With growing use of numerical models to forecast lake conditions under future climates and other stressors, paleo-events in the history of the Great Lakes have greater potential for relevance. Past events and history may extend records of observations, provide estimates of the sensitivity of the lake system to stressing conditions, and contribute scenarios for model validation. Here we describe four examples that hold promise for improving understanding of the present and future Great Lakes: 1) using an event of lake closure to derive climate-hydrology sensitivity, 2) extending the record of lake-level history by examining beach ridge sequences, 3) investigating sedimentary black bands to indicate past anoxia at the lakebed in deep basins, and 4) deriving evidence of lake process teleconnections with atmospheric circulation.
In a recent assessment of the readiness of the scientific community to meet long term challenges for the Great Lakes, Brant (2007) noted, among other points, that improved technologies are needed to expand the time, space and parameter scales by which the ecosystem is observed, and that research should be focused more on prediction than on explanation. The greater focus on prediction implies an increasing use and dependence on numerical models to forecast future limnological and ecological conditions under the impact of climate change and other stressors. With a greater dependence on modeling, there comes a likely greater relevance for past events in Great Lakes history because such events potentially can extend records of observations, contribute evidence of the sensitivity of the lake system to stress, as well as provide scenarios for model validation. The history of the Great Lakes (Figure 1) since their formation during the last deglaciation provides many examples of changing climate, hydrology, and limnology. In this paper, we present four examples of lake history that could lead to a better understanding of present and future lake conditions: (1) using an event of lake closure to derive climate-hydrology sensitivity, (2) extending the record of lake-level history by examining beach ridge sequences, (3) investigating sedimentary black bands to indicate past anoxia at the lakebed in deep basins, and (4) deriving evidence of lake process teleconnections with atmospheric circulation.
As a proxy for Great Lakes history, we use the history of lake levels in the Huron basin over the past 12 000 14C (∼14 000 cal) years, shown by the lowest plot in Figure 2. Original lake levels are based on a recent analysis of glacial isostatic adjustment (Lewis et al., 2007a, b). The upper two curves of Figure 2 illustrate relative changes in air temperature and moisture, as determined for the southeastern Great Lakes using the δ18O isotopic composition of cellulose in fossil trees and lake sediments as a proxy (Edwards et al., 1996).
Hydrologic Lake Closure
Future climates under a regime of global warming may lower Great Lakes levels beyond known variations in the instrumentally-observed data set (Mortsch and Quinn, 1996; Mortsch et al., 2000). By deriving the climate-hydrology relationship of the Great Lakes for a severe variation of the paleoclimate-lake system, we will provide new evidence of climate-hydrology sensitivity, and add confidence to modeled projections of future lake conditions. We have recently found that lake levels in the Huron Basin fell tens of metres below the overflow outlet (near North Bay, Ontario, Figure 1), about 8000 14C (8790–8990 cal) BP, possibly forming isolated lakes within bathymetric depressions (Lewis et al., 2007a; 2007b; Figure 3). This condition began when upstream glacial meltwater and glacial lakes discharged directly to the Ottawa River (Figure 1) and no longer supported water levels in the basins of the upper Great Lakes. The climate at this time was drier than present (Figure 2). As lake level was then below its outlet, water loss by evaporation must have exceeded water supply by precipitation and runoff.
Validation of this lowstand phase and estimation of parameters of the dry climate from proxy paleoclimate evidence are works in progress. The aim of our current work is to determine the relationship of climate to lake levels during this unusual paleo-event of hydrologic closure. Estimation of the hydrology-climate sensitivity of the lake system will aid in modeling and projecting lake levels under future climates (Croley and Lewis, 2006; Lewis et al., 2007c).
Beach ridges are shore-parallel ridges of sand, about 0.5 to 3 m high, which are constructed by fair-weather wave deposition on a sandy gravel beach. Commonly, the ridges are capped with a blanket of sand deposited by wind action. Where sediment supply is ample and the slope of the pre-depositional surface is minor, a series of up to about 100 ridges has been constructed and preserved in a corrugated strandplain (Figure 4). In subsurface, the coarse sandy foreshore deposits built by wave swash at lake level are overlain by dune sand, and are underlain by sands deposited previously in the upper shoreface. Organic peat often accumulates in low areas or swales between beach ridges (Thompson, 1992; Thompson and Baedke, 1995, 1997; Baedke and Thompson, 2000).
Beach ridges are formed in areas receiving a positive rate of sediment supply. These areas commonly are embayments in the shoreline. Study of the internal architecture of the ridges indicates that they form in the final stages of long-term rises in lake levels through the vertical aggradation of the shoreline (Johnston et al., 2007; Thompson and Baedke, 1995). A berm formed during this period of aggradation becomes colonized with vegetation that traps wind-blown sand. A foredune develops that grows in size with progradation of the shoreline. Beach ridges, therefore, are records of highstands in lake level. Past lake-level data are acquired by coring the ridges, then identifying and measuring the elevations of the relatively coarse swash-zone foreshore sediments, the proxies of former lake-levels. Ages of the ridges are determined from the 14C age of basal organic matter in adjacent swales; these data show that some ridge sequences have been forming periodically for 4700 years (Baedke and Thompson, 2000). A newer dating technique, optically stimulated luminescence (OSL), determines ages of mineral grains from the foreshore sediment and is now also being applied (Argyilan et al., 2005).
Beach-ridge forming fluctuations in lake level are likely climate-driven, resulting from changes in net water supply to the lake basin (Thompson and Baedke, 1997). Analysis, using radiocarbon chronology of five strandplains around Lake Michigan shows that beach ridges delineate a 0.5- to 0.6-m rise and fall of lake level about every 30 years (32 ± 6.6 years) (Baedke and Thompson, 2000; Figure 5a). Four to six ridges form groups marking lake-level fluctuations of 0.5 to 1.5 m about every 160 years (120 to 200 years in Thompson and Baedke, 1997). Not only does beach ridge information extend the historic instrumental record of a century and a half by several millennia, but it also adds information about the occurrence of oscillations of longer periods that are not yet evident in the relatively short modern record (Figure 5b).
Sedimentary Black Bands
Black bands are common in the Holocene offshore silty clay basin sediments of the Great Lakes. Known since 1925 in Lake Ontario (Kindle, 1925), black bands in the grey silty clay mud of a freshly sliced lake sediment core disappear after exposure to air for thirty minutes to 1 hour. Hough (1958) reported results from differential thermal analysis that the bands showed no appreciable organic matter, but that they did contain iron sulfide. The jet-black color was due to a state of reduction in the sediments.
Odegaard et al. (2003) have shown from a study of cores in Lake Huron and Lake Michigan that black bands can be correlated between cores within each basin, but not always between basins. This result suggests that the deep water in each basin developed its own unique physical and geochemical conditions.
The black bands in Great Lakes sediments indicate times of suboxic to anoxic conditions at or below the lake floor. The most distinctive bands occur in sediments deposited during the warmest time in Great Lakes history. In Lake Michigan core 16PC, the black bands occur with an oxygen isotope record from benthic ostracode valves (Figure 6). Increased bottom water temperatures are indicated by the most positive δ18O values between 4000 and 7000 BP, the period during which the black bands were best developed in the core. This development could be an analog for future basin conditions under global warming, and could be useful for validating numerical models to be used for projecting future limnologic conditions.
These observations are consistent with a one-dimensional exploratory model of future warming in southern Lake Michigan under the two-times greenhouse gas scenario (McCormick, 1990). This model suggests summer stratification would be increased by up to two months, and the lake might no longer turn over fully during most winters. A reduction in large-scale mixing could lead to a loss of dissolved oxygen in bottom waters and increased anoxia as a result of global warming.
During deglaciation of the Great Lakes basins, seasonal melting in the adjacent ice sheet produced distinctive annual inflows and deposits in the lake basins that are recognized today as couplets of silt and/or clay layers, commonly of different color and grain size. The thickness of these rhythmites or varves is thought to vary with the climate-modulated annual variation in glacial melting and input of suspended sediment into the lake basins.
In a study of 264 annual Huron Basin varves, dating from about 10,000 14C (11,400 cal) BP, highest concentrations of pollen produced during flowering of summer plants were found near the base of each couplet, supporting the model of an annual period of couplet (varve) deposition (Godsey et al., 1999). An evolutionary spectrum for varve thicknesses (Figure 7) shows high spectral density of 2- to 2.5-year periodicity suggesting quasi-biennial oscillation (QBO) influence, and spectral density of 3.5- to 4.0-year periodicity suggesting teleconnection with atmospheric circulation during the Southern Oscillation (El Nino or ENSO) events. The last part of the sequence after varve-year 176 indicates a possible influence associated with both the ENSO and North Atlantic Oscillation (NAO) indices (6- to 7-year periodicity).
This glacial regime of sedimentation no longer exists in the Great Lakes basins, but other annual depositional regimes do. These regimes are influenced by geochemical or biological processes and occur in a few small lake basins with annually laminated sediment within the Great Lakes watershed. Fayetteville-Green Lake, N.Y., is a small lake in the southeastern Great Lakes region (Figure 1) that preserves postglacial varved sediments composed of carbonate laminations formed during annual whiting events, and dark organic-rich laminations composed of detritus washed into the lake from the watershed (Figure 8). The carbonate lamination thickness of each varve is directly correlated to precipitation in the region (annual: r = 0.37, p < 0.0001; decadal: r = 0.53, p < 0.01), presumably owing to geochemical interactions between groundwater and the lake (Hubeny, 2006). In addition, the carbonate-thickness time-series is correlated to both the Pacific/North American (PNA) and Southern Oscillation Index (SOI) teleconnection indices (r = 0.80, p < 0.01; r = −0.47, p < 0.01, respectively) at decadal time scales. We have produced a record for the last millennium, and work is ongoing to extend this record over the last few millennia. Spectral analyses confirm significant cyclicities in the ENSO and PNA bands (Figure 8), which may assist in water-budget planning in the Great Lakes watershed owing to the periodic nature of these teleconnection patterns.
Evidence of teleconnections with atmospheric circulation indices from these geological archives could provide perspective and depth to the recognition and understanding of oscillations in modern records of Great Lakes phenomena. One of these, for example, is the fluctuation in the severity of Great Lakes winter ice cover, which has been found by Rodionov and Assel (2000) to relate to atmospheric circulation indices, in particular to combinations of the Polar/Eurasian, ENSO, Pacific North American, Tropical Northern Hemisphere, and North Atlantic Oscillation indices.
Four examples have been described from the past history of the Great Lakes that could benefit present and future understanding of the lake system. Benefits of past lake history for understanding future lake changes include:
lengthening the existing record of variability, e.g. to reveal long-period oscillations of lake level;
contributing to modeling efforts that project conditions under future climates. The past lake history may be used possibly as an analog scenario, to validate models, or to derive sensitivities from extreme events;
illustrating the sensitivity of the Great Lakes to different conditions, valuable for informing the public about the likelihood of future change;
connecting Great Lake variability to larger atmospheric circulation periodicity.
We thank Mohi Munawar for his interest and encouragement in presenting this topic. Editorial assistance by the Indiana Geological Survey, and reviews by B.J. Todd and G.D.M. Cameron (Geological Survey of Canada), and two anonymous reviewers for the journal, helped improve the manuscript. We appreciate assistance in preparation of the figures by P. Melbourne and G. Grant, Geological Survey of Canada (Atlantic). This is contribution number 20070274 of the Earth Sciences Sector (Geological Survey of Canada) of Natural Resources Canada.