Antarctic Peninsula (AP)

Deglaciation history and the Pleistocene-Holocene transition: The pattern and mechanisms of Holocene palaeoenvironmental change in the AP region have recently been reviewed (Bentley et al., 2009). The early Holocene climate optimum detected in ice cores lasted to around 9.2 ka BP (Masson et al., 2000; Masson-Delmotte et al., 2004) occurred at the same time as the continued deglaciation of the AP continental shelf (Bentley, 1999; Ingólfsson et al., 2003; Bentley et al., 2006). Ice sheet retreat around the AP probably began c. 14-13 ka BP (Evans et al., 2005; Heroy and Anderson, 2005), and continued through the Holocene (Figure 3.26). The early Holocene optimum is however absent in the offshore Palmer Deep record, which is characterized by an apparent ‘cold’ proxy record at that time (11.5 – 9.1 ka BP) (Domack, 2002; Sjunneskog and Taylor, 2002; Taylor and Sjunneskog, 2002). There is little evidence in the terrestrial record in the Peninsula for an early Holocene climate optimum, because most currently ice-free areas were probably still ice-covered before c. 9.5 ka BP (Ingólfsson et al., 1998, 2003).

Figure 3.26 Selected Holocene environmental changes – Antarctic Peninsula

After the early Holocene: The period after the early Holocene optimum shows complex patterns in the AP region. Ice shelves on the western side had collapsed, whilst those on the east were still stable (Hodgson et al., 2006b). The onset of marine conditions in epishelf lake sediments on Alexander Island shows that George VI Ice Shelf collapsed at c. 9.6 ka BP, immediately following the early Holocene optimum (Bentley et al., 2005b). At the same time, ocean records from the Palmer Deep, a basin on the continental shelf of the AP, indicate a dramatic increase in the presence of warmer surface waters over the AP’s continental shelf (Leventer et al., 2002) so it is likely that the ice shelf was attacked both from above

165

(atmospheric temperatures) and below (warm ocean currents) (Smith J A et al., 2007). The ice shelf reformed from c. 7945 yr BP (Bentley et al., 2005b; Smith et al., 2007; Roberts et al., 2008). In contrast, evidence from the Larsen B Ice Shelf, east of the AP, shows that it was stable throughout the Holocene (from 11.5 ka BP), but has now collapsed (in 2002) due to a combination of long-term (postglacial) thinning and cracking combined with rapid recent warming (Domack et al., 2005). This suggests that there was an intensification of the climate contrast between the two sides of the AP in the early Holocene, with a steepening of the thermal gradients to the north and west (Figure 3.27) (Hodgson et al., 2006b). This is backed up by data on the historical retreat of the Peninsula ice shelves as well as by differences in the timing of deglaciation of middle- to inner-continental shelf sites between the west (~13.3 ka B.P. at Palmer Deep and 15.7 ka BP at Lafond Trough) and the east of the Antarctic Peninsula (~10.6 ka BP at Erebus and Terror Gulf, 10.7 ka B.P. at Greenpeace Trough, and 10.5 ka BP at Larsen B embayment, all using conventional ¹⁴C ages) (Domack et al., 2005;

Heroy and Anderson, 2005). According to this hypothesis, the earlier deglaciation of the northern and western side may have made the glacial system more susceptible to the advection of warmer ocean currents. This is consistent with the evidence that at least some ice shelves there retreated in periods of early and mid-Holocene atmospheric and ocean warmth, while the thicker ice shelves on the east, such as Larsen B Ice Shelf, remained buffered against these warm periods.

Figure 3.27 Recent retreats of Antarctic Peninsula ice shelves shown together with the –9ºC mean annual isotherm that marks the southern limit of ice shelf stability. Ice shelves that have retreated or collapsed are marked in red and extant ice shelves in blue. Locations of ice shelves studied are marked together with known Holocene retreat events (Hodgson et al., 2006b).

Deglaciation of the currently ice-free regions similarly showed regional differences in time and duration. Sedimentation began from c. 9.5 ka BP onwards in newly exposed lake basins in the north-eastern part of the AP and some islands to the north (Ingólfsson et al.,

166

1998; Jones et al., 2000; Ingólfsson et al., 2003). Coastal areas in Marguerite Bay and parts of the coast on King George Island in the South Shetland Islands, were ice-free immediately after the early Holocene climate optimum (c. 9.5 ka BP) and some lake basins began to accumulate sediments c. 9.5-9.0 ka BP (Mäusbacher et al., 1989; Schmidt et al., 1990; Hjört et al., 1998; Hjört et al., 2003; Bentley et al., 2005a), but other areas in the South Shetland Islands did not become free of ice until much later in the Holocene (Björck et al., 1996;

Gibson and Zale, 2006). In general, on the west side of the AP significant glacier thinning and ice margin retreat continued until at least c. 7-8 ka BP (Bentley et al., 2006). The transition from glacial to interglacial conditions was broadly completed by around c. 6 ka BP, when most ice-free areas were colonized by biota (Ingólfsson et al., 2003), but Byers Peninsula on Livingston Island deglaciated as late as c. 5-3 ka BP (Björck et al., 1996).

Evidence for mid Holocene glacier readvances are present on some islands, such as Brabant Island after c. 5.3 ka BP, and northern James Ross Island around c. 4.6 ka BP (Hjört et al., 1997; Ingólfsson et al., 1998). The glacial expansion coincided with cold and arid conditions on land from 5 ka BP onwards as detected in peat and lake sediment cores (Björck et al., 1991a,b; Björck et al. 1996) and cold marine waters with extensive sea ice cover in the bays of King George Island (Yoon et al., 2000) and in Lallemand Fjord beginning at 4.4 ka BP and peaking at 3 ka BP (Taylor et al., 2001).

In summary, there seems to be a regionally different response along the western and eastern coast of the AP, with ice shelf collapse restricted mainly to the west during the early Holocene. In addition, while most East Antarctic oases and nunataks were ice-free at the beginning of the Holocene, different parts of the AP were still ice-covered, and some did not deglaciate until as late as c. 5-3 ka BP. There is some evidence for a mid Holocene glacier readvance, coincident with cold marine water and extensive sea ice cover in the coastal areas.

Mid Holocene warm period – Hypsithermal: It was not until the mid-Holocene that the next period of significant warmth occurred in the AP. This interval is reviewed in detail in Hodgson et al. (2004a). The best-dated records place it between either c. 3.2 to 2.7 ¹⁴C ka BP (c. 3.5-3.2 to 2.9-2.7 ka BP) in the AP region (Björck et al., 1991a) or c. 3.3 to 1.2 ¹⁴C ka BP (c. 3.6-3.4 to 1.2-0.9 ka BP) just to the north of the AP (Jones et al., 2000; Hodgson and Convey, 2005). This Mid-Holocene Hypsithermal (MHH) is detected as a period of rapid sedimentation, high organic productivity, and increased species diversity in lake sediments ranging from the South Shetland Islands (Schmidt et al., 1990; Björck S. et al., 1996) and James Ross Island (Björck et al., 1996) to Signy Island in the South Orkney Islands (Jones et al., 2000; Hodgson and Convey, 2005). Sites in the northern AP show increased amounts of South American pollen in lake sediments during this period (Björck et al., 1993). It has also been associated with collapse of the Prince Gustav Channel ice shelf in the northern AP between c. 5 and 2 ka BP (Pudsey and Evans, 2001), and fluctuations of the Larsen-A Ice Shelf between c. 4 and 1.4 ka BP (Brachfeld et al., 2003), while the Larsen B Ice Shelf remained stable. This suggests that the steepening of the thermal gradients to the north and west between the two sides of the AP is common to both the early and the mid-Holocene (Figure 3.27).

In summary, whilst there is widespread agreement on the presence of some sort of warm period in the mid-Holocene, the exact timing often varies by hundreds of years, either because the timing varied spatially, or because there are insufficient numbers of dates or dating uncertainty is high, implying that, in the AP region as in EA, there is a need for further well-dated, high resolution sedimentary records.

Past 2,000 years - Neoglacial cooling, the Little Ice Age and recent climate change: The end of the MHH was marked by colder climate conditions. Numerous studies have identified Late Holocene glacier advances but most are poorly dated or even undated. Some of the

167

putative Neoglacial advances may belong to a Little Ice Age (see Ingólfsson et al., 1998 for review). There is good evidence that the Prince Gustav Channel Ice Shelf started to reform after c. 1.9 ka BP, but due to a variable and sometimes large reservoir effect (6,000 years), this date is far from certain (Pudsey and Evans, 2001). Around c. 1.4 ka BP, as the climate began to cool, the Larsen-A Ice Shelf reformed, but here as well, dating uncertainty is high (Brachfeld et al., 2003). Numerous well-dated biological proxy records in lakes and other sites show a temperature-related decline in production at about this time (Björck et al., 1991b;

Jones et al., 2000; Hodgson and Convey, 2005). Following the MHH, Midge Lake (Livingston Island) records a gradually deteriorating environment with both warm and cold pulses (Björck et al., 1991a). There was one warm event at c. 2 ka BP, and conditions were generally colder than present between c. 1.5 ka BP and 0.5 ka BP. Lake Åsa (Livingston Island) shows a distinct climate deterioration, with cold, dry conditions starting at c. 2.5 ka BP and continuing until close to the present day (Björck et al., 1993). Penguin populations declined between c. 1.3 to 0.9 ka BP and from c. 2.3 to 1.8 ka BP on Ardley Island (Sun et al., 2000; Liu et al., 2006).

Various outlet glaciers or ice shelves such as Rotch Dome, Livingston Island (Björck et al., 1996), and the Muller Ice Shelf (Domack et al., 1995) are thought to have advanced during a period roughly corresponding to the Northern Hemisphere Little Ice Age. However, the precise timing of those advances is well-constrained at only a few sites, and many of the terrestrial records of glacier advances are as yet undated. There is limited evidence of a LIA from lake proxy evidence. Liu et al. (2005) do however show a decline in penguin populations on Ardley Island, South Shetland Islands between 1790 to 1860 AD.

Instrumental measurements show the spatial pattern and magnitude of the recent rapid regional warming, and in particular the pronounced contrast between west (more warming) and east (less warming) sides of the AP. In proxy records, the warming is seen in increased sediment accumulation rates in some AP lake cores (Appleby et al., 1995), and some high-resolution marine cores (e.g. Domack et al., 2003b). Warming was further detected in a monitoring study of lakes in Signy Island where an increase in air temperature resulted in a significant increase in the amount of ice-free days and 4-fold increase in chlorophyll a content, which approximates lake productivity (Quayle et al., 2002, 2004). Few studies have focussed on this period in the proxy records.

In summary, climate conditions probably deteriorated after the Mid Holocene Hypsithermal, coincident with glacier readvance in some regions, yet these are poorly constrained in terms of dating; this is also the case for glacial readvance during the period of the Little Ice Age. Recently rapid climate warming has been observed in various regions of the AP, with glacier fronts retreating (Cook et al., 2005) and the lakes on Signy Island showing a remarkably rapid and magnified response in ecosystem functioning.