LIAlike cycles in the Holocene

As the search for forcing mechanisms to explain glacial history on Holocene time-scales intensifies there is a need to ask: are there

Warmer or fresher

less sea-ice

1200

1000

Warmer or fresher

less sea-ice

1900 Year AD

1850

1900 Year AD

Stykkisholmur

1970 1980

Year AD

1960

1990

2000

1970 1980

Year AD

Figure 21.7 Data from Iceland on Little Ice Age (LIA) time-scales. (A) Mann et al. (1999), (B) sea ice and S18O in B997-328, (C) Stykkisholmur winter and summer temperature trends, (D) Siglunes water column temperature and Stykkisholmur mean annual temperature (MAT).

periodic LIA-like events within the Holocene (Bond et al., 1999, 2001), and are there recent analogues for the LIA within the instrumental record? Lamb (1979), for example, suggested that the GSA of the late 1960s off Iceland and elsewhere in the North Atlantic (Dickson et al., 1988; Belkin et al., 1998) was an analogue for the climate of the LIA. This suggestion clearly advocates changes in the ocean, and specifically in sea-surface temperatures, as a potent force in regional climate changes and associated glacier response. Changes in ocean salinity (Fig. 21.4A) in key areas north of Iceland have the potential to disrupt the global THC (Jonsson, 1992; Clark et al., 2002b) (Fig. 21.3; such changes are frequently called on to explain abrupt climate change on the H- and D-O (= longer) time-scales.

However, a persistent problem with the analogue approach to retrodicting past climatic conditions (e.g. Lamb, 1979) is the issue of'persistence'. How does the climate system switch from decadal-scale oscillations, such as the NAO in either positive or negative mode, to series that have a century- or millennial-scale persistence?

Once we move in time over 1000 yr then the glacier-climate interactions become more complex to interpret because we enter a period when boundary conditions change, such as rates of

1200

1000

1850

2000

1950

1960

1990

2000

glacial isostatic recovery, relative sea level or solar insolation, and the picture (Fig. 21.6b) becomes increasingly removed from the application of modern analogue conditions (Lamb, 1979). Summer insolation was higher than present at northern latitudes in the early to mid-Holocene (Berger & Loutre, 1991; Kutzbach et al., 1996) but these conditions were opposed by the melting of the LIS and the export of freshwater to the North Atlantic (Giraudeau et al., 2000; Teller et al., 2002). For example, between 7 and 12cal.kyr the retreating LIS (Dyke & Prest, 1987a; Dyke et al., 2002) was sending copious volumes of meltwater into the North Atlantic via Gulf of Mexico, Hudson Strait or the Arctic Ocean (Teller, 1995a; Barber et al., 1999; Fisher et al., 2002; Teller et al., 2002; Aharon, 2003; Nesje et al., 2004). In addition, the depths of the Canadian Arctic channels, leading from the Arctic Ocean into Baffin Bay, were 100 m or more deeper due to glacial isostatic depression, hence the 'freshwater' outflow from the Arctic Ocean to the North Atlantic would be affected by this time-dependant relaxation (Williams et al., 1995; Dyke & Peltier, 2000) (Fig. 21.4A). During the early Holocene in particular, as ice sheets retreated, large tidewater glaciers (Fig. 21.6b) lay within most fjords in the Northern Hemisphere and their response to changes in mass balance (see earlier) may not always be equated with climate.

The 'natural' climate of the early and mid-Holocene is now being questioned (Ruddiman, 2003a); comparisons between atmospheric greenhouse gas interglacial records contained in ice cores and the Holocene suggest that both CO2 and methane records reflect human occupation. Thus the natural decrease in solar insolation (Fig. 21.8A) and the onset of neoglaciation might have been mitigated by the rise in greenhouse gases 5-8cal.kyr (Ruddiman, 2003a). In effect, Ruddiman (2003a) argued that the severity of neoglaciation was reduced by anthropogenic changes to the climate system.

The 'climate' part of the glacier relationship (Fig. 21.2) for the past 12,000 yr is based on a variety of proxy data from ice, lake and marine cores. The 'glacier' response is often archived in a series of terminal moraines at a variety of distances from present-day glacier margins, changes in the past ELAs of glaciers (Porter, 1981), or variations in ice-rafted debris (IRD). Thus changes in sediment type, such as grain-size, in both lake (Leonard, 1986; Nesje et al., 1991) and marine settings (Andrews et al., 1997; Bond et al., 1999; Andrews, 2000) are also used in this interval to deduce 'glacial' activity. In reality there is little specific 'climate' data that has been developed that applies to variations in glacier mass balances during the past 12,000 yr. The most direct estimates are those derived from ice cores where changes in accumulation (Meese et al., 1994), stable isotopes of the precipitation (assumed to reflect largely air temperature) and borehole temperatures (Cuffey & Clow, 1997; Dahl-Jensen et al., 1998) (which reflect a long-term integrated measure of air temperatures at ca.3km over central Greenland) have been derived.

Dating of glacial moraines in front of existing valley and cirque glaciers in the Northern Hemisphere led to the concept of 'neoglaciation' (Porter & Denton, 1967; Davis, 1985; Karlen, 1988; Karlen & Denton, 1976), an interval of renewed glacial expansion, starting 5-6cal.kyr after a 'thermal optimum' (Kaufman et al., 2004) (Fig. 21.8B). This was most frequently associated with the orbitally driven reduction in summer insolation at high northern latitudes (Berger & Loutre, 1991) (Fig. 21.8A). During this same interval there was a noticeable increase in IRD on the East Greenland margin, suggesting increased calving from Greenland tidewater outlet glaciers (Andrews et al., 1997; Jennings et al., 2002b). Inferences on the climate of the neoglacial interval over Iceland over the past 5-6cal.kyr (Gudmundsson, 1997; Stotter et al., 1999) are drawn from intervals of moraine formation (Fig. 21.8C), although these may underestimate the real number of glacier advances because of overriding of older moraine systems (Schomacker et al., 2003).

At present, the most direct proxy climate record that exists for north Iceland is the S18O record from benthic foraminifera from cores B997-328 and -330 (Fig. 21.3) (Castaneda et al., 2004; Smith et al., 2005). An error of ±0.5°C has been inferred based on a Monte Carlo simulation on coefficient errors in the calibration equation (Shackleton, 1974). The seafloor temperature data (Fig. 21.8B) show a peak in warmth ca. 7cal.kyr, and there are marked oscillations in temperature, especially over the past 6cal.kyr. Previously it was shown that the record from B997-330 has a significant correlation with changes in AC14 (Andrews et al., 2003), suggesting a link with changes in the strength of the THC (Clark et al., 2002b). A 1400-1500-yr climate oscillation has been detected in several Holocene records and also in the GISP2 ice-core data (Stuiver et al., 1991; Bianchi & McCave, 1999; Bond et al., 2001) (but see Wunsch, 2000). The cause of the 1500-yr oscillation is not known with certainty (Schulz et al., 1999). It has been attributed to stochastic resonance (Alley et al., 2001); attention has also been drawn to a 1470-yr lunar cycle (Berger et al., 2002); and it has also been attributed to aliasing (Wunsch, 2000). A 1470-yr filter was fitted to the B997-330 seafloor temperature data (Andrews & Giraudeau, 2003; Smith et al., 2005) (Fig. 21.8B). This suggests 'LIA-like' events off north Iceland occurring ca.500, 2000, 3500 and 5000 cal. yr BP. However, there is little evidence in the temperature series for older LIA-like events; there is some evidence to signal the 8.2cal.kyr cold event (Alley et al., 1997b; Barber et al., 1999; Andrews & Giraudeau, 2003).

Changes in the flux of IRD (either contributed from sea ice and/or icebergs) into the northern North Atlantic (Bond et al., 1999, 2001) might imply climate-glacier feed-backs sketched on Fig. 21.4A &B. However, Wastl et al. (2001) recognized six neoglacial advances in north Iceland (Fig. 21.8C), but these do not fit neatly into a ca.1500-yr oscillation, nor is there a clear temporal correlation between moraine formation and north Iceland seafloor temperature, nor the variations in haematite-stained quartz in core V28-14 just to the west in Denmark Strait (Figs 21.3 & 21.8C) (Bond et al., 1999). In-point-of-fact, neither the north Iceland temperature record nor the Denmark Strait data (Fig. 21.8C) bear a close temporal correspondence to the north Iceland moraine intervals, which appear to be grouped between inferred 'cold' events.

The 'glacial' origin of distal Holocene IRD sediments in the North Atlantic is difficult to prove versus an alternative hypothesis that they represent sand-size particles carried on or within sea-ice. There is some correlation between the distal IRD events and the IRD record closer to the East Greenland coast (Jennings et al., 2002b), but it is far from clear why the delivery and transport of Icelandic volcanic shards to sites to the south and east of Iceland is not primarily a wind-driven process that is independent of sea

Figure 21.8 (A) Solar insolation and IRD history of East Greenland shelf, (B) 330 temp estimates and a 1470 filter response (note difference in scales), (C) peaks in haematite grains in core V28-14 from Denmark Strait (Bond et al., 1999) (Fig. 21.3) and intervals of moraine formation around north Iceland glaciers (Wastl et al., 2001). The rectangular grey areas represent groups of moraines between inferred cold events.

Figure 21.8 (A) Solar insolation and IRD history of East Greenland shelf, (B) 330 temp estimates and a 1470 filter response (note difference in scales), (C) peaks in haematite grains in core V28-14 from Denmark Strait (Bond et al., 1999) (Fig. 21.3) and intervals of moraine formation around north Iceland glaciers (Wastl et al., 2001). The rectangular grey areas represent groups of moraines between inferred cold events.

ice as a transport agent (Lacasse, 2001). There is increasing evidence for the presence of Icelandic tephras in lake and peat sites on Ireland, Britain, Norway and Sweden (Bjorck & Wastegard, 1999; Lacasse, 2001; Hall & Pilcher, 2002; Wastegard, 2002) as well as in the Greenland ice core (Zielinski et al., 1994; Gronvold et al., 1995), and sediments on the east Greenland shelf (Jennings et al., 2002a). Studies are required of the continuous flux of Icelandic glass shards at several sites around Iceland so that the variability in tephra production can be factored out of the possible IRD contribution. Such a study is not yet available.

21.3.3.1 Meltwater pulses and deglaciation

The evidence for late marine isotope stage (MIS) 2 and 1 'global meltwater pulses', as deduced from the relative sea-level records from sites well-distant from the ice sheets (Fairbanks, 1989; Blan-chon & Shaw, 1995; Aharon, 2004), does not coincide with the evidence for 'meltwater events' deduced from the planktonic iso-topic evidence of foraminifera ocean-core records for the north ern North Atlantic (Andrews et al., 1994; Jones & Keigwin, 1988; Hald & Aspeli, 1997). The former show two massive meltwater pulses that are associated with rapid rises in global sea level. Surprisingly, given the wealth of data that exists for the interval of deglaciation (19-10cal.kyr), no-one has been able to identify the ice sheet(s) associated with MWP1a and Antarctica has been called on to supply the required sea-level rise (Clark et al., 1996a, 2002a; Weaver et al., 2003), although that scenario is not without problems (Licht, 2004). The global meltwater pulses presumably represent increased melting on the ice sheets associated with the rise in summer insolation (Fig. 21.9), although the pulse-like nature of the process indicates other processes must have dominated at times, such as calving into pro-glacial lakes (Andrews, 1973). The planktonic isotope data from the northern North Atlantic indicates that surface melting on the LIS, Greenland and Fennoscandian ice sheets preceded MWP1a and may reflect the collapse and rapid retreat of ice across glacierized continental margins (Jennings et al., 2002a), probably associated with rapid calving of tidewater margins.

#11 interstitial events

480 470 460 450 440 430 420 410

Was this article helpful?

0 0

Post a comment