single purpose of the project.
B. The Madera Irrigation District Contract Expressly Provides For Flood Water To Be Released Into Madera Canal
Because of the highly-variable water supply that flows into Millerton Lake, the Bureau of Reclamation has entered into contracts reflecting the historical reality that flood waters must be diverted into the Madera and Friant-Kern Canals to optimize the multiple-use objectives of the project. The contract with the Madera Irrigation District, for example, defines Class 1 water as "a dependable water supply during each Year" and Class 2 water as "undependable in character and will be furnished only if, as, and when it can be made available as determined by the [Bureau of Reclamation]." Interim Renewal Contract Between the United States and Madera Irrigation District Providing for Project Water Service (MID Contract), Art. 1(c) and (d). Class 2 water is "available for irrigation purposes during spring months of wet years," Friant Flood Report 14 n.2, because it is "necessary to evacuate Project Water from Millerton Lake to prevent or minimize a spill or to meet flood control criteria," MID Contract, Art. 3(h)(1).19
The price charged for Class 1 water historically has been at about twice the rate charged for Class 2 water. Because of its undependable character, Class 2 water may be delivered at times that are inconvenient to all members of the irrigation district, but whose individual members nonetheless may wish to purchase that water to store it for later use, to recharge groundwater supplies, or to irrigate in off-peak periods. Article 3(h)(1) of the MID Contract specifically provides for the contingency of flood releases as Class 2 water: "If, during the months of March through September, the [Bureau of Reclamation official] determines it will be necessary to evacuate Project Water from Millerton Lake to prevent or minimize a spill or to meet flood control criteria, he may establish" periods of time in which irrigators may be obligated to purchase Class 2 water.
Precisely because the availability of flood waters typically begins in the months immediately preceding the irrigation season, the contract provides incentives for the Madera Irrigation District to purchase Class 2 flood water by obligating it either to pay for such water released into the Madera Canal or to forfeit a certain percentage right to purchase Class 2 water at a later date. Thus, if the irrigation contractor does not purchase flood waters released into the Madera Canal in March, the contract provides for forfeiture of up to 7% of the total Class 2 water supply it has contracted to purchase for that year. Unpurchased Class 2 flood water in April may lead to forfeiture of 12% of the contractor's Class 2 total annual water supply, with the percentage rising to 16% for unpurchased Class 2 supply for May, and to 20% for "any other month." MID Contract, Art. 3(h)(3). Flood water released into Madera Canal that is unpurchased as Class 2 water simply flows through the canal and is ultimately released into the Fresno and/or Chowchilla Rivers. The contractual arrangement requires irrigators to accept flood water in the canal, thereby demonstrating that the irrigation purpose of the Madera Canal must coexist with, and at times be subordinate to, the flood control purpose of the facility.
C. The Madera Canal Receives Water Intended Solely For Flood Control Purposes
1. In addition to the dual flood control and irrigation purposes that may be served simultaneously by water releases into the Madera and Friant-Kern Canals, releases serve solely a flood control function in periods of high water flow or when capacity in Millerton Lake must be made available for anticipated inflows. Without citation to any authority, petitioner asserts (Br. 25) that all flood releases from Friant Dam automatically are made into the San Joaquin River. The 1999 Corps report, however, conclusively refutes that assertion: "Typically, during high snowmelt the Madera Canal can be used to convey up to 1,200 cfs into the Fresno-Chowchilla River system." Post-Flood Assessment 3-32. That report notes some of the important discretionary considerations that the Corps and Bureau must take into account that influence a decision to release flood water not intended for irrigation into the Madera and Friant-Kern Canals. Even when a release of water downstream could be made without exceeding the maximum release level of 8000 cfs, for example, the Bureau (pursuant to Corps' rules) releases flood waters into the canals for other reasons, including to ensure safer recreational usage of the San Joaquin, to protect areas downstream, to avert flooding downstream just below Friant Dam, and to avoid flooding at the Mendota Dam. Id. 3-32 to 3-33. Releases into the Madera Canal serve those purposes even when downstream releases would not exceed 8000 cfs. Id. at 3-33.
Petitioner thus erroneously describes the Field Working Agreement between the Corps and the Bureau as "establish[ing] that a substantial proportion of the water that damaged Petitioner's farm has no relationship to flood control whatsoever" and in requiring releases "only into the river bed, not the Madera Canal." Br. 25. Petitioner cites nothing in the agreement itself to support those assertions, and the foregoing description of the operations of the Friant Division is consistent with the proposal made by the Corps that was adopted by Congress in 1944, the Corps' implementing regulations in 1955, and more than four decades of experience of flood control operations in the Friant Division.
2. The history of water releases from Friant Dam, as documented by publicly-available data from the U.S. Geological Survey, also supports the conclusion that the Madera and Friant-Kern Canals are routinely used for flood releases, except in the most severe drought years. As of 1980, according to the Corps of Engineers, 138,000 acre-feet of water each year is released into the Madera Canal to meet requirements for supplying Class 1 irrigation water.20 During the twenty-year period between 1979 and 1999, however, the amount of acre-feet of water released annually into the Madera Canal well exceeded the 138,000 acre-feet intended predominantly for irrigation use (Class 1 water) in every year except the severe drought years of 1989-1990 (111,757 acre-feet) and 1991-1992 (126,750 acre-feet).21 In all other years (18) of that twenty-year period, the amount of water released into the Madera Canal exceeded the amount dedicated predominantly for irrigation. Moreover, in eleven of those eighteen years, the amount of water released was far greater than the amount needed to meet irrigation obligations by the Bureau of Reclamation.22 Nor were the years in those two decades anomalous. Based on a calculation of the mean total acre-feet per month released into the Madera Canal between 1949 and 1999, the total yearly mean released in those years was 264,740 acre-feet, or approximately double the amount of water dedicated predominantly for irrigation purposes. Indeed, contrary to petitioner's undocumented assertion that "a large proportion of the water that has damaged Petitioner's farm has no relationship to flood control at all" (Br. 22), data collected from the past twenty years establish that more water has been released into the Madera Canal for flood control purposes than for predominantly irrigation purposes.23
The Court thus need not go beyond the test it articulated in James to resolve this case. All parts of the CVP and its Friant Division, including the Madera Canal, operate as a federal flood control project, and the waters that allegedly escaped from the canal and damaged petitioner's property both had a direct relationship to flood control and constituted waters that the project could not control. Petitioner simply misapprehends (Br. 19) the practical operation of the Friant Division as demonstrated by statutory requirements, regulatory implementation, and historical practice when asserting that "the Madera Canal has at most an exceedingly attenuated relationship to flood control activities."
IV. PETITIONER'S PROPOSAL TO LIMIT SECTION 702c TO DAMAGES CAUSED BY PROJECTS WITH A PRIMARY PURPOSE OF FLOOD CONTROL IS FLAWED
Petitioner proposes a new test for Section 702c immunity: "[W]hether the waters [causing the damage] were carried in a flood control project 'primarily for the purpose of flood control.'" Br. 19. Until its brief on the merits in this Court, petitioner had not advanced that position in this litigation; it had instead argued that "for water to be 'flood water' it would have to (1) pass through a flood control project and (2) be for the purposes of or related to flood control." Pet. App. 3. Petitioner's new approach, which has not been adopted by any court of appeals, is flawed.
A. The "Primary Purpose" Test Conflicts With The Text Of Section 702c As Construed In James
1. Petitioner argues that its "primary purpose" test is the "most consistent with Congress' intent." Br. 19. As this Court recognized in James, however, the "starting point" for determining that intent is "the language of the statute itself." 478 U.S. at 604 (quotation omitted). In James the Court emphasized the "sweeping terms" of Section 702c, noting the statute's reference to "'[n]o liability of any kind * * * for any damage from or by floods or flood waters at any place.'" Ibid. (quoting Section 702c). As the Court recognized, "[i]t is difficult to imagine broader language." Ibid. That language, and the "equally broad and emphatic language found in the legislative history," showed that "Congress clearly sought to ensure beyond doubt that sovereign immunity would protect the Government from 'any' liability associated with flood control." Id. at 608. Far from being the "most consistent" (Pet. Br. 19) with that intent, petitioner's reading of Section 702c would considerably narrow the broad immunity set forth in the statutory language and recognized in James.
2. Petitioner cites other statutory provisions that ostensibly would be affected by a failure of this Court to adopt petitioner's approach. See Pet. Br. 11-12 & n.2 (discussing 33 U.S.C. 702j-2), 14-16 (discussing 33 U.S.C. 702j). By their plain terms, however, those statutory references are inapposite.
Petitioner cites (Br. 12) 33 U.S.C. 702j-2, for example, which was enacted in 1936 as a specific immunity provision for the White River Levee District. The provision afforded immunity from any damages-whether caused by flooding or not-"on account of the use of said area [both submerged and unsubmerged land] for reservoir purposes during said emergency." 33 U.S.C. 702j-2, Act of June 15, 1936, ch. 548, § 5, 49 Stat. 1509. That provision does not amend or limit Section 702c in any way, and its distinctive scope and purposes are hardly rendered "superfluous" (Pet. Br. 12) by rejection of petitioner's newly proposed "primary purpose" test.
Petitioner is similarly mistaken in asserting (Br. 14-16) that 33 U.S.C. 702j provides support for the "primary purpose" test. That provision, which was enacted in 1928, directed the Corps of Engineers to submit plans to Congress for flood control projects on tributaries to the Mississippi River and to report on the possible beneficial effects of building a system of reservoirs on those tributaries. 33 U.S.C. 702j. Petitioner points out that Congress referred in Section 702j to waters, once they were contained in reservoirs, as "reservoired waters." According to petitioner, Congress thus must have meant to distinguish "reservoired waters" in Section 702j from "flood waters" in Section 702c, and specifically must have intended the latter term to exclude "waters stored and then diverted for some beneficial purpose other than flood control." Br. 14 (emphasis added).
Petitioner's suggested dichotomy between "reservoired waters" and "flood waters" goes too far, however, because flood control dams accomplish all of their beneficial pur-poses, including flood control, by converting the entire flow of the river into "reservoired waters." Temporary storage of water in reservoirs permits its use for recreation, irrigation, power generation, and the prevention of floods. Indeed, in James the Court found that the government's immunity under Section 702c applied to damage from reservoired waters that had been used for recreational purposes. In James, as here, reservoired waters serving multiple uses were released "to prevent flooding." 478 U.S. at 605-606 n.7. To distinguish James, petitioner must ask the Court to hold that water entering a reservoir as "flood water" retains that character if it is put to recreational uses, but not if it is used for irrigation, or perhaps to hold that immunity is defeated by beneficial use apart from flood control after release from the reservoir but not by such non-flood beneficial use before release. Those distinctions should be rejected. They have no relationship to the purpose of the immunity in facilitating federal investment in flood control projects by shielding the United States from liability for damages caused by flood waters in these projects.
Petitioner also invokes the Federal Tort Claims Act (FTCA), 28 U.S.C. 2671 et seq., to support its proposed "primary purpose" test, but James definitively rebuts that argument. Petitioner points out (Br. 16-18) that Congress considered, and failed to enact, exceptions to the FTCA that would have explicitly covered flood control and irrigation projects. The result, according to petitioner (Br. 18), is that negligent operation of federal flood control projects is excluded from the FTCA's waiver of sovereign immunity only if the negligent conduct is within the scope of the FTCA's "discretionary function exception." 28 U.S.C. 2680(a). If that were so, however, the Court in James would have had to determine first whether the negligence alleged in that case was within the discretionary function exception. The Court did not do so, but instead assumed that the alleged government negligence fell outside the scope of any exception to the FTCA: "This case presents the question whether [Section 702c] * * * bars recovery where the Federal Government would otherwise be liable under the Federal Tort Claims Act." 478 U.S. at 598-599 (emphasis added). By determining that Section 702c immunity applied to such a case, the Court held that the FTCA did not amend or displace Section 702c.24
B. Petitioner's "Primary Purpose" Test Finds No Support In Court Of Appeals' Decisions
The "primary purpose" test petitioner now proposes has not been adopted by any court of appeals in the wake of James. See notes 26 & 27, infra. Nor would petitioner prevail (see Pet. 18, Pet. Reply 1) under the approaches that have been adopted by other circuits. See Br. in Opp. 11-16. Contrary to petitioner's suggestion (Br. 7), Henderson v. United States, 965 F.2d 1488 (8th Cir. 1992), did not articulate a "primary purpose" test. On the contrary, the court of appeals in Henderson stated that "section 702c immunity applies if 'governmental control of flood waters was a substantial factor in causing [the plaintiff's] injuries.'" Id. at 1492 (quoting Dewitt Bank & Trust Co. v. United States, 878 F.2d 246, 247 (8th Cir. 1989), cert. denied, 494 U.S. 1016 (1990)) (emphasis added). The Eighth Circuit rejected immunity not because flood control was "the primary purpose" of the project in question (which was power generation), but rather because flood control was not even a substantial factor in causing the injury.25
Unlike the "primary purpose" test advanced by petitioners, the courts of appeals have not inquired into whether the government's flood control activities were the proximate cause of the damages. Rather, courts have examined the factual allegations of the complaint to determine whether they describe a connection between the damages and flood control. Most often, as here, that determination can be made on the pleadings alone. Where the relation between the injury and flood control has been absent, courts have denied immunity.26 Virtually every circuit, however, also has issued a reported decision in which a district court judgment dismissing the complaint has been affirmed because the factual relation to flood control is sufficiently evident from the pleadings to justify the government's immunity under Section 702c.27 For that reason, and those stated in Parts II and III, the judgment dismissing petitioner's complaint would have been uniformly affirmed by the courts of appeals.
C. The "Primary Purpose" Test Would Deny The Government Flood Immunity In Cases Involving Multiple-Use Projects
Adoption of the "primary purpose" test proposed by petitioner would sharply limit the immunity created by Section 702c. Petitioner's test casts doubt on whether the projects designated as flood control projects by the Corps of Engineers pursuant to Section 7 of the Flood Control Act of 1944 (see 33 C.F.R. 208.11(e)) would be covered by Section 702c.28 Indeed, under petitioner's view, no part of the Central Valley Project, which is owned and operated by the Bureau of Reclamation, would qualify for flood immunity, since it was expressly authorized by Congress to be a multiple-use project and Congress did not specify that flood control was the project's "primary purpose." See p. 7, supra. Petitioner's rule thus threatens to erase decades of settled law.29 Only projects owned and operated by the Corps itself would likely have flood control as their primary purpose and, even as to those, a congressional mandate that such projects also serve multiple uses would introduce uncertainty into the scope of immunity.
Petitioner asserts (Br. 21) that its "primary purpose" test promises "ease of administration," but that contention is incorrect. On the contrary, it would require courts to calibrate precisely the respective purposes served by multiple-use projects, a task that Congress itself does not undertake when it authorizes such projects or when it subsequently amends an original authorizing statute.30 Even if petitioner's proposed test were limited to cases involving releases of water from federal flood control reservoirs, it would still in many instances require courts to establish priorities among multiple purposes without any evidence that Congress ranked their relative importance. As the court in Morici concluded, "it is very difficult, if not impossible, as a practical matter to segregate particular . . . water releases into precise categories of purpose; a single release may well serve multiple purposes, just as the project itself serves multiple purposes." 681 F.2d at 648 (quotation omitted). Furthermore, petitioner's proposed test would require complex evidentiary showings that Congress could not have intended when it enacted Section 702c to "ensure beyond doubt that sovereign immunity would protect the Govern-ment from 'any' liability associated with flood control." James, 478 U.S. at 608.31 Petitioner's approach also promotes inefficient government, because it denies an immunity for damages when multiple objectives may be obtained and permits immunity only when flood control is the primary purpose of the project.32 Finally, adoption of petitioner's approach would not resolve the central issue in this case: whether the leakage alleged to have been the result of accreted water damage over time was caused by the flood control purpose or the irrigation purpose of the Madera Canal.
Cases that give rise to assertions of immunity under the Flood Control Act of 1928 involve a diverse array of factual characteristics that make articulating a legal standard more specific than the James test a difficult task. Those include: the nature of the claim (property damage versus personal injury); the nature of the government's alleged negligence (e.g., structural or design, human operational error, failure to warn); the relationship of specific governmental actions to the injury (e.g., long-term operations versus discrete actions); and the traceability of the injury to a particular purpose of the project. None of those distinctions alone provides a sound basis for determining that Section 702c does or does not apply. On the other hand, the requirement of a nexus or relation between flood control and the injury that has been applied by the courts of appeals has yielded consistent results that implement Congress's intent to create a broad immunity for federal flood control activities.
As we have shown, the Madera Canal is an integral part of the flood control operations of the CVP's Friant Division, in its routine uses to serve irrigation needs and to convey heavy flows in the San Joaquin River to downstream locations. Because by regulation and historical practice a substantial quantity of flood water has been released into the Madera Canal, as a matter of law the United States is entitled to immunity from petitioner's suit.
The judgment of the court of appeals should be affirmed.
SETH P. WAXMAN
DAVID W. OGDEN
Assistant Attorney General
BARBARA D. UNDERWOOD
Deputy Solicitor General
DAVID C. FREDERICK
Assistant to the Solicitor General
ROBERT S. GREENSPAN
IRENE M. SOLET
1 This document is a compilation of state and federal materials pertaining to the Central Valley Project. For ease of reference, we cite this document rather than the original sources.
2 The Commission was abolished in 1986 and its functions transferred to the Secretary of the Army. Water Resources Development Act of 1986, Pub. L. No. 99-662, Tit. XI, § 1106, 100 Stat. 4229.
3 The Senate Committee on Irrigation and Reclamation recognized that California's water plan as a whole, and Friant reservoir in particular, would serve an important flood control function. In a 1933 report, the Committee stated: "One of the important proposed objectives of the great central valley project of California is the provision of additional flood protection on the Sacramento and San Joaquin Rivers by the regulation of floods in the two proposed major storage reservoirs, Kennett Reservoir on the Sacramento River and Friant Reservoir on the San Joaquin River." H.R. Doc. 416, at 500. The Committee compared the problem of flooding on the Sacramento and San Joaquin Rivers "to that on the Mississippi river." Id. at 499. It recognized that "[a]lthough of smaller magnitude than that on the Mississippi River, the problem [of flood control on the Sacramento and San Joaquin] is even more intensive, involving maximum flood flows ten times as great per unit of drainage area than the maximum flood flow of the Mississippi below Cairo or Arkansas City." Ibid. The Committee thus predicted that the Kennett and Friant reservoirs "can be operated to materially decrease the flood flows and will, therefore, greatly lessen the frequency of floods and possible damage resulting therefrom." Ibid.
4 Under the Act of June 25, 1910, ch. 407, § 4, 36 Stat. 836, no reclamation project could be undertaken unless recommended by the Secretary of the Interior and approved by the President. Similarly, under the Act of Dec. 5, 1924, ch. 4, § 4(B), 43 Stat. 702, no reclamation project could be approved for construction until the Secretary of the Interior made a written finding that the project was feasible and that its cost was likely to be repaid to the United States.
5 In 1997, California experienced "one of the largest and most extensive flood disasters in the State's history." Post-Flood Assessment, at ES-1. The House Report on the bill ultimately enacted as the Energy and Water Development Appropriations Act, 1998, Pub. L. No. 105-62, 111 Stat. 1320, states that, "[i]n response to the devastating floods of 1997, the Committee has added funds and directs the Corps of Engineers to conduct * * * a comprehensive post-flood assessment for the California Central Valley (Sacramento River Basin and San Joaquin River Basin)." H.R. Rep. No. 190, 105th Cong., 1st Sess. 24 (1997). The Post-Flood Assessment report responded to that directive.
6 Operation of the Friant Division to serve the functions of flood control and irrigation requires extensive weather forecasting efforts by federal and state agencies. Friant Flood Report 16. Accurate predictions of water runoff historically have been difficult to make consistently. Between 1953 and 1978, for example, the error rate varied greatly, with underestimates and overestimates of the forecasted runoff into Millerton Lake routinely as great as 100 percent of the actual runoff. See id. Chart A-9. Given those large discrepancies, which by no means have been limited to those years, the task of managing the water level at Friant Dam requires great flexibility and the need to make frequent, often daily adjustments to the amount of water released.
7 Although for purposes of the motion for judgment on the pleadings the allegations of the complaint must be accepted as true, the government denies that leakage from the canal has caused harm to petitioners. See J.A. 58. At times when the canal is completely dry, petitioner's land has been flooded, which suggests that the problem in fact stems not from leakage from the canal but from subterranean ground water or over-irrigation by petitioner from the water it obtains from the San Joaquin River.
8 Protection of the United States from liability for flooding of land caused by alleged construction failures of a federal flood control project was central to Congress's concern in creating an immunity for federal flood control activities. See James, 478 U.S. at 606-608; see also id. at 617 (Stevens, J., dissenting) (agreeing with majority that immunity under Section 702c should extend to "overflow damage to land").
9 The Court in James (478 U.S. at 605) adopted virtually verbatim the position set forth in the government's brief: "Since Congress was legislating with respect to flood control projects designed to carry flood waters, the conclusion is inescapable that the statute refers to all waters contained in or carried through such structures as well as waters the structures could not retain." 85-434 U.S. Br. 17. The government noted that the Fifth Circuit in James had viewed the statutory language as ambiguous "because the damage that formed the basis for the tort claim might not be related to the operation of a flood control project." Id. at 17 n.7. Rather than broadly arguing for a test that reached well beyond the facts presented in James, as petitioner asserts (Br. 10), the government in James specifically noted that the Court did not need to decide whether the immunity applied in other circumstances, because the damage in James "did result from the release of flood waters from a federal flood control project." Id. at 17 n.7. The Court itself agreed with that conclusion. See 478 U.S. at 610 ("the release of the waters at the Millwood Reservoir and at the Courtableau Basin was clearly related to flood control"). As we establish on pp. 32-40, infra, that is also true in this case.
10 See p. 7, supra. Subsequent statutes authorizing additional features of the CVP repeated or reincorporated that language. See, e.g., Act of Sept. 26, 1950, ch. 1047, § 1, 64 Stat. 1036 ("the entire Central Valley project * * * is hereby reauthorized and declared to be for the purposes of * * * controlling floods"); Act of Aug. 27, 1954, ch. 1012, § 1, 68 Stat. 879 ("the entire Central Valley project * * * is hereby reauthorized" and declared to be for the purposes previously authorized).
11 Section 2 of the 1937 Act mentions the purpose of the CVP to be "controlling floods" in two provisos, only the second of which is addressed by petitioner. See 50 Stat. 850. The first proviso authorizes "the entire Central Valley project" for the purpose of "controlling floods." Ibid. Petitioner attempts to draw significance (Br. 27) from the fact that the second proviso of Section 2 states that "the said dam and reservoirs shall be used, first, for river regulation, improvement of navigation, and flood control; second, for irrigation and domestic uses; and, third, for power." Ibid. (emphasis added). The absence of any mention of "canals" in the priority to be given to the respective uses of the project is not surprising, since canals were thought to serve only irrigation and flood control functions at that time. In any event, the omission of "canals" in the second proviso cannot blunt the significance of the first proviso's unambiguous intent that the "entire Central Valley Project"-which unquestionably includes its canals-is for the purpose of "controlling floods."
12 An additional complicating factor in release calculations from Friant Dam is that the Little Dry, Big Dry, and Cottonwood Creeks empty into the San Joaquin River just below Friant Dam, so their flow at peak levels must be subtracted from the amount of water that may safely be released from Friant Dam. That amount of water, in turn, must either be stored in Millerton Lake until it can be safely released downstream or released into the Madera and/or Friant-Kern Canals. See Post-Flood Assessment 3-33 to 3-34.
13 In its 1944 estimates, which were made prior to completion of Friant Dam, the Corps had put this number at 7000 cubic feet per second. See Flood Control Doc. No. 2, at 39, Tab. 22.
14 Even prior to the promulgation of those rules by the Corps, the Secretary of the Interior recognized the Friant Division's important flood-control role, reporting to the President in 1947 that "operation of Millerton Lake for flood-control purposes will be such as to prevent the maximum discharge under the worst flood conditions on the San Joaquin River, except one, between 1902 and 1943, from exceeding 8,000 cubic feet per second at Skaggs Bridge." H.R. Doc. No. 146, 80th Cong., 1st Sess. 10-11 (1947). Similarly, the Interior Department's 1949 report to Congress on the "Central Valley Basin" contained repeated references to the flood control function of the CVP's Friant component. S. Doc. No. 113, at 36 (referring to "flood-control benefits from works in Sacramento and lower San Joaquin Valleys * * * including the benefits from Shasta and Friant Reservoirs"); id. at 78 (referring to "flood-control benefits from the comprehensive plan, includeing those from the operation of Shasta and Friant Reservoirs"); id. at 93 (stating that "Millerton Lake, formed by Friant Dam * * * will regulate floods and store San Joaquin River run-off for diversion into the Madera and Friant-Kern canals"); id. at 132 (describing one of the functions of Millerton Lake as "to provide flood control on San Joaquin River"); id. at 164 (referring to "additional protection" against flooding along the San Joaquin River and in the Sacramento-San Joaquin delta provided by "Millerton Lake (Friant) Reservoir"); id. at 167 (describing future flood control benefits on the San Joaquin River and its tributaries "attributable to Friant Reservoir, the various prospective reservoirs, channel improvements, and other flood-control works").
15 The fact that the structures authorized in the Flood Control Act of 1928 to control floods on the Mississippi River may not have included the kind of multiple-use projects later undertaken in the Western states and in other parts of the country (see Pet. Br. 14 n.4) does not undercut the application of Section 702c to such projects. The courts of appeals have consistently held since long before James, and petitioner does argue to the contrary, that Section 702c applies to flood control projects authorized by separate statutes that succeeded the 1928 Act. See Aetna Ins. Co. v. United States, 628 F.2d 1201, 1204 (9th Cir. 1980), cert. denied, 450 U.S. 1025 (1981); Lenoir v. Porters Creek Watershed Dist., 586 F.2d 1081, 1086 & n.4 (6th Cir. 1978); cf. National Mfg. Co., 210 F.2d at 270. The direct link between the Flood Control Act of 1928 and Congress' subsequent enactments in the area of flood control supports the longstanding interpretation of Section 702c to apply to projects in addition to those authorized in the 1928 statute. See Flood Control Act of 1936, ch. 688, § 8, 49 Stat. 1596 (specifically saving all provisions of the 1928 Act); Flood Control Act of 1944, ch. 665 § 2, 58 Stat. 889 (referencing the 1936 Act). See also James, supra (applying Section 702c to project authorized under the Flood Control Act of 1946, ch. 596, 60 Stat. 647).
16 Petitioner places significance (Br. 38) on the statement in Gerlach that "[t]he Central Valley basin development envisions, in one sense, an integrated undertaking, but also an aggregate of many subsidiary projects, each of which is of first magnitude." 339 U.S. at 733. In the context of that case, however, the Court's reference to treating the subdivisions of the CVP as each of "first magnitude" is most naturally understood to mean that Congress had specifically directed that the reclamation laws apply to the Friant Division. The Court in Gerlach did not address how such a focus on the CVP's "subsidiary projects"-here, the Friant Division-would affect the application of Section 702c. We have shown, however, that flood control is a central purpose of the Friant Division, so even an enquiry focused on that component of the CVP would not undercut application of the immunity here.
17 Petitioner's complaint asserts only that the Madera Canal is used for irrigation purposes, and not that its exclusive use is for irrigation. See J.A. 8. Thus, accepting the allegations as true does not foreclose the government's defense that the canal also carries flood waters, or that the irrigation waters also serve a flood control function. In any event, even if the complaint is liberally construed in petitioner's favor, the government is entitled to judgment on the pleadings when "it appears beyond doubt that the plaintiff can prove no set of facts in support of his claim which would entitle him to relief." Conley v. Gibson, 355 U.S. 41, 45-46 (1957). A court may take judicial notice of matters of public record-such as statements in judicial opinions, regulatory documents, and historical records-that refute the allegations in the complaint. See, e.g., Papasan v. Allain, 478 U.S. 265, 268 n.1 (1986) (in reviewing motion to dismiss under Fed. R. Civ. P. 12, "we are not precluded in our review of the complaint from taking notice of items in the public record"); Bowen v. Johnston, 306 U.S. 19, 23 (1939); Lamar v. Micou, 114 U.S. 218, 223 (1885); Menominee Indian Tribe v. Thompson, 161 F.3d 449, 455-462 (7th Cir. 1998), cert. denied, 526 U.S. 1066 (1999).
18 Petitioner seemingly acknowledges that some of the water that flowed through the Madera Canal was stored in Millerton Lake for the purpose of controlling floods. See Br. 8 ("[a] large proportion of the waters that damaged Petitioner were not 'flood waters' even when originally stored behind Friant Dam") (emphasis added); id. at 22 ("a large proportion of the water that has damaged Petitioner's farm has no relationship to flood control at all") (emphasis added); id. at 23 ("a still much smaller proportion of the contents of the Madera Canal could include those waters," referring to waters stored behind the dam that could otherwise have caused flooding) (emphasis added).
19 In the district court petitioner stipulated to the fact that "[t]he United States entered into an agreement with [Madera Irrigation District (MID)] whereby MID was allowed to flow water through the Madera Canal on the terms and conditions specifically set forth in the agreement." J.A. 72; see also J.A. 81 (scheduling conference order confirming factual stipulation). We have lodged copies of this contract with the Clerk.
20 "Class 1" water is that amount of water the irrigators can rely on the Bureau to deliver except in the most severe drought years. As such, the government commits to supplying at least that amount of water each year to irrigators. Any other water supplied is conditional upon flood flows. See Friant Flood Report 14 n.1; see also MID Contract, Art. 1(c).
21 The U.S. Geological Survey monitors water releases throughout California (and elsewhere in the country). All of the data cited in the text and notes is compiled from historical stream flow data for releases from Friant Dam to the Madera Canal monitored at Gaging Station No. 11249500, which is available on the U.S. Geological Survey web site <<http://water.wr.usgs.gov/index.html>>.
22 The following data, based on adding the monthly totals available from USGS, are total acre-feet of water per year released into the Madera Canal: 1979-1980 (483,660); 1981-1982 (456,440); 1982-1983 (532,600); 1983-1984 (340,450); 1985-1986 (513,030); 1992-1993 (390,310); 1994-1995 (389,110); 1995-1996 (363,380); 1996-1997 (391,980); 1997-1998 (355,290); and 1998-1999 (323,150). Thus, even if one were to assume that the 138,000 acre-feet per year released for Class 1 irrigation water supply has had no relation whatsoever to flood control, most of the water released in those years had a flood control purpose. E.g., 1979-1980 (483,660 - 138,000 = 345,660 acre-feet).
23 Between 1979 and 1999, a total of 5,965,161 acre-feet of water has been released into the Madera Canal. Assuming the figure of 138,000 acre-feet of Class 1 water per year, the total amount of water released to serve a predominantly irrigation purpose was 2,760,000 acre-feet. The remainder-3,205,161 acre-feet-was water that served a flood control purpose.
24 Not only did this Court reject that argument in James, but every court of appeals to consider that contention prior to James had also rejected it. See, e.g., Aetna Ins. Co., 628 F.2d at 1204-1205; Callaway v. United States, 568 F.2d 684 (10th Cir. 1978); Florida E. Coast Ry. v. United States, 519 F.2d 1184 (5th Cir. 1975); Parks v. United States, 370 F.2d 92 (2d Cir. 1966); Stover v. United States, 332 F.2d 204 (9th Cir.), cert. denied, 379 U.S. 922 (1964); National Mfg. Co. v. United States, 210 F.2d 263 (8th Cir.), cert. denied, 347 U.S. 967 (1954).
25 The Henderson court used language more akin to the "wholly unrelated" test: "We do not believe that section 702c bars Henderson's cause of action in this case because the dam activity here was related to generating electricity and not to flood control." 965 F.2d at 1492. The Corps of Engineers released the water that caused Henderson's death on the order of an electric utility, which under a contract with the United States had the prerogative to make such a request of the Corps. Ibid. The precise claim involved a failure by the Corps to warn recreational users of rapidly rising water levels when water was released from the dam. Id. at 1491. As the court made clear, its rationale rested not on the "primary purpose" of the project or the discharge, but rather on the ground that the court could not "conclude that 'governmental control of flood waters was a substantial factor in causing [Robert Henderson's] injuries.'" Id. at 1492 (citation omitted).
26 See Kennedy v. Texas Utils., 179 F.3d 258 (5th Cir. 1999) (no immunity where injury occurred on dry land due to a condition with no association to flood control); Cantrell v. United States, 89 F.3d 268 (6th Cir. 1996) (no immunity from claim by recreational boater injured by allegedly negligent driver of Army Corps of Engineers' boat); Henderson v. United States, 965 F.2d 1488 (8th Cir. 1992) (no immunity in death of fisherman where drowning was caused by release of water, at direction of private power company, from dam operated for hydroelectric power generation); Boyd v. United States, 881 F.2d 895 (10th Cir. 1989) (no immunity for allegedly negligent failure to warn swimmers of hazard from boats, in death of snorkeler struck by privately operated power boat at flood control lake); E. Ritter & Co. v. Department of the Army, 874 F.2d 1236 (8th Cir. 1989) (no immunity for erosion caused by rain waters that had not yet come in contact with flood control project); see also Williams v. United States, 957 F.2d 742 (10th Cir. 1992) (concluding, on basis of record evidence and language of authorizing statute, that particular project was not a flood control project).
27 Reese v. South Fla. Water Management Dist., 59 F.3d 1128, 1130 (11th Cir. 1995) (per curiam) (government immune from suit in drowning at federal flood control lake caused by opening of water control device, noting that "periodic release of water is fundamental to the operation of a flood control project"); Boudreau v. United States, 53 F.3d 81, 84 (5th Cir. 1995) (based on facts of case, government immune from suit where there was a "sufficient association" between injury to recreational boater during Coast Guard rescue and activities of flood control), cert. denied, 516 U.S. 1071 (1996); Holt v. United States, 46 F.3d 1000, 1004 (10th Cir. 1995) (government immune from suit where there was a "sufficient nexus" between car accident and mist, which was created by water released from flood control project's dam and which caused an ice slick on an adjacent road); Fisher v. United States Army Corps of Engineers, 31 F.3d 683, 685 (8th Cir. 1994) (government immune from suit where shallow level of water as a result of operation of flood control project was a "substantial factor" in a recreational diving accident); Thomas v. United States, 959 F.2d 232 (4th Cir. 1992) (Table) (government immune from suit in recreational diving accident occurring at lake which, despite its commercial uses, had flood control uses as well and which was "monitored and maintained daily by the Army for the purpose of controlling floods"); Fryman v. United States, 901 F.2d 79, 82 (7th Cir.) (government immune from suit for injuries sustained at lake created as part of a flood control project which "increase[d] the probability" of injury), cert. denied, 498 U.S. 920 (1990); Dawson v. United States, 894 F.2d 70, 74 (3d Cir. 1990) (government immune from suit for recreational swimming accident caused by unsafe depth of water due in part to releases of water for flood control purposes); Mocklin v. Orleans Levee Dist., 877 F.2d 427, 430 (5th Cir. 1989) (government immune from suit for drowning caused by deep water in a flotation channel that had been excavated for a flood control project).
28 That regulation lists 112 projects operated by entities other than the Corps for a variety of purposes. By law those projects must follow the regulations promulgated by the Corps. See 33 C.F.R. 208.11(e) (Table).
29 Not only has the Ninth Circuit consistently held that multiple-use projects operated by agencies other than the Corps of Engineers are entitled to flood immunity, see, e.g., Aetna Ins. Co., supra; Morici, supra; but other courts have likewise upheld flood immunity for multiple-purpose projects operated by the Corps, the Bureau, and the Soil Conservation Service. See, e.g., Lenoir, 586 F.2d at 1084; Taylor v. United States, 590 F.2d 263 (8th Cir. 1979); Callaway v. United States, 568 F.2d 684 (10th Cir. 1978); Hedrick v. United States, 184 F. Supp. 927 (D.N.M. 1960).
30 The following are examples of statutes that specify multiple purposes of projects, including flood control, without indicating any priority among the purposes specified: Act of Sept. 20, 1966, Pub. L. No. 89-596, 80 Stat. 822 (Scoggins Dam); Act of Aug. 24, 1962, Pub. L. No. 87-594, 76 Stat. 395 (Arbuckle Dam); Act of June 3, 1960, Pub. L. No. 86-488, 74 Stat. 156 (B.F. Sisk Dam); Act of Aug. 1, 1956, ch. 809, 70 Stat. 775 (Prosser Creek Dam, Stampede Dam, and Marble Bluff Dam); Act of Dec. 29, 1950, ch. 1183, 64 Stat. 1124 (Sanford Dam); Act of June 5, 1944, ch. 234, 58 Stat. 270 (Hungry Horse Dam); Act of Aug. 11, 1939, ch. 699, 53 Stat. 1414 (Sumner Dam).
The following are examples of statutes in which Congress has specified that the "principal purposes" of the projects include, inter alia, flood control: Act of Aug. 16, 1957, Pub. L. No. 85-152, 71 Stat. 372 (Twin Buttes Dam); Act of Aug. 6, 1956, ch. 981, 70 Stat. 1059 (Little Wood River Dam); Act of Feb. 25, 1956, ch. 71, 70 Stat. 28 (Fort Cobb Dam). Even in those statutes, however, Congress did not prioritize among the various "principal purposes." The Corps of Engineers has included all of the projects referenced above in its flood control regulations pursuant to 33 C.F.R. 208.11.
31 For that reason as well, there is no basis for thinking that the Court's application of immunity in this case will cause any of the hypothesized harms suggested by petitioner (Br. 22) under 42 U.S.C. 4022(a)(2)(B), 22 U.S.C. 277d-12, or 30 U.S.C. 877(f). The absence of any case citations in petitioner's brief suggests that the likelihood of such harm is remote. In any event, if Section 702c "is to be changed, it should be by Congress and not by this Court." James, 478 U.S. at 612.
32 For that reason as well, petitioner's suggestion (Br. 31-33) that the applicability of Section 702c immunity should turn in part on whether the government is reimbursed for the construction costs of the project in question should be rejected.
33 U.S.C. 702c, provides, in pertinent part:
No liability of any kind shall attach to or rest upon the United States for any damage from or by floods or flood waters at any place: Provided, however, That if in carrying out the purposes of sections 702a, 702b to 702d, 702e to 702g, 702h, 702i, 702j, 702k, 702l, 702m, and 704 of this title it shall be found that upon any stretch of the banks of the Mississippi River it is impracticable to construct levees, either because such construction is not economically justified or because such construction would unreasonably restrict the flood channel, and lands in such stretch of the river are subjected to overflow and damage which are not now overflowed or damaged by reason of the construction of levees on the opposite banks of the river it shall be the duty of the Secretary of the Army and the Chief of Engineers to institute proceedings on behalf of the United States Government to acquire either the absolute ownership of the lands so subjected to overflow and damage or floodage rights over such lands.
[The last page of the printed brief is a map of the San Joaquin River System, Flood Channel Design Flows and Project Levees, which is not available in postable electronic form.]