J. Great Lakes Res. 13(4):826-840Internal. Assoc. Great Lakes Res., 1987

LAKE ERIE RESEARCH: RECENT RESULTS,REMAINING GAPS

F. M. Boyce, M. N. CharltonLakes Research Branch

National Water Research Institute,Burlington, Ontario L7R 4A6

D. RathkeCLEAR

Ohio State University,Columbus, Ohio 43210

C. H. MortimerCenter for Great Lakes Research

University of Wisconsin,Afilwauke~ W~consin53201

J. BennettEnvironmental Research Institute of Afichigan,

Ann Arbor, Afichigan 48105

ABSTRACT. Evidence that oxygen depletion rates in the central basin hypolimnion were increas­ing and would lead to increasingly severe late summer anoxia resulted in a large public investment inboth the U. S. and Canada to construct sewage treatment plants in order to reduce the phosphorusload to Lake Erie. Improvements in the water quality indicators, phosphorus and chlorophyll, havebeen achieved in the decade since these measures went into effect. There is no clear evidence,however, for a corresponding "recovery" in hypolimnetic oxygen depletion. It is now recognized thatphysical processes, driven by winds and solar heating with strong interannual fluctuations, influencethe timing and severity of late-summer anoxia, and make it difficult to detect long-term trends. As aresult of the 1979 and 1980 intensive field experiments, many of the physical processes are betterunderstood, better documented, and new possibilities exist for improved numerical hydrodynamicalmodeling. Systems modelers have exploited the voluminous Lake Erie data to generate convincingwater quality models that account for the physical variability as well as simulating biochemicalparameters. In some respects the modelers have pushed beyond the frontiers of established knowl­edge of biochemical processes and thus challenge process-oriented researchers to confirm or rejecttheir findings. A consensus isforming among Lake Erie researchers that the key to understanding thelake's response to changing external loading lies in a detailed understanding of the sediment/waterinteraction. Study of sediment-water interactions calls for interdisciplinary efforts that will involvephysicists, biologists, chemists, and modelers. An appendix to this paper lists specific researchrecommendations.ADDITIONAL INDEX WORDS: Research needs, mathematical models, sediment-water interfaces,oxygen consumption, thermal stratification.

THE OBSERVATIONAL RECORD:MONITORING CHANGE IN LAKE ERIE

SINCE 1970

Of all the Great Lakes, Lake Erie has been mostseriously impacted by eutrophication. The effects,among which numbered extensive floating algalblooms and disappearance of certain benthic orga-

826

nisms, were most clearly evident in the western andcentral basins, while the eastern basin remainedless impaired. Studies during the 1950s and 1960sconfirmed that the bottom waters of the centralbasin and more rarely the western basin were sub­ject to anoxic conditions through the stratifiedperiod. The reader is referred to Mortimer's (1987)

LAKE ERIE RESULTS 827

review of the historical records in this issue. How­ever it was not until 1970 when two intensive studyprograms were undertaken that the scientific com­munity began to understand the extent and magni­tude of the problem. The open lake survey span­ning the entire field season (April-December, 1970)remains as a bench mark and data base (physical,chemical, and biological) here used as a referencepoint to evaluate the current status of the lake. Thesecond 1970 study, "Project Hypo," was a bi­national effort designed specifically to examine theprocesses involved in depleting the oxygen fromthe central basin hypolimnion (Burns and Ross1972). Those combined efforts provided the infor­mation necessary to develop a strategy for retard­ing and perhaps reversing eutrophication in LakeErie. Specifically the Hypo study concluded:

"Phosphorus input to Lake Erie must be reducedimmediately; if this is done, a quick improvementin the condition of the lake can be expected; if it isnot done, the rate of deterioration of the lake willbe much greater than it has been in recent years."

Based upon results from the 1970 effort, a pro­gram was implemented under the 1972 USICanadaWater Quality Agreement to reduce phosphorusloading to the Great Lakes. For Lake Erie, the goalof this effort was to reestablish year-round aerobicconditions in the bottom waters such that fishcould inhabit them. A phosphorus target load of11,000 metric tonnes per year was proclaimed inthe 1978 Great Lakes Water Quality Agreement,based upon model predictions (Fifth Year Review,1978). It was projected that at the target level,open lake phosphorus concentrations would bereduced sufficiently to curtail phytoplanktondevelopment. This in turn would reduce the oxy­gen depletion of the bottom waters (hypolimnion)far enough to eliminate the occurrence of anoxicconditions.

Evaluation of the phosphorus abatement pro­gram is an ongoing activity. The first and mostimportant step involved the programs designed toreduce phosphorus loading to the lake and to mon­itor the progress of the reduction effort. To followthe results of the reduction effort, phosphorusloadings to Lake Erie from all sources were esti­mated (Fraser 1987). By 1982 the total load hadbeen reduced to 58010 of that in 1968. The actualcalculated load has been reduced from over 20,000metric tonnes to values currently approaching thetarget load. Fraser points out that the reduction of

phosphorus entering from the Detroit River hasbeen responsible, more than any other single fac­tor, for the reduction in phosphorus entering thelake. The municipal phosphorus abatement pro­grams together with the limitation on detergentphosphate concentration (exception: Ohio) havebrought about this reduction.

Since the current loading of total phosphorus toLake Erie is nearing, but has not stabilized upon,the target level (11,000 metric tonnes per year) setforth in the Great Lakes Water Quality Agree­ment, further measures need to be taken to reachcompliance. An additional 2-3 thousand metrictonnes need to be removed to ensure complianceduring both wet and dry years. Since the targetwas not reached with controls on point sources,i.e., municipal sewage treatment facilities, thefocus for further reduction will be on controllingphosphorus loading from agricultural activities.Studies on the diffuse source contributions (IJC1978, COE 1982) to Lake Erie have providedloading estimates and strategies to reduce agricul­tural sources. As discussed by Logan (1987), dif­fuse source loading from farms contributes to theinput of nutrients (both phosphorus and nitro­gen), sediments, herbicides, and pesticides. Thefocus of the reduction effort should be on thedrainage basins servicing the most intensively cul­tivated regions such as the Maumee River basin inOhio, and that of the Grand River in Ontario.Programs designed to reduce nonpoint agricul­tural loadings of phosphorus should also aim atreducing the accompanying loading of soil, nitro­gen, and contaminants.

It was thought that reduction of phosphorusloading would cause a decrease in concentrationsof phosphorus in the open lake. Papers presentedby F. Rosa and A. EI-Shaarawi evaluate thisresponse (Rosa 1987, EI Shaarawi 1987). Focusingon the central basin, Rosa reports a decrease inmean concentrations of total phosphorus from 22p.g/L in 1968 to 12p.g/L in 1982. For everythousand-tonne reduction in annual loading in theinterval 1967 through 1982, the mean phosphorusconcentration has decreased almost 0.5 p.g/L.Similarily, EI-Shaarawi examined phosphorustrends for all three basins. His results indicate adecrease in total phosphorus concentrations in thewestern basin (1968 to 1978) and in the centralbasin (1968-1980). The eastern basin showed asmaller decline.

The phosphorus reduction program was imple­mented specifically to reduce the nutrient that con-

828 BOYCE et al.

trolled the growth of phytoplankton biomass (1978Great Lakes Water Quality Agreement). Conse­quently, the success of the phosphorus reductionprogram would be evident in the decreasing con­centrations of chlorophyll, itself an indicator ofphytoplankton biomass. EI-Shaarawi examinedthe chlorophyll trend for the three basins andfound, as with phosphorus, a decreasing concen­tration from 1968-1980 for both the central andeastern basins. Since data from the western basinwere limited (1968-1972) it was not possible to pro­ject a similar trend. In addition, EI-Shaarawi wasable to show that the relationship between phos­phorus and chlorophyll can be reasonablydescribed with a straight line, further substantiat­ing the conclusion that the phosphorus reductionprogram has been effective in reducing phyto­plankton biomass.

The ultimate objective of the phosphorus reduc­tion program was to restore oxic conditions to thebottom waters of the central basin during summerstratification (1978 Great Lakes Water QualityAgreement). This would result.in two benefits tothe lake. First, it would restore the use of the sedi­ments and overlying waters to biota intolerant toanoxic conditions. Second, the rapid internal recy­cling of phosphorus, ammonia, and silica whichoccurs during periods of anoxia would be elimi­nated. The recycling of phosphorus from the sedi­ments, also known as internal loading, underanoxic conditions may be as much as several thou­sand metric tonnes (Lam et al. 1983) and mighteventually overshadow external loadings if theduration of anoxic conditions were prolonged.Finally, and as a means of emphasizing that thesystem embodies complicated feedback and cannotbe viewed as a simple summing network, we referto the optical study by Jerome et al. (1987) where itis suggested that a decrease in the primary produc­tivity of the epilimnion, with concomitant increasein water-column transparency, may increase theproductivity in the hypolimnion, enough perhapsto offset some of the oxygen demand.

The oxygen depletion rates for the central andeastern basins have been measured annually sincethe 1970 "Project Hypo". In an effort to determinethe amount of degradation that the lake has suf­fered since the influx of European settlers andindustrialization, historical data have been used tocalculate oxygen depletion rates prior to the 1970study. In recent years, several attempts have beenmade to assess the historical change in the oxygendepletion, refining the earlier work of Dobson and

Gilbertson (1972). Students of this particularaspect of Lake Erie will be aware that scientificopinion has ranged from the bleakest predictionsof impending disaster to assertions that the centralbasin has exhibited occasional late summer anoxiasince prehistoric times and that a reliable trend foreither improvement or worsening cannot be estab­lished. Discussion continues. Details of individualapproaches and methods are provided in the threepapers included in this issue (Charlton 1987, EI­Shaarawi 1987, Rosa and Burns 1987). In general,the authors agree that the oxygen depletion ratehas increased since the early 1950s in response toincreased eutrophication and that, if phosphoruslevels are reduced, the probability of anoxia alsobe reduced.

To date we have seen unmistakable and signifi­cant reductions to open lake concentrations ofphosphorus and chlorophyll; but the anticipateddecrease in the measured hypolimnetic oxygendepletion rate has not materialized in either thecentral or the eastern basin.·If one were to assumea linear relationship between depletion rate andphosphorous loading, then the small observedtrends in depletion rate suggest there is a largebackground depletion rate as yet not influenced bythe control measures. If this is true, then recoverywill be limited too, and we will have to accept spo­radic anoxia as part of the central basin environ­ment. That the reduction of phosphorus and chlo­rophyll in the water is not yet accompanied by areduction of the oxygen depletion rate suggeststhat the depletion rate is controlled by other fac­tors, possibly by processes occurring within thehypolimnion or at the bottom. It is then possiblethat the oxygen depletion rate exhibits a laggedresponse to changing load, and that we are notmeasuring now, nor have we ever measured anequilibrium situation. The latter interpretationseems more plausible, but what are the underlyingmechanisms? In the introductory paper to thisissue, Mortimer (1987) points out the large anddynamic sediment movements that are characteris­tic of Lake Erie. As we shall see, the role of bottomsediments is only partially understood. An addi­tional consideration is the effect of changing waterlevels on the hypolimnetic volume. Each year thethe thermocline forms at roughly the same depthbelow the surface. If that surface is elevated, thehypolimnion thickness will be greater. A I-mchange in hypolimnetic thickness represents a largefractional change of order 20070 in the central basinof Lake Erie.

LAKE ERIE RESULTS 829

HYDRODYNAMICS OF LAKE ERIE:A POST HYPO UPDATE

Throughout Project Hypo it was recognized thatphysical processes mediate biochemical processesand that observed changes in biological or chemi­cal parameters may be the result of water move­ments as well as locally-acting biochemical proc­esses. Mortimer's (1987) review provides a wealthof examples. Among the physical processes consid­ered to influence this budget were the mechanics ofthermocline formation and decay, surface fluxes ofheat and momentum, and particularly the mecha­nism of the entrainment of non-turbulent but strat­ified fluid into an actively stirred layer. An array ofcurrent meters provided evidence of a mean north­west drift of hypolimnion water across the centralbasin but gave few details as to how this movementwas compensated at the edge of the hypolimnion.The flow of subsurface water from the easternbasin into the central basin hypolimnion was con­sidered as a mechanism for the resupply of oxygen­ated water. Diffusion processes within the hypo­limnion were studied as were the mechanisms ofsediment-water interaction.

In the years immediately following ProjectHypo, numerical hydrodynamic modeling wasenergetically pursued and extended to includesome of the biochemical processes. The papers ofSimons, Lam and Jacquet appearing in the specialissue of JFRB are noteworthy (Simons 1976, Lamand Jacquet 1976). Assessments of the modelingstudies to date are included in this issue (Lam andSchertzer 1987, Lam et al. 1987, DiToro et al.1987). This initial activity would not have beenpossible without the 1970 database. The under­standing that the severity of oxygen depletion inthe central basin hypolimnion depended in part onthe physical structure of the hypolimnion - itsthickness and its temperature - was articulated byCharlton (1980) and was boldly synthesized byLam and his colleagues in a report first appearingin 1983 (Lam et al. 1983). The heart of this reportis a thermocline model that is discussed in detail inthis issue (Lam and Schertzer 1987). Inputs to themodel are surface heat fluxes and wind stress; themodel is verified against basin-wide observationsof thermal structure. Climatological reviews of thefluxes and the resultant thermal structure formpart of this issue (Schertzer 1987, Schertzer et al.1987), establishing the interannual variability. Thiswork has alerted the limnological community tothe probabilistic elements of long-term forecast-

ing, a facet approached by EI Shaarawi from astatistical perspective (EI Shaarawi 1987). Mea­surements of the exchange between the central andeastern basins were conducted in the summers of1977 and 1978 and analysed by Boyce et al. (1980)and Chiocchio (1981), providing useful estimatesof the hypolimnetic interchange. The matter ofinterbasin exchanges is reviewed in this issued byBartish (1987).

The intensive experiments of 1979 and 1980 andtheir subsequent analyses have extended substan­tially the Project Hypo findings as well as revealingareas of poorly developed knowledge that willrequire further study. Basin scale circulation ana­lyzed by Saylor and Miller (1987) confirms thefindings of Project Hypo concerning the north­westward drift of hypolimnion water in mid­central basin (Blanton and Winklhofer 1972).Unfortunately, the failure of current meters on thenorthwest shore of the central basin precludes anyfirm evidence of how this circulation is completed,although the pattern seems to consist of closedhorizontal gyres (see below). A modeling study bySchwab and Bennett (1987) confirms the findingsof Saylor and Miller that the basin-wide subsurfacecirculation of the central basin is highly change­able. The two-gyre circulation common to theother, deeper Great Lakes basins occurs duringstrong winds, but the circulation may also com­prise only one gyre rotating in either a clockwise oran anticlockwise direction. Lacking the bottomtopography that steers flow into preferred pat­terns, the central basin circulation is sensitive notonly to the strength of the winds but also to thevariation of the wind stress across the basin.Hamblin's (1987) study of storm-surge modelingpoints to our present unsatisfactory ability to inferthe overlake wind field from shore observationsalone, and this concern is also expressed by Schert­zer (1987). Relationships between land and lakewinds have been developed empirically (Schwaband Morton 1984), but Hamblin recommends thatthe relationship be approached as a research prob­lem in meso-scale meteorology. Onshore/offshoretemperature cross-sections made on the northshore(Chiocchio and Boyce 1987) show persistentupwelling of isotherms within 3 km of the shorealso noted by Mortimer (1987), but provide no evi­dence for the topographic and Kelvin waves thatare evident along the north shore of Lake Ontario(Simons 1983). The seasonal warming of the cen­tral basin hypolimnion is consistent with a meanvertical heat conductance that is slightly greater

830 BOYCE et al.

than the molecular value; this would not be so ifthe bottom transport across the basin were com­pensated by large vertical circulation at the edges.The question of how lateral boundaries affect theinterior flow returns again in considering how theenergetic and coherent mid-basin inertial fre­quency current oscillations adjust to the shoreline(Boyce and Chiocchio 1987b). Nothing in the studyof basin-wide circulations would invalidate theProject Hypo conclusion that the offshore regionof the central basin is one of relative homogeneity.Study of the more dense mid-basin array of currentmeters and thermistor strings (Boyce and Chioc­chio 1987a) confirms the general horizontal uni­formity of the meteorologically-forced and reso­nant motions and demonstrates that a substantialportion of the motion can be ascribed to local forc­ing modified by the closed nature of the basin.Subject to direct confirmation, the evidence indi­cates that the central basin hypolimnion, onceformed, and with the exception of influx from theeastern basin, is subsequently modified throughvertical exchanges with the layers above. A two­dimensional infinite channel model of weekly­averaged thermal structure and circulation(Heinrich et al. 1981) supports the hypothesis thatvertical fluxes dominate the mid-lake zone.

If the central basin hypolimnion is modified pri­marily by vertical exchanges, then the signature ofthese changes ought to be visible in time series ofphysical and biochemical parameters in mid-basin.This reasoning prompted the inclusion of anchorstations into the 1979 and 1980 experiments inwhich physical, chemical, and biological parame­ters were measured at least often enough to trackdiurnal fluctuations. These data (Robertson andBoyce 1987) revealed many instances where short­term changes in bottom temperature or dissolvedoxygen could not easily be ascribed to localchanges, but seemed rather to imply the advectionof water of different quality into the experimentarea. A heat budget of the central basin experimentsite was assembled (Royer et al. 1987b). If verticalfluxes alone were active, then to within experimen­tal error, the estimated surface heat flux shouldequal the time rate of change of the heat content ofthe water column. It was found that the two quan­tities differed on daily timescales by amounts thatare significantly greater than expected error. Theaverage difference, over many days, however, issmall. Large differences between daily heat fluxand daily heat content change should then agreewith estimates of the horizontal flux of heat into

the control volume based on data from currentmeters and thermistor strings. This calculation isgreatly hampered by the lack of flow informationabove the 10 m depth. Closing the heat budget withan estimated advective term proved impossible inany rigorous way despite approximate correspon­dence of major events. It was concluded that distri­butions of temperature, oxygen, and presumablyother parameters of the central basin offshoreregion are undersampled at horizontal separationsof 10 km or more. Estimates of horizontal diffu­sion relative to the spatially averaged flow basedon current meter time series at horizontal separa­tions of 10 km suggest that the diffusion process isrelatively weak, and that patchiness can be sus­tained for long periods of time. Sanderson's (1987)study of the movements of a cluster of droguesconfirms this finding. Thus the homogeneityalluded to earlier refers to statistical distributionsand not instantaneous values.

Bearing in mind the intrinsic variability of themid-basin environment at scales of 10 km or less,we now examine the studies of vertically actingphysical processes based on the 1979 and 1980data. This material is reviewed in detail by Royer etal. (1987a). Entrainment of thermocline water intothe hypolimnion was inferred from Project Hypoship cruise data (Burns 19176) and its potential forresupply of oxygen to the hypolimnion wasassessed. Ivey and Boyce (1982) were able to docu­ment this process in the 1979 current and tempera­ture time series data. The downward entrainmentprocess is clearly visible on two occasions in the1979 records but these two events alone are notsufficient to make a significant long-term impacton the oxygen budget of the hypolimnion. A subse­quent paper by Ivey and Patterson (1984) demon­strated that the mixed layer concepts used to simu­late thermocline development in deep lakes couldbe extended to both the upper and lower mixedlayers in the central basin of Lake Erie to track theprocess of downward entrainment. Boyce andChiocchio (1987b) show how the energetic inertialfrequency component of the mean flow may besimulated from inputs of wind data and thermalstructure. A major conclusion from this study isthat the vertical distribution of mean flow in thelake depends on details of temperature profile thatmight easily be overlooked. Under light-to­moderate winds in mid-summer, the central basinmay have three independently moving layers,reverting to two layers under sustained strongwinds. The thicknesses of the moving layers, par-

LAKE ERIE RESULTS 831

ticularly during the three-layer stage, may varyrapidly. A further important conclusion, articu­lated by Saylor and Miller (1987), and affirmed bysubsequent authors, particularly Royer et al.(1987b) in their description of the profiling currentmeter (GVAPS) data, is the necessity to extend themeasurements of current into the top 10 meters ofthe water column. Finally, an obvious next step inthe simulation of vertical structure is to combinethe mixing model and the mean velocity model intoa model that simulates not only the evolving layerdepths and temperatures but also the mean veloci­ties that contribute importantly to shear-generatedturbulence. The development of thermoclinemodels that allow for the detailed structure of bothcurrents and temperature will permit a more quan­titative assessment, among other things, of theentrainment of overlying water into the centralbasin hypolimnion.

SEDIMENT OXYGEN DEMAND IN THECENTRAL BASIN HYPOLIMNION

The conventional view holds that the oxygendemand of the hypolimnion is in two parts, ademand in the water column (WOO) and a demandat the sediment water interface (SOD). Oxygen isphysically resupplied to the central basin hypolim­nion with the occasional entrainment of thermo­cline water into a turbulent hypolimnion, and thelateral flow into the hypolimnion of water fromthe eastern basin. Both of these processes arereviewed in this issue (Bartish 1987, Royer et al.1987a).

Accounting for the sediment oxygen demand hasproven to be difficult. Again, physical processesare involved as mediators of oxygen transfer acrossthe sediment/water interface when there is noresuspension of bottom sediments, and impor­tantly, through the physical process of resus­pension itself. Using laboratory apparatus to mea­sure deposition and entrainment, Lick and Kang(1987) found that freshly sedimented materialresuspends more readily than material depositedonly a few days earlier. Thus compaction and cohe­siveness are important to the stability of sedi­mented materials. Near active sediment sourcessuch as tributary mouths, the character of surficialsediment is likely to change through the season aseasily resuspended material, loaded in the spring,is moved away, leaving behind coarser material inthe summer. Offshore, resuspended sediment ismore nearly in equilibrium with the bottom in that

average deposition rates exceed resuspension ratesby the (small) net rate of accumulation of sedi­ments. The amount of material that is ultimatelymovable is finite and depends on the stress appliedto the bottom. Sediment trap results reflect thenatural seasonal variation of bottom stress; resus­pension in the fall was increased enough to depositvisible layers in the traps during storm events(Charlton and Lean 1987). Bedford andAbdelrham (1987) indicate that direct measure­ment of sediment resuspension is a difficult task,particularly in the central basin of Lake Erie whereat least three different bottom flow regimes(steady, unsteady, and oscillating) may be encoun­tered singly or in combinations. Progress has beenmade in the measurement and interpretation offlow distributions, but the necessary parallel mea­surements of suspended sediment distributionshave yet to be made. Remote sensing yields usefulsynoptic views of horizontal, near-surface distribu­tions of sediments (Mortimer 1987) but quantita­tive assessment via remote sensing awaits the devel­opment of appropriate optical models.

During SOD experiments, the initial rate of oxy­gen uptake was high due to the disturbance of sedi­ments when the chamber was placed on the bot­tom. As Davis et al. (1987) explained, thoseresuspended sediments have a high percentage ofeasily oxidizable substances that consume dis­solved oxygen rapidly when they are mixed into thewater column. Charlton (unpublished data) foundthat the instantaneous oxygen consumption of 1­cm thickness of surface sediment is equivalent to a1 mg/L depletion of the oxygen in a typical 4-m­thick hypolimnion. These observations are consis­tent with the idea that resuspension of bottom sedi­ments contributes to the magnitude and stability ofthe oxygen depletion rate.

The methods used to measure SOD yield differ­ing estimates depending on the amount of distur­bance of the surficial sediments. Experiments arehampered by an inability to match or to comparethe flow field at the sediment surface in the experi­ments with that in the lake. Although the SODrates measured by Snodgrass and Fay (1987) arelower than those reported by Davis et al. (1987),they are still large enough by themselves to accountfor the total net rate of oxygen consumption calcu­lated from concentration changes between surveys.This is difficult to reconcile with our knowledgethat the consumption of dissolved oxygen in thewater column proper is an important term in thebudget.

832 BOYCE et a1.

The sum of SOD and WaD measurements is atleast twice the observed net depletion rate in thehypolimnion and there is as yet no explanation ofwhere the oxygen comes from to make up the dif­ference. WaD (respiration) estimates from closedbottle experiments appear less uncertain than theSOD estimates and they are consistent withobserved 14C uptake rates and the rate of changeof biomass in the water (Charlton 1983). This lendssome credence to the WaD estimates. Neverthe­less, the only measurement with any real certaintyis the net oxygen depletion rate calculated fromsurvey data.

Sediment trap studies of Bloesch (1982),Charlton (1983), Charlton and Lean (1987), andRosa (1985) have shown that an important sourceof particles in the hypolimnion is the bottom itself.Although some of the particles caught in the trapsmust be from shoreline erosion, the increasing fluxof particles into traps close to the bottom leads tothe conclusion that both horizontal and verticaltransport of sediments results from sediment-waterinteractions. The flux of particles into sedimenttraps increases as the traps are placed closer to thebottom sediment surface of the lake. Generally,particles seem to be correlated with oxygen deple­tion (Charlton and Rao 1983). Thus, when WaD ismeasured in bottles, some portion of the oxygenuptake may be due to particles that came from thebottom.

The particles, solutes, and gases resuspended orentrained in an SOD experiment are preventedfrom mixing with the open water column. If weconsider that the oxygen consumption of the parti­cles suspended in an SOD experiment would nor­mally occur in the open water column and is mea­sured by WaD experiments, then some portion ofthe total hypolimnion oxygen consumption is rep­resented twice in the sum of WaD + SOD derivedexperimentally.

The extent of this double counting of oxygenconsumption is not yet known, but we can conjec­ture from the sediment trap results (Charlton andLean 1987) that it may be important. The doublecounting is not removed by the usual practice ofsubtracting the volumetric WaD from apparentSOD; a correction could be made by subtractingthe areal resuspended consumption from apparentSOD but present information is insufficient to dothis. There must, however, be some oxygen con­sumption associated with the sediments and we donot propose that this is inconsequential. Since the"sediment water interface" is not an interface but is

really a mobile gradient of sediment/water concen­tration, there is a question of how to define what isa water sample and what is a sediment sample.Nevertheless, the results of SOD experiments aredependent on the stirring speed because both diffu­sion and resuspension are augmented by watermotion. As pointed out by Lam et al. (1987b), theoxygen depletion of ice-covered lakes (Mathias andBarica 1980) is related to mean depth and thisallows the sediment related effect to be deter­mined. In eutrophic lakes the sediment effect wasequivalent to 0.23 g O/m-2/day compared to 0.08g 02/m-2/day in oligotrophic lakes. Even afterdoubling these rates to correct for temperature dif­ferences between the lakes, the rate (SOD = 0.88 g02/m-2/day) of Davis et al. (1987) seems too highfor Lake Erie. In Mathias' and Barica's study, chlo­rophyll in the eutrophic lakes was 4-20 timeshigher than in Lake Erie and sediment organicmatter was up to 25010 of the total dry material byweight compared to the 10% in Lake Erie. LakeErie most closely resembles the "oligotrophic"lakes in Mathias and Barica (1980). If we assumethen that the real consumption in Lake Erie sedi­ments is between 0.08 and 0.16 g 02/m-2/day muchof the apparent discepancy between experimentalconsumption rates and the observed oxygen deple­tion is eliminated.

The hypothesis that the bottom sediment actson oxygen partly by supplying oxygen-consumingmaterials to the water column is consistent with:(1) the observations that part of the depth effecton oxygen depletion is due to sediment/waterinteractions; (2) the notion that there is a short­and long-term buffering or stability in oxygendepletion caused by storage of sediment organiccarbon; and (3) observations by Burns (1976) andCharlton and Rao (1983) that the difference inoxygen depletion between the central and easternbasins correlates well with differences in particleconcentrations in the water; and (4) observationsby Charlton and Lean (1987) that sediments areroutinely resuspended and the downflux of neworganic matter into the hypolimnion is insuffi­cient to maintain the oxygen depletion processduring stratification.

If we could assure repeatability of or at leastaccount for changes in the oxygen-consumingpotential of sample sediments, then we might con­sider laboratory oxygen demand experiments inwhich both particle concentrations and turbulenceare measured variables. The problems of recreat­ing "natural" boundary layer turbulence in an arti-

LAKE ERIE RESULTS 833

ficial enclosure may force the adoption of a simplegrid-stirring device such as that utilized by Tsaiand Lick (1986) to compare sediment samples. Theloss of exact dynamic similarity would be offset byrepeatability and easily-measured energy input.

Mortimer (1987) attempts an approximate sedi­ment budget of the three Lake Erie basins and sug­gests how sediment transport among the basinsmay affect nutrient and contaminant distributions.The experiments discussed above relate to animportant local aspect of sediment movement butignore sediment transport at basin scales.

SYSTEMS MODELING OF LAKE ERIE

Readers of Lam and Schertzer (1987), Lam et al.1987a and 1987b, and DiToro et al. 1987, togetherwith their antecedents will be impressed with theenormous task of synthesis undertaken by theseworkers and by the skill of the models they haveproduced in simulating water quality in Lake Erie(in particular the dissolved oxygen concentrationsin the central basin hypolimnion). Not only havethese models reproduced the annual cycle of dis­solved oxygen in any given year, but they have alsobeen able to incorporate the substantial interan­nual variations in weather and significant changesin nutrient loading.

The success of the models demonstrates that thesystem is predictable within the ranges of the forc­ing variables encountered to date. It is not certainin all respects however, that the models are exactanalogues of the system itself, a condition thatshould be met if the models are to be used in apredictive mode with complete accuracy. Providedthat the models are applied to simulate conditionsthat fall within the tested range and that the confi­dence limits of the results are taken into account,useful predictions can be made.

The situation is relatively clear on the physicalside. The extension of the concepts of moleculardiffusion to a turbulent environment provides arobust mixing model that is readily adjustable toparticular systems (Lam et al. 1987a, 1987b). Theturbulent kinetic energy models applied to a mixedlayer, such as that of Ivey and Patterson (1984),provide another approach, equally robust. Bothapproaches are combined in the model of Mellorand Durbin (1975) where the turbulent kineticenergy equation is solved to generate values of theeffective diffusivity. One conclusion of the physi­cal studies reported in this issue is that the majorexchanges between the hypolimnion and the rest of

the central basin take place as vertical fluxes (withthe exception of the horizontal exchanges with theeastern basin that are explicitly accounted for).The driving forces, heat flux and wind stress, varylittle across the basin. On the other hand, the mix­ing model used by DiToro and his colleagues is notas realistic a representation of the physical systemsince interface depths are fixed, except in extremecases, although heat transfer is adequately simu­lated by diffusion.

On the biological and chemical side, each modelrepresents an individual act of synthesis on the partof its creators for there are few clear indications inthe biological literature that a relationship shouldtake one shape and not another. The Lam modelarises out of earlier work by Simons and Lam(Simons 1979) in Lake Ontario and these authorsadmit that some biologically unconventional for­mulations were needed. Clearly, there has to besome grouping of the many recognized nutrientforms, the many active and interactive species inthe food web, and possibly some elimination ofcompartments, species, and interactions of minorimportance. Just as clearly, there is a limit to thedegree of simplification allowable if the complex­ity of response is to be followed. The DiToromodel lists 15 biochemical variables to be simu­lated in each of thre~ layers. The Lam dissolvedoxygen model operates with three. This differencebetween these two models reflects the modelers'purposes. Unlike the physical component of themodels, we do not know whether or not the bio­chemical components are working simplificationsarrived at from a deliberate scaling and lumping ofa set of primitive equations. This state of affairs isa consequence of the lack of quantitative informa­tion on biochemical processes.

Note that the models of Lam and associates pre­dict a mild response of oxygen depletion toreduced phosphorus loadings similar to the 0-2mg/L in the historic response to increased loadingsin Charlton (1987). The response predicted byDiToro et al. (1987) seems to be larger than in Lamet al. and is perhaps somewhat optimistic in thelight of an anoxia occurrence in 1985.

A major advance in approach to the modelingwork has been the translation of the effects of sto­chastic factors such as weather into probabilitiesbased on the historic data. This means that long­term forecasting now can contain, for example, theprobability of certain oxygen levels likely to occurunder steady nutrient loading. Thus, there hasbeen progress from the realization that weather-

834 BOYCE et al.

driven physical factors affect the biology andchemistry (e.g., Burns 1976, Charlton 1980) to theuse of weather and nutrient loading data to predictconditions in the lake.

The modeling results tend to confirm the origi­nal management decisions to reduce nutrientloads; but the need for surveillance monitoring ofwater quality and further research into lake proc­esses remains. For example, the response of oxy­gen to reduced phosphorus loads is so weak, andexperimental data sufficiently unreliable, that themodeled response of Lam et al. is based on fittedparameters of SOD. New research by Manning andMayer (1986) indicates that sediments have becomeincreasingly oxidized and have increased phos­phorus retention capabilities in the last 5 years.Thus, while the effects of lake management areoccurring, the continuing research helps delineatethe changes and provides the information to inter­pret how and why the changes occur. Gaps in mod­eling include the apparent lag in responses tochanging loads and the processes associated withparticle settling, cohesion, resuspension, and car­bon storage. Indeed, these sediment/water interac­tions may be important in the recharge of contami­nants to the water and their subsequentvolatilization or burial.

The use of either model as a predictive tool issubject to the restrictions mentioned above. Witheach season's data comes the opportunity to verifythe models anew and to extend the range of inputsover which they function reliably. Models servefunctions other than the prediction of the system'sresponse to hypothetical inputs. They can provideinsights that are beneficial to many studyelements:1) On the physical side, it has been stated by Royeret al. (1987a) and demonstrated by Boyce andChiocchio (1987b) that in a variable, unevenlyforced environment, cause and effect relations arevery difficult, even impossible to establish fromobservations alone and that models play an essen­tial role in developing and testing hypotheses. Thatthe present system models work with the data nowin hand should be a positive and creative challengeto biologists and chemists to explain why or tooffer for testing alternative, more realisticrelationships.2) The models can be exploited as a guide to aneconomical and effective monitoring strategy.Time series of temperature and dissolved oxygendistributions measured at one or two locations inthe central basin could be extended to the whole

basin using a model Such an approach might repaythe development and deployment costs out of sav­ings in ship usage, and at the same time providedata that are much better suited to surveillancerequirements. An integration of modeling and sur­veillance activity could payoff handsomely.3) A major conclusion of this entire experiment isthat a careful reassessment of the so-called sedi­ment oxygen demand is needed. There is evidencethat the conceptual models through which theavailable data have been interpreted may be incor­rect. The study of sediment-water interactions,crucial not only to the issue of eutrophication ofLake Erie, but also to the recycling and/orremoval of toxic contaminants in Lake Erie andelsewhere, is one where modeling will be essentialto interpreting the difficult but necessary fieldexperiments.

CONCLUSIONS

We hasten to point out that not all of the workpresented in this issue has received equal weight inthis summary. This reflects the biases of thepresent authors as well as their desire to tell acogent, if simplified story. We hope that readersapproaching this issue with many different pointsof view will all find something of value. At thesame time we do not suggest that we have summa­rized all the important Lake Erie work to havebeen completed in the decade since the "Lake Eriein the early seventies" issue of the Journal of theFisheries Research Board of Canada (Vol. 33, No.3).

In that special issue of JFRB, Burns and his col­leagues (Burns et al. 1976) evaluated the scientificknowledge and data available in 1975 in the hopethat an overall model or ". . . at least a qualitativedescription of processes in the lake could be givensuch that water movements, sediment characteris­tics, and nutrient loadings could be linked toexplain phytoplankton, zooplankton, and fish dis­tributions." In their summary discussions they con­cluded that" ... in nearly every study, some neces­sary rate process had not been investigated with theconsequence that the results and findings of thestudy remained inadequate for incorporation intoa truly predictive model." They published a dia­gram summarizing their view of the model compo­nents (Fig, 1) that we reproduce here. At that time,only the physical transport model was consideredto be adequate for predictive purposes. In thatpaper, the authors presented a list of 16 studies

LAKE ERIE RESULTS 835

OUTPUT - CURRENTS,- WAVE HIS,- STORM SURGES,- TEMPERATURES.

OUTPUT - SEDIMENTDISTRIBUTION.

- SHORE EROSION.

OUTPUT- PHYTOPLANKTONDENSITIES

- HYPOLIMNIONOXYGENCONCENTRATIONS

OUTPUT- FISHPOPULATIONS

BENTHICMODEL

SOLUBLE +PARTICULATE

NUTRIENTCONCENTRATION

SETTLING MODELL...--------1 LIVE AND

INANIMATEPARTICLES

HYDRODYNAMIC

MODEL

SEDIMENT-WI\TERINTERACTION

MODEL

SOLUBLE +PARTICULATE

EXTERNAL INPUTS

* MODEL WHERE KNOWLEDGEIS PRESENTLY ADEQUATE FORTHE PREDICTIVE MODELLINGOF LAKE ERIE.

FIG. 1. Schematic outline of the more important data inputs, models, and linkages between models required toformulate a predictive model for a whole lake system as conceived by Burns et al. (1976).

required to complete the scheme. A measure of thefar-sighted care with which the diagram and the listwere drawn is the fact that many of the itemsremain as valid research topics; it is a measure ofthe energy devoted to Lake Erie studies in theintervening decade that useful progress has beenmade with many of them. In Appendix A we dis­play a summary assessment of our progress downthe list of 16 items; in Appendix B we present ourown list.

We do not, however, present an updated versionof Figure 1, for the modeling studies of Lam et al.and DiToro et al. (1987) challenge 'the reductionistviewpoint on every side and demonstrate that real­istic simulations are possible with a much simpli­fied network of processes.

Scattered through this issue, and indeed in thissummarizing paper, are many individual conclu­sions. Holding to the agreed focus on processes

relating to oxygen depletion in the central basin,we would conclude that with the exception of sedi­ment movements and an assessment of the impor­tance of upwelling, physical processes are reasona­bly well accounted for. That is to say they can beordered and scaled. Modeling studies convincinglydemonstrate that vertical exchange processesresponsible for the formation and maintenance ofstratification strongly influence the susceptibilityof the central basin hypolimnion to lake-summeranoxia. Physical studies, notably those of Royer etal. (1987) show the difficulties of inferring verticalfluxes from direct control volume experiments,and underline the importance of simulation model­ing in testing hypotheses over time intervals longenough to average out variability. The non­physical studies seem to converge on the conclu­sion that the sediment/waterIbenthos interactionsrequire better understanding. The pursuit of this

836 BOYCE et al.

understanding will involve further multidiscipli­nary work and eventual synthesis of it throughmodeling.

As a community resource, it is no exaggerationto say that Lake Erie is the most valuable of all theGreat Lakes (Burns 1985). But the population den­sity in its basin both defines and undermines itsvalue. As a result of scientific study of Lake Erie,we are now engaged in one of the most importantlarge-scale ecological experiments ever under­taken - to determine if and when eutrophic condi­tions can be reversed in a large lake through con­trol of loadings. Beyond that experiment therelooms another, perhaps more important still-todetermine the potential for recovery after reduc­tion in inputs of toxic chemicals. Indeed we couldsay that a state of environmental stress may bepersistent over a long time, although the actualstresses may change. As J. Vallentyne (Great LakesFisheries Research Laboratory, Burlington,Ontario; personal communication) has oftenpointed out, it is not just the Great Lakes basinthat is involved in the outcome, it is most of theindustrialized world. For if the wealthy and ener­getic society of the Great Lakes basin cannot findthe will, the resources, and the stamina to followthis venture through, the chances of success else­where are small.

Environmental science, like medicine, in addi­tion to concern with the management of specificcrises or illnesses, should be devoting energy tomaintaining wellbeing. Another role for environ­mental scientists is that of a curator, charged notonly with the protection of the treasure entrustedto his care, but also with fostering public enthusi­asm and respect for it. Unlike the troubleshooter,the curator's role requires continued contact withthe subject. It is oriented to process, not products.While this special issue on Lake Erie is indeed aproduct, it is also part of the process of staying intouch with the lake during a period of rapid, andwe think hopeful change. Although the decision towork together to prepare a joint report with somethought toward its integration has had its cost interms of delay, it has maintained and we hopestrengthened an international network of scientistswho have long-term interests in Lake Erie.

The 1979 and 1980 experiments did not occurspontaneously but were developed in conjunctionwith the I.J .C.-sponsored intensive surveillanceprogram. We feel that such programs have pro­vided excellent incentives for focussing energy onGreat Lakes problems in quantities sufficient to

make real progress. The long-term monitoring andsurveillance of the Great Lakes should allow forboth intensive experiments and synthesizingreviews. The collection of studies presented here isevidence of scientific progress, yet there remainsinteresting and useful work to be done. As we sug­gested earlier, the proper goal may be more thanthe final completion of this work; the questionsremaining also revitalize the process of staying intouch. Environmental science provides an essentialfeedback between human society and the substratethat society depends upon; thus the ecology ofLake Erie should include its students.

ACKNOWLEDGMENTS

The efforts of the many contributors to this issue,particularly their continuing faith and patience,are gratefully acknowledged, as is the support ofthe National Water Research Institute, the GreatLakes Environmental Research Laboratory, andthe Center for Lake Erie Area Studies. We thankthe staff of the International Joint CommissionRegional Office, in Windsor, Ontario, for hostingthe many workshops and meetings necessary to thepreparation of this issue.

REFERENCES

Bartish, T. 1987. A review of exchange processes amongthe three basins of Lake Erie. J. Great Lakes Res.13:607-618.

Bedford, K. W., and Abdelrhman, M. 1987. Analyticaland experimental studies of the benthic boundarylayer and their applicability to near-bottom transportin Lake Erie. J. Great Lakes Res. 13:628-648.

Blanton, J. 0., and Winklhofer, A. R. 1972. Physicalprocesses affecting the hypolimnion of the CentralBasin of Lake Erie. In Project Hypo, ed. N. M.Burns and C. Ross, pp. 9-38. Paper No.6, CanadaCentre for Inland Waters, Burlington, Ontario.USEPA Tech. Rept. TS-05-71-208-24.

Bloesch, J. 1982. Inshore-offshore sedimentation dif­ferences resulting from resuspension in the EasternBasin of Lake Erie. Can. J. Fish. Aquat. Sci. 39:748-759.

Bourbonniere, R. A., VanSickle, B. L., and Mayer T.1986. The Great Lakes Sediment Bank-I. NationalWater Research Institute Contribution No. 86-151.

Boyce, F. M., and Chiocchio, F. 1987a. Water move­ments at a mid-central basin site: time and spacescales, relation to wind, and internal pressure gradi­ents. J. Great Lakes Res. 13:530-541.

____ , and Chiocchio, F. 1987b. Inertial frequencycurrent oscillations in the central basin of Lake Erie.J. Great Lakes Res. 13:542-558.

LAKE ERIE RESULTS 837

____ , Chiocchio, F., Eid, B., Penicka, F., andRosa, F. 1980. Hypolimnion flow between the centraland eastern basins of Lake Erie during 1977. J. GreatLakes Res. 6:290-306.

Burns, N. M. 1976. Oxygen depletion in the Central andEastern Basins of Lake Erie. J. Fish. Res. BoardCan. 33:512-519.

____ . 1985. Erie: The lake that survived. Totowa,N.J.: Rowan and Allanheld.

____ , and Ross, C. 1972. Project Hypo. PaperNo.6, Canada Centre for Inland Waters, Burlington,Ontario. USEPA Tech. Rept. TS-05-71-208-24.

____ , Jacquet, J.-M., Kemp A. L. W., Lam, D.C. L., Leach, J. L., Munawar, M., Simons, T. J.,Sly, P. G., Thomas, R. L., Watson, N. H. F., andWilliams, J. D. H. 1976. Processes within Lake Erie.J. Fish. Res. Board Can. 33:639-643.

Charlton, M. N. 1980. Oxygen depletion in Lake Erie:Has there been any change? Can. J. Fish. Aquat. Sci.37:72-81.

____ . 1983. Downflux of sediment, organic mat­ter, and phosphorus in the Niagara River area ofLake Ontario. J. Great Lakes Res. 9:201-211.

____ . 1987. Lake Erie oxygen revisited. J. GreatLakes Res. 13:697-708.

____ , and Lean, D. R. S. 1987. Sedimentation,resuspension, and oxygen depletion in Lake Erie(1979). J. Great Lakes Res. 13:709-723.

____ , and Rao, S. S. 1983. Oxygen depletion incentral and eastern Lake Erie: relationship with bac­teria, chlorophyll, POC, and morphometry. J. GreatLakes Res. 9:3-8.

Chiocchio, F. 1981. Lake Erie hypolimnion and meso­limnion flow exchange between the Central and East­ern Basins during 1978. National Water ResearchInstitute unpublished report, Internal Report 009,Burlington, Ontario.

____ , and Boyce, F. M. 1987. Upwelling on thenorth shore of Lake Erie's Central Basin during thesummer of 1979. National Water Research Institute,Internal Report, Burlington, Ontario.

Corps of Engineers (COE) 1982. Lake Erie wastewatermanagement study report. Buffalo District, Buffalo,New York.

Davis, W. S., Fay, L. A., and Herdendorf, C. E. 1987.Overview of USEPA/CLEAR Lake Erie sedimentoxygen demand investigations during 1979. J. GreatLakes Res. 13:731-737.

DiToro, D. M., Thomas, N. A., Herdendorf, C. E.,Winfield, R. P., and Connolly, J. P. 1987. A postaudit of a Lake Erie eutrophication model. J. GreatLakes Res. 13:801-825.

Dobson, H. H., and Gilbertson, M. 1972. Oxygendepletion in the hypolimnion of the Central Basin ofLake Erie, 1929 to 1970. In Project Hypo, ed. N. M.Burns and C. Ross. Paper No.6, Canada Centre for

Inland Waters, Burlington, Ontario. USEPA Tech.Rept. TS-05-71-208-24.

El-Shaarawi, A. H. 1987. Water quality changes inLake Erie, 1968-1980. J. Great Lakes Res.13:674-683.

Fraser, A. S. 1987. Tributary and point source totalphosphorus loading to Lake Erie. J. Great LakesRes. 13:659-666.

Hamblin, P. F. 1987. Meteorological forcing and waterlevel fluctuations on Lake Erie. J. Great Lakes Res.13:436-453.

Heinrich, J., Lick, W., and Paul, J. 1081. Temperaturesand currents in a stratified lake: a two-dimensionalanalysis. J. Great Lakes Res. 7: 264-275.

International Joint Commission (He). 1978. Environ­mental management strategy for the Great Lakes sys­tem. Final report from the Pollution from Land UseActivities Reference Group (PLUARG). Windsor,Ontario.

Ivey, G. N., and Boyce, F. M. 1982. Entrainment bybottom currents in Lake Erie. Limnol. Oceanogr. 27:1029-1038.

____ , and Patterson, J. C. 1984. A model of thevertical mixing in Lake Erie in summer. Limnol.Oceanogr. 29: 553-563.

Jerome, J., Bukata, R. P., and Bruton, E. 1987. Exist­ing and anticipated irradiation in the hypolimneticwaters of central Lake Erie. J. Great Lakes Res. 13:568-577.

Lam, D. C. L., and Jacquet, J.-M. 1976. Computationsof physical transport and regeneration of phosphorusin Lake Erie, fall 1970. J. Fish. Res. Board Can. 33:550-563.

____ , and Schertzer, W. M. 1987. Lake Erie ther­mocline model results: comparison with 1967-1982data and relation to anoxic occurrences. J. GreatLakes Res. 13:757-769.

____ , Schertzer, W. M., and Fraser, A. S. 1983.Simulation of Lake Erie water quality responses toloading and weather variations. Environment Can­ada Scientific Series No. 134.

____ , Schertzer, W. M., and Fraser, A. S. 1987a.Oxygen depletion in Lake Erie: modeling the physi­cal, chemical, and biological interactions, 1972 and1979. J. Great Lakes Res. 13:770-781.

____ , Schertzer, W. M., and Fraser, A. S. 1987b.A post-audit analysis of the NWRI nine-box waterquality model for Lake Erie. J. Great Lakes Res.13:782-800.

Lick, W., and Kang, S. W. 1987. Entrainment of sedi­ments and dredged materials in shallow lake waters.J. Great Lakes Res. 13:619-627.

Logan, T. J. 1987. Diffuse [non-point] source loadingof chemicals to Lake Erie. J. Great Lakes Res.13:649-658.

Manning, P. G., and Mayer, T. 1986. Iron-phosphoruslayers in sediments of Lake Ontario. Internal Report,

838 BOYCE et al.

National Water Research Institute, Burlington,Ontario.

Mathias, J. A., and Barica, J. 1980. Factors controllingoxygen depletion in ice-covered lakes. Can. J. Fish.Aquat. Sci. 37: 185-194.

Mellor, G. L., and Durbin, P. A. 1975. The structureand dynamics of the ocean mixed layer. J. PhysicalOceanogr. 5: 718-728.

Mortimer, C. H. 1987. Fifty years of physical investiga­tions and related limnological studies on Lake Erie,1928-1977. J. Great Lakes Res. 13: 407-435.

Oliver, B. G., and Bourbonniere, R. A. 1985. Chlorin­ated contaminants in surficial sediments of LakesHuron, St. Clair, and Erie: implications regardingsources along the St. Clair and Detroit Rivers. J.Great Lakes Res. 11: 366-372.

Robbins, J. A. 1982. Stratigraphic and dynamic effectsof sediment reworking by Great Lakes zoobenthos.Hydrobiol. 92: 611-622.

____ , Krezoski, R., and Mozley, S. C. 1977.Radioactivity in sediments of the Great Lakes: post­depositional redistribution by deposit-feeding orga­nisms. Earth Planet. Sci. Lett. 36: 325-333.~ , Edgington, D. N., and Kemp, A. L. W.

1978. Comparative Pb21O, Cs137, and pollen geo­chronologies of sediments from Lakes Ontario andErie. Quatern. Res. 10: 256-278.

Robertson, D. G., and Boyce, F. M. 1987. Anchor sta­tion experiments in central Lake Erie, 1979 and 1980.National Water Research Institute, Internal Report,Burlington, Ontario.

Rosa, F. 1985. Sedimentation and sediment resus­pension in Lake Ontario. J. \ Great Lakes Res. 11:13-25.

____ . 1987. Lake Erie central basin total phos­phorus trend analysis from 1968 to 1982. J. GreatLakes Res. 13:667-673.

____ , and Burns, N. M. 1987. Lake Erie centralbasin oxygen depletion changes from 1929-1980. J.Great Lakes Res. 13:684-696.

Royer, L., Chiocchio, F., and Boyce, F. M. 1987a.Tracking short-term physical and biological changesin the central basin of Lake Erie. J. Great Lakes Res.13: 587-606.

____ , Hamblin, P. F., and Boyce, F. M. 1987b. Acomparison of drogues, current meters, winds, and avertical profiler in Lake Erie. J. Great Lakes Res. 13:578-586.

Sanderson, B. 1987. An analysis of Lagrangian kine­matics in Lake Erie. J. Great Lakes Res. 13:559-567.

Saylor, J. H., and Miller, G. S. 1987. Studies of large­scale currents in Lake Erie, 1979-80. J. Great LakesRes. 13: 487-514.

Schertzer, W. M. 1987. Heat balance and heat storageestimates for Lake Erie, 1967 to 1982. J. Great LakesRes. 13: 454-467.

____ , Saylor, J. H., Boyce, F. M., Robertson, D.G., and Rosa F. 1987. Seasonal thermal cycle of LakeErie. J. Great Lakes Res. 13: 468-486.

Schwab, D. J., and Bennett, J. R. 1987. Lagrangiancomparison of objectively analyzed and dynamicallymodeled circulation patterns in Lake Erie. J. GreatLakes Res. 13: 515-529.

____ , and Morton, J. A. 1984. Estimation ofoverlake wind speed from overland wind speed: acomparison of three methods. J. Great Lakes Res.10:68-72.

Simons, T. J. 1976. Continuous dynamical computa­tions of water transports in Lake Erie for 1970. J.Fish. Res. Board Can. 33:371-384.

____ . (Ed.) 1979. Assessment of water qualitysimulation capability for Lake Ontario. EnvironmentCanada Scientific Series No. 111.

____ . 1983. Resonant topographic response ofnearshore currents to wind forcing. J. Phys.Oceanog~ 13:513-23.

Snodgrass, W. J., and Fay, L. 1987. Values of sedimentoxygen demand measured in the central basin ofLake Erie, 1979. J. Great Lakes Res. 13:724-730.

Tsai, C.-H., and Lick, W. 1986. A portable device formeasuring sediment resuspension. Paper presented at29th Great Lakes Research Conference, May, 1986,Scarborough, Ontario.

Williams, J. D. H., Shear, H., and Thomas, R. L. 1980.Availability to Scenedesmus quadricauda of differentforms of phosphorus in sedimentary materials fromthe Great Lakes. Limnol. Oceanogr. 25: 1-11.

Young, T. C., J. V. DePinto, S. C. Martin, and Bonner,J. S. 1985. Algal-available particulate phosphorus inthe Great Lakes basin. J. Great Lakes Res. 11 :434-446.

APPENDIX A: Assessment of progress with the16·item list of Burns et al. (1976).

a) Careful measurements of the pattern of watermass exchange between the three basins . . . .

Exchange mechanisms are largely understood andrough quantitative assessments have been made ofthe exchange between the central and easternbasins. More work is needed to delineate theimportance of intrusions of the central basin hypo­limnion into the western basin.

b) An investigation of the resuspension of fine­grained sediments by currents and wind-waves,together with the extraction of nutrients from thesediments as a result of the hyrodynamic action

LAKE ERIE RESULTS 839

These processes continue to be recognized as apotential key to our understanding of the centralbasin response to changing nutrient loading. Mea­surements are difficult but capabilities are grow­ing. Much work has been accomplished on thepotential bioavailability of sediment phosphorus,but little information is available on actual in-lakeutilization (Young et af. 1985). The issue may alsobe central to contaminant/water interactions.

c) An investigation of storm-episode sedimenttransport within the lake in contrast to sedimenttransport under non-storm conditions ....

Related to b) above. Some studies are underway inthe Sandusky area. Remote sensing can describewhole-basin patterns qualitatively; but quantita­tive assessment by this means awaits developmentof appropriate optical models, based on joint fieldwork and image analysis.

d) A careful examination of the sediment-waterinterface with emphasis on the form of nutrientsand toxic materials in the centimeter below theinterface and in the meter of water above ....

Related to b) above. Studies are underway in theSandusky area and in the central basin.

e) River inputs of phosphorus, nitrogen, and sus­pended solids, in particular, should be examined asa function of flow so that accurate loading esti­mates of materials can be calculated; atmosphericinputs of nutrient elements should also be mea­sured .... More emphasis on event-sampling hasincreased the accuracy of loading estimates, butsubstantial uncertainties remain.

o The apatite content of river waters should bemeasured and a test of apatite solubility in LakeErie waters and sediment interstitial waters shouldbe carried out . . . .

Apatite has been found to be incapable of support­ing algal growth in availability experiments(Williams et af. 1980, Young et af. 1985).

g) The loading estimates of fine-grained materialsand associated elements to the sediments should beimproved by ensuring adequate basin coverage inthe cores examined . . . .

Extensive sediment surveys have not been contin­ued. New work reveals substantial downward mix­ing in the sediments (Robbins et af. 1977, Robbinset af. 1978, Robbins, 1982, Oliver and Bourbon­niere 1985, Bourbonniere et af. 1986). Some stud­ies have been carried out by Robbins and Mudroch(Great Lakes Research Laboratory, Ann Arbor,MI; personal communication) on sedimentationrates in each basin as related to the input of metals,particularly lead.

h) A major study of aerobic heterotroph bacterialactivity in the lake should be carried out in con­junction with the studies (i) and (j) described below

Bacterial numbers seem to. be correlated with oxy­gen depletion (Charlton and Rao 1983).

i) Respiration and other oxygen-consuming proc­esses should be studied to the same extent as acomplementary study of primary productivity.These processes should be investigated both in thewater and at the sediment-water interface ....

A report on primary production and respiration isin preparation (Charlton, National WaterResearch Institute, Burlington, Ontario, personalcommunication). Most new carbon is respired inthe epilimnion. Net community production is ofthe same order of magnitude as the central basinoxygen depletion rate. The interpretation of directmeasurements of sediment oxygen demand isunder discussion.

j) The rates of decay and formation of detritusand soluble organic material should be carefullydetermined . . . .

No progress.

k) Phytoplankton (including microplankton) spe­cies and densities, together with their associatedproductivity at all depths in the lake, should bedetermined . . . .

This has not been attempted; expensive, multi-yearstudies are needed. A less reductionist approach isnow popular.

1) The vertical migration of both phytoplankton

840 BOYCE et al.

and zooplankton should be studied in relation toepilimnion nutrient budgets ....

Not attempted.

m) The cause of major algal blooms should bedetermined, if possible ....

Algal blooms have been infrequent since 1970.

n) Rotifers must be included in the study of thezooplankton and the grazing pattern of all themajor species should be examined ....

Not attempted.

0) A lake-wide program to study the benthosshould be carried out . . . .

A program is in the planning stages at the NationalWater Research Institute.

p) The effect of fish on phytoplankton and zoo­plankton populations should be determined ....

Not attempted but is possibly important with theresurgence of the walleye fishery and the findingsconcerning impact of stocked species on the foragefish in Lake Michigan.

APPENDIX B: Specific researchrecommendations.

1) An array of instruments deployed through thestratified period on the northwest shore of the cen­tral basin would improve our understanding of thehypolimnion circulation and the importance ofupwelling on the shore. With the existing shipdata, we are unable to gauge the variability of thisphenomenon under changing wind conditions.

2) Time series measurements of currents and tem­peratures in Pelee Passage using the new currentmeters and temperature recorders are needed todocument the frequency of intrusions of centralbasin hypolimnion water into the western basin.

3) The question of the importance of sedimentresuspension and transport will require much fullerinvestigation, first in an assessment of available,unanalyzed data, and later in the form of a multi­disciplinary experiment. The interaction of the

resuspended sediment with the water column needsto be determined for both uptake and release proc­esses. Furthermore, the possible effects of volatil­ization of contaminants from resuspended sedi­ments should be investigated. Carefully designedin situ experiments aimed at collecting properlysampled, long time series of these processes will berequired, and these in turn will need substantialand long-term engineering development. Tech­niques under consideration are time-lapse photog­raphy, acoustic back-scattering, and micro­electrode profiling of sediments. Physicalcomponents are under active development in thecourse of the Upper Great Lakes ConnectingChannels study referred to above. The central andwestern basins of Lake Erie are good places forthis work.

4) The implications of the lag in oxygen responseto reduced phosphorus loadings should be extrapo­lated to other issues such as toxic contaminants.Since some toxicants accumulate in the sediments,how do sediment mixing and rapid burial causedby basin erosion combine to produce a response insediment characteristics? Is there a linkagebetween trophic state, mean depth, and sedimentinputs that may govern the sensitivity of Lake Erie(and the other Great Lakes) to contaminantloads?

5) What is the in-lake utilization of "bioavailable"sediment-bound phosphorus, and what will be theeffect of reduced erosion of farmland in thebasin?

6) The eastern basin has been relatively unre­searched due to the concentration of effort on thecentral basin oxygen problem. Although the hypo­limnion oxygen is suitable for fish in the easternbasin, there are local depressions in the thermo­cline oxygen that approach levels hazardous tofish. The cause of this phenomenon should bedetermined and other aspects such as sedimentstorage/export and the effect of shoreline develop­ment on nearshore water quality should beinvestigated.

7) The difference between summer and winterprocesses should be studied. Total phosphorustends to disappear from offshore water in the sum­mer. The lake appears to be more productive in thecold water, fully mixed period when the entirewater column is in contact with the sediments.