LAKE ONTARIO REGULATION UNDER TRANSPOSED CLIMATES

JOURNAL OF THE AMERICAN WATER RESOURCES ASSOCIATIONVOL. 33, NO.1 AMERICANWATER RESOURCES ASSOCIATION FEBRUARY 1997

LAKE ONTARIO REGULATION UNDER TRANSPOSED CLIMATES'

Deborah H. Lee, Thomas E. Croley II, and Frank H. Quinn2

ABSTRACT: The implications of Lake Ontario regulation undertransposed climates with changed means and variability are pre-sented for seasonal and annual time scales. The current regulationplan is evaluated with climates other than the climate for which itwas developed and tested. This provides insight into potential con-flicts and management issues, development of regulation criteriafor extreme conditions, and potential modification of the regulationplan. Transposed climates from the southeastern and south centralcontinental United States are applied to thermodynamic models ofthe Great Lakes and hydrologic models of their watersheds; theseclimates provide four alternative scenarios of water supplies toLake Ontario. The scenarios are analyzed with reference to the pre-sent Great Lakes climate. The responses of the Lake Ontario regu-lation plan to the transposed climate scenarios illustrate severalkey issues: (1) historical water supplies should no longer be the solebasis for testing and developing lake regulation plans; (2) duringextreme supply conditions, none of the regulation criteria can bemet simultaneously, priority of interests may change, and newinterests may need to be considered, potentially requiring substan-tial revision to the Boundary Waters Treaty of 1909; (3) revised reg-ulation criteria should be based on ecosystem health andsocio-economic benefits for a wider spectrum of interests and not onfrequencies and ranges of levels and flows of the historical climate;and (4) operational management of the lake should be improvedunder the present climate, and under any future climate with morevariability, through the use of improved water supply forecasts andmonitoring of current hydrologic conditions.(KEY TERMS: Lake Ontario; St. Lawrence River; regulation; cli-mate change; climate variability; surface water hydrology; waterpolicy.)

INTRODUCTION

Assessment of climate-related water resourceimpacts (both mean changes and changes in variabili-ty) on Lake Ontario is of interest because of thesocial, environmental, and economic significance of

this natural resource. More than eight million peoplelive within the lake's drainage basin, with the majori-ty located along the shoreline in large metropolitancommunities (Thorp and Allardice, 1994). The lakeprovides water for residential, commercial, and insti-tutional facilities, agriculture, industry, electric powergeneration (in-stream hydroelectric, fossil fuel, andnuclear plants), navigation, sanitation, recreation,and habitat for wildlife, fish, waterfowl, and otheraquatic life. The sport fishing industry alone demon-strates the environmental and economic importanceof the lake; in 1985, the estimated economic impact ofthe industry was $141 million U.S. and $87 millionCanadian (Michigan Sea Grant, 1990).

Lake Ontario is particularly sensitive to climatebecause it is the most-downstream lake in the chainof the live Laurentian Great Lakes; its level and out-flow are an integration of the climate conditions overeach lake's drainage basin. Previous climate changeimpact studies (Smith and Tirpak, 1989; Croley, 1990,1993; Hartmann, 1990; Lee and Quinn, 1992), usinggeneral circulation model (GCM) outputs of double-carbon-dioxide (2xCO2) climate change scenarios,show significant expected hydrologic impacts. Theseimpacts culminate in projected decreases of LakeOntario average annual outflows of 40 percent and,depending on how outflows are regulated, reducedaverage annual levels of about 1.4 m (Lee and Quinn,1992).

These previous studies applied GCM-generatedcorrections to historical Great Lakes meteorologicaldata to assess impacts of mean changes in climate,but changes in variability could not be assessed. Forthis study, we use transpositions of actual climates

1Paper No. 96099 of the Journal of the American Water Resources Association (formerly Water Resources Bulletin). Discussions are openuntil August 1, 1997.

2Respectively Hydrologist, Research Hydrologist, and Physical Sciences Division Head, Great Lakes Environmental Research Laboratory,2205 Commonwealth Blvd., Ann Arbor, Michigan 48105-1593.

JOURNAL OF THE AMERICAN WATER RESOURCES ASSOCIATION 55 JAWRA

Lee, Croley, and Quinn

from the southeastern and south central continentalUnited States because they incorporate naturalchanges in variability within existing climates, aswell as mean changes. The Great Lakes Environmen-tal Research Laboratory (GLERL) modeled the poten-tial hydrology in the Laurentian Great Lakes underfour transposed historical climate scenarios, rangingfrom 6 south and 0 west to 10 south and iF west ofthe Great Lakes (Croley et al., 1996). Lengthy (atleast 40 years) and detailed records of daily weatherat about 2,000 sites were used to represent physicallyplausible and coherent scenarios of alternative cli-mates. Such data sets incorporate reasonable valuesand frequencies of extreme events, ensuring represen-tation and transposition of desired temporal and spa-tial variability over the Great Lakes. GLERLestimated the Great Lakes hydrology of each trans-posed climate by applying their system of hydrologicmodels to these data and comparing outputs to a basecase derived from historical data. The scenarios repre-sent analog climates that could occur under globalwarming, as suggested by recent 2xCO2 GCM simula-tions (Houghton et al., 1992).

Changes in climate variability have previouslybeen unexplored, yet they could pose significant prob-lems for water resource managers. We focus on theresponse of Lake Ontario regulation to the transposedclimates because the International Joint Commission(IJC) is considering a review of its criteria governingthe management of the levels and outflows of thisinternational boundary water (Canadian and Ameri-can). The criteria ensure that the lake is managed tobest meet the different (and often conflicting) objec-tives of diverse interest groups. The regulation plan,which was designed to satisfy the criteria given thehistorical water supply sequence, is also being consid-ered for modification or replacement by a new plan.Evaluation of the current regulation plan, given cli-mates other than that for which it was developed andtested, can provide valuable insight into potentialconflicts and management issues, into development ofregulation criteria for extreme conditions, and intomodification of the regulation plan.

We present here the changes in Lake Ontariohydrology and the response of the regulation planunder the transposed climates, on seasonal (monthly)and annual time scales. The sections that followdescribe the present climate and dynamics of LakeOntario, the institutional and technical framework ofthe lake's regulation, methodologies used to transposeclimates, and estimates of their hydrological impacts.Lastly, ramifications for regulation that are associat-ed with these impacts are presented.

LAKE ONTARIO DYNAMICS AND CLIMATE

The Lake Ontario basin, shown (transposed) in Fig-ure 1, contains an area of approximately 80,000 km2,19,000 km2 of which is water surface. UpstreamGreat Lakes flows enter Lake Ontario through theNiagara River and the Welland Canal, a diversionbypassing Niagara Falls for navigation andhydropower. There is also a small flow diverted fromLake Erie to Lake Ontario via the New York StateBarge Canal System. Outflows from Lake Ontarioare regulated by structures 169 km downstream fromthe lake's outlet in the St. Lawrence River betweenMassena, New York, and Cornwall, Ontario. Belowthe control structures, the drainage from the OttawaRiver basin enters the St. Lawrence River near Mon-treal, Quebec. Their confluence is Lake St. Louis, thewater levels of which are also considered in LakeOntario regulation. From Lake St. Louis, the waterflows through the St. Lawrence River to the Gulf ofSt. Lawrence and the ocean. Lake Ontario is deep,with an average depth of 86 m and a maximum depthof 244 m.

The behavior of Lake Ontario is governed by itshuge storage of water and energy within the lake andits basin, and by flows from the upstream GreatLakes. Because of the large storage capacity of theupstream lakes and the limited discharge capacity ofthe Niagara River, inflows to Lake Ontario from theupstream lakes vary slowly over time (several years)and exhibit considerable persistence subsequent tochanges in climate. In comparison, watershed andlake surface supplies to the lake (watershed runoffand lake precipitation less lake evaporation) can behighly variable, exhibiting persistence on a time scaleof several months. Both sources of water exhibit defi-nite seasonal trends. Lake Ontario annual precipita-tion is about 93 cm (expressed over the lake), annualrunoff to the lake is about 169 cm, and annual lakeevaporation is about 65 cm. In comparison, the annu-al inflow from the upstream lakes is 1,037 cm. Thechanges in Lake Ontario levels (changes in lake stor-age) are determined by these inflows and by the regu-lated outflows. Since regulation began, the averageseasonal range in lake levels has been about 0.6 m,and the inter-annual range of levels is. about 0.9 m.Prior to regulation, the inter-annual range in levelswas about 1.4 m.

JAWRA 56 JOURNAL OF THE AMERICAN WATER RESOURCES ASSOCIATION

Lake Ontario Regulation Under Transposed Climates

+ 400 N

+ 350 N

+ 300 N

65° W

LAKE ONTARIO REGULATION

Lake Ontario regulation encompasses an institu-tional decision-making framework, regulation objec-tives (i.e., criteria), and a computational responsemodel (i.e., the regulation plan) designed to meet thecriteria with the 1860-1954 supply sequence. Descrip-tions of these elements and their interactions follow.Additionally, problems of the regulation plan exhibit-ed during extreme water supply sequences, and modi-fications to address them, are presented.

Institutional Framework

The International Joint Commission (IJC) estab-lished the International St. Lawrence River Boardof Control in 1952 (IJC, 1952) to ensure that waterdischarges from Lake Ontario and through the

International Rapids (St. Lawrence River) complywith the IJC's Orders of Approval. The St. LawrenceSeaway and Power Project altered river hydraulicsand required regulation of Lake Ontario to maximizebenefits from navigation and hydropower and to pro-tect riparian interests. Lake Ontario regulation beganin 1960 upon completion of the control works built aspart of the Seaway project.

Reporting to the Board are the Regulation Repre-sentatives, the Operations Advisory Group, the Work-ing Committee, and the River Gauging Committee.The two Regulation Representatives monitor andevaluate hydrologic conditions, conduct weekly regu-lation plan calculations, provide forecasts of weeklyoutflows and levels (.updated monthly), advise theBoard on regulation strategies, and relay directionfrom the Board of Control. The Operations AdvisoryGroup (OAG) meets at least weekly to consider hydro-logic information and regulation plan discharge, pro-vided by the Regulation Representatives, and to

Scenario

S x 100 W

Scenario

Scenario

100 S x 110 W

100 S x 50 W

Figure 1. Transposed Climate Shifts to the Lake Ontario Basin.

JOURNAL OF THE AMERICAN WATER RESOURCES ASSOCIATION 57 JAWRA

Lee, Croley, and Quinn

recommend an outflow for the coming period to theBoard. The OAG may propose a flow other than thatspecified by the regulation plan if it deems that condi-tions warrant. If both the OAG and Regulation Repre-sentatives recommend the same outflow, and it isconsistent with the direction given by the Board, thenit is implemented. If not, the matter is sent to theBoard for decision. If the Board cannot reach a con-sensus, the decision is made by the IJC. This has hap-pened during periods of extreme high or low waterlevels and outflows CD. Pay, Great Lakes St. LawrenceRegulation Office, Environment Canada, personalcommunication, 1996). Once an outflow decision ismade, it is implemented by the agencies responsiblefor the operation of the control works.

The Working Committee is chaired by the Regula-tion Representatives with members selected from theOAG and the Board. It reviews proposed regulationplans and other technical matters related to lake reg-ulation. The River Gauging Committee is also chairedby the Regulation Representatives and is responsiblefor ensuring that water levels and flows are reportedaccurately.

Regulation Criteria

There are 12 Criteria in the Orders of Approval forthe regulation of Lake Ontario. The Criteria weredeveloped in reference to "pre-project" conditions —the levels and flows that would have occurred in thepast (1860-1954) with the actual water supplies toLake Ontario adjusted for a diversion of 88 m3s-1 outof the Great Lakes basin at Chicago and for a diver-sion of 142 m3s1 into the Great Lakes basin from theHudson Bay watershed. The Criteria are describedelsewhere (IJC, 1952, 1956; U.S. Army Corps of Engi-neers, 1991) and briefly summarized in Table 1. Theyintend benefits to three specific interests (navigation,hydropower, and riparian) without unacceptableadverse effects to any of them. Seaway navigationinterests desire lake levels above low water chartdatum to utilize full available draft, but do not desirehigh Lake Ontario outflows that result in high St.Lawrence River velocities and hazardous cross cur-rents. Navigation interests at Montreal Harbor, tomake effective use of their deeper draft vessels, desirehigher than average outflows such that levels areabout 1 m above chart datum. Hydropower interestsdesire uniform high minimum flows to maximize theirfirm power capacity on a long-term basis. Ripariansupstream of the control works desire reductions in therange of water levels and frequency of extremes.Downstream riparians desire reductions in the rangeand frequency of extreme outflows, often in conflictwith the desires of upstream riparians.

The Orders of Approval (IJC, 1952, 1956) also statethat the project shall be operated to "safeguard as faras possible the rights of all interests affected by lev-els" on the upper St. Lawrence River and LakeOntario. The Orders of Approval also specify that theoperation of the project should not conflict with usesfor domestic and sanitary purposes.

Regulation Plan

Lake Ontario's regulation plan, Plan 1958-D (Inter-national St. Lawrence River Board of Control, 1963),strives to satisfy the Criteria, with the exception ofCriterion (k). The plan was implemented in 1963, suc-ceeding earlier versions. The plan is operated weeklyby the Regulation Representatives to recommendLake Ontario outflows to the OAG. Maximum andminimum flow limitations in Plan 1958-D restrictLake Ontario outflows to an annual range of 5,320m3s1 to 8,780 m3s1. One additional limitationrestricts weekly outflow changes to 570 m3s'. Theplan satisfies the criteria [except for Criterion (k)land other requirements of the Orders of Approval for1860-1954. Criterion (k) was included to provide forsupplies more extreme than those of 1860-1954. TheIJC or the Board has authorized deviations from theplan to best meet Criterion (k) in the 1960s inresponse to low supplies and during the 1970s, 1980s,and 1990s in response to high supplies.

Problems and Modifications

Sanderson and Wong (1987), Hartmann (1990),Lee and Quinn (1992), and Lee et al. (1994) haveshown that the plan lacks computational robustnessduring simulations with water supply sequences moreextreme than those used in the plan's development.With persistent low supply sequences, regulated out-flows exceed water supplies to the lake, and lake lev-els fall below the lower limit [Criterion (j)]. Withpersistent high supply sequences, water suppliesexceed regulated outflows, and lake levels exceed theupper limit [Criterion (h)]. The plan fails numericallyor returns unreasonable results under these circum-stances.

Lee et al. (1994) improved the plan's robustness forlow supplies by using a pre-project Lake Ontario dis-charge relationship and waiving minimum flow limi-tations when levels fall below 74.15 m. This helpslimit impacts to no worse than under pre-project con-ditions, in the spirit of the Criteria. Lee et al. (1994)modified the plan for high supply conditions by con-sidering Board operations during 1974-1989, a period

JAWRA 58 JOURNAL OF THE AMERICAN WATER RESOURCES ASSOCIATION

Lake Ontario Regulation Under Transposed Climates

TABLE 1. Orders of Approval Lake Ontario Regulation Criteria.

Criteria Description Purpose

A Montreal Harbor levels shall not be below pro-projectconditions April 1-December 15.

Protect navigation interests in Montreal Harbor.

B Lake outflows shall be as large as feasible with stable icecover December 15-March 31.

Benefit winter hydropower operation.

C Lake outflow shall not exceed pre-project conditions duringspring break-up of ice.

Protect Montreal Harbor and downstream interests

D Lake outflow shall not exceed pre-project conditions duringOttawa River spring flooding,

Protect Lake St. Louis, Montreal Harbor, anddownstream riparian interests.

E Lake minimum outflow shall ensure the maximum dependableflow for hydropower.

Protect hydropower interests.

F Lake maximum outflow shall be low as possible. Minimize channel excavation.

G Lake extreme levels shall be reduced from those experienced. Protect Lake Ontario riparian interests.

H Monthly mean lake levels shall not exceed 75.37 ma with thesupplies of 18604954b,

Protect Lake Ontario riparian interests.

I Monthly mean lake levels exceeding 75.O7ma shall be lessfrequent than pro-project.

Protect Lake Ontario riparian interests.

J April 1 lake level shall not be lower than 74.l5ma; monthlymean levels shall not be lower than 74.15 m5 April 1-November 30.

Ensure adequate levels for the navigation season.

K When water supplies exceed those of 18601954b, provide allpossible relief to riparians; when water supplies are less thanthose, provide all possible relief to navigation and powerinterests.

Provide relief to riparian owners upstream anddownstream.

Provide relief to riparian owners upstream and downstream.stream.

* Protect navigation and riparians downstream at leastas much as pre-project conditions.

Ensure past protection of navigation and riparianinterests downstream.

alnternational Great Lakes Datum 1985.bAdjusted for the Chicago and Hudson Bay watershed diversions.CSupplementary Order.

of high supplies and lake levels. Deviations from theplan are reflected in the recorded levels and flows forthis period. Because the complex plan deviation deci-sions of the Board cannot presently be incorporated,modifications were made to closely match simulatedand recorded monthly levels and flows for 1974-1989.Between 74.15 m and 76.35 m, Plan 1958-D outflowspecifications are used subject to modified maximumoutflow limitations. Above 76.35 m, plan outflows areaugmented to reduce storage above this level. Lee etal. (1994) describes plan modifications in detail.

The modified plan monthly average levels agreedwith the 1974-1989 record; root mean square errorswere 0.06 m to 0.15 m. The unmodified plan yieldedroot mean square errors of 0.52 m to 0.67 m. Similar-ly, modified plan flows had root mean square errors of

167 m3s-1 to 634 m3s, while unmodified plan errorswere 320 m3s-1 to 807 m3s' [see Figure 5 and Table 4of Lee et al. (1994)]. We consider the modified planacceptable for simulation; it is not used operationally.

CLIMATE TRANSPOSITION

GCMs predict that continuing increases in atmo-spheric trace gas concentrations will result in warmerconditions over the Great Lakes, comparable to south-ern climates, and drier conditions, comparable towestern climates. We relocated climatic zones in thesoutheastern and south central United States to theGreat Lakes basin to sample climatic differences in

JOURNAL OF THE AMERICAN WATER RESOURCES ASSOCIATION 59 JAWRA

Lee, Croley, and Quinn

fluctuations over time. The major advantage of thisapproach is that the transposed data represent actualclimates. Temporal variability, the frequency andmagnitude of extremes, and spatial relationships areobviously realistic.

This technique takes advantage of the dense net-work of observing sites in the U.S. and Canada. Bychoosing latitudinal and longitudinal shifts, we canmatch meteorology predicted by GCMs or that is moreextreme. There are approximately 2,000 climate (tem-perature and precipitation) stations in the areas ofinterest. For a 40-year period of daily measurements,this corresponds to about 30 million values for eachvariable. The development of such detailed scenarios,by means other than climate transposition, faces amonumental problem in ensuring that the large datasets have physically plausible temporal and spatialvariability.

Daily maximum and minimum air temperatures,precipitation, and snowfall were obtained for the 43-year period of 1948-1990 from the dense array of sta-tions in the National Weather Service's cooperativeobserver network. A subset also has daily wind speed,humidity, and cloud cover, located generally atNational Weather Service offices and airport observ-ing stations. We considered four separate climaticregimes based on published 2xCO2 GCM ranges.Figure 1 depicts the Lake Ontario basin shift for eachclimate scenario. The first two climate regimes corre-spond roughly to the upper range of GCM predictionsfor temperature for the Great Lakes basin (Houghtonet al., 1992). Scenario 1 (from 6S and 10GW) corre-sponds to warmer temperatures and mixed precipita-tion changes over the Great Lakes basin. The LakeOntario climate would be 6.2CC warmer and 26 per-cent wetter than the present, similar to the presentclimate of the lower Ohio River valley. Scenario 2 (6'Sx O'W) corresponds to warmer temperatures andincreased precipitation over the Great Lakes basin.The Lake Ontario climate would be 6.5'C warmer and18 percent wetter than the present. The next two cli-mate regimes went beyond the range of current GCMpredictions to study response to a major climaticshock. Scenario 3 (lU'S x 11'W) corresponds to hightemperatures and mixed precipitation changes overthe Great Lakes basin. The Lake Ontario climatewould be 9.3'C warmer and 49 percent wetter thanthe present. Scenario 4 (10'S x 5'W) corresponds tohigh temperatures and increased precipitation overthe Great Lakes basin. The Lake Ontario climatewould be 9.7'C warmer and 33 percent wetter thanthe present. Detailed descriptions of the scenarios' cli-matology are available elsewhere (Croley et al., 1996).

GLERL relocated all meteorological station datafor each climate scenario, checked and corrected it byremoving obvious outliers, and Thiessen-weighted to

obtain areal averages over the 121 watersheds andseven lake surfaces of the Great Lakes basin for alldays of record (1948-1992). They also reduced all his-torical data (base case) within the Great Lakes (1900-1990). GLERL developed, calibrated, and verifiedconceptual model-based techniques for simulating hy-drological processes in the Laurentian Great Lakes(including Georgian Bay and Lake St. Clair as sepa-rate entities). GLERL integrated the models into asystem to estimate net basin supplies to the lakes,lake levels, whole-lake heat storage, and water andenergy balances for forecasts and for assessment ofimpacts associated with climate change (Croley et al.,1996). These include models for rainfall-runoff (121daily watershed models), over-lake precipitation (adaily estimation model), and one-dimensional (depth)lake thermodynamics (seven daily models for lakesurface flux, thermal structure, and heat storage).The models were assessed partially by comparingmodeled time series to historical time series forrunoff, lake evaporation, water surface temperature,heat balances, and net basin supplies.

GLERL first applied the system of hydrologicalmodels to the (untransposed) historical daily meteoro-logical time series of air temperature, precipitation,wind speed, humidity, and cloud cover within theGreat Lakes basin. They modeled the 40 full yearsbetween 1951 and 1990 by arbitrarily using Janu-ary 1, 1990 modeled values as initial conditions (forsoil moisture, snow pack, ground water storage, andlake heat storage and surface temperature). GLERLrepeated the simulation with end conditions used asinitial conditions until there was no change; thisbecame the estimate of the "base case" hydrology.(This required only one iteration for all sub-basinsand lakes.) GLERL then used these initial conditionsto estimate the Great Lakes hydrology of each trans-posed climate by directly applying the system ofhydrological models to all transposed climate 40-yeardaily meteorological time series. The impacts wereestimated by comparing the outputs for each trans-posed climate to the base case. The section that fol-lows gives annual average (1951-1990 period) andseasonal estimates for both means and standard devi-ations for selected Lake Ontario variables.

HYDROLOGICAL IMPACT ASSESSMENT

Basin Meteorology

The overland air temperatures for all transposedscenarios are higher throughout the annual cyclethan for the base case (Table 2). The differences are

JAWRA 60 JOURNAL OF THE AMERICAN WATER RESOURCES ASSOCIATION

Lake Ontario Regulation Under Transposed Climates

TABLE 2. Average Annual Lake Ontario Climate and Transposed Climate Differences.

Variable BASEScenario (Absolute or Relative Differences)

6Sx1O°W 6SxO'W 1OSx11W 1OSx5W

Overland Air Temperaturea 7.2CC 6.2CC 6.5CC 9.3CC 9.7CC

Overland Precipitationb 934mm 26% 18% 49% 33%

Snow Water Equivalentb 15.7 mm -92% -96% -99% -99%

Soil Moistureb 20.8 mm -28% -29% -28% -34%

Groundwaterb 11 mm -27% .27% -21% -23%

Total Basin Moistureb 61 mm -37% -42% -36% -48%

Overland Evapotranspirationb 473 mm 52% 48% 88% 87%

Runoff as an Overland Depthb 461mm -1% -14% 9% -22%

Over-lake Air Temperaturea 7.8CC 5.8CC 5.3CC 9.9CC 9.2CC

Lake Heatb 8.4 1017 cal 68% 62% 157% 144%

Ice Coverb 0.9% -100% -100% -100% -100%

Water Surface Temperaturea 8.6CC 6.rC 5.6'C 10.2CC 9.2'C

Lake Evaporation Depthb 645 mm 42% 21% 68% 66%

Over-lake Precipitationa 934 mm 242mm 166 mm 461 mm 312 mm

Runoff as Overwater Deptha 1,701 mm -19mm -238mm 151 mm -374 mm

Lake Evaporation Depth3 645 mm 272 mm 138 mm 438 mm 423 mm

Net Basin Supply3 1,990 mm -49 mm -210 mm 174 mm -485 mm

Lake Inflowa,d 10,949 mm -6,119 mm 60 mm -6,446 mm -102 mm

Total Supply3 12,939 mm -6,168 mm -150 mm -6,272 mm -587 mm

Lake Ontario Lavel'c 74.81 m -1.48 m -0.03 m -1.51 m 0.03 m

Lake Ontario Rangeb 0.56 m 32% 7% 64% 38%

Lake Outflowb 7,848 m3s' -48% -1% -48% -4%

Lake St.Louis Level 21.43 m -1.33 m -0.07 m -1.35 m -0.16 m

Lake St.Louis Rangeb 0.91 m 0% 16% 11% 30%

3Changes from Base Case are given as absolute differences.bChanges from Base Case are given as relatiue differences.CLavels are referenced to IGLD 85.dEquivalent to Lake Erie outflow.

greatest for the southernmost scenarios (3 and 4). Thedifference is smallest during the late summer andlargest during the late winter for all transposed sce-narios (see Figure 2 for scenario 2). Changes in annu-al variability of air temperature are remarkablysmall. Table 3 shows the averageannual steady-statestandard deviation of air temperature, depicting thevariability from year to year in the 40-year period. Itappears artificial and is the result of truncation ofannual air temperatures to the nearest 0. 1C beforethe standard deviation was calculated. The smallchange in variability did not warrant recalculation.Variability also changes little throughout the seasonalcycle for all scenarios. The seasonal patterns remain,with more variability associated with cooler tempera-tures.

Overland precipitation shows much more variabili-ty than air temperature among the scenarios. Precipi-tation is greater for all scenarios (Table 2). Thenorthernmost scenarios (1 and 2) increase precipita-tion less than do the southernmost scenarios (3 and4). Throughout the annual cycle, the Lake Ontariobasin shows a shift in seasonal precipitation earlier

for most scenarios; however, this shift is not apparentfor scenario 2 in Figure 2. Changes in annual vari-ability of precipitation are more pronounced than forair temperature (Table 3). Generally, the western sce-narios (1, 3, and 4), which are also the wettest, aremost variable (Table 3). The variation is greater than100 percent for the most southwest scenario (3). Sea-sonally, the variability of precipitation is highest inthe late summer and early fall, similar to the basecase pattern but more pronounced. The pattern ofvariability in these scenarios is consistent with theknown spatial distribution of precipitation variability.For instance, Griffiths and Driscoll (1982, Figure7.22) show that the current Great Lakes region expe-riences a relative minimum of precipitation variabili-ty, which increases to the west and south.

Basin Runoff

Basin runoff in Table 2 decreases for scenarios 1, 2,and 4 as a net effect of increased overland air temper-atures, increased evapotran spiration, and decreased

JOURNAL OF THE AMERICAN WATER RESOURCES ASSOCIATION 61 JAWRA

Lee, Croley, and Quinn

total basin moisture [Snow water equivalent, soilmoisture, ground water, and surface storage (notshown in Table 2)]. Scenario 3 shows an increasebecause moisture reductions are offset by precipita-tion increases. Runoff decreases the most for the east-ernmost scenarios (2 and 4).

Precipitation (mm)120 _

Evapotranspiration (mm over land)110 __________

0 :Basin Moisture (mm)-I

storage in the basin than do the eastemmost (2 and4).

Seasonal peak runoff shifts to earlier in the yearfor all four scenarios. This results from the loss ofsnow moisture storage; more winter runoff than thebase case contributes to the runoff shift. It is also duein part to seasonal shifts in evapotranspiration. Forthe westernmost scenarios (1 and 3), the bulk of theannual evapotranspiration and the peak evapotran-spiration occur earlier in the seasonal cycle. In sce-narios 2 and 4, while evapotranspiration increases,the seasonal pattern is not significantly changed. Fig-ures 2 and 3 depict the typical seasonal behavior ofevapotranspiration and runoff for scenario 2.

Snow water variability is greatly decreased simplybecause snow water is greatly decreased toward itslower bound of zero (Table 3). Changes in variabilityof soil moisture, ground water, surface storage, andtotal basin moisture (-1.1 mm to +1.4 mm) are smallwith respect to both annual and seasonal values.Table 3 also shows more variable evapotranspirationfor the southernmost scenarios (3 and 4) correspond-ing to the greatest and most variable precipitation;(Tables 2 and 3). This results from evapotranspirationbeing moisture limited; more variability in the mois-ture supply translates into more variability in evapo-transpiration. A slight increase in runoff variabilityexists during the winter for all scenarios, correspond-ing to the absence of the snow pack; runoff thenvaries more with precipitation. The greatest consis-tent change, over all Great Lakes, in variability ofboth evapotranspiration and runoff, occurs on theLake Ontario basin; it is exposed to the most-easternpart of each scenario with its most variable precipita-tion.

Lake Evaporation

The increased air temperatures, consequent in alltransposed climates, significantly alter the heat bal-ance of the surface hydrology. The snow pack isalmost completely eliminated and evapotranspirationincreases significantly (Table 2). The greatest evapo-transpiration increases and snow pack decreasesoccur in the southernmost scenarios (3 and 4), whichincrease air temperatures the most and make mois-ture less available in the soil and ground water zones.Table 2 shows a general lowering of soil moisture andground water storage that results from either greaterincreases in evapotranspiration (scenarios 3 and 4) orlower increases in precipitation (scenarios 1 and 2).The total moisture storage is lowered for all scenariosby about one-third to one-half. The westernmost sce-narios (1 and 3) show a slightly lower loss of moisture

All scenarios produce significant increases in lakeevaporation. Table 2 shows increases of 21 percent to68 percent, with the largest increase corresponding tothe southernmost scenarios (3 and 4). The sum ofinsolation, reflection, net long wave exchange, sensi-ble heat exchange, and latent heat exchange increaseheat in storage 62 percent and 157 percent in Table 1,appearing as a constant amount higher throughoutthe seasonal cycle. This means that ice is eliminated(-100 percent for all four scenarios). The averagesteady-state increase in water surface temperatureranges from 5.6'C to 10.2CC with the largest resultingfrom the southernmost scenarios (3 and 4). Water sur-face temperatures peak earlier under the transposedclimates than under the base case.

The variability associated with lake heat balancevariables are summarized in Table 3. The stored heat

Air Temperature (°C)30

—10

130

0

Water Temperature (°C)

J F MA MJ JASON 0Figure 2. Seasonal Lake Ontario Average Meteorology

and Hydrology for Scenario 2 (6'S x OW)(Black — Base Case; Gray —Scenario 2).

JAWRA 62 JOURNAL OF THE AMERICAN WATER RESOURCES ASSOCIATION

Lake Ontario Regulation Under Transposed Climates

TABLE 3. Average Annual Lake Ontario Climate and Transposed Climate Variability.

Variable BASEScenario (Reintive Differences)

6°SxlO°W 6°SxO°W l0°Sxll°W lO°Sx5°W

Overland Air Temperature Std. Dev. 0.60°C 0% -17% 0% 0%

Overland Precipitation Std. Dcv. 89.6 mm 99% 71% 149% 96%

Snow Water Equivalent Std. Dev. 10.3 mm -83% -93% -97% -99%

Soil Moisture Std. Dcv. 2.3 mm 39% 30% 61% 35%

Groundwater Std. Dcv. 1.4 mm 64% 43% 93% 64%

Total Moisture Storage Std. Dcv. 11.2 mm .17% -35% -12% -32%

Overland Evapotranspiration Std. Dcv. 41.3 mm 114% 115% 183% 162%

Runoff as Overland Depth Std. Dcv. 55.2mm 115% 60% 151% 67%

Over-Lake Air Temperature Std. Dcv. 0.90°C -33% -33% -44% -33%

Lake Heat Std. Dcv. 0.80 1017 cal 50% 50% 25% 38%

Ice Cover Std. Dcv. 1.8% -100% -100% -100% -100%

Water Surface Temperature Std. Dcv. 1.00°C -40% -40% -60% -50%

Lake Evaporation Depth Std. Dcv. 60.2 mm 13% 15% -3% 9%

Over-Lake Precipitation Std. Dcv. 90 mm 99% 71% 149% 96%

Runoff as Over-water Depth Std. Dcv. 204mm 115% 60% 151% 67%

Net Basin Supply Std. Dcv. 304 mm 102% 57% 141% 74%

Lake Inflowa Std. Dcv. 1,016 mm 5% -3% -22% 50%

Total Supply Std. Dcv. 1,222 mm 6% 2% 5% 41%

Lake Ontario Level Std. Dcv. 0.09 m 433% 0% 411% 144%

Lake Ontario Range Std. Dcv. 0.14 m 71% 21% 107% 57%

Lake Outflow Std. Dcv. 741 m3r' 1% 2% -5% 42%

Lake St. Louis Level Std. Dcv. 0.23 m 25% 6% 18% 51%

Lake St. Louis Range Std. Dcv. 0.21 m -41% 14% -35% 28%

aEquivalent to Lake Erie outflow.

variability increases some for all scenarios andappears more uniformly spread across the seasonalcycle in every scenario than in the base case since theice pack is eliminated. Since the ice pack is gone, itsvariability is zero, implying a relative change of 100percent in Table 3.

The water surface temperature in Table 3 is lessvariable. Its variability also is spread more uniformlythroughout the season than for the base case, againreflecting ice pack elimination. Peak variability shiftsfrom summer to spring in all except scenario 2 andresults from a change in seasonal heat storage. Thereare significant changes in developments of verticalthermal profiles in the lake and the lake's mixingthroughout the annual cycle; biannual turnoverscease as the lake stays above 4°C throughout the year(Croley et al., 1996). As a result of the loss of ice coverand increased heat storage throughout the annualcycle, evaporation occurs under the transposed cli-mates in 95 percent to 97 percent of the annual cycle,compared with 84.2 percent of the annual cycle underthe base case.

Lake Water Balance

Precipitation, runoff, and lake evaporation sumalgebraically as the net basin supply to the lake andall are presented in Table 2 as over-lake depths andabsolute differences. Net basin supply is less than thebase case for all scenarios except scenario 3, where aprecipitation and runoff increase were sufficient tooffset evaporation losses. Net basin supply decreases2 percent to 24 percent for scenarios 1, 2, and 4 com-pared to the base case, and increases 9 percent forscenario 3.

There is a shift in the peak net basin suppliesunder all transposed climate scenarios from April toMarch (Figure 3). Also, the net basin supplies aregreater from December through March under thewesternmost scenarios (1 and 3) than under the basecase; for the easternmost scenarios (2 and 4), the netbasin supplies are greater than the base case fromJanuary through March, with the minor exception inAugust of scenario 2, which is only slightly higher.

The variability associated with the net basin sup-plies and its components is summarized in Table 3.The annual variability of net basin supplies increasesunder all scenarios, from 57 percent to 141 percent.

JOURNAL OF THE AMERICAN WATER RESOURCES ASSOCIATION 63 JAWRA

Lee, Croley, and Quinn

Seasonally, the variability is distributed across theseasonal cycle approximately as the base case, butlarger. There does appear to be generally greaterincreases in variability over the late summer-fall-win-ter-early spring, relative to the late spring-early sum-mer.

0

11

Net Basin Supply (mm over lake)360

I I I I I I I I

Lake Inflow (mm)

Total Supply (mm)

Lake inflow (Lake Erie outflow) and net basin sup-ply sum algebraically as the total supply; both arepresented in Table 2 as over-lake depths and absolutedifferences. Lake inflows are significantly less thanthe base case for scenarios 1 and 3 (-56 percent and-59 percent, respectively) but are only slightly differ-ent for scenarios 2 and 4 (+1 percent and -1 percent,respectively). Since these flows are over five timesgreater than net basin supplies, they largely deter-mine the total supplies to the lake. Total supplies

drop 48 percent for scenarios 1 and 3 and only 1 per-cent and 5 percent for scenarios 2 and 4, respectively,even though scenarios 2 and 4 showed the greatestdecreases in net basin supplies. Likewise, scenarios 1and 3 have the highest drop in total supplies eventhough they showed the least change in net basinsupplies. These results differ from earlier studies thatused GCMs (Croley, 1990, 1993) in that both lakeinflows and net basin supplies decreased concurrent-ly. Scenarios 1 and 3 results are most similar to theearlier studies, in decreased total supplies.

There is no shift in the May seasonal peak of lakeinflows under any scenario (Figure 3). For scenarios 2and 4, winter-spring outflows are greater than thebase case, and late summer-fall outflows are less. Theseasonal shift in net basin supplies, however, doesshift the peak total supply from April to March for allscenarios (Figure 3). Total supplies are greater thanthe base case during the winter-early spring monthsfor scenarios 2 and 4 (due to higher winter runoff onall Great Lake basins), and lower for the late spring-summer-fall months (due to increased evaporation onall lakes).

The annual variability of lake inflows in Table 3increases under scenarios 1 and 4 (5 percent and 50percent, respectively), and decreases under scenarios2 and 3 (-3 percent and -22 percent, respectively).Small increases in total supply variability occur forscenarios 1 through 3 (2 percent to 6 percent). A largeincrease in variability occurs under scenario 4 (41percent), primarily due to the large variability of thelake inflows. Seasonally, the variability of the totalsupplies is distributed across the seasonal cycleapproximately as it is in the base case, but larger.

LAKE REGULATION RESPONSETO TRANSPOSED SCENARIOS

The response of the Lake Ontario regulation planwas evaluated for each scenario by routing the basecase and transposed climate water supplies for 1951through .1990 through the modified channel routingand lake regulation model of the Great Lakes system(Lee et al., 1994). The water supplies were also routedthrough the unmodified model for comparison. Initialand other system conditions (diversions, consumptiveuses, and connecting channel weed retardation) werethose reported by Lee (1993). Because no ice formedin the climate transposition scenarios, ice retardationwas neglected for the four scenarios. Ottawa Riverflows were treated by applying long-term averagedifferences (between recorded flows in the St.Lawrence River at Montreal and from Lake Ontario)

JAWRA 64 JOURNAL OF THE AMERICAN WATER RESOURCES ASSOCIATION

300Runoff (mm over lake)

Evaporation (mm over lake)

1000

1300

950J F MA MJ JASON 0

Figure 3. Seasonal Lake Ontario Average Total Supplyand Components for Scenario 2 (6S x OW)

(Black — Base Case; Gray — Scenario 2).

Lake Ontario Regulation Under Transposed Climates

to the modeled Lake Ontario outflows. The complexityof, and the lack of a basin runoff model for, the regu-lated Ottawa River system precluded application ofthe transposed climate methodology there. The GreatLakes regulation and routing models were run tosteady-state conditions for each scenario and the basecase.

Scenarios 1 and 3

Average annual Lake Ontario levels are signifi-cantly less than the base case for the westernmostscenarios (1 and 3), -1.48 m and -1.51 m, respectively(Table 2), comparable to earlier studies (Lee andQuinn, 1992). These decreases are due to the 48 per-cent drops in the total supply, which are furtherreflected in lake outflow reductions of 48 percent.However, the average annual Lake Ontario range(defined as the annual maximum level minus theminimum) increased by 32 percent and 64 percent forscenarios 1 and 3, respectively (Table 2). Levels ofLake St. Louis (at Pointe Claire), downstream of LakeOntario's control structures, also decline by 1.33 mand 1.35 m for scenarios 1 and 3.

There is no shift in the seasonal peak levels for sce-nario 1 from that of the base case (Figure 4). Peak lev-els shift from June to May for scenario 3 due to amore rapid decline in total supplies during the latespring and summer than either the base case or sce-nario 1. Peak outflows for both scenarios shift fromMay to June (Figure 4), attributable to the way out-flows are specified entirely by the pre-project relation-ship because of the severe decline in lake levels. Forthe base case, Plan 1958-D releases higher outflows inthe spring months than those specified by the pre-project relationship in order to compress the range ofseasonal levels. There is a corresponding lag in peaklevels from April to May for Lake St. Louis (Figure 4).

Average annual Lake Ontario level variabilityincreased 433 percent and 411 percent, and LakeOntario range variability increased 71 percent and107 percent (Table 3). However, lake outflow variabili-ty changed only 1 percent and -5 percent for scenarios1 and 3. Recall that lake outflows are specified underthe modified regulation plan by the pre-project dis-charge relationship when lake levels fall below 74.15m. Because of the large storage capacity of the lakerelative to the pre-project discharge capacity, outflowsare less variable than lake levels. This response,along with the increased variability of the total sup-plies, results in the greatly increased variability oflake levels while outflows are only slightly more vari-able.

I I I I I I I I I I

Lake St. Louis Levels (m, IGLD85)

191 I I I I I I I I I I I

JFMAMJJASONDFigure4. Seasonal Lake Ontario Average Total Supply and

Regulation Responses for Scenario 1 (6S x lOW)(Black — Base Case; Gray — Scenario 1).

The regulation criteria (a-j) could not be satisfiedsimultaneously. For example, Criterion (j) (maintain-ing adequate navigation depths) was met by furtherreducing outflows in violation of the supplementaryorder (ensure past protection of navigation and ripari-ans downstream). Or conversely, Criterion (a) (main-taining adequate navigation depths in MontrealHarbor) can be met by further reducing Lake Ontariolevels in violation of Criterion (j).

Many of the criteria would be irrelevant under sce-narios 1 and 3. Increased water temperatures and noice negate the need for Criteria (b) and (c), whichaddress ice formation and break-up. The significantdecreases in Lakes Ontario and St. Louis levelsnegate the need for Criteria (d), (g), (h), and (i),addressing extreme high levels and outflows.

JOURNAL OF THE AMERICAN WATER RESOURCES ASSOCIATION 65 JAWRA

Lake Ontario Total Supplies (m3s1)

Lake Ontario Levels (m, IGLD85)

10000

2000

76 -

73

9000 -

3000

22 -

Lake Ontario Outflows (m3s1)

Lee, Croley, and Quinn

Criterion (k) would no longer be sufficient to pro-vide relief to riparians under high levels or to naviga-tion and hydropower under low levels. More decisionswould be referred to the IJC as the regulation becamemore contentious in the OAG and Board of Control.Undoubtedly, other interests would seek entrance tothe decision-making process and demand changed pri-orities. Such a precedent exists in the Board's recentattention to environmental and recreational boatinginterests. The decreases in total basin moisture wouldincrease demand for domestic, sanitary, and irrigationwater. Increasing consumptive use would furtherdecline lake levels and outflows. Domestic and sani-tary uses are given precedent over navigation andhydropower by the Boundary Waters Treaty. Consid-eration of other interests and severe climate changeimpacts would probably force decision making abovethe IJC into higher legislative and executive levels ofthe Canadian and U.S. governments, potentiallyrequiring a revised Boundary Waters Treaty.

Use of the unmodified regulation plan with the sce-narios, computationally "drains" the lake becausespecified outflows exceed total supplies. New regula-tions would be needed. We demonstrated one alterna-tive in our modifications; another might lower themaximum and minimum flow and lake level limita-tions and alter other parameters appropriately for achanged climate. Both alternatives permanently losewater storage. A third alternative would reduce out-flows to equal total supply, maintaining levels typicalof the present climate. The difficulty would be recog-nizing a climate shift before a permanent reduction inlake volume occurs. With any of these alternatives,downstream interests would suffer due to decreasedflows, but some riparian benefits would be preservedwith the third alternative.

Scenario 2

The regulation plan responds to scenario 2 verymuch like the base case for average annual water lev-els and outflows of Lakes Ontario and St. Louis, whilethe average annual ranges increase by 7 percent and16 percent, respectively (Table 2). Seasonal patternsof levels and outflows on both lakes are similar, butspring values are higher than the base case, and fallvalues are lower, reflecting increased spring runoffand fall evaporation. Level and outflow variability onboth lakes do not increase significantly although therange variability does (Table 3).

The response of the modified and unmodified regu-lation plans is shown in Figure 5 for scenario 2. Withthe unmodified plan, total supplies exceed the maxi-mum outflow limits beginning in 1973, causing levelsto rise beyond the upper limit [Criterion (h)]. With the

modified plan, outflow limits are relaxed underextreme supply sequences, and average monthly lev-els below 75.37 m are maintained. A similar responseis observed with the base case.

Scenario 2 illustrates that a total supply regimesimilar to the past 40 years may occur with an alteredclimate as far as Lake Ontario is concerned. One sig-nificant difference would be the absence of ice cover,permitting higher winter outflows than specified inthe plans and reducing seasonal peak outflows andlake levels. Criteria (b) and (c) could be ignored, andwinter and spring outflows increased, so long as Cri-terion (d) is not violated (lake outflow should notexceed pre-project outflows during Ottawa Riverspring flooding).

Under this scenario, the present plan could be mod-ified to better meet the regulation criteria. The exist-ing institutional structure and the Boundary WatersTreaty would not likely change. Minimum base flowsfor hydropower could increase and become seasonallymore uniform. A year-round ice-free season wouldbenefit navigation. Riparians and recreational boaterswould benefit from reduced seasonal fluctuations inlevels and flows. However, environmental interestswould most likely suffer from increased lake tempera-tures, reduced lake turnovers, and reduced ranges.

The outflow capacity could be increased throughadditional channel excavation [Criterion (f)] to helpmanage high supplies. However, the IJC Levels Refer-ence Study (Levels Reference Study Board, 1993) con-cluded the costs probably would be prohibitive.

Scenario 4

Scenario 4 is similar to scenario 2 in terms of itsimpact on annual average lake levels and outflows(Table 2). However, the average annual rangesincrease by 38 percent and 30 percent, respectivelyfor Lakes Ontario and St. Louis. Scenario 4 increasestotal supplies following 1965, but with more variabili-ty and extremes than either the base case or scenario2. It also exhibits much lower total supplies for 1954-1960 and 1967-1973. There are no shifts in the peaklevel and outflow on Lakes Ontario and St. Louis, butthere is greater seasonal variability. Spring values arehigher than the base case, and late fall values arelower, reflecting increased spring runoff and fall evap-oration. The variability of the average annual levels,outflows, and ranges on Lakes Ontario and St. Louisincreases significantly (Table 3). Thus, scenario 4resembles the present regime's average levels andoutflows but has much more variability and extremesin total .supplies.

The response of the modified and unmodifiedregulation plans are similar to that of scenario 2

JAWRA 66 JOURNAL OF THE AMERICAN WATER RESOURCES ASSOCIATION

81

74

Lake Ontario Regulation Under Transposed Climates

Lake Ontario Level (m, IGLD85)

11000

50001951

Lake Outflow (m3s1)

1990

Figure 5. Annual Lake Ontario Average Lake Levels and Outflows for Scenario 2 (6S x lOW)and Regulation Plan Limits (Black — UnmodifiedPlan; Gray — Modified Plan).

(Figure 5). Again, with the unmodified plan, total sup-plies exceed the maximum outflow limits beginning in1973, causing levels to rise beyond the upper limit[Criterion (h)]. With the modified plan, outflow limitsare relaxed and levels are maintained near the upperlimit. During the extreme low supply periods, 1954-1960 and 1967-1973, Lake Ontario levels fall belowthe lower limit [Criterion (j)] with the unmodifiedplan because the minimum outflow limits are greaterthan total supply. However, the modified plan reducesoutflows, and the levels rise correspondingly. Becausebeginning levels in 1954 and 1967 are not below thelower limit (74.15 m), outflows are a function of a

supply indicator integral to Plan 1958-D. The supplyindicator is the difference of the weighted past totalsupplies from the weighted observed supplies (1860-1954). The weights are sensitive to long-term supplychanges, but not to short-term changes (InternationalSt. Lawrence River Board of Control, 1963) and arenot a good index of future supplies during rapid andlarge transitions, particularly for the second half of1957 (Figure 6). Note also the frequent seasonal lagwith total supplies. The low supply indicators resultin low plan outflows and high lake levels, peaking in1957. This artifact of the modified plan would not betolerated in actual practice; Criterion (k) would be

JOURNAL OF THE AMERICAN WATER RESOURCES ASSOCIATION 67 JAWRA

1960 1970 1980

C/)C,)

>1

0CI)

Co•1—'0I-

100004ICo.0

>00U)

—1000

—2000

Figure 6. Lake Ontario Total Supply and Adjusted Supply Indicator, 1954-1959, for Scenario 4(10S x 5W) (Black — Total Supply; Gray — Adjusted Supply Indicator).

invoked to compromise between the minimum outflowlimits and the low plan outflows. Changes to the mod-ified and unmodified plans (a more sensitive supplyindicator and revised outflow limits) to address theseproblems are recommended.

Revision of the regulation plan as discussed underscenario 2 would not be as successful under scenario 4due to the increased variability in total supplies,which has serious implications for lake management.Croley and Lee (1993) show that present water supplyforecasts of the U.S. Army Corps of Engineers andEnvironment Canada (made with the smaller presentvariability in total supplies) are only marginally bet-ter than climatology for one-month outlooks and thesame or worse for six-month outlooks. Skill woulddecrease further with increased variability in totalsupplies. Skill can increase only with improved long-range (> 30 days) quantitative weather forecasts,observations of current hydrologic conditions, and theuse of risk assessment (Lee et al., 1997) and proba-bilistic forecasts (Croley, 1996) in decision-making.Also required are revision of expectations and bettereducation as to the limits of lake regulation and fore-casting, and development a priori of managementplans for extreme conditions.

CONCLUSIONS

Historical water supplies should no longer be thesole basis for the testing and developing of lake regu-lation plans. We must consider alternate climates tomore fully understand plan response. Also, the regu-lation criteria cannot be met simultaneously duringextreme supply conditions. Interests and their priori-ties may change under alternate climates. The insti-tutional structure of lake regulation may be affected;decisions may be relegated to higher governmentauthorities when issues cannot be resolved within thecurrent structure. Revision of the Boundary WatersTreaty may be one result. The criteria in the Ordersof Approval established engineering design require-ments for the Seaway and hydropower project giventhe 1860-1954 climate regime(s). Revisions to the cri-teria (and any regulation plan that embodies them)must now focus on operational and adaptive require-ments for existing and alternative climate regimes,such as ecosystem health, economics, and social bene-fits to a wider spectrum of interests.

Plan improvements require better weather fore-casting and observation of current hydrologic condi-tions, risk assessment and probabilistic forecasts indecision-making, revision of expectations and bettereducation as to the limits of lake regulation and fore-casting, and development a priori of managementplans for extreme conditions. Hydrologic predictions

JAWRA 68 JOURNAL OF THE AMERICAN WATER RESOURCES ASSOCIATION

Lee, Croley, and Quinn

1954 1955 1956 1957 1955 1959

Lake Ontario Regulation Under Transposed Climates

are now available that utilize new long-range meteo-rological forecasts and probabilistic techniques (Cro-ley, 1996); evaluation of the predictions remains. Leeet al. (1997) show probabilistic forecast use to assessrisk in operational lake level decisions. Education onlake regulation limitations is an ongoing activity thatrecently suffered from budgetary reductions. Develop-ment of management plans for extreme conditionsremains to be implemented. These initiatives shouldbe explored by the International St. Lawrence RiverBoard of Control and its Working Committee.

ACKNOWLEDGMENTS

The authors thank Mr. David Fay of the Great Lakes-St.Lawrence Regulation Office, Environment Canada, for his insight-ful comments. This paper is GLERL Contribution No. 1005.

LITERATURE CITED

Croley, T. E., II, 1990. Laurentian Great Lakes Double-CO2 Cli-mate Change Hydrological Impacts. Climatic Change 17:27-47.

Croley, T. E., II, 1993. CCC GCM 2xCO2 Hydrological Impacts onthe Great Lakes. In: Climate, Climate Change, Water LevelForecasting and Frequency Analysis, Supporting Documents,Volume 1 - Water Supply Scenarios, D. H. Lee (Editor). Interna-tional Joint Commission, Washington, D.C., Section 9, 33 pp.

Croley, T. E., II, 1996. Using NOAA's New Climate Outlooks inOperational Hydrology. Journal of. Hydrologic Engineering1(3):93-102.

Croley, T. E., U and D. H. Lee, 1993. Evaluation of Great Lakes NetBasin Supply Forecasts. Water Resources Bulletin 29(2):267-282.

Croley, T. E., II, F. H. Quinn, K. E. Kunkel, and S. A. Changnon,1996. Climate Transposition Effects on the Great Lakes Hydro-logical Cycle. NOAA Technical Memorandum ERL GLERL-89,Great Lakes Environmental Research Laboratory, Ann Arbor,Michigan.

Griffiths, J. F. and D. M. Driscoll, 1982. Survey of Climatology. Belland Howell, Columbus, Ohio.

Hartmann, H. C., 1990. Climate Change Impacts on LaurentianGreat Lakes Levels. Climatic Change 17:49-67.

Houghton, J. T., B. A. Callander, and S. K. Varney (Editors), 1992.The Supplementary Report to the IPCC Scientific Assessment.Intergovernmental Panel on Climate Change, Cambridge Uni-versity Press, Cambridge, Great Britain, 200 pp.

International Joint Commission, 1952. An Order of Approval of theConstruction of Certain Works for the Development of Power inthe International Rapids Section of the St. Lawrence Riverdated October 29, 1952. International Joint Commission, Wash-ington, D.C., and Ottawa, Canada.

International Joint Commission, 1956. An Order of Approval of theConstruction of Certain Works for the Development of Power inthe International Rapids Section of the St. Lawrence River, asamended by a supplementary Order dated July 2, 1956. Interna-tional Joint Commission, Washington, D.C., and Ottawa, Cana-da.

International St. Lawrence River Board of Control, 1963. Regula-tion of Lake Ontario Plan 1958-D. International Joint Commis-sion, Washington, D.C., and Ottawa, Canada.

Lee, D. H., 1993. Basis of Comparison, Great Lakes-St. LawrenceRiver System. NOAA Technical Memorandum ERL GLERL-79,Great Lakes Environmental Research Laboratory, Ann Arbor,Michigan.

Lee, D. H., A. H. Clites, and J. P. Keillor, 1997. Assessing Risk inOperational Decisions Using Great Lakes Probabilistic WaterLevel Forecasts. Environmental Management 21( 1):43-58.

Lee, D. H. and F. H. Quinn, 1992. Climate Change Impacts onGreat Lakes Levels and Flows. In: Proceedings of 28th AnnualConference and Symposia of the American Water ResourcesAssociation on Managing Water Resources During GlobalChange. American Water Resources Association, Reno, Nevada,pp. 387-396.

Lee, D. H., F. H. Quinn, D. Sparks, and J. C. Rassam, 1994. Simu-lation of Maximum Lake Ontario Outflows. Journal of GreatLakes Research 20(3) 569-582.

Levels Reference Study Board, 1993. Levels Reference Study, GreatLakes-St. Lawrence River Basin. International Joint Commis-sion, Washington, D.C., and Ottawa, Canada.

Michigan Sea Grant, 1990. Lake Ontario. Extension Bulletin E-1870, Michigan Sea Grant Extension Office, Lansing, Michigan.

Sanderson, M. and L. Wong, 1987. Climatic Change and GreatLakes Water Levels. In:. The Influence of Climate Change andClimatic Variability on the Hydrologic Regime and WaterResources. International Association of Hydrologic SciencesPublication. No. 168, pp. 447-487.

Smith, J. B. and D. A. Tirpak (Editors), 1989. The Potential Effectsof Global Climate Change on the United States, Report toCongress. Report EPA-230-05-89-050, U.S. Environmental Pro-tection Agency, Office of Policy, Planning, and Evaluation, Wash-ington, D.C.

Thorp, S. and D. R. Allardice, 1994. A Changing Great Lakes Econ-omy: Economic and Environmental Linkages. In: Proceedings ofState of the Lakes Ecosystem Conference. Great Lakes Commis-sion, Ann Arbor, Michigan, 37 pp.

U.S. Army Corps of Engineers, 1991. Lake Ontario Regulation PlanImprovements. Great Lakes Levels, Update Letter No. 77,North Central Division, Chicago, Illinois.

JOURNAL 01 THE AMERICAN WATER RESOURCES ASSOCIATION 69 JAWRA