Glaciers and climatehow simple is the relationship

The cryosphere is an integral part of the climate system, thus emphasis is rightly placed on the response of ice bodies to changes in the biosphere, atmosphere and ocean climate. There is also a positive feed-back (Fig. 21.1) between the extent, thickness and spatial geometry of the large ice caps and ice sheets, especially on the atmospheric circulation and the waveforms in the upper atmospheric circulation (Felzer et al., 1994). At the simplest level, a change in the extent of snow and ice over an area of the LIS (ca. 12 x 106km2) would have an impact on the hemispheric surface reflectivity and on the net radiation budget, whereas changes in the height of ice sheets have the possibility of affecting the geometry of the upper atmospheric waves. Global circulation models (GCMs) are being used to evaluate the importance of ice sheets on the climate of the planet during glacial intervals (Felzer et al., 1994; Pollard & Ingersoll, 1980; Calov et al, 2002).

Another important feed-back is the impact of glacial meltwa-ter on the global ocean circulation (Broecker & Denton, 1989; Broecker, 1997). The outburst of large volumes of glacial melt-water from the retreating LIS into the North Atlantic (Teller, 1995b; Barber et al., 1999) is theorized to curtail the thermohaline circulation (THC) and trigger abrupt cold intervals. Less attention, however, has been given to the impact of freshwater during intervals of glacial build-up. One would be forced to assume that as the ice sheets grew across the Northern Hemisphere the amount of freshwater reaching the Arctic Ocean (presently around 1700km3yr-1) would be reduced, as would the freshwater entering Hudson Bay and the Labrador Sea. In addition, as global ice volume increased then the inflow of relatively fresh Pacific Water into the Arctic Ocean would be reduced and would cease when sea level fell to about -50 m (Elias et al., 1996). However, the global atmospheric and ocean circulation changes associated with the build-up of ice sheets are not greatly commented upon (Johnson, 1997).

Glacial geologists and geomorphologists frequently equate the retreat or advance of ice bodies with specific climatic parameters, especially summer temperature, and the notions of'glacial intervals = cold and interglacial intervals = warm' are instinctive (Knight, 1999) interpretations, or so it appears. It is worth redrawing Meier's (1965) diagram (Fig. 21.2A), however, which shows the steps in the chain of processes between changes in climate and changes in the response (advance, retreat, or stillstand) of the glacier (Porter, 1981; Knight, 1999). This diagram shows a unidirectional cascade of relationships, but there are feed-back loops between the different elements. For example, the overall energy balance is partly related to position of the equilibrium line altitude (ELA), which is a function of the net mass balance. It is clear from Fig. 21.2 that the 'regional' climate is archived by an ice mass through a series of elements. First is the 'local' climate for that particular cirque or valley glacier, where elements such as orientation, direction related to dominant winter snow storms, shading from summer insolation, etc., come into play (Williams, 1975; Porter, 1981). The critical mass balances (bn) affecting a glacier on a yearly or longer cycle are a product of the energy balance on the glacier's surface, where for the majority of the world's ice bodies (i.e. those with surface characteristics that are classified as either subpolar or polar) the critical issue in summer losses is not temperature per se but the balance of net radiation (Paterson, 1981).

The position of the glacier's frontal margin is controlled on a year-to-year basis by mass loss, the result of a particular summer's climate, and on the forward motion of the glacier, which represents the integrated flow response to events in the accumulation and ablation zones over some interval of time (Knight, 1999). The flow component has a time-constant that depends on the mass exchange of the glacier (which is largely a function of whether it is temperate, subpolar, or polar (Andrews, 1975; Knight, 1999) and the flow law, especially the temperature sensitive coefficient which varies by two orders of magnitude between ice at 0° and -20°C (Paterson, 1981). This parameter is balanced by the mass turnover of the glaciers—the accumulation at the ELA in polar glaciers is around 0.1m whereas it can reach 10 m in some temperate areas. Thus the response of a glacier's margin to a change in climate is not immediate and will lag by years, decades or even centuries. The response of the margins of polar, subpolar and temperate glaciers to a given climatic perturbation might not be synchronous because of these factors.

Meier's diagram (Fig. 21.2A) of the links between climate and glacier response is designed primarily to reflect the realities of the situation for terrestrial mountain glaciers. The situation is dramatically different when glaciers terminate in marine or lacustrine basins (Meier & Post, 1987; Meier et al., 1994) (Fig. 21.2B). There is now an important, non-climatic, control on the mass balance and resultant glacier response (Mercer, 1961; Brown et al., 1982; Mann, 1986; Alley, 1991a). Most frequently the effect on glaciers or ice streams that terminate in fresh or seawater is an additional mass loss to the glacier system by the mechanical removal of icebergs by calving—it must be admitted that the controls on calving are not well known (Thomas, 1977; Warren, 1992; Kenneally & Hughes, 1995-96). However, in the right circumstances investigations in Antarctica have shown that mass can be gained by the freezing-on of seawater to the base of the ice shelf. In East Greenland, Syvitski et al. (1996) noted that the ocean dynamics in Kangerlussuaq Fjord (68°N) lead to the importation of relatively warm, salty water at depth to balance the surface freshwater outflow from the summer ablation. There is thus a positive feed-back mechanism between the atmospheric climate and the oceanic estuarine circulation (Fig. 21.2B). If, for example, cold summers resulted in reduced meltwater this would dampen the importation of warmer waters from the deep shelf trough (ultimately being fed by Atlantic Water in the Irminger Current (Malmberg, 1985; Azetsu-Scott & Tan, 1997)), which would then reduce basal melting at the ice front.

During intervals of positive mass balance, tidewater glaciers face the problem of advancing along fjords with water depths between 100 and 1000m (Syvitski et al., 1987). As the ice advances into deep water there is a tendency for calving to increase (Brown et al., 1982) and thus there is a question as to how outlet glaciers advance along deep-water fjords. Observations in the temperate Alaskan fjords indicate that the advancing ice constructs a shoal (moraine) by transporting overrun fjord sediments to the ice front (Alley, 1991a). In subpolar and polar areas Andrews (1990) suggested an alternative mechanism based on the rate of calving being retarded by the formation of a siqussaq (a mélange of sea ice and icebergs) that extends down fjord and anchors on the sill. In all areas, however, once tidewater glacier retreat starts it can be irreversible unless there are changes in the three-dimensional geometry of the fjord (Mercer, 1961).

Terrestrial and marine studies by earth scientists focus on the sedimentary records for glacier response to climate (Fig. 21.2), and records need to be interpreted in ways that clearly identify what is known or what is assumed about depositional mechanisms (Porter, 1981). Take, for example, the issue of dating moraines from the temperate glacier forelands of Norway versus the subpolar areas of the Canadian Arctic. In the first case the moraines are built-up by sediments entrained in the lower traction zone with some additional materials supplied from rockfall. On an annual basis the margins retreat somewhat during the summer months when ablation is high and then readvanced during the winter when the forward motion is not counterbalanced by surface melting. When the summer ablation is much greater than the winter forward movement the moraines are isolated and lichens could colonize the surface boulders (Matthews, 1977). In contrast in the subpolar landscape, sediments are brought to the surface of the ice sheet, and a process of thermal protection and sediment redistribution produces ice-cored moraines (Boulton, 1972). In such an environment the linkages between moraine formation and 'climate' are even more attenuated than in the temperate glacier situation, and the established ages of the moraines are times of ice-core stabilization (Davis, 1985)—which lag the timing of the climate change by some decades or even centuries (0strem, 1965).

The discussion of glaciers and climate (Figs 21.1 & 21.2) would not be complete without some mention of the role of surging glaciers and the forcing of glacier response by non-climatic controls. Surging glaciers and glaciers that end in marine or lake waters (Fig. 21.2B) can have responses that are out-of-phase with climate changes. Mercer (1961) showed that changes in the three-dimensional geometry of fjords determine where an ice margin can be stable or can be in retreat. Hillaire-Marcel et al. (1981) also argued that large moraines associated with the LIS can be formed in association with the marine limit because of the reduction in the ice sheet's mass balance as the margin moved from a calving-dominated regime to one where summer conditions forced mass loss (Andrews, 1987b). Surging glaciers occur today in most glac-ierized areas of the world (Meier & Post, 1969; Canadian Journal of Earth Sciences, 1969) and are typified by quasi-periodic periods of rapid flow following a longer interval of slow flow and buildup. Their contorted medial moraines identify them but it is unclear whether past surges leave any distinctive signature in terms of depositional architecture or characteristic sediments. Surging glaciers respond to some self-triggering mechanism(s) and the relationship to climate and changes in mass balance is thus tenuous.

An important, but climatically ambiguous element in the reconstructing of past climates is the determination of the elevation of former ELA (sometimes erroneously referred to as the 'snowline') (0strem, 1964; Andrews & Miller, 1972; Porter, 1981; Furbish & Andrews, 1984). On past and present mountain glaciers the ELA can be defined as the upper limit of lateral moraines of a particular age. Changes in elevation of the ELA through time have been measured by a number of different techniques (Furbish & Andrews, 1984). Whereas the ELA is glacier-specific, the delimitation of the glaciation threshold (GT) or limit is a method for determining the regional climatic controls on the distribution of ice bodies (Porter, 1977; Williams, 1978b; 0strem et al., 1981). In most cases, however, changes in the elevation of an ELA or GT are expressed as a change in temperature; this approach does not attach any importance to the effect of winter accumulation on a glacier's mass balance.

There are relatively few efforts to link glaciological indexes of climate, such as changes in bn, in a regional synthesis. One fundamental question is: how far do the changes in ELA or mass balance from a single glacier correlate with glaciers in a 10, 50, 100,1000 km or more radius (Dugdale, 1972; Cogley et al., 1995)? With the interest in the recent increase of temperatures on small glaciers this has become an important issue (Dyurgerov & Dwyer, 2001; Dyurgerov & Meier, 2000). Cogley et al. (1995) presented an important synthesis (and data set, their appendix F) of existing northern high-latitude glacier mass balance data and the degree to which the 50 or so glaciers showed similar mass balance trends since ca. ad 1960. Their focus was on the long glacier-climate records from the White Glacier, Axel Heiberg, Arctic Canada (Fig. 21.3). Although the lengths of the individual data sets are varied an important finding (their fig. 4.7) is that glaciers more than 500-700 km apart show little temporal correlation in their mass balances. This suggests some spatial limit to the 'regional climate' in Fig. 21.2.

Diagrams such as Figs 21.1 & 21.2 are useful conceptual tools but in terms of glacial history these diagrams need to be adjusted for changes in the time domain. Thus the next endeavour is to consider the glacier climate system at successively longer time-scales, starting with multidecadal to century scale oscillations, such as the Little Ice Age, to millennial-scale features such as Heinrich (H-) (Heinrich, 1988), Dansgaard-Oeschger (D-O) (Johnsen et al., 1992), and Younger Dryas-like events (Peteet, 1995), and finally to orbital-scale Milankovitch events (Hays et al., 1976; Weertman, 1976; Imbrie et al., 1992; Ruddiman, 2003b).

A repetitive question through all temporal scales of glacier-climate relationships is what do we mean by synchrony in response? Are correlations between glacial events synchronous if they differ by ±10, ±50 and ±100 yr? Although this question sometimes gets involved in semantics there has to be some reason or rule for correlating a glacial event within a region, hemisphere or globally. Thus the designation of glacial events in New Zealand as 'Younger Dryas' (Denton & Hendy, 1994) raises a question because the type events in Norway and Scotland are 400-600 yr younger. This issue becomes even more important with the current interest in 'see-saw' climate systems, whether it is the 'Greenland above/Europe below' model operating at North Atlantic Oscillation (NAO) time-scales (multiyearly) (Rogers & van Loon, 1979) (Fig. 21.3) or the bipolar oscillation (Broecker, 1998; Rind et al., 2001b), which operates on longer (millennial) scales.

It is important, especially for researchers starting in the field, to realize that the human species is endowed with both a facility and need for 'pattern recognition'. The reader of the literature on the 'correlation of glacial events,' on all spatial and temporal scales, will invariably find that the paper in question does indeed

Figure 21.3 Map showing North Atlantic sites (Iceland as an insert—Siglunes = *S, Stykkisholmur (Sty) B997-328 and -330): W, White Glacier; *H, core HU87-009; *V, core V28-14; *G, Greenland Summit ice cores; P, core PS2644; solid black line = Cock-burn Moraines; black squares = convection sites; L and H, the general locations of the Iceland Low Pressure system and the Azores High Pressure cell, which define the North Atlantic Oscillation.

Figure 21.3 Map showing North Atlantic sites (Iceland as an insert—Siglunes = *S, Stykkisholmur (Sty) B997-328 and -330): W, White Glacier; *H, core HU87-009; *V, core V28-14; *G, Greenland Summit ice cores; P, core PS2644; solid black line = Cock-burn Moraines; black squares = convection sites; L and H, the general locations of the Iceland Low Pressure system and the Azores High Pressure cell, which define the North Atlantic Oscillation.

find a correlation between those records and others in the literature! Very, very rarely does a researcher go out of his or her way to argue that they have a glacial record that has no correlative! The term 'correlative' is of course the key. At what timeresolution, given the complexities of the glacier/ice sheet's response to the integrated climate system, are glacial responses correlative? There is no simple answer to this question that I am aware of, but I suspect that many correlations of glacial events are forced into an existing paradigm.

Was this article helpful?

0 0

Post a comment