KANSAS v. COLORADO on exceptions to report of special master

543US1 Unit: $$U6 [03-06-07 12:48:44] PAGES PGT: OPIN
86 OCTOBER TERM, 2004
Syllabus
KANSAS v. COLORADO
on exceptions to report of special master
No. 105, Orig. Argued October 4, 2004—Decided December 7, 2004
Kansas and Colorado entered into the Arkansas River Compact (Compact)
in 1949, but disagreements over the equitable distribution of the river’s
upper waters persisted. In 1985, Kansas charged that Colorado had
violated the Compact by drilling new irrigation wells that, in Compact
Art. IV–D’s words, “materially depleted” the river water otherwise
available “for use” by Kansas’ “water users.” Accepting the recommen-
dation set forth in the First Report of the Special Master to find that
Colorado had unlawfully depleted the river in violation of Art. IV–D,
this Court remanded the case for remedies. Kansas v. Colorado, 514
U. S. 673, 694. In proposing remedies in his Second and Third Reports,
the Master said that Colorado’s Compact violation had occurred from
1950 through 1994; recommended that Colorado pay Kansas damages;
divided the water losses into six categories, calculating damages some-
what differently for each; and urged that Kansas be awarded prejudg-
ment interest on damages for losses incurred from 1969 through 1994.
The Court subsequently adopted these recommendations with one ex-
ception: It held that prejudgment interest would run from 1985
(not 1969). Kansas v. Colorado, 533 U. S. 1, 15–16 (Kansas III). The
Master has now filed a Fourth Report setting forth his resolution of
certain remaining issues. Kansas takes exception to several of his
recommendations.
Held:
1. Kansas’ request to appoint a River Master to decide various techni-
cal disputes related to decree enforcement is denied. This Court has
appointed River Masters to help resolve States’ water-related disputes
only twice before, Texas v. New Mexico, 482 U. S. 124, and New Jersey
v. New York, 347 U. S. 995, each time on the Special Master’s recommen-
dation, always as a discretionary matter, and only when convinced that
such an appointment would significantly aid resolution of further dis-
putes, see Vermont v. New York, 417 U. S. 270, 275. The Court is not
convinced that such an appointment is appropriate here. For one thing,
further disputes in this case, while technical, may well require discre-
tionary, policy-oriented decisionmaking directly and importantly related
to the underlying legal issues. These potential disputes differ at least
in degree from those that the Court has asked River Masters to resolve
in past cases. See, e. g., Texas v. New Mexico, supra, at 134, 135–136.

543US1 Unit: $$U6 [03-06-07 12:48:44] PAGES PGT: OPIN
87 Cite as: 543 U. S. 86 (2004)
Syllabus
Administration of the present decree will involve the highly complex
computer-run Hydrologic-Institutional Model (H-I Model or Model), and
resolution of many modeling disputes may well call for highly judg-
mental determinations of matters that are more importantly related to
the parties’ basic legal claims. For another thing, the need for a River
Master here is diminished by the fact that the parties may be able to
resolve future technical disputes through binding arbitration under
Compact Art. VII or through less formal dispute-resolution methods
like joint consultation with experts, negotiation, and informal mediation.
The Special Master recommended all of these alternatives, while oppos-
ing appointment of a River Master because it would “simply” make it
“easier to continue this litigation.” Fourth Report 136. Pp. 92–94.
2. Kansas’ exception to the Special Master’s prejudgment interest
calculation is overruled. The calculation and Kansas’ objection grow
out of this litigation’s special history. The Master initially calculated
prejudgment interest on the basis of “considerations of fairness,” Third
Report 97, dividing the prejudgment period into three temporal
subcategories: (1) an Early Period from 1950, when Colorado’s un-
lawful water depletion began, through 1968, when Colorado should first
have known about it; (2) a Middle Period from 1969 through 1984; and
(3) a Late Period from 1985, when Kansas filed its complaint, through
1994, the last year for which evidence was available at the time of the
trial on damages. The Master adjusted damages from all three periods
for inflation, but he awarded additional prejudgment interest only from
1969 to the judgment date, for a total damages award, including pre-
judgment interest, of $38 million. Id., at 107. The Kansas III Court
accepted the Master’s equitable approach, 533 U. S., at 11, but applied
its own “considerations of fairness” in concluding that “prejudgment
interest should begin to accrue” as of 1985, id., at 12–15, and n. 5. On
remand, the Master therefore calculated prejudgment interest from
1985 onward on Late Damages alone. Kansas’ argument that the
Master should have calculated prejudgment interest (from 1985) on all
damages—i. e., on Early, Middle, and Late Damages—would make good
sense in an ordinary case. But the question here is not about the ordi-
nary case, but rather what Kansas III’s prejudgment interest determi-
nation meant in that case’s special context. For one thing, the Court
there did not seek to provide compensation for all of Kansas’ lost invest-
ment opportunities; rather, it sought to weigh the equities. For
another, it was apparent that the Master’s earlier determination
involved both a decision about when to begin to calculate interest (1969)
and what to calculate that interest upon (Middle and Late Damages
only). Saying nothing about the Master’s total exemption of Early
Damages, id., at 14, the Court changed the when (from 1969 to
1985), but not the methodology for calculating the what. In context, the

543US1 Unit: $$U6 [03-06-07 12:48:44] PAGES PGT: OPIN
88 KANSAS v. COLORADO
Syllabus
Court’s silence fairly implies acceptance, not rejection, of the Master’s
underlying methodology, which now yields a post-1985 interest calcula-
tion based upon Late Damages only. This view is reinforced by the
resulting numbers. Were the Court now to accept Kansas’ argument,
the final damages award would be roughly $53 million (in 2002 dollars),
not the $38 million originally calculated by the Master (in 1998 dollars).
The Court cannot reconcile that numerical result with its acceptance in
Kansas III of the Master’s equitable approach and with its own equita-
ble determination, which implied a modest adjustment of the $38 million
award in Colorado’s favor, not, as Kansas now seeks, a major adjust-
ment of the award in Kansas’ favor. Ibid. Pp. 95–99.
3. Kansas’ exception to the Special Master’s recommendation that the
H–I Model be used with a 10-year measurement period to determine
Colorado’s future Compact compliance is overruled. Kansas seeks, in
place of the 10-year period, a 1-year period. Kansas points to Compact
Art. V–E(5), which says that there “shall be no allowance or accumula-
tion of credits or debits for or against either State.” Kansas argues
that a 10-year period averages out oversupply and undersupply during
the interim years, with the likely effect of awarding Colorado a “credit”
in dry years for oversupply in wet years. Adding that Art. IV–D for-
bids Colorado to deplete the river water’s “availability for use,” Kansas
says that the 10-year period effectively frees Colorado from the obliga-
tion to compensate Kansas for years (within the 10-year period) when
overpumping may have made water “unavailable” for Kansas’ use.
Kansas also notes that the parties and the Master have heretofore used
a 1-year measuring period in calculating past damages. The Court is
not persuaded by these arguments. The Compact’s literal words are
not determinative. Its language essentially forbids offsetting debits
with “credits,” but it does not define the length of time over which
a “credit” is measured. Any measurement period inevitably averages
interim period flows just as it overlooks interim period lack of water
“availability.” At the same time, practical considerations favor the
Master’s approach. The Master found that Model results over measure-
ment periods less than 10 years are highly inaccurate, but that the
Model functioned with acceptable accuracy over longer periods of time.
Moreover, Kansas is unlikely to suffer serious harm through use of a
10-year period because Colorado has developed a river water replace-
ment plan to minimize depletions. Assuming, as Kansas argues, that
the Compact’s framers expected annual measurement with no carryover
from year to year, those framers were likely unaware of the modern
difficulties of complex computer modeling and, in any event, would have
preferred accurate measurement. The fact that both parties earlier
agreed to use annual measurement is not determinative here because

543US1 Unit: $$U6 [03-06-07 12:48:44] PAGES PGT: OPIN
89 Cite as: 543 U. S. 86 (2004)
Syllabus
that stipulation was made before the Master fully examined the Model’s
accuracy. Pp. 99–103.
4. Also overruled is Kansas’ exception to the Special Master’s recom-
mendation that the final amounts of water replacement plan credits to
be applied toward Colorado’s Compact obligations be determined by the
Colorado Water Court and appeals therefrom. Kansas argues that the
Water Court is a state court, that Colorado cannot be its own judge in
a dispute with a sister State, West Virginia ex rel. Dyer v. Sims, 341
U. S. 22, 28, and that this Court must pass on every essential question,
e. g., Oklahoma v. New Mexico, 501 U. S. 221, 241. Kansas’ objection
founders, however, upon additional language in the Master’s full recom-
mendation—and his attendant analysis—making clear that all replace-
ment credits are subject to Kansas’ right to seek relief under this
Court’s original jurisdiction; that Colorado’s replacement plan rules af-
fect the rights, not only of Kansas water users, but also of Colorado
senior water users; that both groups have similar litigation incentives;
and that permitting the Colorado Water Court initially to consider chal-
lenges to credit allocations will help prevent inconsistent determina-
tions. The full recommendation will help avoid potential conflict and
adequately preserves Kansas’ rights to contest any adverse Water Court
determination. Pp. 103–104.
5. Kansas’ exception to the Special Master’s finding that Colorado
complied with the Compact between 1997 and 1999 is overruled. Kan-
sas’ objection rests on its claim that the Master cannot use an accounting
period longer than one year. This Court has already found against
Kansas on that matter. P. 104.
6. Kansas’ exception to the Special Master’s refusal to make recom-
mendations on 15 disputed issues is overruled. As the Master found,
there are good reasons not to decide these issues immediately. The
issues in the second category, which involves challenges to the accuracy
of the figures used to determine whether Colorado depleted the river
between 1997 and 1999, are mostly moot. Moreover, the passage of
time will produce more accurate resolution of disputes in the first and
third categories (and any future second-category disputes). Pp. 104–
106.
Kansas’ exceptions overruled; Special Master’s recommendations ac-
cepted; and case recommitted to Special Master.
Breyer, J., delivered the opinion of the Court, in which Rehnquist,
C. J., and O’Connor, Scalia, Kennedy, Souter, and Ginsburg, JJ.,
joined, and in which Stevens and Thomas, JJ., joined except for Part
II. Thomas, J., filed an opinion concurring in part and concurring in the
judgment, post, p. 106. Stevens, J., filed an opinion concurring in part
and dissenting in part, post, p. 107.

543US1 Unit: $$U6 [03-06-07 12:48:44] PAGES PGT: OPIN
90 KANSAS v. COLORADO
Opinion of the Court
John B. Draper, Special Assistant Attorney General of
Kansas, argued the cause for plaintiff. With him on the
briefs were Phill Kline, Attorney General, Eric Rucker,
Chief Deputy Attorney General, David Davies, Deputy At-
torney General, Harry Kennedy, Assistant Attorney Gen-
eral, Leland E. Rolfs, Special Assistant Attorney General,
and Andrew S. Montgomery.
David W. Robbins, Special Assistant Attorney General of
Colorado, argued the cause for defendant. With him on the
brief were Ken Salazar, Attorney General, Carol D. Angel,
First Assistant Attorney General, and Dennis M. Montgom-
ery, Special Assistant Attorney General.
James A. Feldman argued the cause for the United States.
With him on the brief were former Solicitor General Olson,
Assistant Attorney General Sansonetti, Deputy Solicitor
General Kneedler, Jeffrey P. Minear, and Patricia Weiss.
Justice Breyer delivered the opinion of the Court.
We again consider a long-running water dispute between
Colorado and Kansas. The water is that of the Arkansas
River, once proudly called the “Nile of America.” The river
originates high in the Rocky Mountains. It runs eastward
through Colorado, Kansas, Oklahoma, and Arkansas, before
joining the Mississippi near the town of Arkansas Post. For
decades, Kansas and Colorado disagreed about the division
of its upper waters. See Kansas v. Colorado, 206 U. S. 46
(1907); Colorado v. Kansas, 320 U. S. 383 (1943). In 1949,
they entered into an interstate compact. See Arkansas
River Compact (Compact), 63 Stat. 145 (agreeing to “[e]qui-
tably divide and apportion” the waters (internal quotation
marks omitted)). But the disagreements have persisted.
Present proceedings began in 1985, when Kansas charged
that Colorado had violated the Compact. Kansas pointed
out that Compact Art. IV–D says:
“This Compact is not intended to impede or prevent fu-
ture beneficial development of the Arkansas River basin

543US1 Unit: $$U6 [03-06-07 12:48:44] PAGES PGT: OPIN
91 Cite as: 543 U. S. 86 (2004)
Opinion of the Court
in Colorado and Kansas by Federal or State agencies, by
private enterprise, or by combinations thereof, which
may involve construction of dams, reservoir, and other
works for the purposes of water utilization and control,
as well as the improved or prolonged functioning of ex-
isting works: Provided, that the waters of the Arkansas
River, as defined in Article III, shall not be materially
depleted in usable quantity or availability for use to
the water users in Colorado and Kansas under this
Compact by such future development or construction.”
Id., at 147 (emphasis added and internal quotation
marks omitted).
Kansas submitted that Colorado “development,” in particular
increases in ground water consumption through new and ex-
isting irrigation wells, had “materially depleted” the water
otherwise available “for use” by Kansas’ “water users.”
Our appointed Special Master agreed, recommending that we
find that Colorado had unlawfully depleted the river in vio-
lation of Art. IV–D. 2 First Report of Special Master
336 (hereinafter Report). We accepted the Special Master’s
recommendations and remanded the case for remedies.
Kansas v. Colorado, 514 U. S. 673, 694 (1995) (Kansas I).
The Special Master set forth proposed remedies in his Sec-
ond and Third Reports. He said that Colorado had over-
depleted more than 400,000 acre-feet of usable river flow
from 1950 through 1994. Second Report 112. He recom-
mended that Colorado pay Kansas monetary damages to
make up for the depletions. Third Report 119. He divided
losses into six categories, calculating damages somewhat dif-
ferently in each category. See id., at 120. And he recom-
mended that Kansas be awarded prejudgment interest on
damages reflecting losses incurred from 1969 through 1994.
Id., at 107. We subsequently adopted the Special Master’s
recommendations with one exception; we held prejudgment
interest would run from 1985 (not 1969). Kansas v. Colo-

543US1 Unit: $$U6 [03-06-07 12:48:44] PAGES PGT: OPIN
92 KANSAS v. COLORADO
Opinion of the Court
rado, 533 U. S. 1, 15–16 (2001) (Kansas III). See infra, at
95–97. And we remanded the case. 533 U. S., at 20.
The Master has now filed a Fourth Report setting forth
his resolution of certain remaining issues. Kansas takes ex-
ception to several of the Fourth Report’s recommendations.
We overrule Kansas’ exceptions and adopt all of the Special
Master’s recommendations.
I
Kansas asked the Special Master to recommend that we
appoint a River Master with authority to decide (within clear
error limits) various technical disputes related to decree en-
forcement. See Texas v. New Mexico, 482 U. S. 124, 134
(1987) (appointing a River Master to “make the calculations
provided for in [a] decree” concerning the Pecos River).
The Special Master rejected Kansas’ request, recommending
instead that “the Court retain continuing jurisdiction in this
case for a limited period of time” to permit the Special Mas-
ter himself to resolve any lingering issues (subject, of course,
to this Court’s review). Fourth Report 135. Kansas here
renews its request for appointment of a River Master.
We recognize that this Court has previously appointed a
River Master to help resolve water-related disputes among
States. Texas v. New Mexico, supra, at 134–135; New Jer-
sey v. New York, 347 U. S. 995, 1002–1004 (1954). But it has
done so only twice before, each time on recommendation of
the Special Master, always as a discretionary matter, and
only because it was convinced that such an appointment
would significantly aid resolution of further disputes. See
Vermont v. New York, 417 U. S. 270, 275 (1974) (per curiam)
(“[I]t is a rare case” where we will install a River Master).
We are not convinced that such an appointment is appro-
priate here.
For one thing, further disputes in this case, while techni-
cal, may well require discretionary, policy-oriented decision-
making directly and importantly related to the underlying
legal issues. In this respect, potential disputes in this case

543US1 Unit: $$U6 [03-06-07 12:48:44] PAGES PGT: OPIN
93 Cite as: 543 U. S. 86 (2004)
Opinion of the Court
differ at least in degree from those that we have asked River
Masters to resolve. Implementation of the Pecos River De-
cree, for example, involved application of a largely noncon-
troversial mathematical curve. The curve correlates inflows
at various New Mexico River locations with expected out-
flows so that engineers can estimate, for any given inflow,
the amount of water that is required to be available for
Texas’ use. See Texas v. New Mexico, 462 U. S. 554, 572–573
(1983); see also Texas v. New Mexico, 446 U. S. 540 (1980)
(per curiam). Lingering disputes between Texas and New
Mexico, we thought, would involve not the curve’s shape but
whether officials had properly measured the flows. 482
U. S., at 134–135. Although these disputes might call for a
“degree of judgment,” they would often prove capable of me-
chanical resolution and would usually involve marginal calcu-
lation adjustments. Id., at 134; see id., at 135–136; Fourth
Report 128 (The Pecos River Master “does not adjudicate
the kinds of disputes” potentially at issue here).
Administration of the decree in this case, by contrast, will
involve not a simple curve but a highly complex computer
model, the Hydrologic-Institutional Model (H–I Model or
Model). The H–I Model seeks to determine just what the
precise water flows into Kansas would have been had Colo-
rado not allowed increased consumption of ground water
after 1949. See 2 First Report 231. Modeling disputes—
and there have been many—involve not just measurement
inputs, but basic assumptions underlying the Model. See,
e. g., Kansas I, supra, at 685–687; 2 First Report 237–240;
Fourth Report 123–124. Their resolution may well call for
highly judgmental decisionmaking about matters that (com-
pared to the Pecos) are more importantly related to the par-
ties’ basic legal claims. See id., at 128.
Moreover, the need for a River Master is diminished by
the fact that the parties may find it possible to resolve future
technical disputes through arbitration. The interstate com-
pact itself creates an Arkansas River Compact Administra-

543US1 Unit: $$U6 [03-06-07 12:48:44] PAGES PGT: OPIN
94 KANSAS v. COLORADO
Opinion of the Court
tion (Administration) empowered to resolve differences aris-
ing under the Compact. Art. VIII, 63 Stat. 149. The
Administration consists of three representatives from each
State and a representative of the United States acting as
chair. Art. VIII–C. Each State has one vote; the United
States has no vote. Art. VIII–D. In case of an equally di-
vided vote, the Administration (with the consent of both
States) may refer a matter for resolution to the “Representa-
tive of the United States or other arbitrator or arbitrators.”
Ibid. (internal quotation marks omitted). The arbitrator’s
determinations are binding. Ibid.
At oral argument, counsel for Kansas suggested a willing-
ness to use arbitration, noting that “in the one case [he was]
aware of, Kansas’ suggestion of doing an arbitration was re-
jected by Colorado.” Tr. of Oral Arg. 17. Colorado’s coun-
sel responded that Colorado had proposed “that binding arbi-
tration be used and has committed itself to participate in
that.” Id., at 26; see also Reply Brief for Colorado Opposing
Exceptions 15. These comments suggest that neither party
opposes arbitration, and indeed that Colorado would accept
it. Nor have the parties expressed any opposition to the use
of other less formal means to resolve disputes, such as joint
consultation with experts, negotiation, and informal media-
tion. See, e. g., Kansas v. Nebraska, 538 U. S. 720 (2003)
(Kansas, Colorado, and Nebraska resolved Republican River
dispute by settlement and stipulation); Fourth Report 134
(discussing ongoing “joint efforts” and “cooperation” among
the States to resolve lingering disputes over the waters of
the Republican River).
The Special Master recommended both binding arbitration
and these other less formal methods as alternatives, while
opposing appointment of a River Master and observing that
such an appointment would “simply” make it “easier to con-
tinue this litigation.” Id., at 136.
For all of these reasons, we deny Kansas’ River Master
request.

543US1 Unit: $$U6 [03-06-07 12:48:44] PAGES PGT: OPIN
95 Cite as: 543 U. S. 86 (2004)
Opinion of the Court
II
Kansas takes exception to the Special Master ’s pre-
judgment interest calculation. The calculation and the ob-
jection grow out of the special history of this litigation.
After we initially remanded this case for remedial deter-
minations, see Kansas I, 514 U. S. 673, the Special Master
found that Colorado’s unlawful water depletion had harmed
Kansas beginning in 1950 and that Colorado must pay mone-
tary damages reflecting that harm. Kansas asked the Spe-
cial Master to award prejudgment interest on those damages
incurred through 1994. Colorado replied that the Com-
pact—like the common law—did not foresee interest pay-
ments in respect to unliquidated claims, particularly where,
as here, damages were highly speculative. And even with
the best of good will, said Colorado, it still could not have
known prior to the filing of the complaint (in 1985) how much
it owed Kansas. See Third Report 92–94; Kansas III, 533
U. S., at 11–13; Brief for Defendant in Kansas III, O. T. 2000,
No. 105, Orig., pp. 28–32.
The Special Master resolved the argument by deciding to
calculate prejudgment interest on the basis of what he called
“ ‘considerations of fairness.’ ” Third Report 97 (quoting
Board of Comm’rs of Jackson Cty. v. United States, 308 U. S.
343, 352 (1939)). In a kind of Solomonic compromise, he di-
vided the prejudgment period into three temporal subcate-
gories: (1) an Early Period, the period from 1950, when Colo-
rado’s unlawful water depletion began, through 1968, when
Colorado should first have known about it; (2) a Middle Pe-
riod, the period from 1969 through 1984; and (3) a Late Pe-
riod, the period from 1985, when Kansas filed its complaint,
through 1994, the last year for which evidence was available
at the time of the trial on damages. He adjusted damages
from all three periods (Early, Middle, and Late) for inflation.
But he awarded additional prejudgment interest, reflecting
Kansas’ loss of use of the money, “only from 1969 to the date
of judgment.” Third Report 107. Both Kansas and Colo-

543US1 Unit: $$U6 [03-06-07 12:48:44] PAGES PGT: OPIN
96 KANSAS v. COLORADO
Opinion of the Court
rado interpreted his order as awarding interest only on Mid-
dle and Late Damages (1969–1994), not on Early Damages
(1950–1968). Kansas III, Exception and Brief for Plaintiff
Kansas 9; App. to Fourth Report 12–13. The resulting total
damages award, including prejudgment interest, came to $38
million. Ibid.
On appeal to this Court, Colorado attacked the award of
any prejudgment interest, while Kansas called for full pre-
judgment interest. We accepted the Special Master’s equi-
table approach. We were unable to conclude that Colorado
should have known that prejudgment interest would “auto-
matically” be imposed “in order to achieve full compen-
sation.” 533 U. S., at 14. But, we added, Colorado did
believe (or should have believed) that we would assess “ ‘con-
siderations of fairness’ ” in order to achieve a just and equita-
ble remedy. Ibid. Hence “the Special Master acted prop-
erly . . . in only awarding as much prejudgment interest as
was required by a balancing of the equities.” Ibid.
The Special Master, we found, properly refused to “award
prejudgment interest for any years before either party was
aware of the excessive pumping in Colorado.” Id., at 15.
We then applied our own “considerations of fairness” and
concluded that “prejudgment interest should begin to ac-
crue,” not as of 1969 (the Special Master’s date), but as
of 1985. Id., at 14–15. We wrote in an accompanying
footnote:
“Justice O’Connor, Justice Scalia, and Justice
Thomas would not allow any prejudgment interest. . . .
Justice Kennedy and The Chief Justice are of the
opinion that prejudgment interest should run from the
date of the filing of the complaint [1985]. Justice Sou-
ter, Justice Ginsburg, Justice Breyer, and [Jus-
tice Stevens] . . . agree with the Special Master’s view
that interest should run from the time when Colorado
knew or should have known that it was violating the
Compact [1969]. In order to produce a majority for a

543US1 Unit: $$U6 [03-06-07 12:48:44] PAGES PGT: OPIN
97 Cite as: 543 U. S. 86 (2004)
Opinion of the Court
judgment, the four Justices who agree with the Special
Master have voted to endorse the position expressed in
the text.” Id., at 15, n. 5.
On remand, the Special Master, seeking to remain faithful
to our determination, calculated prejudgment interest from
1985 onward, and calculated that interest on (post-1984) Late
Damages alone, i. e., completely exempting both Early Dam-
ages and Middle Damages from prejudgment interest. Kan-
sas now objects to this last-mentioned limitation; it chal-
lenges the sum upon which post-1984 interest runs. Kansas
says the Special Master should have calculated prejudgment
interest (from 1985) on all damages, i. e., on Early Damages,
Middle Damages, and Late Damages alike. After all, says
Kansas, “[p]rejudgment interest serves to compensate for
the loss of use of money due as damages . . . thereby achiev-
ing full compensation for the injury those damages are in-
tended to redress,” West Virginia v. United States, 479 U. S.
305, 310–311, n. 2 (1987) (citing Comment, Prejudgment In-
terest: Survey and Suggestion, 77 Nw. U. L. Rev. 192 (1982)).
See Exceptions and Brief for Plaintiff 29. Kansas lost the
“use of ” all the “money due as damages,” i. e., Early and
Middle Damages as well (which were “due” at least by 1985).
Why then, asks Kansas, calculate post-1984 interest on only
some of the damages then due?
Kansas’ argument would make good sense in an ordinary
case. But the question here is not about the ordinary
case, but rather what the Kansas III paragraph we quoted
above means in context. And the Kansas III context is a
special one.
For one thing, like the Special Master, we did not seek to
provide compensation for all lost investment opportunities;
rather, we sought to weigh the equities. For another, it was
apparent that the Special Master’s earlier determination in-
volved both a decision about when to begin to calculate inter-
est (1969) and what to calculate that interest upon (Middle
Damages and Late Damages only). Brief for Plaintiff in

543US1 Unit: $$U6 [03-06-07 12:48:44] PAGES PGT: OPIN
98 KANSAS v. COLORADO
Opinion of the Court
Kansas v. Colorado, O. T. 2000, No. 105, Orig., pp. 9, 25, n. 8.
All damages incurred before his selected date were totally
exempt from interest. Kansas contested the when by ar-
guing that we should award interest for the entire period.
Kansas also contested the what by arguing that, even accept-
ing the Special Master’s preferred date, interest should run
on Early Damages as well as Middle and Late Damages.
See id., at 25, n. 8 (“Even if a defendant’s good-faith igno-
rance of its breach were a valid reason to deny prejudgment
interest, it would not justify the Special Master’s recommen-
dation to deny Kansas compensation for its loss of use of
money [reflecting Early Damages] after 1968”).
In overruling Kansas’ exception and sustaining Colorado’s
exception, we said nothing about the Special Master’s total
exemption of Early Damages. 533 U. S., at 14. Thus, we
changed the when (from 1969 to 1985) in Kansas III, but
(despite Kansas’ argument) we did not change the methodol-
ogy for calculating the what. In context, our silence fairly
implies acceptance, not rejection, of the Special Master’s
underlying methodology. Moving the date forward thus
meant moving the exemption period forward as well. And
that methodology now yields a post-1985 interest calculation
based upon Late Damages only.
This view of our prior opinion is reinforced by the result-
ing numbers. The Special Master’s original 1969 date (and
methodology) produced a total damages award to Kansas,
including prejudgment interest, of about $38 million (in 1998
dollars). Were we to accept Kansas’ argument (and calcu-
late post-1984 interest on all damages), the final damages
award would be roughly $53 million (in 2002 dollars). App.
to Fourth Report 12. We cannot reconcile that numerical
result with our acceptance in Kansas III of the Special Mas-
ter’s equitable approach and with our own equitable determi-
nation. That determination implied a modest adjustment of
the $38 million award in Colorado’s favor, not, as Kansas now

543US1 Unit: $$U6 [03-06-07 12:48:44] PAGES PGT: OPIN
99 Cite as: 543 U. S. 86 (2004)
Opinion of the Court
seeks, an adjustment of the award in its own favor. App. to
Fourth Report 12.
Consequently, we overrule Kansas’ objection.
III
Kansas and Colorado have agreed to use a computer
model, the H–I Model, to measure Colorado’s future Compact
compliance. This highly complex set of computer programs
determines whether Colorado’s post-1949 wells deplete the
river of usable water that the Compact makes available for
Kansas. It does so by trying to account for almost every
Arkansas-River-connected drop of water that arrives in,
stays in, or leaves Colorado, whether by way of rain, snow,
high mountain streams, well pumping of underground water,
evaporation, canal seepage, transmountain imports, reser-
voir storage, or otherwise. 2 First Report 233–235. With
all “switches” turned on, the Model predicts how much river
water will leave Colorado for Kansas during a given month.
Id., at 234–235. To obtain a figure representing an unlawful
depletion (or lawful accretion) under the Compact, the Model
subtracts from this figure (the actual flow) a number repre-
senting a hypothetical prediction of how much water would
have flowed into Kansas had Colorado not dug and operated
post-1949 wells. The Model obtains this prediction through
a computer rerun with the Model’s “post-1949 well” switch
turned off. Ibid. The final figure is then adjusted to reflect
depletions to usable, as opposed to total, flow. App. to Sec-
ond Report 37.
Not surprisingly, the Model’s ability to calculate depletions
has proved highly controversial, leading to many Model mod-
ifications during this litigation. See, e. g., 2 First Report
236–240 (describing Colorado’s objections to the original
Model). The Special Master has recommended use of the
Model together with a 10-year measurement period to deter-
mine the amounts of any future depletions. Fourth Report
139. That is to say, a determination of whether Colorado

543US1 Unit: $$U6 [03-06-07 12:48:44] PAGES PGT: OPIN
100 KANSAS v. COLORADO
Opinion of the Court
owes Kansas water in Year 11 will be made by taking the
Model’s total result for Years 1–10, for year 12 by the Model’s
total result for Years 2–11, and so forth. Id., at 117; App.
to Fourth Report 86, Exh. 14. Kansas takes exception to
the 10-year measurement period.
Kansas seeks a measurement period of one year. In sup-
port, Kansas points to Compact Art. V–E(5), 63 Stat. 148,
which says that there “shall be no allowance or accumulation
of credits or debits for or against either State.” (Internal
quotation marks omitted.) Kansas argues that a 10-year pe-
riod averages out oversupply and undersupply during the
interim years, with the likely effect of awarding Colorado a
“credit” in dry years for oversupply in wet years. Kansas
adds that Art. IV–D, 63 Stat. 147, forbids Colorado to deplete
the river water’s “availability for use.” (Internal quotation
marks omitted.) Kansas says that the 10-year measurement
period in effect frees Colorado from the obligation to com-
pensate Kansas for years (within the 10-year period) when
overpumping may have made water “unavailable” for Kan-
sas’ use. (Internal quotation marks omitted.) Kansas also
notes that the parties and the Special Master have used a
1-year measuring period in this litigation for purposes of cal-
culating past damages. See Exceptions and Brief for Plain-
tiff 37–40, 43–44.
Like the Special Master, we are not persuaded by Kansas’
arguments. The literal words of the Compact are not deter-
minative. The Compact’s language essentially forbids off-
setting debits with “credits,” but it does not define the
length of time over which a “credit” is measured. Any pe-
riod of measurement inevitably averages interim period
flows just as it overlooks interim period lack of water “avail-
ability.” Thus annual measurement offsets and overlooks
seasonal differences; seasonal measurement, monthly differ-
ences; monthly measurement, weekly differences, and so
forth.

543US1 Unit: $$U6 [03-06-07 12:48:44] PAGES PGT: OPIN
101 Cite as: 543 U. S. 86 (2004)
Opinion of the Court
At the same time, practical considerations favor the Spe-
cial Master’s measurement approach. Model results over
measurement periods of less than 10 years are highly inaccu-
rate. The Special Master found, for example, that the cur-
rent iteration of the Model, if used to project river diversions
(including well pumping) during a single year, produces fig-
ures that overpredict actual diversions in some years and
underpredict them in others by as much as 22%. Fourth
Report 111. Similar inaccuracies plague the Model’s projec-
tion of actual river flows. Id., at 112. If projected diver-
sions and flows deviate substantially in this way from actual
measured diversions and flows, 1-year estimates of final
depletions to usable flow—the figure that determines Kan-
sas’ damages—cannot be accurate. Id., at 115 (“I find that
the H–I model is not sufficiently accurate on a short-term
basis to be used to determine compact compliance on a
monthly or annual basis”). But measured over long periods
of time, say, the full 540 months between 1950 and 1994, the
Model’s predicted and observed diversions “matched almost
perfectly.” Id., at 114. For this reason, the Master con-
cluded that “[o]nly by using longer term averages do the
model simulations more closely match historic data.” Id.,
at 115. Thus, the 10-year measurement period is needed to
ensure Model accuracy.
Nor is Kansas likely to suffer serious harm through use of
a 10-year measuring period. That is because Colorado has
developed a water replacement program designed to mini-
mize depletions. See Amended Rules and Regulations Gov-
erning the Diversion and Use of Tributary Ground Water in
the Arkansas River Basin (Use Rules), App. to Fourth Re-
port 36, Exh. 6; Fourth Report 8–13. The program protects
both Kansas water users and senior Colorado users by insist-
ing that Colorado users with junior rights (and in particular
those who obtain water from post-1949 wells) replace the
river water that they use. They must either (1) buy replace-

543US1 Unit: $$U6 [03-06-07 12:48:44] PAGES PGT: OPIN
102 KANSAS v. COLORADO
Opinion of the Court
ment water, say, from the Rockies’ western slope or (2) buy
land irrigated under pre-1949 water rights and remove it
from cultivation. Id., at 10–13. In practice, junior users
belong to one of three associations that conduct these trans-
actions, reporting the details monthly to the Colorado State
Engineer ’s Office, and receiving replacement “credits,”
which they divide among their members. Id., at 13.
Were the replacement program and the H–I Model both
to work perfectly, the Model’s net depletion figure, whether
determined each month, each year, or each decade, would be
zero (that is, there would be no difference between actual
flow and what the flow would have been under precompact
conditions). Of course, perfection is impossible; and Kansas
claims certain defects in the Use Rules. See id., at 27. But
operation of the Rules should help to diminish the real
amount of any depletion, thereby limiting any negative effect
that a 10-year measurement period might have upon Kansas.
See id., at 119–120; see also id., at 32. The 1997–1999 re-
sults, showing essentially no aggregate depletion, suggest
the water replacement program will have this effect. Ibid.
Kansas argues that the Compact’s framers expected an-
nual measurement. And they quote a Colorado Commis-
sioner as recognizing that there would be “ ‘no carry-over
from year to year,’ ” see Exceptions and Brief for Plaintiff
39 (quoting Joint Exhibit 3, pp. 14–84). Assuming, argu-
endo, that the framers opposed such carryover, they were
likely unaware of the modern difficulties of complex com-
puter modeling. And we believe that those framers, in any
event, would have preferred accurate measurement. After
all, a “credit” for surplus water that rests upon inaccurate
measurement is not really a credit at all.
Kansas also points out that earlier in this litigation both
parties agreed to the use of annual measurement for pur-
poses of calculating past damages. The parties made that
stipulation, however, before the Special Master fully exam-

543US1 Unit: $$U6 [03-06-07 12:48:44] PAGES PGT: OPIN
103 Cite as: 543 U. S. 86 (2004)
Opinion of the Court
ined the Model’s accuracy. In any event, their previous
agreements do not govern this determination.
We overrule Kansas’ exception.
IV
As we just mentioned, measuring the depletion caused by
Colorado’s post-1949 wells involves taking account of Colora-
do’s water replacement program, which credits Colorado
with non-Arkansas water pumped into the Arkansas and
with Arkansas water not used because farmers have re-
moved from cultivation lands previously irrigated under
pre-1949 water rights. The Special Master has recom-
mended that “the final amounts of Replacement Plan credits
to be applied toward Colorado’s Compact obligations shall be
the amounts determined by the Colorado Water Court, and
any appeals therefrom.” Fourth Report 138, ¶ 9. Kansas
takes exception to this recommendation.
Kansas points out that the Colorado Water Court is a state
court. It says that a “ ‘State cannot be its own ultimate
judge in a controversy with a sister State,’ ” Exceptions and
Brief for Plaintiff 45–46 (quoting West Virginia ex rel. Dyer
v. Sims, 341 U. S. 22, 28 (1951)), and that this Court must
“ ‘pass upon every question essential’ ” to resolving the dis-
pute, Exceptions and Brief for Plaintiff 46 (quoting Okla-
homa v. New Mexico, 501 U. S. 221, 241 (1991), in turn quot-
ing Kentucky v. Indiana, 281 U. S. 163, 176–177 (1930)).
Kansas believes that the Special Master’s recommendation
violates these well-established principles.
Kansas’ objection founders, however, upon additional lan-
guage in the Master’s full recommendation. The recommen-
dation adds:
“This is not to say, however, that the Colorado Water
Courts are empowered to make a final determination on
any matter essential to compact compliance at the State-
line, or that Colorado’s reliance on such Water Court ac-
tions will necessarily satisfy its compact obligations. . . .

543US1 Unit: $$U6 [03-06-07 12:48:44] PAGES PGT: OPIN
104 KANSAS v. COLORADO
Opinion of the Court
All replacement credits, no matter how determined, are
subject to the right of Kansas to seek relief under the
Court’s original jurisdiction [as set forth in] Section
VIII.” Fourth Report 138–139, ¶ 9.
In the cross-referenced Section VIII, the Special Master
makes clear that Colorado’s replacement plan rules affect the
rights, not only of Kansas water users, but also of Colorado
senior water users; that both groups of water users have
similar litigation incentives; and that permitting the Colo-
rado Water Court initially to consider challenges to credit
allocations will help prevent inconsistent determinations.
Id., at 93–95.
In our view, the Special Master’s full recommendation
will help to avoid the potential conflict he mentioned. It
also adequately preserves Kansas’ rights to contest any
adverse Water Court determination. We overrule Kansas’
exception.
V
The Special Master found that Colorado complied with the
Compact for the period 1997–1999. Kansas takes exception
on the ground that the Special Master used a measurement
period “greater than one year.” Exceptions and Brief for
Plaintiff 47. Kansas concedes that its objection rests upon
its claim that the Special Master cannot use “an accounting
period longer than one year.” Ibid. Having found against
Kansas on that matter, supra, at 100–103, we must overrule
this exception.
VI
At the end of its brief, Kansas lists 15 disputed issues that
the Special Master has not yet decided. It groups them into
three categories:
1. “Disputed H–I Model Calibration Issues” (“[c]alibra-
tion procedures, parameters and criteria,” “[c]anal ca-
pacities,” “altered diversion records,” “statistical out-
liers,” “Sisson-Stubbs water right” representation);

543US1 Unit: $$U6 [03-06-07 12:48:44] PAGES PGT: OPIN
105 Cite as: 543 U. S. 86 (2004)
Opinion of the Court
2. “Disputed 1997–1999 Accounting Issues” (“[d]ry-up
acreage,” “Sisson-Stubbs credit,” “winter water book-
overs” credit);
3. “Disputed Future Compliance Issues” (“[d]ry-up
acreage monitoring,” “[d]ry-up credits” and external
source “return flow obligations,” credit beyond “precom-
pact uses,” “[s]pecial waters monitoring,” winter water
release credit timing, “[o]ffset [a]ccount” accounting pro-
cedures, “consumptive use credit and return flow obliga-
tions”). Exceptions and Brief for Plaintiff 48–49.
Kansas takes exception to the Special Master’s refusal to
make recommendations on these issues now. It points out
that we cannot leave unanswered important questions “ ‘es-
sential’ ” to our “ ‘determination of a controversy’ ” between
the States. Id., at 49 (quoting Oklahoma v. New Mexico,
supra, at 241). And Kansas asks us to require the Special
Master to decide them.
As the Special Master found, however, there are good rea-
sons not to decide these issues immediately. There is no
need to resolve most of the issues in the second category.
They involve challenges to the accuracy of the figures used
to determine whether Colorado depleted the river between
1997 and 1999. The Special Master concluded that Colorado
was in compliance during 1997–1999, in the process relying
upon Kansas’ own figures. Fourth Report 30–31. As far as
we can tell from the briefs, these issues are mostly moot.
The passage of time will produce more accurate resolution
of disputes in the first and third categories (and any of those
in the second that arise again in the future). The parties
will learn more about matters relevant to their resolution,
namely, the H–I Model’s strengths, weaknesses, and methods
of monitoring and measurement. That is why the Special
Master recommended that we retain jurisdiction over this
case and permit him to take up lingering issues at a future

543US1 Unit: $$U6 [03-06-07 12:48:44] PAGES PGT: OPIN
106 KANSAS v. COLORADO
Opinion of Thomas, J.
date. Id., at 135–136, 139. We accept that recommendation
and overrule Kansas’ objection.
The Special Master also recommended that experts for the
two parties confer, e. g., id., at 91–92, and he expressed the
hope that expert discussion, negotiation, and, if necessary,
binding arbitration would lead to resolution of any remaining
disputes. Id., at 135–136. We express that hope as well.
VII
For these reasons, we overrule all Kansas’ exceptions.
We accept the Special Master’s recommendations and recom-
mit the case to the Special Master for preparation of a decree
consistent with this opinion.
It is so ordered.
Justice Thomas, concurring in part and concurring in
the judgment.
I join the Court’s opinion with the exception of Part II,
which concerns whether prejudgment interest should begin
accruing in 1985 only on damages thereafter arising (post-
1985 damages) or also on damages then owing (pre-1985 dam-
ages). As Justice O’Connor explained in Kansas v. Colo-
rado, 533 U. S. 1 (2001) (Kansas III), neither the Arkansas
River Compact itself nor the common law at the time of the
compact’s formation allows Kansas to recover any prejudg-
ment interest. See id., at 21–25 (opinion, joined by Scalia
and Thomas, JJ., dissenting in part). The Court did not
adopt that view in Kansas III, but neither did it adopt the
now-familiar rule that Kansas should be made whole with
an award of prejudgment interest spanning the duration of
Colorado’s breach, from 1950 to the present. See, e. g., Mil-
waukee v. Cement Div., National Gypsum Co., 515 U. S. 189,
195–196, and n. 7 (1995); West Virginia v. United States, 479
U. S. 305, 310–311, n. 2 (1987).
The Court instead crafted what it viewed as an equitable
compromise, designed to apply sui generis to these States

543US1 Unit: $$U6 [03-06-07 12:48:44] PAGES PGT: OPIN
107 Cite as: 543 U. S. 86 (2004)
Opinion of Stevens, J.
and their particular dispute, in which prejudgment interest
would begin to accrue in 1985. See Kansas III, supra, at
14–16. Its compromise left open the door to the present liti-
gation, for saying when prejudgment interest began to ac-
crue did not answer on what the interest was accruing. The
Court therefore must again decide what is too little or too
much compensation for Colorado’s depletion of the Arkansas.
That weighing is as unnecessary now as it was before. Kan-
sas is not entitled to prejudgment interest, and its exception
seeks only to compound the windfall it received in Kan-
sas III. I therefore agree with the Court that Kansas’ sec-
ond exception to the Special Master’s Report should be
overruled.
Justice Stevens, concurring in part and dissenting in
part.
With the exception of Part II, I join the Court’s opinion.
In dissenting from Part II, I adhere to the views that we
expressed in Kansas v. Colorado, 533 U. S. 1, 13–16 (2001)
(Kansas III).1 In Kansas III, in a compromise that was
required in order to issue a judgment of the Court, we ac-
cepted the views of The Chief Justice and Justice Ken-
nedy that prejudgment interest should run from 1985, the
date the complaint was filed. Ibid. Like today’s majority,
I adhere to the judgment reflecting that compromise. Un-
like the majority, however, I believe that prejudgment inter-
est should run, starting in 1985, on all damages that accrued
after Colorado “knew or should have known that it was vio-
lating” its compact with Kansas—i. e., from 1969. Id., at 15,
n. 5. Such a result best respects the reasoning behind our
conclusion in Kansas III that prejudgment interest is an ap-
propriate component of the award of damages.
In Kansas III, recognizing that a monetary award does
not fully compensate for an injury unless it includes an inter-
1 Kansas III was predated by Kansas v. Colorado, 514 U. S. 673 (1995),
and Kansas v. Colorado, 522 U. S. 1073 (1998).

543US1 Unit: $$U6 [03-06-07 12:48:44] PAGES PGT: OPIN
108 KANSAS v. COLORADO
Opinion of Stevens, J.
est component, we affirmed the Special Master’s determina-
tion that the unliquidated nature of Kansas’ claim did not
by itself bar an award of prejudgment interest. Id., at 14.
Nevertheless, equitable concerns persuaded a majority of the
Court to overrule the Special Master’s determination that
prejudgment interest should begin to run in 1969, the date
on which Colorado first knew, or should have known, that it
was violating the Arkansas River Compact. Although we
did not explicitly discuss the point in our opinion, we also
agreed with the Special Master’s decision to exclude from
the principal amount on which interest would run any dam-
ages that had accrued prior to 1969.2
The methodology that led to that conclusion was the Mas-
ter’s appraisal of the equities—in his judgment, interest
should not be imposed on the portion of the damages award
that was attributable to relatively innocent conduct that oc-
curred before 1969. See Report 106–107 (“The general lack
of knowledge in the early years about pumping in Colorado
and its impacts along the Arkansas River served to pro-
tect Kansas during the liability phase of the case against a
claim of laches. The same degree of fairness, I believe,
should now relieve Colorado of the obligation to pay full
interest rates on damages from depletions during 1950–68
2 Kansas had objected to the Master’s refusal to award interest on all
damages accruing after 1950. See Brief for Plaintiff in Kansas III, O. T.
2000, No. 105, Orig., p. 25, n. 8. Although we did not discuss Kansas’
exception to the Special Master’s determination regarding the total
amount of damages on which interest would run, we overruled the objec-
tion and thereby approved the Master’s selection of the period after 1968
as the appropriate measure of damages on which interest should be paid.
See Kansas III, 533 U. S. 1, 14 (2001); see also Third Report of Special
Master 106–107 (hereinafter Report) (explaining that Colorado’s aware-
ness of its breach was central to the determination that interest should
run on post-1968 damages). Today, the Court explains why it would be
inequitable to give Kansas the relief that would be the equivalent of sus-
taining an objection that we overruled three years ago, but does not ex-
plain why we should not accept the Special Master’s original determina-
tion that all post-1968 damages should bear interest.

543US1 Unit: $$U6 [03-06-07 12:48:44] PAGES PGT: OPIN
109 Cite as: 543 U. S. 86 (2004)
Opinion of Stevens, J.
period . . .”). But the Master did find that Colorado was
required to pay interest on damages that occurred between
1969 and 1985. See ibid.; see also Brief for United States
in Opposition to the Exceptions of Kansas and Colorado in
Kansas III, O. T. 2000, No. 105, Orig., p. 27 (“For the period
from 1969 to the date of judgment, the Master recommended
that Kansas be awarded prejudgment interest”). Our opin-
ion did not reject that portion of his judgment, and did not
contain any suggestion that he had erred in that respect.
See 533 U. S., at 12, 14. The happenstance that we selected,
as a compromise, the date the complaint was filed as the date
on which interest should begin to accrue should have no
bearing on the principal amount of damages that gave rise
to the interest obligation. Thus, I believe that the Special
Master’s Fourth Report erred in its conclusion that we
meant to limit the principal amount of damages to those that
occurred after 1985.
Surely if this were an ordinary tort case involving a sin-
gle harm-causing event, an award of prejudgment interest
would apply to the entire damages recovery, not just to the
portion that resulted from events occurring after interest
began to accrue. See Funkhouser v. J. B. Preston Co., 290
U. S. 163, 168 (1933). Indeed, were this an ordinary case,
we would no doubt have awarded prejudgment interest in
the entire amount that Kansas requested in Kansas III.
This, however, is a unique case in which unusual equities
necessitated a compromise designed to resolve a dispute be-
tween two States. Thus, I agree with the majority that the
Special Master was correct in rejecting Kansas’ argument
that the principal on which interest should run should be
“the nominal damages occurring from 1950 through 1984.”
App. to Fourth Report 15.
However, the fact that Kansas’ request represents too
large a measure of damages does not convince me that Kan-
sas is entitled to no interest for damages prior to 1985.
Nothing in our Kansas III opinion compels such a result.

543US1 Unit: $$U6 [03-06-07 12:48:44] PAGES PGT: OPIN
110 KANSAS v. COLORADO
Opinion of Stevens, J.
In my view, the proper measure of damages on which Colo-
rado owes Kansas interest is the entire amount attributable
to the time that Colorado knew, or should have known, that
it was violating the compact. That date is 1969—the date
that the Special Master initially chose and that we implicitly
accepted as appropriate in Kansas III. Choosing 1969 as
the initial date for the damages period not only has the bene-
fit of respecting our affirmation of the methodology in the
Special Master’s Third Report, it also results in a total dam-
ages sum that is less than the $38 million the Special Master
originally awarded.
Accordingly, I would sustain Kansas’ second objection to
the Special Master’s Report, but only insofar as it applies to
post-1968 damages.