Section 4

HYDROMETEOROLOGY: CLIMATE AND HYDROLOGYOF THE GREAT LAKES

Jan A. Derecki

4.1 Introduction

4.1.1 Description and Scope

The climate of the Great Lakes Region isdetermined by the general westerly atmos­pheric circulation, the latitude, and the localmodifying influence of the lakes. Due to thelake effect, the regional climate alternates be­tween continental and semi-marine. Thesemi-marine climate is more consistent con­tiguous to the lakes, but with favorablemeterological conditions, it may penetratedeeply inland. The Great Lakes climate andhydrology are closely related. Variations inmean lake levels, and consequently lake out­flows, are controlled by the imbalance be­tween precipitation and evaporation.

The Great Lakes drainage basin is discussedin other appendixes and a brief summary ofthe climatic and hydrologic elements of thedrainage basin is given in Section 1 of thisappendix. The major storm tracks affectingthe Great Lakes Region are indicated in Fig­ure 4-13. Distributions of mean annual valuesfor air temperature, precipitation, runoff, andwater losses over the land areas of the Basin,showing latitudinal and lake-effect variations,are indicated in Figures 4-15, 17, 19, and 20,respectively. The average monthly means,highs, and lows of overland air temperaturefor the individual lake basins and for the totalGreat Lakes Basin are shown in Figure 4-16.

4.1.2 Lake Effect

With a total water volume of 22,813 km3

(5,473 cu. mi.) stored in the lakes, varying from484 km3 in Lake Erie to 12,234 km3 in LakeSuperior, the Great Lakes have a tremendousheat storage capacity. Through air-water in­teraction, the lakes influence the climate overthem and over adjacent land areas. Because ofthe lake effect, air tempt!!ratures are moder­ated, winds and humidities are increased, andprecipitation patterns are modified. Althoughthese phenomena have been recognized fordecades, it is only in recent years that inten­sive programs have been undertaken to de­termine the more exact nature and mag­nitudes of these processes.

The Great Lakes moderate temperatures ofthe overlying air masses and surroundingland areas by acting as heat sinks or sources.The process of heat exchange between thelakes and atmosphere is both seasonal anddiurnal. During spring and summer, the lakesare generally colder than air above and have acooling effect on the atmosphere. During falland winter, the lakes are generally warmerthan the atmosphere and serve as a heatsource. However, during the winter months,the ice cover reduces the lake effect.

A daily pattern of heat exchange issuperimposed upon the seasonal pattern. Thisdaily pattern is produced by land-water tem­perature differences. Because lakes are moreefficient than land areas in storing heat, laketemperatures have a tendency to remain sta­ble, while land temperatures undergo dailyvariations that are more in line with the airtemperatures. When the land is warmer thanwater, the relatively warmer air over adj acentland areas tends to rise and is replaced bycolder, heavier air from the lakes. When theland is colder, the process is reversed. This

Jan A. Derecki, Lake Survey Center, National Oceanic and Atmospheric Administration, Detroit,Michigan.

71

71 Appendi~ 4

process ofheat exchange produces light winds,which are known as lake breezes. The offshoreand onshore lake breezes are illustrated inFigure 4-14. The direction of lake breezes isgoverned by the land-water temperature dif­ferences and is independent of the general at­mospheric circulation. However, lake breezesoccur only during relatively calm weather andaffect a limited air mass along the shoreline,rarely extending more than several kilome­ters (2-3 miles) inland. The moderation of tem­peratures by the lakes affects regional ag­riculture by reducing frost hazards in theearly spring and in the fall, thus lengtheningthe growing season, especially in coastalareas. Examples of this effect are the cherryorchards of the Door Peninsula in Wisconsinand the Grand Traverse Bay area in Michigan,and the vineyards of the western Lake Erieislands in Ohio.

Because lake breezes have a limited rangeand require special conditions, the lake effecton winds is minor. A much more importanteffect is the considerable increase in geo­strophic wind speed over the lakes. This in­crease is caused by reduced frictional resis­tance to air movement over the relativelysmooth water surface, and by the difference inatmospheric stability created by air-watertemperature differences. Recent studies indi­cate that the increase in overwater wind speedvaries from approximately 15 percent in mid­summer to as much as 100 percent in late falland early winter. The average annual in­crease is approximately 60 percent.

The Great Lakes also cause an increase inoverwater humidity by releasing large quan­tities of moisture through evaporation. On anannual basis the humidity over the lakes av­erages 10 percent to 15 percent higher thanthat over the land. Seasonal changes inhumidity over the lake compared to that overland vary from a decrease of approximately 10percent due to overwater condensation in thelate spring, to an increase of approximately 10percent in the summer, 15 percent in the fall,and 30 percent in the winter.

The Great Lakes also influence the distribu­tion ofcloud cover and precipitation. Modifica­tion ofpreeipitation patterns due to lake effectis caused by the changes in atmospheric stabil­ity in combination with prevailing wind direc­tion and topographic effects. During summer,the air undergoes overland warming beforepassing over the lakes. The warm air over rel­atively cold water results in the developmentof stable atmospheric conditions, which dis­courage formation of air-mass showers and

thunderstorms. During winter, the conditionsare reversed, and the cold, inland air passingover relatively warm water becomesles8 sta­ble and picks up moisture, which encouragessnow flurries. As the winter air masses moveover the lakes, the moisture that accumulatesin the air produces heavy snowfalls on the leesides of the lakes, due to orographic effects ofthe land mass. The fact is well documentedthat heavy snowbelt areas result from the lakeeffect. These areas include Houghton on LakeSuperior, Owen Sound on Lake Huron, Buffaloon Lake Erie, and Oswego on Lake Ontario.

4.1.3 Measurement Networks

Basic meteorological data in the GreatLakes Basin are available from regularobser­vation networks operated by the Nation­al Weather Service and the CanadianMeteorological Service. The networks consistof a limited number of first order stations thatprovide hourly observations for air tempera­ture, precipitation (total and snow), windspeed and direction, humidity, and duration ofsunshine and cloud cover. More numerous co­operative stations provide daily observationsfor air temperature and/or precipitation. Cer­tain more specialized stations collect addi­tional data on solar radiation, radiosonde in­formation, weather radar, and pan evapora­tion. Other regularly observed data useful inGreat Lakes climatology include water tem­peratures recorded by municipalities at theirwater intake structures and by Federal agen­cies at ~elected lake perimeter location&

In addition to the regular networks, moresophisticated data are collected periodically orseasonally for research on lake climatology.These include special precipitation networks,established and operated on lake islands andadjacent shorelines; synoptic surveys con­ducted by research vessels that take observa­tions for the whole range ofhydrometeorologicparameters; lake towers that give meas­urements with vertical profiles for selectedparameters for air-water interaction studies;aerial surveys by conventional aircraft for icereconnaissance and water surface tempera­tures, using infrared and airborne radiationthermometer techniques; and weather satel­lites that provide useful information for theinvestigation of doud and ice cover on thelakes.

Hydrologic data on the Great Lakes Basinare compiled and pu~lished by several agen­cies. Records of tributary streamflow to the

lakes are available from the U.S. GeologicalSurvey and the Canada Centre for Inland Wa­ters, Department of the Environment,Canada. The extent of gaged area increasedsubstantially in the late 1930s, giving cover­age to approximately 50 percent of the Basin.At present approximately 64 percent of theBasin is gaged; gaged areas for Lakes Supe­rior, Michigan, Huron, Erie, and Ontario ba­sins represent approximately 53, 71, 66, 67, and63 percent of their respective basins. In addi­tion to surface water data, these agencies andthe Geological Survey of Canada publish ob­servation well records providing informationon ground-water conditions. However, thenetwork of observation wells useful in deter­mining ground-water flow to the lakes is ex­tremely limited.

The Great Lakes levels and outflows areavailable from the Lake Survey Center. Lakelevels are determined from a network of waterlevel gages maintained by the Lake Survey inthe United States and the Fisheries andMarine Service, Department of the Environ­ment, in Canada. Flows in the connectingchannels are determined by the Lake Survey

HrdrofMteorolon 78

from appropriate water level gage ratingsbased on periodic current meter flow meas­urements.

4.2 Radiation

4.2.1 Total Radiation Spectrum

Total radiation received at the surface of theearth consists of shortwave radiation comingdirectly from the sun or scattered downward,and longwave radiation, emitted from the at­mosphere. Portions of the incoming radiationin both short and long wavelengths are re­flected and additional longwave radiation isemitted to the atmosphere. The main radia­tion exchange processes taking place withinthe terrestrial system (space-atmosphere­earth) are illustrated in Figure 4-93, present­ing annual radiation balance, which is basedlargely on information provided by London.503

The net effect of shortwave radiation is thesolar heating of the earth, while longwaveradiation results in cooling.

8\1//

...... ...-- .... -....- .......

/ ,I I \

T014LSMORTwAVE

Al.eEDO

SPACE

UPPER ATMOSPHERE

LOWERATMOSPHERE

TOTAl.1.0NGWAVE

'''''DIATION 1.055

AeSCReEOI'lIlOM

SUNLIGHT

SOLAR (SHORTWAVE) RADIATION TERRESTRIAL (LONGWAVE) RADIATION

After London, 19117

FIGURE 4-93 Annual Atmospheric Heat Budget. Shows percentage distribution of radiationcomponents for northern hemisphere.

7.. A'PPendi~"

There is some duplication of names used forthe same radiation components. Shortwave orglobal radiation is generally referred to assolar radiation and both names are used inter­changeably in this report. Insolation is inci­dent or incoming solar radiation. Similarly,terrestrial radiation is used synonymouslywith longwave radiation, while longwaveradiation from the atmosphere is called at­mospheric radiation.

A regular network for measuring shortwaveradiation has been established but only few ofthese are in the Great Lakes Region. A regularnetwork for total radiation measurements(shortwave and longwave) has not been estab­lished, since there are only a limited number ofradiometers in operation at various researchinstallations. Available information indicatesthat the average monthly all-wave incidentradiation in the Great Lakes Region variesfrom a winter low of approximately 400langleys per day (ly/day) in December to asummer high of approximately 1400 ly/day inJune or July.

4.2.2 Solar Radiation

Solar radiation is reduced by the atmos­phere before reaching the earth's surface. At­tenuation of the extraterrestrial solar radia­tion is caused by scattering, reflection, andabsorption by gas molecules, water vapor,clouds, and suspended dust particles (Figure4-93). As a result of the attenuation, the in­coming shortwave radiation on a horizontalsurface arrives partly as direct solar radiationand partly as sky radiation (scattered down­ward by atmosphere and diffused through theclouds). Sky radiation is a high percentage ofthe total incident radiation during low decli­nation of the sun and on overcast days.

Part of the incoming solar radiation is re­flected from the receiving surface (clouds andearth) back to the atmosphere, the amount ofreflection depending on the surface albedo orthe ratio of reflected to incident radiation. Al­bedo values for a water surface depend on thesolar altitude (angle ofthe sun above the hori­zon), cloud cover, and the roughness of thewater surface, but for many practical pur­poses these factors can be assumed to be con­stant for daily or longer periods. During theLake Hefner study, Anderson 16 developed em­pirical curves, which interpret water surfacealbedo as a function of sun altitude for variouscloud cover conditions. Based on results ofthat study, Kohler and Parmele 464 recom-

mended an average daily albedo for water sur­face of 6 percent. The relatively low albedo foropen water conditions increases drasticaHywith ice and snow cover. Bolsenga18 gives al­bedo values for various types of ice common onthe Great Lakes. These values range from 10percent for clear ice to 46 percent for snow ice,both free of snow cover, The presence of par­tial or complete snow cover on the ice can sig­nificantly increase these values.

There is a limited network of regular sta­tions that measure incident solar radiation inthe Great Lakes Region. The average monthlyvalues from these stations are shown in Fig­ure 4-94. Periods ofrecord for the stations varyfrom 10 to 50 years. Based on records from theradiation network, the average monthly in­coming solar radiation in the Great LakesBasin varies from a low of approximately 100ly/day in December (winter solstice) to a highof approximately 550 ly/day in June and/orJuly (near the summer solstiee), with an aver­age annual value of about 320 ly/day. The avoerage monthly extremes for reflected solarradiation from the lakes represent from 6 to 33ly/day (6 percent water surface albedo).

Beginning in the last decade, direct overwa·ter measurements of solar radiation were in·cluded in the synoptic surveys of the GreatLakes conducted by several research organi·zations. These measurements are generallylimited to the navigation season (April.December), are not continuous, and are some·what biased towards fair weather conditions,but nevertheless they represent actual condi­tions over the lakes and provide a basis forcomparison of the overwater and overlandradiation. Richards and Loewen 653 conducteda preliminary study of this type, which showsthat incident solar radiation over the lakes isgreater than that recorded on adjacent landstations during summer and smaller duringwinter months. This confirms the physicalconcepts of the lake effect. Their study is lim­ited to four years of data during the April­December periods and shows that overwaterradiation at the beginning and end of theperiod amounts to 90 percent of the overlandradiation. The overwater radiation increasesgradually during spring and summer to an av­erage high ofapproximately 140 percent oftheoverland radiation in the late summer, then itdecreases rapidly in the fall.

Other recent studies of solar radiation onthe Great Lakes include determination of theradiation balance for Lake Ontario (Bruce andRodgers 108 and Rodgers and Anderson 615). De­termination of the total atmospheric water

oJ'MAMJJASOND

Io

HJldrometeorolol1'11 75

GREAT LAKES BASIN

\eGO .......

.::.:......~

10

J'MAIUJASOND

FIGURE 4-94 Average Daily Solar Radiation (Langleys) in the Great Lakes BasinFrom Phillips. 1969

vapor over the Great Lakes Basin and deriva­tion of a relationship between atmosphericwater vapor and surface dew point(Bolsenga 74.77) could contribute to parame­terization of the solar radiation term. Thismight compensate for the lack of recordingstations in the lakes.

4.2.3 Terrestrial Radiation

Terrestrial radiation over a body of water(or over land) consists of the incident atmos­pheric radiation, reflected atmospheric radia­tion, radiation emitted by the water body, andenergy released through the processes ofevaporation, condensation, and precipitation(latent heat), and turbulent heat transfer(sensible heat). The net result of the incident,reflected, and emitted radiation components is

the net back radiation, a longwave radiationloss to the atmosphere (Figure 4-83). The netback radiation is primarily a function of thetemperature of the water surface, which con­trols emitted radiation, and the water vaporcontent of the air, which controls atmosphericradiation. Other factors that affect the netback radiation include the emissivity of water(relative power of a surface to emit heat byradiation), which reduces emitted radiationbelow that of black body (an ideal surface thatemits maximum radiation); the reflectivity ofwater, which controls reflected atmosphericradiation; and the concentration of carbondioxide and ozone in the atmosphere, whichare minor contributors to the atmosphericradiation. The reflectivity of a water surfacefor atmospheric radiation is 3 percent (Ander­son 18), only about half as much as for the solarradiation.

76 Appendix"

The earth and the atmosphere can absorband emit more than 100 percent of radiation,exceeding the original input from the sun.This is possible because of the so-called green­house effect of the atmosphere. By blockingterrestrial radiation (very small direct loss tospace), the atmosphere forces the earth sur­face temperature to rise above the value thatwould occur in the absence ofthe atmosphere,which in turn produces upward vertical trans­fer of both latent and sensible heat.

Terrestrial radiation may be determined in­directly from the total (all-wave) and solar(shortwave) radiation measurements, butall-wave measurements are too sparse for thispurpose. Atmospheric radiation may also becomputed utilizing various radiation indices(temperature, percent of sunshine or cloudcover, vapor pressure). Anderson and Baker 15

present a method of computing incident ter­restrial radiation under all atmospheric con­ditions from observations of surface air tem­perature, vapor pressure, and incident solarradiation. Emitted radiation is determinedfrom water surface temperatures. Based onavailable information, estimates ofterrestrialradiation for the Great Lakes are as follows:monthly incident atmospheric radiation var­ies from a winter low of approximately 300ly/day in December to a summer high of ap­proximately 800 ly/day in June or July; re­flected atmospheric radiation for thesemonths represents 9 to 24 lylday (3 percentreflectivity); monthly emitted radiation fromthe lakes varies from a low of approximately400 lylday during winter to a high of 900 lyldayduring summer; monthly net back radiation(longwave radiation loss) is roughly 100 lyldaythroughout the year (Figure 4-100).

4.3 Winds

4.3.1 Lake Perimeter Winds

Winds are a critical factor of lake climatebecause they provide energy for lake waves,constitute a principal force for driving lakecurrents and shifting ofice cover, and throughair movement provide means for the regula­tion of thermal budget over the lakes and ad­jacent land areas by heat dissipation andtransfer. In the Great Lakes Region the globalatmospheric circulation with prevailing west­erly winds is of particular importance on thelower lakes where it coincides with the lon­gitudinal axes of the lakes, exposing the full

lengths of the lakes to the winds and the leeshores to the maximum lake effect.

Because of the lake effect on adjacent landareas, wind data from stations located arOundthe perimeter of the lakes are of particularinterest to the Great Lakes. Since more repre.sentative data were not available, these datahave often been used in past studies as over­water winds, frequently without adjustmentfor anemometer height or the increase in windspeed over the lakes. Average monthlyperimeter wind speeds for the Great Lakes aregiven in Table 4-7. The average annualperimeter wind speed generally increasesfrom north to south, from approximately 4.5mls (10 mph) to 5.0 mls (11 mph). Averagemonthly wind speeds increase from the sum­mer low of 3.5 mls to 4.0 mJs (~9 mph) to thewinter high of 4.5 mls to 5.5 mls (10-12 mph).

A summary of wind direction for selectedstations around the lakes and the St. Law­rence River is presented in Figure 4-12. Thewind roses in this figure show wind directionfrequencies for the months of February, May,August, and November, indicative of the fourseasonal periods.

Actual wind conditions on the lakes and fur­ther inland vary somewhat from those indi­cated by shore stations, which are affected to avarying degree by the lakes and lake-land in­teraction. The perimeter weather stations arelocated at some distance inland, and may gen­erally be unaffected by lake breezes, but thestations located on the lee sides of the lakesare certainly affected by the lakes duringwinds from the prevailing wind directions.

The long east-west axis of Lake Superior isdivided by the Keweenaw Peninsula, whichseparates the lake into two basins wherewinds are frequently of opposite direction. Onthe western end of the lake (Duluth) the winds'are predominantly from the west and north­west during cold months and from the east andnortheast during the warm months. On theeastern end ofthe lake (Sault Ste. Marie) thereare predominantly easterly winds in the coldmonths and westerly winds during warmmonths. In the middle section of the lake(Marquette) predominant winds are from thenorthern and southern quadrants. The meanmonthly wind speed at these stations variesfrom 3 mls (7 mph) in the summer to 6 mJs (14mph) in the winter (for average perimeterwind speeds for the whole lake see Table 4-7).The maximum recorded wind velocity was 41

TABLE 4-7 Average Perimeter Wind Speedsfor the Great Lakes (mls)

ad Superior Michigan Huron Erie Ontario!,!r1 _

January 4.6 5.4 4.8 5.5 5.0february 4.5 5.3 4.4 5.5 5.0March 4.6 5.5 4.6 5.5 5.0APril 4.8 5.5 4.6 5.4 4.8

4.6 5.0 4.2 4.7 4.3MaY4.0 4.3 3.7 4.2 3.9June

July 3.8 3.8 3.6 3.8 3.8A!J&ust 3.8 3.8 3.5 3.8 3.6September 4.1 4.3 4.0 4.1 3.8OCtober 4.4 4.9 4.3 4.4 4.0!Iovelllher 4.6 5.4 4.8 5.2 4.6Dec_ber 4.5 5.3 4.8 5.3 4.8

Annual 4.4 4.9 4.3 4.8 4.4

Values are based on mean data published in 1969 for thefollowing stations:

Superior: Sault Ste. Marie. Marquette. Duluth. andthunder Bay.

Michigan: Milwaukee, Muskegon. and Green Bay.Huron: Alpena. Gore Bay. and Wiarton.Erie: Toledo, Cleveland, !uffalo. and London.Ontario: Rochester, Syracuse, Trenton. and Toronto.

mls (91 mph) from the south at Marquette inMay, 1934.

Around Lake Michigan the predominantwind direction is from the western quadrant,perpendicular to the long axis of the lake. Be­cause of the north-south lake orientation, thehighest seas generally coincide with strongnortherly and southerly winds. Prevailingwinds from these directions are reported atsome locations, in contrast to the general pre­dominant westerly direction. The variation inprevailing winds is evident in northern LakeMichigan where winds in Traverse City comefrom the south during fall, while Green Bay,on the opposite (western) shore, is assailed bywesterly winds. Around the southern portionof the lake prevailing winds in the spring atMilwaukee are from the north, while at Chi­cago they are from the southwest. The meanmonthly wind speed at these stations variesfrom 3m/s to 6 mls (7-14 mph), which issimilar to Lake Superior, but the annual windspeed around Lake Michigan is higher. Thehighest wind velocity recorded on all the GreatLakes, 49 mls (109 mph) from the southwest,occurred at Green Bay in May 1950.

Winds on Lake Huron may be equally effec­tive on the sea state from all directions due tothe lake configuration. There is considerablevariation in wind direction around the lake,but in general, prevailing winds are from thewestern quadrant. In some locations prevail­ing winds shift seasonally to the south during

H"drom.tlteorolotrl 11

fall <Wiarton, Ontario), and a large percentageof winds along the western shore come fromthe eastern quadrant during warmer months,as indicated at Alpena and Bay City (SaginawRiver Light) in Michigan. The range of meanmonthly wind speed at these stations variesfrom the summer low of 4 mls (8 mph) to thewinter high oU mls (13 mph). The highest vel­ocity recorded was 27 mls (61 mph) from thesouthwest at Alpena in November 1940.

The highest monthly wind speeds aroundthe Great Lakes occur on Lake Erie, whichalso has the largest range between themonthly values of wind speed. The meanmonthly wind speed at stations located aroundthe lake varies from 4 mls to 8 mls (8-18 mph).These winds are predominantly from thewestern quadrant with a prevailing directionfrom the southwest. which coincides roughlywith the long axis of the lake. This fact, alongwith the relative shallowness of the lake,makes Lake Erie highly susceptible to large­scale water level motions, especially at theeastern and western extremes of the lake. Be­cause of the prevailing wind direction, lakeeffect on the lee shores is quite pronouncedand the monthly wind speeds at Buffalo arenormally somewhat higher than at other sta­tions around the lake. Prevailing winds atsome locations are from the eastern quadrant,and in the middle section of the lake (Cleve­land), prevailing winds during warmermonths shift to the north and south directions.The maximum velocity recorded was 41 mls (91mph) from the southwest at Buffalo inJanuary 1950.

The predominant wind direction aroundLake Ontario is similar to that of Lake Erie,with prevailing winds during most monthsfrom the southwest (Rochester, Trenton),which approaches the direction of the longaxis of the lake. During winter months thepredominant wind direction shifts to the west.On the northwestern end ofthe lake (Toronto)winds frequently prevail from the west, and attimes from the north. The mean monthly windspeed at these stations varies from 3 mls to 6mls (7-13 mph). The highest wind velocity re­corded was 33 mls (73 mph) from the west atRochester in January 1950. The prevailingwinds along the St. Lawrence River are paral­lel to the river. primarily from the southwestand secondarily from the northeast.

4.3.2 Overwater Winds

Overwater winds differ from overland

18 Appendiz"

winds, both daily and seasonally, because ofdifferences in air stability conditions and fric­tional resistance. Daily variation is causedmainly by diurnal heating and cooling, whichare more pronounced over land areas thanover water and result in larger daily wind var­iations over land than over water. Seasonalvariation is caused by the winter heating andsummer cooling effects of the lakes. The lakesoffer less resistance to wind movement, result­ing in considerably higher overwater windspeeds regardless of the season.

The highest wind speeds (one-minute windgusts) on the Great Lakes, reported fromanemometer-equipped vessels since 1940, arelisted for each lake as follows: Lake Superior,42 mls (93 mph) from the northwest in June1950; Lake Michigan, 30 mls (67 mph) from thewest-southwest in November 1955; Lake Hu­ron, 49 mls (109 mph) from the west-northwestin August 1965; Lake Erie, 38 mls (85 mph)from the north-northwest in June 1963; LakeOntario, 26 mls (57 mph) from the west­northwest in November 1964. These velocitieswere observed during the navigation seasonand are based largely on observations takenfour times daily during synoptic hours (0100,0700, 1300, and 1900 hours, EST). Higher windspeeds may have occurred during wintermonths and at times other than synoptichours. Most of the shipboard wind directionslisted by the National Weather Service verifythe predominantly westerly wind direction in­dicated by the perimeter stations.

The first intensive effort to determine over­water winds utilized various ships-of­opportunity programs, which were conductedperiodically and consisted initially of commer­cial vessels making wind observations fourtimes daily. The data collection program hasnow been expanded to off-shore towers, buoys,research vessels, and research stations lo­cated on small islands in the Great Lakes. Be­cause of practical limitations imposed onmeasurement of wind data for prolongedperiods oftime over the lakes, the primary aimof wind measurement programs was to deter­mine the relationship between overland andoverwater winds. Several studies of this typehave been conducted, relating shore data withobservations from ships and islands located onthe Great Lakes (Hunt,397.398 Lemire,492 Bruceand Rodgers,I08 and Richards et aI.649).

The ratios of wind speed over water to windspeed over land vary diurnally and seasonally,and are a function mainly ofthe stability oftheair. For unstable atmospheric conditions, withwater temperature much higher than air,

lake-land wind ratios are about two, and forstable atmospheric conditions, with air muchwarmer than water, wind speed ratio valuesare near one; for the adiabatic or neutral sta­bility conditions values are intermediate.Hunt's investigation was conducted mainlyfor Lake Erie during navigation season(April-November), with results grouped intothe spring and fall periods. Bruce and Rod­gers l08 prepared a similar study for Lake On­tario. Their investigation was extended byLemire 492 who included data from some oftheother Great Lakes and derived monthly windspeed ratios for the spring, summer, and fallmonths (March-October). Richards 848 ex­tended these ratios for the winter monthsusing partial results determined by Lemireand extrapolation based on the air-water tem­perature difference, along with limited windobservations on Lake Ontario. The variationof wind speed ratios determined in thesestudies is shown in Table 4-8. Monthly ratiosvary from 1.2 to 2.1, with low values duringsummer and high during winter, and an over­all annual average of about 1.7.

The effects of overwater fetch (length ofopen water) on lake winds, besides atmos­pheric stability, were studied by Richards etal.,849 who utilized wind data collected duringsynoptic surveys on Lakes Erie and Ontario.Their analysis included five stability ranges(from very unstable to very stable), four windspeed classes (3 mls to 8 m/s), and five fetchranges (10 km to 65 km). They found that thelake-land wind ratio increases with the at­mospheric instability, but the increase is mostpronounced in light winds. Under very unsta­ble atmospheric conditions (large negative airtemperature-water temperature difference,T A - Tw) the wind ratio increases graduallyfrom 1.4 for strong winds to 3.0 for light winds,with a 2.2 value for all winds. Under very sta­ble atmospheric conditions (large positive TA ­

Tw difference) the wind ratio increases gradu­ally from 0.8 for strong winds to 1.4 for lightwinds, with a value of 0.9 for all winds. Thus,under very stable conditions the lakes mayreduce the wind speed, especially in strongwinds. The effect ofoverwater fetch was not aspronounced and somewhat erratic. Under un­stable atmospheric conditions the wind speedratio increases with the overwater fetch, butonly for lengths smaller than 50 km (25 nauti­cal miles). Under stable atmospheric condi­tions the relationship between wind ratios andoverwater fetch was highly erratic. Sum­marized results of this study are listed to­gether with other wind studies in Table 4-8.

Hydrometeorology 79

TABLE 4-8 Lake-Land Wind Speed Ratios for the Great Lakes

Hunt Lamire Richards, Dragert, McIntyre(1958) (1961) (1966)

Stability Range

Period Ratio Period Ratio TA-Tw (OC) Ratio

January 1.961February 1.941March 1.88April 1.81 •

Spring 1. 35 May 1.71 >-12.6 2.24June 1.31 -12.5 to - 4.1 1.88July 1.16 - 4.0 to 4.0 1.44August 1. 39 4.1 to 12.5 1.06=Fall 1.82 September 1.78 > 12.6 0.92October 1.99November 2.091December 1.981

Navigation Season 1.58 1.63

Annual 1.75 1.51

1Va1ues for winter months were extended by Richards (1964) through extrapolation.

4.4 Air Temperature

4.4.1 Lake Perimeter Temperature

Temperature is one of the principal indi­cators of climate and exerts a large influenceon other climatic elements, such as precipita­tion and evaporation. The vast water ex­panses of the Great Lakes moderate air tem­perature over the lakes, which in turn has amoderating effect on adjacent land areas. Anindication of the lake effect on shoreline tem­peratures is given in Figure 4-95 prepared byPond,822 which compares mean hourly airtemperatures for March, June, and Septemberat Douglas Point, located on the easternshores of Lake Huron, to those at Paisleyclimatological station, some 20 km (12 miles)inland. During late winter (March), the lake­shore station is consistently warmer by 2°C to4°C (3-T'F) than the inland station. This effectchanges gradually during spring to the sum­mer effect, providing daytime cooling of thelakeshore station by as much as 4°C (T'F) andnighttime warming by 2°C (3°F) in June. In theearly fall the effect reverses again and thelakeshore station is consistently warmerthroughout the day.

Because of the lake effect and lack of directoverwater measurements for any longerperiod oftime, various investigators used datafrom perimeter stations to estimate air tem­perature over the lakes. The average monthlyand annual air temperatures for the individuallakes, based on perimeter data for the 1931-69period, are listed in Table 4-9. Average annualtemperatures vary from 4°C (39°F) on LakeSuperior to 9°C (48"F) on Lake Erie, with in­termediate values on other lakes. Averagemonthly temperature extremes also occur onLakes Superior and Erie, and vary frommonthly lows of approximately -HOC and-4°C (13 and 26°F) in January to monthlyhighs of about 18"C and 22"C (65 and 71°F) inJuly, on the two lakes, respectively.

The air temperature decreases as latitudeincreases, being lowest for Lake Superior andhighest for Lake Erie. Disregarding minorlocal variations,the average annual tempera­ture on Lake Superior varies from 1°C (34°F)along the extreme northern shore to 5°C (41°F)along the southern shoreline (see Figure 4-15).Distribution of average annual temperatureon Lake Michigan varies from approximately~C (43°F) in the north to lOoC (50°F) in thesouth. On Lake Huron, the average annualtemperature increases southward from 5°C to

80 Appendix"

4.4.2 Overwater Temperature

The difference between air temperatureover lake and land areas in the Great LakesBasin has been studied by several inves­tigators. In describing the climate of SouthBass Island in western Lake Erie, Verber 848

compared island records (Put-in-Bay) withperimeter and inland stations and concludedthat the mean mid-summer temperature(July) decreases gradually from Put-in-Bay toperimeter stations (Sandusky and Toledo) andto inland stations (Tiffin and Bucyrus) locatedwithin 80 km (50 miles) south of the lake. Heattributes the higher overlake temperaturesto the increased solar radiation due to lessprecipitation and cloud cover over the lake.During mid-Winter (January) the reverse istrue, because the portion of Lake Erie aroundthe island region is shallow and normallyfreezes over and is largely ice covered, thusreducing the lake effect. Although Put-in-Bayhas the highest mean July temperature andthe lowest mean January temperature of thefive stations, its average monthly range dur­ing the year is the smallest, because thermalstability over the lake acts as a damperagainst sudden heating or cooling. The aver-

values for the average conditions on the lakes(Hunt,·11 Powers et al.,1S4 Snyder,151 Rodgersand Anderson,·'711 Derecki,114 and Richards848).

1615 SEPTEMBER /'--........

~ ,/ \13 /' \12 / ,

/ "-11 I ...........10 I _AN CAlCE TEIIP£RllTUII£ ..., ........

9 ....,1800 02 01 01 oa 10 12 IA .. • zo 22 M

TIME IESTI

-F. "C.

10 MARCH

-1 ,,---,-2 "." ,-3 " ....,

// '-,-4 / ' ....-5 /

...._---_/ 10£'" CAllE TE_~ATUIII 13'"-6

00 02 01 oa 01 lO 12 IA .. • zo zz 2A

2221

___UPOINT

,-.".-. ......"20 --......lY

,,/ , JUNE

19 / \/ \

18 I \17

I/

16 I151413 _AN LAKE TEMPERATURE: ....,

1211

00 02 01 oa oa lO 12 IA .. I' zo zz M

FIGURE 4-95 Mean Hourly Temperatures forDouglas Point and Paisley, Lake Huron, for theMonths of March, June, and September, 1962

From Pond. 1964

8"C (41 to 4'T'F). The annual temperature onLake Erie increases from a low of goe (47"F)along the northeastern shore to a high of 11°C(52"F) along the southwestern shoreline. OnLake Ontario, the annual air temperature in­creases from 7°C to goe (45 to 48°F) between thenorthern and southern shores.

It should be noted that the air temperaturesdiscussed above are based on lake perimeterstations and may be different from those rep­resenting mid-lake conditions. Some differ­ence in these temperatures is introduced bythe land effect, which takes place not only atshore stations but also in the shallow coastalwaters. Furthermore, most first order perime­ter stations used to derive temperature esti­mates are located some distance inland, wherethe land effect is more pronounced. Neverthe­less, because of data limitations, air tempera­tures from the perimeter stations around thelakes are generally used as representative

TABLE 4-9 Average Perimeter Air Tempera-ture for the Great Lakes, 1931-1969 (DegreesCentigrade)

LakePet"iod Superior Michigan Huron Erie Ontario

January -11.2 - 6.4 - 7.4 - 3.8 - 5.2February -10.4 - 5.6 - 8.0 - 3.7 - 5.2Karch - 4.7 - 0.4 - 3.5 0.9 - 0.1April 3.1 6.8 4.1 7.4 6.8May 9.2 12.7 10.2 13.6 13.1June 14.6 18.4 15.8 19.2 18.7July 18.1 21.3 18.9 21.8 21.4August 17 .4 20.4 18.7 20.9 20.3September 13.0 16.4 14.2 17.2 16.4October 7.2 10.3 8.7 11.1 10.2November - 0.7 2.8 2.0 4.3 3.8December - 7. 7 - 3.6 - 4.2 - 1. 7 - 2.8

Annual 4.0 7.8 5.8 8.9 8.1

Values are based on data for the following stations:Superior: Sault Ste. Marie, Marquette, Duluth, and

Thunder Bay.Michigan: Milwaukee. Muskegon, and Green Bay.Huron: Alpena, Gore Bay, and Wiarton.Erie: Toledo, Cleveland, Buffalo, and London.Ontario: Rochester, Syracuse, Trenton, and Toronto.

age monthly maximum-minimum tempera­ture range increases gradually inland fromapproximately 8°C (14°F) at Put-in-Bay to ap­proximately 12°C (2~F) at Bucyrus, and thefrost-free season decreases gradually inlandfrom more than 200 days to approximately 150days. Also, the hottest days in July showhigher temperatures on mainland stationsthan at Put-in-Bay. During the winter an icecover around the island region, usually form­ing in January and lasting through February,acts as an insulator between the warm waterand cold air, producing enough change in thenormal temperature pattern to make Feb­ruary colder than January. On exceptional oc­casions when the lake is free of ice duringthese two months, temperature was approxi­mately 3°C (5°F) higher.

Summer temperature conditions for west­ern Lake Erie may be assumed to be indicativeof temperature modification by the otherGreat Lakes, although mid-lake modificationis undoubtedly more pronounced since the is­land itself produces some effect. Winter condi­tions, on the other hand, may not be compara­ble because other lakes have much greaterdepths and different ice-cover conditions. in acomparison of summer data for Fort Williamand Caribou Island 075 miles away) made inconnection with a synoptic survey of Lake Su-.perior, Anderson and Rodgers 14 show that airtemperature on the island is much more stablethan and differs considerably from that atFort William, on the perimeter of the lake.They state that measurements from the islandare extremely valuable since they representan entirely maritime situation, which iscaused by modification of low level air massesby the lake. However, there are only a fewisland stations measuring air temperatureson the lakes and most are operated only duringthe navigation season.

Measurement of air temperature on laketowers or buoys, and synoptic surveys by ves­sels initiated in the late 1950s, provide morereliable data by eliminating possible island ef­fects. Based on synoptic survey data collectedby the research vessel Porte Dauphine onLake Ontario, Bruce and Rodgers 108 observedthat air temperatures at 3 mOO ft) above thewater surface are much closer to water sur­face temperatures than to land temperaturesat the lake perimeter (mean of temperaturesat Toronto and Rochester). Rodgers and An­derson,875 utilizing these data in the energybudget study of Lake Ontario, made similarobservations and showed that air tempera­ture over water in June is about 6°C OO°F)

Hydrometeorology 81

higher than water surface temperature, whileair temperature at Toronto displays a differ­ent pattern and is on the average approxi­mately 1rc (30°F) higher than water temper­ature. These figures are based on only threedays of data from a single cruise, and theirmagnitudes may not be valid for longerperiods. Rodgers and Anderson875 state thatthere are insufficient data to provide a reliableconversion of land station temperatures to theoverwater air temperatures. At present, withapproximately a decade of data available, thisdifficulty has been overcome but the conver­sion factors have not been developed. Thereare no published reports presenting overwa­ter temperatures on the lakes, other than datareports for the individual surveys.

4.5 Water Temperature

4.5.1 Water Surface Temperature

The oldest sources of water surface temper­ature data in the Great Lakes are the recordsobtained at various marine structures, such asdocks, breakwaters, and lighthouses. Thesestations were later replaced by the somewhatmore sophisticated sites offered by the intakestructures of the water treatment plants lo­cated around the lakes. Water temperature atthe intake stations is obtained in the coastalwaters, a few hundred to a few thousand me­ters off shore, and at depths of3 to 15 meters (l0to 50 feet) below the surface. These data obvi­ously do not represent the temperature at thesurface and require adjustments for open lakeconditions. Initially, open lake measurementswere made by commercial vessels along theirnavigation routes, and more recently by re­search vessels engaged in synoptic surveys ofthe lakes. The latest development in measur­ing surface water temperatures involves theuse of airborne infrared thermometers. Theuse of airborne radiation thermometers per­mits fast and regular observations of surfacetemperatures over large areas. Informationon surface temperatures is also provided bysatellite imagery, but in the present state ofart this information cannot be used for quan­titative temperature determination.

Among the earlier studies of the water sur­face temperatures in the Great Lakes werethose by Freeman 271 and by Horton andGrunsky.376 In both studies water tempera­ture records from daily observations at vari­ous harbor locations for the 1874-86 period

8t Appendix ~

TABLE 4-10 Comparison of Great Lakes Water Surface Temperature (Degrees Centigrade)

Jan. Feb. Mar. Apr. May June July Aug. Sept. Oct. Nov. Dec. Annual

Lalte SuperiorLake Survey (1944) 1904-43 1 0 0 0 3 4 8 12 11 8 5 1 4Millar (1952) 1935-39 2 4 7 13 12 9 6Richards & Irbe (1969) 1959-68 2 0 0 1 2 4 7 12 12 9 6 4 4

Lalte MichiganLake Survey (1944) 1904-43 1 0 0 1 4 7 12 17 18 16 11 7 2 8Hillar (1952) 1935-41 5 11 16 21 18 12 8

Lalte HuronLake Survey (1944) 1904-431 0 0 1 3 6 12 18 19 17 12 7 2 8Millar (1952) 1935-41 4 9 18 20 16 12 7Richards & Irbe (1969) 1959-68 3 2 1 1 4 8 15 18 16 12 8 6 8

Lalte ErieLalte Survey (1944) 1904-43 1 0 0 3 6 9 18 22 22 21 14 7 1 10Millar (1952) 1937-41 10 17 21 23 19 IS 9Richards & Irbe (1969) 1950-68 1 1 1 3 9 17 21 22 19 15 9 4 10

Lalte OntarioLalte Survey (1944) 1904-431 0 0 2 5 9 14 18 19 17 13 7 1 9Millar (1952) 1936-46 3 2 2 3 6 12 19 21 18 13 7 4 9Richards & lrbe (1969) 1950-68 3 2 2 3 6 12 19 21 18 13 7 4 9

lperiod shown for Lalte Survey study indicates extreme limits and not actual length of data.

1

were used as basic data. Monthly tempera­tures for the individual lakes were derived byapplying correction factors for time of obser­vations and some adjustment for open lakeconditions.

The U.S. Lake Survey825 compiled monthlywater surface temperatures for each lake fromtemperature data collected at various loca­tions by the field parties from 1904 through1943. Derived values differ somewhat fromthose given by Freeman and are considerablydifferent from Horton and Grunsky's values.

Probably the best known and most oftenused Great Lakes water surface temperaturesare those determined by Millar.s43 Millar'sstudy is based on the data obtained from con­tinuous recordings of water temperaturetaken by thermographs installed on the con­denser intakes of steamships. Data collectioncovers the 193~6period for Lake Ontario andthe 1935-41 period for all other lakes. Millardeveloped temperature distributions on eachlake by months and derived average monthlyvalues for each lake. Due to restricted naviga­tion, winter temperatures for lakes other thanOntario were either not available or gave in­sufficient coverage to derive reliable monthlymeans. Many investigators in recent yearshave used Millar's temperatures, most fre­quently to adjust surface temperatures de­rived for various periods from the water in­takes or other sources. Studies of this typeinclude Hunt,39S Snyder,751 Rodgers and An­derson,87s and Richards and Rodgers.854

Determinations of Great Lakes water sur­face temperatures, limited to a single lake,were made by several investigators.Church 142.143.144 analyzed water temper­atures for Lake Michigan, based on bathy­thermograph observations obtained during1941-44 period. He showed that irregularitiesin the water temperature distribution are theresult of strong winds and upwelling, both ofwhich act to lower the water temperature atthe surface. Another presentation of LakeMichigan temperatures for the summermonths is given by Ayers et a1.29 Their valuesare based on synoptic surveys conducted in1955. Ayers et al.28 also determined watertemperatures for Lake Huron from a similarsurvey conducted on that lake in 1954. The lasttwo studies are summarized by Ayers.25

Monthly water temperatures on Lake Erie aregiven by Powers et a1.824 who present a com­parison of long-term water intake records tooffshore cruise data.

The most recent determination of water sur­face temperatures was made by Richards andIrbe.851 Their study covers the 1950-68 periodfor Lakes Erie and Ontario, and the 1959-68period for Lakes Huron and Superior. Monthlytemperatures determined for each year on theindividual lakes were based on available in­formation from airborne radiation thermome~

ter surveys, ship observations, water intakestations, and subjective adjustments of meanlake temperatures based on mean air temper­atures from shoreline stations. The subjective

HlIdrometeorolof/1/ 88

'HERMOCLINE

.('HIMNION i

-----~

10 12 14 \6 l' 20 22

IfM'lta'Ulf .C"

6o

100

adjultments of mean water temperatures.,ere used primarily during winter months forlakes lacking sufficient temperature meas­urements.

A comparison of the Great Lakes mean.ater surface temperatures (Table 4-10),howS that surface temperatures vary withlatitude and depth of the lakes. The averageannual surface temperatures vary from 4°C(40"F) for Lake Superior, the northernmostand deepest lake, to lOoC (600F) for Lake Erie,the southernmost and shallowest lake. Aver­age annual surface temperatures on the otherlakes, with intermediate latitudes and depths,amount to SoC (46°F) for Lakes Michigan andHuron and 9°C (48"F) for Lake Ontario. Theaverage monthly surface temperatures varyfrom the winter lows of O"C (32"F) on Lake Su­perior and 2°C (36°F) on Lake Ontario to thesummer highs of 1aoC (56°F) on Lake Superiorand 2SOC (73°F) on Lake Erie.

4.5.2 Temperature at Depth

Many of the studies mentioned in the pre­ceding discussion on water surface tempera­ture also deal with the vertical temperaturedistribution in the Great Lakes. Church 1~.143

showed that the annual temperature of LakeMichigan undergoes four basic seasonal cycleswith distinct characteristics, namely, thespring warming, summer stationary, autumncooling, and winter stationary. Because ofdif­ferent latitudes and depths, the timing andduration in the lakes of these cycles are notsynchronous, but all lakes display these fourbasic seasonal periods. Occurrence of theperiods is governed by the lake temperature­water density relationship.

The seasonal changes of thermal structurein large deep lakes of mid-latitudes, such asthe Great Lakes, are shown graphically inFigure 4-96, and discussed in Section 6. As thewarming season progresses, the top layers ofwater absorb heat from the atmosphere, be­coming progressively warmer, and throughconduction of heat downward and mixing in­duced by winds, the warm layer becomesdeeper. The warm upper water (epilimnion)is separated from the cold deeper water(hypolimnion) by the thermocline. Theepilimnion is less dense and literally floats ontop of the hypolimnion. In the early stages ofdevelopment, the thermocline is rather weakand is easily broken down by wind action,which readily mixes the thin surface layer ofheated water with colder water below, thus

FIGURE 4-96 Seasonal Changes in the Ther­mal Structure of • Large, Deep Lake of Mid­latitudes

producing further development and deepen­ing of the thermocline. As the thermocline ap­proaches its maximum depth, the mixing be­comes progressively less. The final deepeningoccurs during summer storms with relativelylittle activity during calmer periods.

The summer stationary period is charac­terized by nearly stationary lake surface tem­peratures in their maximum range. At the be­ginning of this period the thermocline de­scends to its maximum depth, where it re­mains relatively stable. The establishment ofa strong thermocline at constant depth indi­cates that there is only a negligible transfer ofheat by conduction, either above or below thethermocline. Anderson and Rodgers 14 showedthat the maximum depth of the thermoclineduring this period of peak heat content isabout 15 m (50 ft) in all the Great Lakes, inspite of marked difference in their transpar­ency, configuration, size, orientation, andlatitude.

Lake cooling begins in the fall, with sub­stantial net loss of heat resulting from theinteraction of cooling and heating processes,such as radiation, evaporation, conduction,precipitation, and condensation. In the fullydeveloped cooling period of late fall, with

84 Appendi%.4

water surface temperature substantiallyabove that of maximum density, cooling of thehomogeneous surface layer proceeds rapidly,since only this layer is affected due to stabilityof the thermocline. At the same time, with theincreased frequency of storms in this seasonthe upper layer becomes deeper as well. As thecooling continues, temperature of the upperlayer approaches that of maximum density,the thermocline becomes less stable, and thewind-stirring may be complete from top to bot­tom with the destruction of the thermocline.Church1" indicates this critical temperatureof the homogeneous surface layer to be 6°C(43°F) or slightly above. In the advancedstages of cooling, depth exerts an importantcontrol on temperatures in the lakes. With theloss of the thermocline the entire column ofwater becomes isothermal and further coolingat the surface proceeds slowly because the en­tire water column loses heat.

During the winter stationary period the lakewaters are again at less than maximumdensity, but in this case the surface is colderthan 4°C. In coastal areas and lakes with shal­lower depths temperatures are generallyisothermal vertically. In deep waters an in­verse stratification develops, with colderepilimnion and slightly warmer hypolimnion.This stratification is not as pronounced as dur­ing the summer, because the downward mix­ing process is aided by strong winds and con­vection produced by daytime heating duringwinter. Water temperature in the entire col­umn or the deep upper layer, whichever thecase, approaches freezing point on the lakeswith extensive ice cover, and stays at aboutZOC (3~F) on the lakes which are largely free ofice. Because great depths are affected, watertemperature changes during this period arenaturally slow. Thus, Lake Erie with its shal­low depths freezes sooner and more often thanLake Superior, which because of its greatdepths is capable of sustaining tremendousheat losses. By the same token, ice breakup onLake Erie is much faster.

Since each lake consists of a whole range ofdepths, which generally increase from shallowcoastal waters to a maximum depth in mid­lake, each lake contains water masses withdistinct thermal structures characteristic oftheir depth, particularly during warming andcooling periods. In the beginning stages of thespring warming period the shallow watersalong the shore warm up much faster thanmid-lake areas, producing large horizontaltemperature gradients near the boundary be­tween the warm and cold surface tempera-

tures. The coastal waters are well above 4°C atthe surface and vertically stratified, Whilethose in mid-lake are less than 4°C and uni­form in temperature from top to bottom indepths as great as 180 m (600 ft) (Church IQ

and Rodgers8 'l'2). The most extensive investi_gation of this phenomenon, named the ther­mal bar, was made on Lake Ontario by Rod­gers.lI13 As the warming season progresses thethermal bar zone moves lakeward from theshores and eventually disappears. During thefan cooling period the process is repeated, withthe cold temperatures moving offshore to­wards the center of the lake.

In the above discussion of seasonal temper­ature changes within a column of water, noconsideration was given to any movement ofwater into or out of the column due to thenormal lake currents. Actually, water tem­peratures and currents in a lake are closelyrelated. Any lake is subject to water move­ment resulting from the inflow-outflow bal­ance and the general water circulation in­duced by winds and geostrophic forces. Thus,once established, a thermocline seldom staysstill, but fluctuates in a wave-like motion (Sec­tion 6). Rodgers 8 '72 states that the thermoclinemoves up and down through a distance of10 to20 m (30 to 60 ft), with wavelengths of tens ofkilometers. He also states that these internalwaves are seldom evident to the casual ob­server and their detection requires continu­ous observation of temperatures at one ormore locations within the lake.

Periodically, strong winds blowing steadilyfrom one direction may tilt the thermocline,deepening the epiIimnion on the downwindshore and reducing its depth on the upwindshore. Prolonged strong winds pile up warmsurface water and produce sinking on thedownwind shore, while at the same time theyremove warm surface water and produce cor­responding upwelling of cold water fromdeeper layers on the upwind shore. Upwellingoccurs frequently on the northwest shorelineof Lake Ontario and the west shores of LakeMichigan. Tilting of the thermocline can beobserved from the water temperature recordsof municipal water intakes located on the op­posite shores of the lake.

4.5.3 Air-Water Temperature Relationship

The difference between air and watersurface temperatures is the primary indicatorof the atmospheric stability over the lakes.When the air is warmer than the water s~r-

Hlldro1'MteoroloW 85

-8

.10 '-- --'

FIGURE 4-97 The Relationship of Air-WaterTemperature Differences in the Great Lakes

tario, Hunt" used mean values from threeperimeter stations (Toronto, Oswego, andTrenton) for the 1937-56 period to obtain aver­age air temperature over the lake. Water sur­face temperatures for this lake were obtainedby adjusting values given by Freeman,lnU.S. Lake Survey,815 and Millar. l43 In thesame study, for the northern Lake Michiganair temperature, Hunt used St. James andBeaver Island data from 1911 to 1956; watersurface temperatures were obtained from rec­ords in the vicinity of Beaver Island from un­stated sources. In a study on Lake Erie,Hunt 398 used the water surface temperaturesgiven by Freeman, U.S. Lake Survey, and Mil­lar, and the normal air temperatures fromfour perimeter stations (Detroit, Cleveland,Erie, and Toledo). In a second study for LakeOntario, Rodgers and Anderson 875 used Mil­lar's water surface temperatures and normalair temperatures from Toronto and Rochester.

In all cases except Lake Michigan, air tem­perature over the lakes was determined fromperimeter stations and may be considerablydifferent than at mid-lake, as pointed out byRodgers and Anderson in their study. Becauseland area is more sensitive to both heating andcooling, these air temperatures obtained atperimeter stations would tend to intensify theextreme conditions, being higher than at mid­lake during summer and lower during winterperiods. Air temperature data for Lake Michi­gan represent conditions over an island andmay be more representative. The water sur­face temperatures, except those by Millar,were also determined mostly from perimeterstations and would tend to have the same de­ficiencies, but probably not to the same de­gree. Thus, the air-water relationships shownprobably over-accentuate the magnitudes be­tween these temperatures.

The monthly values of air-water tempera­ture difference indicate that there are fourperiods of prevailing atmospheric conditionsover the Great Lakes. Duration and timing ofthese periods vary for different lakes, depend­ing primarily on latitude. Generally, the air isnormally warmer than water during springmonths of April, May, and June (also July innorthern areas), and the atmospheric condi­tions are stable. During mid-summer monthsofJuly and August the air and water tempera­tures are approximately the same, thus indif­ferent or adiabatic (occurring without loss orgain of heat) equilibrium conditions exist inthe atmosphere. From early fan to mid·wintermonths (September through February) the airis normally colder than water and the atmos-

o N 0.. " M

face, the air will tend to be stable. Conversely,when the air is cooler than the water surface,the air will tend to be unstable. Thus, otherthings being equal, the greater the positivedifference between the air and water surfacetemperatures, the more stable the atmos­phere and smaller the opportunity for the oc­currence of overwater precipitation, highwinds, and evaporation. On the other hand,the greater the negative difference betweenthe air and water surface temperatures, themore unstable the atmosphere and greaterthe opportunity for the occurrence of theabove climatic processes.

Atmospheric stability over the Great Lakesvaries appreciably during the year. The air isnormally warmer than the water surface dur­ing spring and colder for a somewhat longerperiod in fan (Figure 4-97). Because of theshortcomings in data used for air-water tem­perature differences, their magnitude maynot be representative of the actual mid-lakeconditions, and considerable variation in themagnitude shown at times for the same lakeseems to indicate this deficiency. For Lake On-

86 Appenditr '"

phere is unstable. However, during wintermonths the presence of ice and snow cover hasa significant modifying effect on the air-watertemperature relationship. Finally, during latewinter in March (also April in northern areas)the air and water temperatures are againabout the same, and the adiabatic equilibriumconditions prevail. The adiabatic equilibriumconditions occur during relatively short,transitory periods and variation in the atmos­pheric conditions during these two periodsmay be considerable. In contrast, the springstable conditions and the fall unstable condi­tions present long, well-established periods.

4.6 Humidity

4.6.1 Lake Perimeter Humidity

The influence of the lakes produces higherand more stable humidity in the Great Lakesarea than at similar latitudes in the mid­continent. Since warmer atmosphere is capa­ble ofholding more water vapor, the amount ofwater vapor present in the atmosphere variesconstantly with temperature and availabilityof moisture. However, this discussion is con­cerned with the relative humidity or the ratiobetween the actual vapor pressure and thesaturation vapor pressure at the same tem­perature. Since all lakes provide large quan­tities of moisture through evaporation, theupper Great Lakes with lower temperaturesand corresponding lower dew points attainsomewhat higher values of relative humidity.Prevailing winds and lake breezes are impor­tant factors in raising or lowering humidityvalues on land areas adjacent to the lakes.

Humidity measurements on lake perimetersare provided by the first order meteorologicalstations located around the lakes. Humidityvalues at these stations are generally pub­lished for the four daily synoptic hours (1:00and 7:00, a.m. and p.m.), but hourly values arealso available. Data from these stations arethe sole source ofcontinuous humidity recordsfor extended periods of time. Because of thelake effect on adjacent land areas, various in­vestigators have utilized these data to obtainestimates of humidity over the lakes by av­eraging records from several perimeter sta­tions. Some of the more recent studies alsoemployed correction factors to adjust these es­timates to overlake conditions. The correctionfactors were derived from infrequent overwa­ter measurements, similar to those used forwinds, and are discussed later.

TABLE 4-11 Avenge Perimeter Humidity forthe Great Lakes (Percent)

LakePeriod Superior Michigan Huron Erie DDtario

January 77 76 81 77 78February 76 73 79 77 77March 74 73 76 74 74AprH 69 69 73 70 69May 68 66 70 69 68June 72 69 74 70 70July 74 71 74 71 68August 76 74 77 74 72September 79 76 7. 75 74October 76 73 79 75 74Nov"'er 78 76 83 78 78Dec"'er 79 78 83 79 78

Annual 75 73 77 74 73

Values are based on mean data published in 1969 for thefollowing stations:

Superior: Sault Ste. Marie. Marquette. Duluth. aDd:'bUDder Bay.

Michillan: Milwaukee. Huskegon, and Green Bay.Huron: Alpena, Gore Bay. and Wianon.Erie: Toledo, Cleveland. Buffalo. aad London.Ontario: Rochester, Syracuse, Trenton, and Toronto.

An estimate of the average monthly and an­nual humidity values for the individual GreatLakes, based on data from perimeter stations,is given in Table 4-11. The perimeter humidityfor all lakes increases from a low of approxi­mately 70 percent in the spring to a high ofapproximately 80 percent during the late fall.Average annual humidity varies from 73 per­cent for Lake Michigan to 78 percent for LakeHuron, with intermediate values for otherlakes.

Examination of records for the individualstations around the lakes indicates a daily var­iation and a general northward increase inhumidity. Based on four observations a day,highest humidity normally occurs late atnight and during early morning hours <1:00and 7:00 a.m. readings), while lowest humidityoccurs in the early afternoon (1:00 p.m. read­ing). At most stations average annual relativehumidity values for the night and morningreadings range between 75 and 85 percent,with the maximum values occurring duringsummer. The afternoon readings range be­tween 60 and 70 percent, and are lowest in thespring and summer. At most locations averagedaily range in humidity is from 5 to 10 percentduring winter and from 15 to 20 percent duringsummer.

.(.6.2 Overwater Humidity

The humidity records from perimeter sta­tions contain both lake and land effects and

TABLE 4-13 Average Perimeter Precipita-tion for the Great Lakes, 1937-1969 (cm)

Lalr.ePeriod Superior Michigan Huron Erie Ontario

January 5.5 4.8 6.7 6.5 6.eF"bruarj 4.1 3.9 5.1 5.7 6.4Karch 4.4- 5.0 5.3 6.9 6.5April 6.0 7.3 6.5 8.5 7.2Kay 7.6 7.6 7.0 8.2 7.5June 9.0 8.6 7.2 8.3 6.3July 7.2 7.5 6.7 7.7 7.2August 8.7 7.7 7.5 8.1 7.4September 8.6 8.4 8.2 7.1 7.0October 6.4 6.2 6.9 6.8 6.9November 6.8 6.3 7.7 7.4 7.5December 5.5 4.8 7.3 6.5 7.0

Arlnua1 79.8 78.1 82.3 87.7 83.7

4.7.1 Lake Perimeter Precipitation

4.7 Precipitation

H1IdrometeoroloVII 81

ference in time is related to the number ofdaily observations, and indicates that four ob­servations a day are not sufficient to obtainthe daily humidity distribution.

!(ate: Based on data assembled "w the Lake Survey Center.NOAA.

Precipitation includes all forms of moisturedeposited on the earth surface from the at­mosphere. The principal forms ofprecipitationinclude rain, hail, sleet, and snow, all of whichare readily measurable. Of particular interestin the Great Lakes Basin is the precipitationmeasured at lake perimeters. Investigationsof precipitation distribution indicate thatperimeter stations show marked variationfrom precipitation further inland, and sincedirect overwater measurements are generallynot available, it is assumed that perimeter ob­servations are sufficiently representative ofoverwater conditions (e.g., Freeman 271).

Estimates for the average monthly and an­nual precipitation on individual lakes duringthe 1937-69 period are shown in Table 4-13.The annual precipitation varies from 78 cm(30.8 inches) for Lake Michigan to 88 cm (34.5inches) for Lake Erie, with an overall averagefor all the lakes of 81 cm (32.0 inches). Annualprecipitation increases from north to southand from west to east. The southward increasein precipitation is climatic, since warmer at­mosphere is capable of sustaining more mois­ture, while the eastward increase is caused bythe lake effect, since additional moisture issupplied to the atmosphere by the lakes as

Jackson(1963)

1959-1962

Richards & Fortin(1962)

1959-1961..riod-January 1.33 1. 25

february 1.30 1.24

.rcb 1.21 1.22

April 1.14 1.04.86 .89MaY.94 .94June

July 1.09 1.10

August 1.09 1.10September 1.11 1.09October 1.15 1.14November 1.15 1.13December 1.31 1. 28

Annual 1.14 1.12

_ABLE 4-12 Lake-Land Humidity Ratios for..Great Lakes

are not necessarily representative of the ex­tensive water areas included in the GreatLakes. The major controlling factor of humid­ity is the air temperature, and temperaturesover lake and land areas differ. Overwaterhumidity data are obtained either from directoverwater measurements, which are availableonly on an intermittent basis, or from empiri­cal relationships derived from those meas­urements. Such relationships combine manyof the differences between overwater andoverland conditions into a single correctionfactor, a lake-land humidity ratio (Richardsand Fortin,650 Jackson 420). The monthlyhumidity ratios derived in these studies areshown in Table 4-12. They indicate that on anannual basis overwater humidity is some 10 to15 percent higher than overland humidity atperimeter stations. During spring, overwaterhumidity is approximately 10 percent lowerthan perimeter humidity, but during the restof the year overwater humidity is higher, witha maximum difference of approximately 30percent in the winter.

The average daily variation of the humidityratios presented in the studies shows highhumidity ratios during the night and lowratios during daytime hours. The nighttimemaximum occurs generally between 1:00 and4:00 a.m. and the daytime minimum occursaround noon. Richards and Fortin, based onfour daily observations, indicate lowesthumidity ratios at 1:00 p.m., while Jackson,using eight daily observations, shows the low­est ratio at 10:00 a.m. Inspection of theirdiurnal variation curves shows that the dif-

88 Appendix '"

..

4··I ---i---::---).'-4-_-...A

•*.,

... ...

N· ... n·

FIGURE 4-98 Mean Monthly Number of Days with Measurable Precipitation (open bars) andThunderstorms (shaded) in the Great Lakes Basin

From U.S. Weather Bureau. 1959

they are exposed to the prevailing westerlywinds.

The seasonal precipitation pattern showswell-distributed and abundant precipitationthroughout the year, although a larger por­tion of the annual supply falls during thesummer months, a characteristic of continen­tal climates. The relatively high summer rain­fall is especially pronounced on the westernand northern lakes. The average monthly pre­cipitation increases from a winter low of 4 cmto 5 cm (1.6 to 2.0 inches) in the upper lakes and6 cm to 7 cm (2.4 to 2.8 inches) on the lowerlakes to a summer high of8 cm to 9 cm (3.1 to 3.5inches) on all lakes.

The mean number of days per month withmeasurable precipitation at perimeter sta­tions increases from the windward to the leesides of the lakes, and also increases generallyfrom summer to winter months. Exceptions tothis seasonal distribution occur along thewestern and northern shores of Lakes Michi­gan and Superior. The number of days with

measurable precipitation and thunderstormsat perimeter stations is shown in Figure 4-98.Highest thunderstorm frequencies occuralong the western shore of Lake Michigan andthe southern shore of Lake Erie.

During the winter months precipitation inthe Great Lakes Basin is largely in the form ofsnow. In the northern areas it generally con­sists of snowfall exclusively, with permanentsnow cover throughout the winter. In thesouthern areas precipitation alternates be­tween snowfall and rainfall, with intermittentsnow cover on the ground.

4.7.2 Overwater Precipitation

Observations have indicated that largebodies of water, such as the Great Lakes, mod­ify the atmosphere above them, including pre­cipitation patterns. One theory explaining thereduction of overwater precipitation in the

·summeris that the water cools the air above it.

. r precipitation measurements on then(~il to confirm this in all cases. Horton

GI nsky 376 and Verber 848 suggested thate::: overwater precipitation in the sum­ay be due to less thunderstorm activ­

;':is results from the cooling effect of the• which produces a more stable atmos­

I, over the water than over surrounding~reas.Byers and Braham 117 .ma.de a simi­

observation about Lake MIchIgan and~ed that in the winter the warm~n~ef~ectoftil lakes encourages greater precIpItatIOn on

e lee shore than on the windward shore.~arson 599 from a study of precipitation ob­rvatio~s by radar, found that the formation

=air-mass showers over Lake Michigan in the.ummer was inhibited.

The elements affecting overwater precipita·tion in the winter are less definite. It is appar­ent that the warming effect of the lakes en­courages snow flurries, but whether the pre­cipitation is less, equal to, or more than thatfalling on the adjacent shorelines is a matterof controversy. Light precipitation belts onwindward sides and heavy precipitation beltson the lee sides of the lakes are well estab­lished, but the quantities recorded at theselocations do not necessarily yield representa­tive overwater precipitation. The observedheavy snowfall on the lee shores of the lakesmight be confined to the lake perimeters andwould then represent an accumulation of pre­cipitation resulting from the lake-land in­teraction. The elevation of the air mass as itmoves from the water to the land surface,coupled with the air movement from the warmwater to the cold land during winter, combineto cause more precipitation on the lee shores.This may apply to the islands as well. Wintermeasurements are also less accurate becausesnow is more sensitive to wind, increasing theeffects of exposure, so the gages do not ade­quately measure winter precipitation. Freez­ing of the lakes complicates the process fur­ther.- Recognizing the shortcomings of perimeterobservations for estimating overwater pre­cipitation, several investigators derived rela­tionships of overwater to perimeter precipita­tion, utilizing data from islands to representoverwater conditions. However, island datamay not be reliable for this purpose, and re­sults of the studies are often contradictory.

Based on records from Beaver Island in theLake Michigan and North Bass Island in LakeErie, Horton and Grunsky 376 concluded thatprecipitation on the lakes was lower than atperimeter stations. They calculated seasonal

Hrdrometeorolon 89

lake-land precipitation ratios, which indicatethat precipitation on Lakes Superior, Michi­gan, and Huron (Beaver Island) averages 93percent of that measured at shore stationsduring winter months, and 94 percent duringsummer months. For Lakes Erie (North BassIsland) and St. Clair their overwater precipi­tation amounts to 84 percent of perimeter pre­cipitation for winter and 85 percent for sum·mer months. However, Day205 suggests thatthe differences in island and shore precipita­tion are due to wind reduction of precipitationgage catch at the more exposed island sites,rather than any real deficit in precipitation onthe lakes.

To study overwater precipitation a storage­type precipitation network was established in1952 on a number of islands in northern LakeMichigan by the U.S. Lake Survey and theU.S. Weather Bureau. Based on twice-a-yearprecipitation records from this network andmonthly records from Beaver Island and adja­cent shore stations, Hunt 395 concluded thatann ual overwater precipitation on Lakes Mich­igan and Ontario averages 79 percent of thatmeasured at perimeter stations, with monthlyvalues varying from 60 percent in August to 91percent in November. Hunt assumed precipi­tation at the smallest, exposed island to betrue overwater precipitation. Kohler"l ques­tioned that assumption and indicated that therelative catch of the island gages is highly cor­related with windiness, and that virtually allthe differences in precipitation catches couldbe explained by relative windiness at the gagesites. In another precipitation study of thenorthern Lake Michigan island network,Kresge, Blust, and Ropes 476 agreed withKohler and used the higher measuredamounts at both island and land gages as trueoverwater and shore precipitation and cor­rected the other gage records for gage expo­sure. The annual overwater precipitation wasabout the same (102 percent) as precipitationfrom shore stations. Seasonally, overwaterprecipitation ranged from 3 to 10 percent lessthan perimeter precipitation in the summer,depending on the offshore and onshore winds,respectively; and 9 percent more in the winter.

In a Lake Michigan precipitation studybased on records from the Four-Mile Crib inthe southern tip of the lake and land gages inthe Chicago area, Changnon137 determined alake-land precipitation relationship similar tothat derived by Hunt. Changnon's lake-landprecipitation ratios show monthly variationsranging from 78 percent in October to 95 per-

90 Appendiz 4

IUpChurch ratios are ba.ed on inland stations located 160 km westof the lake (thiS appendix).

TABLE 4-14 Lake-Land Precipitation Ratiosfor Lake Michigan and Lake Erie

Irelae I

Upchurch1

Hunt Chanlnon Et aJ. Derecki(1959.) (191)1) (1963) (1970) (191)4)

.l5 years 11 years 30 years 6 years 13 yearsPeriod 1911-56 194~-51> 1906-62 1963-68 1920-46

January .88 .95 1.13 1.39 .95February .85 .81> 1.13 2.26 .89llarch .81 .84 1.07 1.05 1.03April .78 .81 1.01 .89 1.04llay .75 .87 .96 .83 1.07June .73 .79 .93 .49 .92July .69 .82 .90 .70 1.04August .60 .87 .91 .64 1.03September .74 .83 .98 1.10 .95October .89 .78 1.05 .96 .89November .91 .79 1.10 1.62 1.02Occ~1Ilber .89 .89 1.13 1.32 1.03

Wlr.ter(Nov-Apr) .85 .86 1.09 1.42 .99SUIl"'JDer(llay-Oct) .73 .83 .96 • :9 .98

Annual .79 .84 1.02 1.10 .99

cent in January, with an annual average of 84percent.

The lake-land precipitation ratios calcu­lated for Lake Erie (Derecki 214) show consid­erable monthly variations, without a definiteseasonal trend. Overwater precipitation wasbased on records from Pelee, South Bass (Put­in-Bay), and Catawba Islands, and land pre­cipitation on records from surroundingperimeter stations. The resultanf"annual ratiowas 99 percent with monthly ratios that var­ied from 89 percent in February and October to107 percent in May.

In 1963 the U.S. Lake Survey, in cooperationwith the U.S. Weather Bureau, modified theirexisting precipitation network in northernLake Michigan by replacing the storage-typegages, which were read twice a year, with pre­cipitation recorders producing hourly read­ings. Similar gage networks were establishedin western Lake Erie in 1964 and eastern LakeOntario in 1969. Using recorded data fromLake Michigan islands and land stations lo­cated some 160 km (100 miles) west of the lake(Figure 4-18), Upchurch 808 derived lake-landratios which vary from 49 percent in June to226 percent in February, with an annual aver­age of 110 percent.

The monthly precipitation ratios derived inthe above studies are tabulated in Table 4-14.Monthly ratios given by Hunt 395 and byKresge et a1. 478 represent values fromsmoothed annual graphs, while those ofothersare arithmetic averages for the number of

Present methods of measuring precipitationhave many shortcomings. such as use of pointmeasurements to represent an area, varia­tions in gage catch, accuracy, effects ofwindi­ness and exposure on the catch, and access orinstallation difficulties in remote areas onlarge bodies of water. A potentially powerfultool for eliminating some of these problemsand obtaining more truly representative pre­cipitation data may be the use of weatherradar. Weather radar is applicable to bothland and water areas, but it is of particularinterest for large lakes because of gage instal­lation and access difficulties.

The use of weather radar for obtainingquantitative precipitation data requiresclimatological analysis of photographed pre­cipitation echo patterns. The process involvesuse of a grid overlay on radar photographs,counting echo occurrences, and correlatingwith measured precipitation. Current use ofradar precipitation observations, althoughpromising, is not advanced sufficiently to pro­vide usable, quantitative data. Poor perform­ance is attributed mainly to weak radarequipment, which displays a decrease of echooccurrences per volume of rainfall outwardfrom the radar location, thus limiting the ef­fective range (Bruce and Rodgers 108). Morepowerful radar equipment is required. Fur­ther developments in radar technology incombination with high-speed computers mayprovide an ideal answer to the overwater pre­cipitation measurement problem.

years used in their derivations. The Upchurchdata are not comparable directly with otherdata in the table, because Upchurch used re­mote inland gages while all others used lakeperimeter gages. The inland stations, how­ever, indicate more correctly the lake effectson precipitation distribution throughout theyear.

4.7.3 Weather Radar

4.8 Evaporation

4.8.1 Determination of Evaporation

Evaporation from the lakes is the loss ofwater from the lake surface to the atmospherein the form of water vapor. Considering lakeand land areas of the Great Lakes Basin, two­thirds of the water supplied by precipitation is

Lake ErieLake 1I1ch1&an

lost by evaporation. Evaporation losses fromthe water surface of the lakes, where the sup­ply of moisture is continuous, are substan­tially higher than from the land and amount toapproximately three-quarters of the overwa­ter precipitation. Thus, it is readily apparentthat evaporation has a great effect on theavailability of water and on the heat budget ofthe lakes, since evaporation is basically a cool­ing process.

There is no direct method to measure evap­oration from large water bodies. Since the ac­tual evaporation losses are dependent directlyon meteorological factors, it is possible to de­velop methods that use hydrologic andmeteorologic data and permit determinationof evaporation losses from the lakes with ac­ceptable accuracy. These methods includewater budget, mass transfer, energy budget,evaporation-pan observations, atmospherichumidity budget, and momentum transfer.The first four ofthese methods have been usedto compute evaporation from the Great Lakes.

The water budget method consists of solvingthe water budget or mass-balance equation,described at the end of this section, for theunknown evaporation component. All othermajor components of the water budget neces­sary to compute evaporation frem the GreatLakes are either measured directly or can beestimated from related measurements. This isthe only direct method of computing evapora­tion estimates and has been used in variousstudies to provide control for other methods,which require determination of empirical con­stants. Evaporation, as determined by thewater budget method, is a residual of severallarge factors and includes the errors of thesefactors, which may affect the computed evap­oration values considerably. Care must beexercised to reduce these errors to a minimumby using all available data and considering theeffects of the lakes on some of them, such as onoverwater precipitation.

The mass transfer method of computingevaporation is a modified application of Dal­ton's law, where evaporation is considered tobe a function of the wind speed and the differ­ence between the vapor pressure of saturatedair at the water surface and the vapor pres­sure of the air above. A summary of thetheoretical development of the mass transfermethod and a review of the equations de­veloped by various investigators is given byAnderson et a1.17 The more promising of theseequations were tested by Marciano and Har­beck.512 The problem in applying this methodto the Great Lakes is that climatological data

Hydrometeorology 91

for any appreciable period of time are almostexclusively restricted to the perimeter landstations, which may be and often are not rep­resentative of open-lake conditions. Varia­tions in air stability that affect both wind andvapor pressure are essentially diurnal incharacter over land and seasonal over water.The required adjustments for perimeter data,or lake-land ratios for wind and humidity havebeen made in recent years and are being im­proved as more overwater data are collected.Thus, the mass transfer method of computingevaporation from the Great Lakes holds greatpromise for the future. Its primary advantageis the elimination of the main objections of thewater budget method, namely, uncertaintieswith respect to ground water and dependenceofcomputed evaporation on large factors, suchas inflow and outflow from the lakes.

The energy budget method requires deter­mination of the energy exchange between abody of water and the atmosphere, which in­cludes such factors as the net solar and atmos­pheric radiation, conduction of sensible heatto the atmosphere, energy utilized by evap­oration, net advective energy, and energystorage within the body ofwater, disregardingsome minor heat sources or sinks (chemical,biological, exchange with bottom sediments).The water loss is determined from the relatedenergy or heat loss by evaporation. Detaileddiscussion of various terms comprising theenergy budget and its practical application ispresented by Anderson.16 However, in theGreat Lakes this method has been used in­frequently because of the difficulty in obtain­~ng data on energy components.

A convenient and inexpensive method of ob­taining evaporation estimates, although fre­quently questioned on theoretical grounds, isthat of evaporation-pan observations, whichutilize observed water losses from evapora­tion pans and experimentally determinedpan-to-lake relationships. The ratios of panevaporation to lake evaporation, or pan coeffi­cients, vary depending on pan characteristics.An extensive investigation of the relation­ships between pan and lake evaporation wasreported by Kohler et a1.462

4.8.2 Evaporation from Lakes

To compute evaporation from the GreatLakes various investigators have used one ormore of the four methods described above. Thewater budget and mass transfer methods wereused most frequently. There are only two

9£ Appendix.+

known studies which utilize the energy budgetapproach. The evaporation-pan observationmethod has also been used infrequently. Re­sults obtained by various investigators oftendiffer considerably, especially for shorter,monthly periods. Some of this variation is nat­ural, reflecting variability of evaporation be­tween the various periods of record involved,but some of it, especially in extreme cases, isdue to computation procedures and inac­curacies of basic data. More recent determina­tions use better basic data and should be moreaccurate.

Among the earliest methods used to deter­mine evaporation from the Great Lakes is thatof evaporation-pan observations. Henry841concluded that the ratio of evaporation fromfloating pans to land pans was approximately0.5 and used this ratio to estimate lake evap­oration. Hickman 355 described experiments inwhich water temperature in the pan wasmaintained at the lake surface water temper­ature, and the pan-lake ratio or pan coefficientwas assumed to be 1.0. The latest estimates ofevaporation from the lakes by this method(U.S. Weather Bureau828) give pan coefficientsranging from approximately 0.75 in the south­ern areas of the Basin to 0.80 in the northernareas. The above studies give annual evapora­tion estimates, without. breakdown into sea­sonal or monthly amounts.

The water budget is one of the traditionalmethods of determining lake evaporation(Russell,692 Freeman,271 Pettis,60S Hunt,395Brunk,lll and Derecki 213.214). In some studies(Hunt, Derecki) precipitation from perimeterstations was adjusted by lake-land ratios torepresent overwater conditions. Others usedunadjusted perimeter precipitation. Laterstudies showed that Hunt's reduction wasprobably too extreme, resulting in somewhatlower evaporation than indicated by most ofthe other recent studies for Lake Ontario. Be­cause Lakes Michigan-Huron have a commonoutlet (St. Clair River), water budget determi­nations for these lakes can be made only forboth lakes as a single unit.

Probably the first mass transfer determina­tion of Great Lakes evaporation is that givenby Freeman 271 who used a relatively simpleformula similar to those used in all subsequentGreat Lakes mass transfer studies. Otherevaporation studies, mostly of Lake Ontario,which employed mass transfer methods arethose by Horton and Grunsky,376 Hunt,395Kohler,461 Snyder,751 Bruce and Rodgers,IOsRichards,us Richards and Rodgers,654 andRichards and Irbe.651 Richards648 modified the

mass transfer equation by introducingmonthly wind and humidity ratios for adjust­ing data obtained from land stations. In previ­ous studies wind data from perimeter stationswere being adjusted to overwater winds basedon seasonal periods. Because of data limita­tions, most studies mentioned above employedinconsistent periods of record to determinevarious factors of the mass transfer equation.In some of them, the average evaporation val­ues are based on only a few years of record,which is much too short to establish reliablelong-term trends. The first mass transfer de­termination of evaporation with consistentperiods of record for all factors was made forLake Ontario by Richards and Rodgers.654This study was extended to include other lakesbordering on Canada by Richards and I rbe.651

Evaporation studies by the energy budgetmethod conducted on the Great Lakes are lim­ited to Lake Ontario (Bruce and Rodgers,lOSand Rodgers and Anderson 675). The results ofthese studies are practically identical. The au­thors concede that their estimates are highdue to inaccuracies of data used, and they dis­cuss the possible errors. Continuing researchon interaction between the atmosphere andlake surface should enable more direct evalua­tion of the energy exchange factors, reducetheir dependence on empirical relationships,and improve the accuracy of evaporation es­timates determined by the energy budgetmethod.

Evaporation from the Great Lakes varieswith latitude and depth. The warmer, lowerlatitudes provide greater evaporation oppor­tunity, and the lake depths govern heat stor­age capacity. The influence of lake depths ismainly seasonal. Deeper lakes warm and coolmore slowly, retarding the seasonal low andhigh evaporation losses. The depths of theGreat Lakes coincide with latitude; Lake Su­perior is deepest (410 m) and Lake Erie is theshallowest (65 m), while the centrally locatedlakes have approximately similar inter­mediate depths (230 to 280 m). Thus, evapora­tion from the lakes increases from north tosouth, being lowest for Lake Superior andhighest for Lake Erie. The centrally locatedlakes, Michigan, Huron, and Ontario, have in­termediate evaporation rates.

The average evaporation values for individ­uallakes, Obtained in some ofthe better knownor more recent studies mentioned in the pre­ceding paragraphs, are listed in Table 4-15.The table also shows the methods used, thesource, and the periods of record, although, formethods other than water budget and the last

H 1Idrometeorolog71 98

TABLE 4-15 Comparison of Great Lakes Evaporation (Centimeters)

Jan. Feb. Mar. Apr. May June July Aug. Sept. Oct. 1Iov. Dec. Annual

8.9 7.9 6.6 5.8 0.3 0.3 1.5 -0.3 3.38.9 6.6 5.8 -0.8 -1.0 -1.0 -0.5 2.8 5.8

9.1 8.4 6.9 2.5 0.0 0.0 0.0 3.0 4.37.6 6.4 4.8 1.8 -1.8 -5.8 -7.4 -0.5 5.6

11.7 9.1 5.8 1.5 1.0 -3.6 -8.1 -4.1 3.3

5.3 7.9 9.7 57.26.9 9.7 9.7 52.8

7.1 6.9 8.1 56.46.4 6.9 8.4 32.36.4 10.4 12.2 45.7

73.257.9

64.810.4 10.2 10.7

8.6 6.6 6.68.6 6.9 7.6

7.9 9.93.0 7.9

0.5 4.8

7.6 7.1 4.1 2.3 1.5 1.87.6 6.6 4.1 1.0 -2.3 -3.0

8.4 2.8 4.3 3.6 -0.3 -0.3

1M! SU1'1II0Il

tam IIUIIGETrr...-a (1926) 1921-25Derseti (1965) 1939-59

IMSS TIAIl5Fnrr_ (1926) 1900-24Sny....r (1960) 1921-50IichArds , Irbe (1969) 1959-68

~ MICHIGAN

IWlS TRANSFERPr_n (1926) 1900-24Snyder (1960) 1921-50

~ MICHIGAN-BURON

WATIlll lIIlIlCETrreeaan (1926) 1915-24

~ lIUR01'IWlS TRANSFERPre_ (1926) 1900-24Snyder (1960) 1921-50Richards , Irbe (1969) 1959-68

7.9 7.9 5.1 2.0 2.0 2.57.9 6.9 4.3 1.3 -1.5 -2.8

11.9 8.9 4.6 -0.8 0.0 -2.0

7.9 10.23.3 8.30.5 5.1

7.1 6.1 7.18.1 6.9 8.4

11.4 11.7 11.7

74.961.211.6

~~

WATEll. BtllGI:THunt (1959a) 1934-53 6.1Morton , ~berl (1959) 1934-52 8.1

MASS TRANSFEll.Pre_ (1926) 1900-24 6.6Hunt (1959a) 1937-52 7.1Snyder (1960) 1921-50 7.9Richards' lrbe (1969) 1950-68 9.7

ENERGY BllDGETRogers' Anderson (1'61) 1958-60 12.4

2.8 12.71.5 1.5

6.6 2.~

5.1 2.54.1 1.3

0.0 11.7 8.9 10.2

83.384.3

63.879.5

86.9

80.562.769.671.6

106.485.990.9

9.17.1

6.18.97.6

6.16.17.68.9

9.711.2

10.7

7.19.7

13.7

12.212.2

9.4 10.9

13.510.915.7

15.013.7

10.9 6.18.1 7.19.4 6.69.1 8.9

11.4 10.411.7 11.4

13.716.0

16.012.714.2

4.8 H.49.4 13.7

17.0 18.39.4 12.27.1 11.2

0.82.8

2.8 9.4 13.0 11.70.3 4.8 6.4 9.70.3 5.8 10.2 10.70.3 4.8 8.1 10.9

-0.5 2.8 7.9 9.70.5 4.6 10.2 10.7

3.6 1.91.3 2.0

-3.0 5.8

-3.8 0.00.0 1.0

1.8 1.0 0.52.0 2.0 1.3

5.1 0.5 -1.8

4.1 2.5 0.85.6 2.0 0.04.3 1.5 -1.54.3 -1.8 1.0

3.05.8

8.9

6.65.66.97.4

4.65.3

~!!!!

WATIlll BllDGETPreesan (1926) 1951-24Derecki (1964) 1937-59

MASS TRANSFERPre...n (1926) 1900-24Snyder (1960) 1921-50Richards , Irbe (1969) 1950-68

NOTE: EnerBY budget and _liS transfer IItudiell (except IUcbards , lrbe. 1969) uad vadable periods of record for variousfactors involved.

mass transfer study (Richards and Irbe,651)these periods are only approximations. Re­sults of most studies show reasonable agree­ment between different computations.

The average long-term seasonal variation ofevaporation from the Great Lakes is shown inFigure 4-99, presented as smoothed graphsbased on studies indicated in Table 4-15.Lakes Michigan and Huron are included in asingle graph, because ofcommon water budgetcomputations and limited determinations byother methods. The available information in­dicates that the evaporation from these lakesis similar and compares with that of Lake On­tario. In view of the latitude and depth dis­tributions of these three lakes, this similarityappears to be quite reasonable.

The long-term average annual evaporation

from the Great Lakes amounts approximatelyto 65 cm (25 cm (25 in.) The following are esti­mates ofannual evaporation for the individuallakes: Superior, 55 cm (21 in.); Michigan­Huron, 65 cm (26 in.); Erie, 85 cm (38 in.); andOntario, 70 cm (28 in.). These estimates werebased on several studies and should be consid­ered as indicators of the long-term averagevalue. During individual years annual evap­oration may vary considerably from the long­term values listed above.

Seasonally the low evaporation normally oc­curs in the spring, when the water tempera­ture is close to or even below the dew pointtemperature of the air. These low evaporationrates vary from slight evaporation to conden­sation. With rising water temperatures evap­oration increases until it reaches the high

94 Appendix"

II>ANNu"t f:VAPORATION

/- •sUPUIOlil-S~c:fftl'''" l I \I. MltH - HURON -65c:ftl,(26,n 1

ERIE - esc,," (331ft) / \\ONtARIO -1Oc",(28lftl

'2 / \.Il~\10 "'/ ................. ' ,

", /~/ >\;.'"" / 1'1 " 3 i:" I' I \~ ~\ lot / \

,\. \ / Id I \\ \\ \ / 'i! I

" J<'\ ~ \ "J: I\ .~ \ I II I\ \i I / k,,\')/ l' If EVAPOIATION0 0

~ ICONOENSATiON

....._".,,'-2 . .

-IJAN MAR MAY JUt SIP NOV JAN

MONTHS

FIGURE 4-99 Evaporation from the GreatLakes. The smoothed Jines are based on severalstudies from 1926-1969.

evaporation season, which for most lakes is inthe fall, when the water temperature is con­siderably above the dew point temperature ofthe air. Because of its great depth and tre­mendous heat storage capacity, highest evap­oration from Lake Superior occurs in the latefall and early winter period. The long-term av­erage values of the highest monthly evapora­tion vary from approximately 9 cm (3.5 in.) forLake Superior to 15 cm (6.0 in.) for Lake Erie.These high evaporation rates, coupled withlow air temperatures, cause rapid dissipationof heat from the water surface. With thesharply falling water temperatures, evapora­tion begins to decrease.

4.9 Runoff from Drainage Basin

4.9.1 Surface Runoff

Water enters the lakes from the drainagebasin mainly through the tributary streams.When rainfall exceeds the infiltration capac­ity of the soil the excess water reaches thetributary streams either as direct surfacerunoff or as interflow. Interflow is defined asthat water which percolates through the soilabove the phreatic or permanently saturated

ground-water zone. Although interflowreaches the streams with more delay than di­rect surface runoff, the two are difficult to dis­tinguish, and together are referred to as sur­face runoff. The infiltration capacity of soilsvaries widely depending upon precipitationfactors such as intensity, duration, and type ofprecipitation, and soil factors such as antece­dent soil moisture, type ofsoil, shape and slopeof the basin, and vegetation. During wintermonths, the ground is usually frozen, andinterflow is reduced to the periods of groundthaws, but surface runoff still occurs duringintermittent snowmelts. Basin hydrology isconsidered in Appendix 2, Surface Water Hy­drology.

Water which infiltrates into the soil istransmitted downward to the ground-watertable and becomes a part of the ground-waterreservoir. Ground-water flow is discussedlater in this section.

The storage of water on the drainage basinin whatever form, be it ground water, channelstorage, or snowpack, is a fundamental hydro­logic factor of runoff delay. Natural regula­tion by upland lakes and artificial regulationby man-made reservoirs and control struc­tures are also common in the basins of all theGreat Lakes. Lake regulation delays, pro­longs, and diminishes the peak runoff by stor­ing water during periods of high runoff.

Runoff is measured at gaging stations onmany of the tributary streams and togetherwith measured flows in connecting rivers isthe most accurate data in the hydrologic cycle.Unlike measurements of other components,which sample only points within an area,gaged runoff effectively integrates the entirearea about the point of measurement. How­ever, runoff data contain some uncertainties,but these are considered to be random. Errorsmay be introduced in the measurement ofrunoff, particularly during winter months dueto ice effects, and by extrapolating the gagedrunoff to the nearby ungaged areas to obtaincoverage for the entire drainage basin.

Routine computations of the total averagesurface runoffto the lakes are not made by anyagency associated with the Great Lakes, butrunoffestimates have been made in the courseof hydrologic studies. The values ofrunoffpre­sented in this report were obtained by directareal extrapolation of the available gagedstreamflow records to the nearby ungagedareas, and summation of the total runoffamounts from individual basins of the tribu­tary streams within each lake basin, which intum were combined to obtain total runoff from

Hydrometeorology 95

TABLE 4-16 Runoff·Precipitation Ratios, TABLE 4-17 Average Runoff into the Great1937-1969 Lakes, 1937-1969

LakeRunoff in cm depth on lake surface

Period Superior Michigan Huron Erie Ontario

Period Superior1 Michigan Huron 2 Ontario.49 .47 .43 .55 .55 ErieJanuaryFebruary .50 .66 .411 .68 .55 January 3.4 4.3 5.5 8.2 13.2March .61 .62 .75 .1l6 .95 February 2.9 5.1 5.2 8.9 12.5April 1.01 .61 .93 .63 1.13 March 3.8 6.4 8.4 14.1 22.8May .85 .42 .67 .35 .60 April 8.6 9.2 12.7 12.8 30.3June .46 .26 .37 .19 .34 May 9.8 7.2 10.2 7.2 17.7July .36 .24 .27 .12 .20 June 6.5 5.2 6.1 4.2 9.2August .26 .18 .18 .08 .15 July 4.5 3.9 4.2 2.4 5.7September .28 .18 .16 .08 .16 August 3.6 3.1 3.0 1.5 4.4October .40 .28 .28 .13 .25 September 3.6 3.2 3.0 1.3 4.6November .44 .32 .33 .20 .34 October 3.7 3.7 4.1 2.0 6.5December .52 .44 .43 .40 .48 November 4.2 4.1 5.3 3.3 9.8

Annual .52 .39 .44 .35 .48 December 3.7 4.2 6.0 5.9 12.8

Annual 58.3 59.6 73.7 71.8 149.5

the entire drainage area of the Great Lakes.Similar methods were employed in previousstudies (Freeman,271 Horton and Grunsky,376Pettis,60li Hunt39li). Results obtained by vari­ous investigators generally differ somewhatbecause of different periods of record andvarying coverage of the gaged tributary area,which increases steadily. Peak runoff to all thelakes usually occurs in the early spring as aresult of melting snow augmented by rainfalland lack of growing vegetation. The low pointusually occurs in the late summer.

To facilitate comparison with other compo­nents of the hydrologic cycle, such as precipi­tation or evaporation, runoff values are fre­quently given in units of depth over respectiveareas, and this practice is retained in this re­port. Runoff distribution on the drainagebasin is shown in Figure 4-19. Comparison ofrunoff with the corresponding overland pre­cipitation for both monthly and annualperiods, expressed as runoff-precipitationratios, is shown in Table 4-16. These ra!iosindicate that average annual runoff dUringthe period from 1937 to 1969 represents from35 to 49 percent of the overland precipitation,while average monthly runoffs vary from ahigh of 113 percent in the spring to a low of 8percent during the summer. The substantialdifferences in the retention of precipitation onthe basins of the individual lakes are dueprimarily to higher evapotranspirationlosses in the south, but other factors such asinfiltration capacity of soils, forest cover, landcultivation, and urbanization, also affect therelationship between runoff and precipitation.Although precipitation on Lake Erie is amongthe highest and is comparable with that ofLake Ontario, runoff into Lake Erie is low be­cause of the high water losses by evapotrans-

1Inc1u3es diversions into Lake Superior (about 5cm/yr -

2140 m Is). See DIVERSIONS.Excluding drainage area of Lake St. Clair.

piration. Similarly, the Lake Superior basin,with the lowest precipitation, has a high runoffbecause of the reduced water losses in thenorth. High runoff in the spring accounts for30 to 40 percent of the annual amount.

The importance of surface runoff to the hy­drology ofthe Great Lakes depends not only onthe absolute magnitude of flow but also on therelative magnitude of runoff with respect tosurface area of each lake. Runoff representsone-third to one-halfof the overland precipita­tion and is almost equal to overwater precipi­tation on all the lakes except Ontario, whererunoff is almost twice as high. The averagerunoff values for monthly and annual periods,expressed in centimeter depth on the lake sur­face, are listed in Table 4-17. Average annualrunoff during the 33-year period (1937-69) var­ied from 58 cm (23 in.) for Lake Superior to 150cm (59 in.) for Lake Ontario, 60 cm for LakeMichigan, 74 cm for Lake Huron and 72 cm forLake Erie. The annual value for Lake Supe­rior includes approximately 5 cm (2 in.) fromthe Ogoki and Long Lake Diversions, whichare channeled and measured in the tributarystreams, while Lake Erie runoff excludesstreamflow tributary to Lake St. Clair, since itis measured in the Detroit River and thus be­comes part of the inflow to that lake. The aver­age monthly runoffvaried from a high of30 cm(12 in.) during April on Lake Ontario to a low of1.3 cm (0.5 in.) during September on Lake Erie.

4.9.2 Underground Flow

Underground flow from ground water

96 AppeMiz.4

reaches the lakes by percolation. either di­rectly through the lake bottom or throughtributary streams. Since most streams in theGreat Lakes Basin derive their base flow fromground water. the underground flow reachingthe lakes through tributary streams is mea­sured and considered with the surface runoff.Thus, direct contribution from ground waterto the lakes is of primary interest to the lakehydrology. Depending on the geohydrologiccharacteristics of the system and the waterlevels in lakes, the underground flow could beeither into or out of the lakes.

There is very little information on the un­derground flow in the Great Lakes Basin. par­ticularly on direct ground-water supplies, andknowledge in this field should be expanded.The evaluation of direct underground flow re­quires ground-water profiles. which can be de­rived from a network of observation wells lo­cated around the lakes. There is only a handfulof wells located within 15 km (10 mi.) of thelakes, and wells located further inland may bepoor indicators of ground-water flow. In addi­tion to lake perimeters. special attentionshould also be concentrated on any areaswhere, because of geologic structure,ground-water divides may be significantly dif­ferent from topographic divides.

There are many geological studies ofground-water conditions for specific areas ofthe Basin. However, these studies cover rela­tively small areas (one section of a lake basin)and are not related directly to the un­derground flow into or out of the lakes. Theyare approached from the geological point ofview and consider areal geology, water use.economics of water development. and theusual ground-water data obtained from ob­servation wells, pumping tests. and chemicalanalysis. Thus. they reveal only limited infor­mation, which is not readily adaptable to largeareas encompassing one or more lake basins.Large area hydrologic studies dealing withground water are more important for the un­derground flow aspect. Because of the varyinggeologic and climatic conditions over suchlarge areas, estimation of the undergroundflow into anyone lake is very difficult, andground-water data are not sufficient to de­termine this flow by direct methods.

In most hydrologic studies undergroundflow was considered negligible and assumed tobe zero (Freeman,271 Horton and GrunskY,376Hunt.395 Morton and Rosenberg,561 Der­ecki 214). Horton and Grunsky state that there isa marginal belt around the perimeters ofLakes Michigan and Huron. from which water

reaches the lakes mainly through groundwater and not through surface streams. Theyalso state that there are numerous instancesof watershed leakage from one drainage basinto another and probably from the higher-lyingdrainage basins directly into the lakes. Basedon these considerations, they concluded thatthe summer runoffto the lakes estimated frommeasured streamflow may be less than theactual runoff by an amount exceeding 13 cm (6in.) on the lakes, but in their study they as­sumed the underground flow to be zero. Huntacknowledges a possibility of a considerableunderground flow between Lakes Erie andOntario. due to the very large potential headbetween them. He investigated this matterand concluded that there is no appreciableground-water flow into Lake Ontario. and thatwhat flow might exist is probably steadythroughout the year.

In some hydrologic studies undergroundflow was estimated by various forms of waterbudget computations. Russe1l 6l12 estimatedmonthly amounts of water entering the lakesfrom underground sources, which resulted inthe annual values 65 em (26 in.) on Lake Supe­rior, 83 cm (33 in.) on Lakes Michigan-Huron,51 cm (20 in.) on Lake Erie, and 130 cm (61 in.)on Lake Ontario. However, Russell did not dis­tinguish between ground-water flow intostreams and ground-water flow directly intolakes, so values listed above include base flowof streams. Pettis 605 estimated direct un­derground flow into the upper lakes as 42 em(16 in.) to Lake Superior and 68 cm (23 in.) toLakes Michigan-Huron. Pettis also states thata considerable part of a large undergroundwater body in the northern part of the State ofOhio has a definite motion towards Lake Erie.

In contrast to the high underground flowclaimed by the above investigators, Berg­strom and Hanson 62 computed inflow to LakesMichigan-Huron from the ground-watersources to be approximately 0.5 cm (0.2 in.).They suggest that the actual discharge alongthe shore could be several times that givenabove, but the amount would still be relativelysmall. and probably less than the error inher­ent in extrapolation of runoff, lake evapora­tion, and precipitation. Snyder7S1 indicatesthat there are underground outflows fromLake Superior and Lakes Michigan-Huronthat amount to approximately 22 cm (9 in.) and5 em (2 in.) on these lakes. respectively. He alsosuggests that there are underground inflowsto Lakes Erie and Ontario of approximately 19em (8 in.) and 8 cm (3 in.), respectively. Snyderbases his indicated groundwater flow on the

differences in evaporation estimates com­puted by mass transfer and water budgetmethods.

Widely divergent opinions on the amountand direction ofunderground flow to the GreatLakes leave the subject in a state of con­troversy. For example, Pettis 605 claims theunderground inflow to Lake Superior is 42 cmon lake surface per year and Snyder 751 indi­cates an outflow from the same lake of approx­imately 22 cm.1t is apparent from the cautiouswording and the conflicts that exist in the pre­ceding studies that the estimates are littlemore than conjectures.

4.10 Lake Inflow and Outflow

4.10.1 Inflow

The inflow to any of the Great Lakes is thequantity of water supplied by the lake above,modified by the local inflow to the connectingriver. Local inflow is usually less than one-halfpercent of the total flow and generally is dis­regarded in computing lake outflows and thecorresponding inflows to the lakes below.However, the Lake Huron outflow and LakeErie inflow normally differ by approximately 2percent and occasionally the difference ismuch higher. Therefore, flow at the head ofSt.Clair River is used to determine outflow fromLakes Michigan-Huron, while the flow of theDetroit River is used to obtain inflow to LakeErie. The importance of inflow to the lakesincreases progressively in descending orderthrough the lakes. The smaller lower lakes areaffected by these flows to a much greater ex­tent than the upper lakes. Inflow to LakesMichigan-Huron is of the same order ofmagnitude as the overwater precipitation, butin Lakes Erie and Ontario, it is an order ofmagnitude greater. During the period of hy­drologic study, 1937-69, the average annual in­flow was equivalent to 60 cm (24 in.) for LakesMichigan-Huron, 640 cm (250 in.) for LakeErie, and 927 cm (365 in.) for Lake Ontario. Thevariation in the annual inflow during this timehad a range of approximately 40 cm (15 in.) forLakes Michigan-Huron, and approximately240 cm (95 in.) for Lakes Erie and Ontario.Thus, in the lower lakes the variation in theannual inflow is three times greater than theaverage annual overwater precipitation. Be­cause of the magnitude and fluctuation of theinflows to the lower lakes, accuracy ofinflow is

Hydrometeorology 97

extremely important in studies of the hy­drologic cycle.

4.10.2 Outflow

The outflow ofeach of the Great Lakes is thequantity of water flowing from a given lakethrough its natural outflow river and throughman-made outlets. As a factor in the hy­drologic cycle of the lakes the outflow of anylake, including that through man-made diver­sions, represents the water yield of the entirebasin above the point of outflow measure­ment.

The outflow from Lake Superior through theSt. Marys River has been artificially con­trolled since 1922 by a gated dam structure,which allows diversions for the generation ofpower. Release of water through the controlstructure and for the generation of power ismade in accordance with a regulation plan,designed to maintain the level of Lake Supe­rior within specified limits.

The outflow from Lakes Michigan-Huronthrough the St. Clair and Detroit Rivers,which together with Lake St. Clair constitutethe natural outlet of these lakes, is controlledlargely by the level of Lake Huron at the headofthe St. Clair River and the level ofLake Erieat the mouth of the Detroit River. No man­made control of this flow exists, except for thefixed remedial control provided by the dikesconstructed in the lower Detroit River to com­pensate for the effects of deepening the navi­gation channels in that river. These compen­sating controls do not regulate lake levels.However, they are designed to provide thesame net discharge capacity in the rivers asexisted before the improvements, so that thelevels upstream are maintained. The naviga­tion improvements in the St. Clair River, in­cluding some works recently completed, causea lowering of the levels of Lakes Michigan­Huron, and remedial works are being evalu­ated. In the winter period ice normally slightlyreduces the open water flow of the St. Clairand Detroit Rivers.

The outflow from Lake Erie through theNiagara River is controlled largely by thelevel of Lake Erie at the head of the river andthe level of the Chippawa-Grass Island Poolabove Niagara Falls. If the diversions ofwaterfrom the Chippawa-Grass Island Pool for thegeneration of power are not compensated for,they can lower the levels of Lake Erie. How­ever, a submerged weir was constructed dur­ing the period 1942-47 at the downstream end

98 Appendix ..

TABLE 4-18 Average Flows in ConnectingRivers of the Great Lakes, 1937-1969 (m3/s)

St. St. St.Period Marys Clair Detroit Niagara Lawrence

January 1,980 4,420 4,620 5,240 6,:!30February 1,980 4,250 4,420 5,240 6,230!'larch 1,940 4,810 4,960 5,320 6,400April 2,020 5,130 5,270 5,640 6,830!'lay 2,180 5,240 5,350 5,920 7,020June 2,310 5,350 5,410 5,950 7,190July 2,450 5,410 5,490 5,830 7,140August 2,570 5,440 5,490 5,720 7,000September 2,550 5,380 5,440 5,580 6,770October 2,510 5,320 5,380 5,470 6,570November 2,430 5,270 5,300 5,470 6,460December 2,110 5,130 5,240 5,470 6,460

Annual 2,250 5,100 5,200 5,570 6,690

of the pool, as the initial phase of the remedialworks designed in part to counteract this ef­fect. Presently, a gated control structure ex­tending partially into the river from theCanadian shore provides compensation for thepower diversion, which was increased by the1950 treaty between the United States andCanada. This second control structure is lo­cated downstream and runs parallel to theweir.

The outflow from Lake Ontario through theSt. Lawrence River since 1958 has beenlargely controlled by the release of waterthrough the Iroquois Dam near Iroquois, On­tario. Beginning in April 1960 the release ofwater has been made in accordance with aregulation plan, which provides for weeklyflow changes throughout the year. Thus, LakeOntario levels are fully regulated and are in­dependent of any channel changes or diver­sions. Prior to the construction of the IroquoisDam in 1958, the Galop Rapids, a short dis­tance downstream from Ogdensburg, NewYork, constituted a natural weir, the flow overwhich was controlled substantially by thelevel of Lake Ontario.

The average monthly and annual flows ofthe outflow rivers for the 1937-69 period arelisted in Table 4-18. Prior to 1957 the flows ofthe St. Clair River were based on published rec­ords of combined St. Clair-Detroit Riverflows, since the flows in both rivers were con­sidered to be essentially equal. During theperiod of study, annual flow from the St.Marys, St. Clair, Detroit, Niagara, and St.Lawrence Rivers averaged 2,250 m 3/S , 5,100m 3/S , 5,210 m 3/S , 5,580 m 3/S, and 6,680 m 3/S (79,180, 184, 197, and 236 thousand cfs), respec­tively. Low flows occur in the winter and highflows in the summer, with a progressive delayof the summer highs in the upstream rivers.

The range in the average monthly flows wasapproximately 700 m 3/S (25,000 ds) for the St.Marys and Niagara Rivers, and 1,300 m 3/s(45,000 cfs) for the other rivers. The differencebetween the high and low annual flows duringthe 33-year period varied from approximately1,400 m 3/S (50,000 cfs) on the St. Marys River to2,000 m 3/S (70,000 cfs) on the Detroit Ri ver. Therelatively large variations in flow of the St.Clair and Detroit Rivers, in comparison withother rivers, are due primarily to ice retarda­tion of winter flows.

The importance of outflow to lake hydrologyincreases progressively with downstreamlakes. The average annual outflows (riverflows and diversions), expressed in units depthon the lake surface, represent 86 cm (34 in.) onLake Superior, 139 cm (55 in.) on LakesMichigan-Huron, 706 cm (278 in.) on Lake Erieand 1,077 cm (424 in.) on Lake Ontario. Varia­tion in the annual outflow during the period ofstudy had a range of approximately 50 cm (20in.) for Lakes Superior and Michigan-Huron,190 cm (75 in.) for Lake Erie, and 300 cm (120in.) for Lake Ontario.

Numerous studies of the connecting riverflows have been made. These studies haveanalyzed the effects on flow equations of reg­imen changes, formation of ice, weed retarda­tion, and water temperatures. Detailed dis­cussion of outflows is given in Appendix 11,Levels and Flows. Flow equations and themeans ofrevising them when channel changesoccur, and turbine ratings used to computeflows through power structures are generallywell established. As a result, data on the flowsin the connecting rivers ofthe Great Lakes areconsidered to have a greater accuracy than forany other hydrologic factor, except the changein lake storage.

4.10.3 Diversions

Water diversions in the Great Lakes Basinmay be broadly divided into two types: outsidediversions, which take water into or out of thesystem; and inside diversions, which retainwater entirely within the system. Diversionsof water into the Basin have the effect of rais­ing water levels of the lake into which the di­verted water is discharged and the levels ofthe lakes downstream through which the di­verted water must pass on its way to the sea.Diversions of water from the Basin have theconverse effect on the levels ofthe lakes at anddownstream from the point of diversion. Di­versions from one point to another within the

rBasin may have no effect on the lake levels ifwithin the same lake basin; or, if not compen­sated for, diversions may lower the levels ofthe lake upstream and temporarily raise thelevels downstream. The temporary rise inlevels downstream is due to increased dis­charge rates while the levels of the lake abovedrop to adjust to the larger outlet capacity dueto the diversion. The rate of adjustment inlake levels decreases exponentially with time,and the period of adjustment depends on thesize of the lake involved and the capacity of itsoutlet. For example, Lakes Michigan-Huronreach 90 percent adjustment in approximatelyseven years and Lake Erie in less than oneyear (Bajorunas 3lia).

There are five major diversions in the GreatLakes Basin (Figure 4-1): the Ogoki and LongLake Projects divert water into Lake Supe­rior; the Chicago Sanitary and Ship Canal di­verts water out of Lake Michigan; and theWelland Canal and the New York State BargeCanal divert water from Lake Erie and Niag­ara River into Lake Ontario. All other diver­sions presently in existence on the St. Marys,Niagara, and St. Lawrence Rivers divertwater from one point to another within thesame river and with remedial structures haveno effect on lake water supplies and lakelevels.

The Ogoki and Long Lake Projects divertwater into Lake Superior from the AlbanyRiver drainage basin in the Hudson Bay wa­tershed. The Ogoki diversion diverts waterfrom the Ogoki River into the Nipigon River.The Long Lake diversion diverts water fromLong Lake at the head of the Kenogami Riverto the Aguasabon River. These diversionshave increased the supply of water to LakeSuperior by an average rate of approximately142 m 3/s (5,000 cfs), which is equivalent to ap­proximately 5 cm (2 in.) on the lake surface peryear.

The Chicago Sanitary and Ship Canal, alongwith the Calumet Sag Canal, a branch whichconnects with Lake Michigan south of Chi­cago, diverts water from Lake Michiganthrough the Des Plaines and Illinois Rivers tothe Mississippi River. This diversion, com­monly referred to as the Chicago diversion.represents the amount of water diverted fromLake Michigan for navigation purposes andfor domestic use by the City of Chicago. Thetotal diversion from the lake at Chicagoamounts to 88 m 3/s (3,100 cfs), which repre­sents an annual amount of water exceeding 2cm (about 1 in.) on the surface of LakesMichigan-Huron. The net effect of all three

Hydrometeorology 99

outside diversions, the inputs to Lake Supe­rior and the output from Lake Michigan, is toincrease the supplies to Lake Michigan-Huronand the downstream lakes by approximately54 m 3/S (1,900 cfs).

The diversion of water through the WeIlandCanal, from Lake Erie at Port Colborne toLake Ontario at Port Weller, includes waterused in the DeCew Falls power plant, whichamounts to most of this diversion, plus diver­sions for navigation purposes. The total WeI­land Canal diversion amounts to approxi­mately 198 m 3/s (7,000 cfs), which is equivalentto approximately 25 cm (10 in.) on the LakeErie surface per year.

The New York State Barge Canal withdrawswater from the Niagara River at Tonawanda,New York, for navigation purposes, but thewater diverted into the canal is returned toLake Ontario at Oswego, New York. TheBarge Canal diverts approximately 31 m 3/s(1,100 cfs) during the navigation season. Since1956 there has been no diversion during win­ter months.

The average monthly and annual flows dur­ingthe period ofstudy, 1937-69, for the majordiversions described above are listed in Table4-19. Annual values listed in the table aresomewhat different from the normal valuespresented in the discussion because of periodicvariations from the normal.

4.11 Lake Level Fluctuations

4.11.1 Lake Levels

The elevations of the water surface of theGreat Lakes are tied to the mean sea level atFather Point, Quebec, on the Gulf of St. Law­rence. This plane of reference, establishedespecially for the Great Lakes in 1955, is calledthe International Great Lakes Datum. Theaverage monthly and annual lake levels forthe 33-year period, 1937-69, are given in Table4-20. These levels are based on mean lake leveltabulations published by the Lake Survey,and represent records from master gages,each lake having a single master gage locatedat a strategic point. Approximate water sur­face elevations of the lakes are 183 m (601 ft.)for Lake Superior; 176 m (578 ft.) for LakesMichigan-Huron; 174 m (570 ft.) for Lake Erie;and 75 m (245 ft.) for Lake Ontario.

Of primary interest in lake hydrology arethe variations of lake levels caused by thechanging volume of water in the lakes. These

100 Appeftdiz 4.

TABLE 4-19 M~or DiveraioRB in the Great Lakes Buin, 1937-1969 (tubic Meter. per Second)

Ogoki Long Lake Chicago WeIland N. Y. StateProject Project Diversion Canal Barge Canalinto into out of from L. Erie from Niagara R.

L. Superior L. Super~or L. Michi~an to L. Ontario to L. OntarioPeriod 1943-691 1939-69 1937-69 1937-694 1937-695

January 94 35 90 160 9February 75 35 88 162 9March 65 29 85 162 4April 67 28 95 170 22May 148 54 99 177 31June 216 65 107 178 31July 152 48 110 172 31August 121 39 114 181 31September 111 34 105 180 31October 105 33 91 182 31November 120 36 85 181 31December 111 35 98 170 16

Annual Average 115 39 97 173 23

1period of record starts in July 1943. Since 1945 total amount of Ogoki and Long2Lake diversions has averaged 142 m3/s (5,000 cfs).3Period of record starts in July 1939. 3Since 1938 total diversion (navigation plus domestic pumpage) has averaged 88 m Is(3,100 cfs). However, higher flows were authorized on two occasions by the U. S.

4Supreme Court. 3Since 1950 total diversion (navigation and hydropower) has averaged 198 m /s

5(7,000 cfs). 3Since 1929 during navigation season this diversion has amounted to 31 m /s (1,100cfs). Since 1956 there has been no diversion during winter months.

volumetric changes are generally referred toas lake level fluctuations and apply to the en­tire lake. They involve time periods of suffi­cient duration to allow absorption of any localshort-period variations, so that entire watersurface can be assumed to be level. The localshort-period variations, classified as waterlevel disturbances, do not involve volumetricchanges but displacement of water levelcaused primarily by winds and variations inbarometric pressure. Water level disturb­ances are discussed in Section 6, while detaileddiscussion of lake levels is given in Appendix11, Leve18 and Flow,.

The water level fluctuations represent stor­age or depletion ofwater in the lakes. Seasonalfluctuations undergo a relatively regular cy­cle; high levels usually occur in the summerand low in the winter.

TABLE 4-20 Average Levels of the GreatLakes, IGLD (1955). 1937-1969 (Meters)

Superior lUclligan-lluron Erie Ontarioat at at at

Period Marquette Barbor Beach Cleveland 08mo

Ja1lWlry 183.00 176.01 173.66 74.40February 182.93 176.01 173.68 74.42March 182.89 176.02 173.76 74.50April 182.92 176.09 173.92 74.71Hay 183.04 176.19 174.03 74.85June 183.13 172.26 174.07 74.92July 183.20 176.32 174.06 74.88August 183.23 176.30 174.00 74.77Septl!lllber 183.23 176.25 173.90 74.64October 183.20 176.19 173.79 74.51Novlllllber 183.15 176.13 173.70 74.44Ileclllllber 183.08 176.08 173.68 74.42

Annual 183.08 176.15 173.85 74.62

4.11.2 Change in Storage

The change in storage on the lakes for any

,TABLE 4-21 Average Chap il. Storage onthe Great Lakes, 1937-1"9 (Centimeters)

1 Michtaau-Erie3 Ontario4

p~riod Superior Huron 2

January -6.7 -2.1 0.6 0.9February -5.2 0.0 2.4 3.4Karch -2.1 4.0 14.0 15.2April 9.1 11.3 15.5 21.3KaY 11. 3 8.2 6.4 11.0June 8.8 6.7 1.8 0.3July 4.0 0.9 -4.3 -7.6August 1.8 -3.4 -8.5 -12.8September -1.8 -5.8 -10.7 -13.4October -4.9 -6.7 -9.4 -10.7November -5.8 -4.6 -4.9 -4.3December -8.5 -5.8 0.0 -1.8

Annual 0.0 2.7 2.9 1.5

Note: Change in storage determined from 10-daymeans (5 at end and 5 at beginning of fol­lowing month) by averaging records from the

1 following gages:Thunder Bay, Duluth, Hichipicoten, Marquette.

2and Pt. Iroquois.Milwaukee. Ludington. Mackinaw City. Harbor Beach.

3Thessalon, and Godericb.4Cleveland and Port Stanley.Oswego, Kinaston. Cobourg. Toronto, Port Weller,and Rochester.

given period is determined from the change inlake levels. Mean lake levels for several daysare used for the determination of beginning­of-period levels. This minimizes the effect ofexternal forces such as winds or barometricpressure. The mean level of a lake at any giventime is determined by averaging recordedlevels of several gages. situated at pointsaround the lakes in a pattern selected to pro­vide good approximation of the whole lakelevel.

In recent years, the gage patterns used fordetermination of lake storage have been coor­dinated by the Lake Survey and Canadianagencies to provide consistent values in bothcountries. These gage patterns consist of fivegages for Lake Superior, six gages for LakesMichigan-Huron and Ontario, and two gagesfor Lake Erie. Each determination is based onten days of recorded levels (five at the end ofone month and five at the beginning of thenext month). This determination period israther long for the beginning-of-month levels.In other determinations four (two plus two) ortwo (one plus one) days were normally used.The most recent determination is based on twodays of recorded levels (one at end and one atbeginning of month), employing more gagesweighted by the Thiesson polygon method to

HJldromeuorologJl 101

reduce possible effects of short-term waterlevel disturbances (Quinn,lalla Quinn and'Todd-b).

The coordinated average change in storagefor monthly and annual periods on each lakeduring the 1937-69 period is shown in Table4-21. Since the change in lake storage isprimarily a seasonal phenomenon, the long­term annual values should be small due tobalancing of rising and falling lake levels, asindicated in the table. The average seasonalchange in storage varies with latitude. Thelower lakes have rising lake levels during win­ter and spring, and falling lake levels duringsummer and fall. This distribution is delayedby approximately one month on LakesMichigan-Huron and by a full season (3months) on Lake Superior. The highest aver­age monthly rise was approximately 11 cm onLakes Superior and Michigan-Huron, 16 cm onLake Erie, and 21 cm on Lake Ontario; thehighest average monthly decline was approx­imately 8 cm for Lake Superior, 7 cm for LakesMichigan-Huron, 11 cm for Lake Erie, and 13cm for Lake Ontario. During individual yearsthe variation in annual and monthly change inlake storage may be considerable. In extremecases this range may exceed several times thehighest average monthly change in storage oneach lake.

The change in storage discussed above in­cludes volumetric changes, which are affectedby water density variations. A mass of waterexpands or contracts as it is heated or cooled.The amount of expansion or contraction de­pends on the change in temperature and depthto which this change becomes effective (ther­mocline depth). Investigations of the thermalexpansion ofwater in the Great Lakes indicatethat thermal expansion is insignificant andmay be disregarded (Hunt,3lI5 Derecki 214).

4.12 Heat Budget

The interaction of the various climatic andhydrologic elements results in heating andcooling processes within the lakes. Some proc­esses take place at the lake surface and aretransmitted through the water body whileothers produce heat changes by mixing of thewater masses. Meteorological factors such asradiation, air temperature, precipitation, andevaporation affect surface temperature, whilewinds contribute to the deepening of the sur­face layer. Hydrologic factors such as runoff,inflow, and outflow cause temperaturechanges by horizontal movement of watermass.

lOt Appendi:r"

From Rodgers, IH9

Qh = conduction of sensible heat to orfrom the atmosphere

Q., = energy utilized by evaporationQt = energy storage within the body of

water.

4.13.1 Water Budget Computations

4.13 Water Budget

The heat budget for Lake Ontario, the onlylake for which such determination has beenmade (Rodgers and Anderson,8'75 Bruce andRodgers,tOl and Rodgers872), is presented inFigure 4-100. The largest energy change isproduced by the absorption and loss of heat bythe lake water mass; the lake gains heat dur­ing spring and summer months and loses heatin the fall and winter. The radiation processesproduce both gain and loss of heat; the lakeabsorbs heat from solar radiation and losesheat through the longwave radiation ex­change between water surface and atmos­phere. Evaporation cools the water surfaceand produces heat loss, except during springwhen slight condensation produces small heatgain. The transfer of sensible heat to the at­mosphere results from the air-water tempera­ture differences; the lake surface is coolerthan air and gains heat in the spring andsummer, and the process is reversed in the falland winter. The net effect of all these proc­esses is to produce heat gain during thespring-summer period and heat loss duringfall and winter months.

The heat budgets for the other lakes wouldfollow generally similar patterns, althoughthe amounts of energy contained in variousprocesses would differ depending on the hy­drometeorological conditions on each lake.The accuracy of heat budget presented forLake Ontario may be sufficient to indicategeneral trends for various energy processes,but evaporation studies show that accuracyshould be improved for successful applicationto the solution of practical problems. This wasone ofthe objectives ofthe International FieldYear for the Great Lakes, an intensive fieldobservation program conducted on Lake On­tario in 1972.

The water budget of the Great Lakes is anaccounting of all incoming and outgoing wa­ter, such as inflow and outflow by the rivers,supply from and storage in the ground, over­water precipitation, evaporation, and varia-

•oo ......

!

o

.i

o ..!

o

+400

- 000

+000

- ZOO =

-2110

"AT AlS_EOCIIl lIlIT ITLAIlE OlIITAIlID

LONG-WAVEIlAllIATION LOSS

The heating and cooling processes are sum­marized in the heat budget of the lakes, whichre presents the amount ofenergy gained or lostby the lakes during various temperaturechanges. There are five basic energy or heatprocesses affecting the Great Lakes. The fourmajor processes include energy produced byradiation, sensible heat transfer to or from theatmosphere, heat loss by evaporation, andenergy storage within the lake. A fifth processof net advected energy may be important lo­cally, especially at the mouths of the inflowrivers and near the effluents of sewage dis­posal or cooling water from power plants.However, this process has very little effect onthe total heat content of the lakes becausesuch inflows with substantial difference intemperatures are relatively small. The energyexchange may be expressed by the equation:

O. + Qv =~ + Qh+ Q., + Qt

where Q. = net solar radiation (incidentminus reflected)

Qv = net advected energy (heat due towater input minus output andsnow melt)

Qb = net terrestrial radiation (emittedminus atmospheric)

IJIFI .. lal .. IJIJlalsloINIDI

FIGURE 4-100 The Heat Budget of Lake On­tario

,TABLE 4-22 Aver.,e Water Budget. 1937-1969 (Centimeters)

BalanceWater Supply Water Loss Storage Needed

Lake p R I 0 E t£, B

Superior 80 58 0 86 55 0 -3Michigan-Huron 80 67 60 139 65 3 0Erie 88 72 640 706 85 3 6Ontario 84 150 927 1,077 70 2 12

P + R + I - 0 - E - 6S ... ±B

P precipitation on the lake surfaceR runoff from drainage area (surface and underground)I = inflow from the upstream lakeso outflow to the lake belowE evaporation from the lake surfacef)S change in storage of water in the lakeB balance needed

NOTE: Diversions are included in runoff, inflow, or outflow, whereapplicable.Evaporation values are the long-term estimates, not necessarilyapplicable to this 33-year period.

4.13.2 Importance of Water Budget

Lake levels and outflows of the Great Lakes

tion of these factors for each lake. The differ­ences needed for balancing of the major fac­tors represent a combination of any possibleground-water flow and cumulative errors inestimating other factors. For most lakes thesedifferences are quite small for the average an­nual values, and are well within the limits oferror. The largest difference, for Lake On­tario, is approximately equal to 15 percent ofprecipitation or evaporation, 8 percent ofrunoff, or 1 percent of inflow or outflow. Forshorter monthly periods and for individualyears, the percent differences should increasesignificantly, because the effect ofcompensat­ing reduction of random errors would besmaller.

Further studies pertaining to the waterbudget of the Great Lakes should be directedtowards elimination of existing gaps in pres­ent knowledge, improvement ofdata collectionnetworks, comparability of measurement ac­curacies for various factors, development andimplementation of new measurement meth­ods, and closer coordination of these efforts inboth countries.

I =oE

as

tion ofwater storage in the lakes. These waterbudget factors are interrelated in the hy­drologic cycle, which is composed of a per­petual sequence of events governing the de­pletion and replenishment of water in the Ba­sin. The Great Lakes water budget may beexpressed by the equation:

P + R + I = 0 + E ::t aswhere P = precipitation on the lake surface

R = runoff from drainage area (sur­face and underground)inflow from the upstream lakesoutflow to the lake below

= evaporation from the lake surface= change in storage of water in the

lake (plus if storage increases,minus if decreases)

In practical applications the water budgetequation may be modified by eliminating allfactors that are negligible or not applicable toindividual lakes (e.g., inflow for Lake Supe­rior). Factors other than those listed may alsobe included. For example, runoff and groundwater may be treated separately, and diver­sions may be included as a separate factor.

The average annual water budget for the1937-69 period is shown in Table 4-22, whichcontains groupings of water !SupplYJ waterlosses, lake storage, and algebraic aceumula-

104 Appendiz"

effectively integrate all other components ofthe water budget and are of primary interestto lake users. However, growth of the popula­tion and economy of the area has resulted inan increase in and diversification of demandsfor lake water, and the competition for its useis increasing rapidly. Use of the lakes for navi­gation, water power, municipal and industrialwater supplies, sanitation, irrigation, fish andwildlife, recreation, and other riparian inter­ests frequently results in conflietingdemands,some of which are detrimental to water qual­ity. To provide optimum utilization and pres­ervation of the lakes, a thorough understand­ing of the entire hydrologic cycle ofthe systemis necessary.

The importance of the water budget to lakewater resources has been recognized in many

studies and investigations (e.g., Freeman,ITI

Horton and Grunsky,n. U.S. Congress­Senate ll7•IIJ). Knowledge of the magnitudesand variations of the individual water budgetcomponents is needed for the improvement offorecasts of lake levels and outflows, for therefinement of lake regulation plans, and fordetermination of the effects of diversions intoand out ofthe system. Because ofthe vastnessofthe Great Lakes, changes in lake levels takeplace rather slowly and advanced informationon the expected stages is of great interest tonavigation, hydropower, and for shore protec­tion. For Lakes Superior and Ontario, the onlylakes presently regulated, accurate forecastsare even more important to permit planningfor the most beneficial operation of the reg­ulating structures.