Australia: The Land Where Time Began
Early Holocene Ice-Sheet Decay, Rising Sea Level and Abrupt Climate Change
During the Holocene the current interglacial period began and was a critical part of the transition from the climatic conditions of the glacial to the interglacial. Beginning about 12,000 BP and continuing until 7,000 BP this period was marked by continuing ice sheet retreat, the ice sheets having extended throughout the polar and temperate regions during the previous glacial period. A dramatic sea level rise resulted from this meltdown, associated with sudden jumps of short duration that were associated with catastrophic collapse of ice sheets. Between 8,500 BP and 8,200 BP there were 2 events that have been linked to the final drainage of Lake Agassiz in north-central North America. In the deltas of the Mississippi and Rhine-Meuse record the water release from this lake in the form of 2 sudden rises of sea level, 0.4 m in the Mississippi Delta and about 2.1 m in the Rhine-Meuse Delta. According to Tornqvist et al. these jumps in sea level can be related to the 8.2 Kyr event, an abrupt cooling of the Northern Hemisphere, and it has been suggested that the release from Lake Agassiz of fresh water into the North Atlantic was of a large enough volume to perturb the meridional overturning circulation of the North Atlantic. It is now becoming apparent that during the Early Holocene abrupt climate changes that were associated with perturbations of the circulation of the North Atlantic requires a sustained release into the ocean of freshwater, as rising sea level, on the order of decimetres to metres, that can now be detected with confidence and linked to climate records.
The most recent part of the history of the Earth, the Early Holocene, in which there was rapid ice sheet retreat and rising sea level under conditions of an interglacial, which, as suggested by the authors1, makes it more relevant in the context of ice-sheet - sea level integrations in the future. In spite of this, and the rates of eustatic sea level rise during this time, on the order of 1 cm/year or more (Fleming et al., 1998; Stanford et al., 2011), which is well within the range of predictions that have been made for later in the 21st century (Pfeffer, Harper & O'Neel, 2008), the authors1 say it remains poorly studied. The relatively warm conditions of the Early Holocene in the high latitudes of the Northern Hemisphere, which is a critical region in relation to global climate change, were primarily driven by orbital, and not by greenhouse gas forcing, the conclusion in a recent paper quoted by the authors1 (PALSEA (PALeo SEA level working group, 2010) is that "the early to mid-Holocene may be key to understanding future sea-level change."
The onset of the Holocene has been linked to a sudden bout of rising sea level as a result of meltwater pulse 1B, that is now believed to involve millennial-scale period of high rates of sea level rise, up to about 2.5 cm/yr, (Stanford et al; Bard et al., 2010), instead of an originally postulated short-lived pulse (Fairbanks, 1989). According to the authors1 they define the end of the Early Holocene as being at 7,000 BP, (all the ages in their paper are in calendar years before AD 1950) given that after this time there has been a marked deceleration of rates of eustatic rise of sea level (Fleming et al., 1998). This also corresponds to the end of melting of the significant Laurentide Ice Sheet (LIS) which has been dated to 6.8 ± 0.3 ka BP (Carlson et al., 2008). It has been stated (Smith, Harrison, Firth & Jordan, 2011) that there are few empirically based detailed RSL (relative sea level) graphs that cover the full period of the Early Holocene sea level rise. The authors1 say there are 2 reasons for this, coral-based reconstructions that are often used for the last deglaciation are less effective for resolving the last few 10s of metres of sea level rise (after 10-9 ka BP) as a result of their limited resolution (>5 m). The other reason is that coastal peat records have been used extensively to study RSL changes in the Holocene. As a result of limited peat formation and preservation, as well as many of these records being located offshore, data density from these records is dramatically reduced before 8-6 ka BP (Engelhart, Peltier& Horton, 2011).
If the response of the solid Earth to redistributions of mass, glacial isostatic adjustment (GIA) has not been fully accounted for in studies of the interactions between ice sheets and sea level cannot be successful. As a result of the critical role of GIA corrections in the interpretations of sea level changes of the present that is obtained from instrumental records such as tide gauges, satellite altimetry and GRACE (Milne et al., 2009), it is extremely important that the GIA component of these records is removed, that ultimately must occur by the use of GIA modelling. The Early Holocene has been shown by studies (Engelhart, Peltier& Horton, 2011; Shennan & Horton, 2002) to commonly exhibit offsets between predictions of a GIA model and RSL reconstructions, therefore presenting an obstacle to progress. The authors1 have attributed this to rapid meltwater transfer from the remaining ice sheets to the global ocean during this time, which drives GIA responses that are rapid and spatially complex during the culmination of the last deglaciation.
According to Tornqvist et al. the prospect of significant advances towards the unravelling of the ocean-cryosphere-atmosphere system during the Early Holocene is highlighted by the present contribution. They stress, among others, the need to better understand the abrupt changes of climate during interglacials, the key role being for RSL records of high resolution, as well as concerted efforts to synthesise existing data on sea level. A comprehensive review of RSL records of the Early Holocene has been provided elsewhere. The focus of this study is mainly on abrupt sea level events (sea level jumps), in some cases being linked to abrupt climate change between about 9.5-6.5 ka BP, a period for which detailed RSL sea level records have become available recently. The authors1 also stress how important GIA modelling is to spur continued progress to understanding the RSL change during the Early Holocene.
Ice dammed lakes and ice sheets
It is suggested by reconstructions that are coral-based (Fairbanks, 1989) that about half, about 50-60 m, of the global sea level rise that has occurred since the Last Glacial Maximum did so during the Early Holocene. The LIS was several times the size of the Greenland Ice Sheet (GIS) at the onset of the Holocene, though it was essentially gone by the end of the Holocene (Carlson, 2008). Before 9,000 BP, the Fennoscandian Ice Sheet had completely melted, so the eustatic sea level rise was largely controlled by loss of ice from the LIS and Antarctic Ice Sheet (AIS) during the remainder of the Early Holocene, though their relative contributions is still being debated vigorously.
The retreat of the LIS is reasonably well-restrained spatially when compared with the AIS, though there is a problem with the lack of ice thickness data. The result is estimates of the contribution to sea level from LIS that are markedly different. It has been suggested by recent estimates that between 11-7 ka the LIS added about 30 m of meltwater, though the associated ice sheet reconstruction (Licciardi, 1998) differs substantially from the ICE-5G model that is widely used (Peltier, 2004), (and the new ICE-6G model is not publically available yet) that assumes an ice sheet that is much thinner for this time interval, notably to the east of Hudson Bay, which results in a LIS contribution of about 20 m. Ice loss from the AIS in the Early Holocene was substantial (Mackintosh et al., 2011; Hall, 2009). Model studies that were in part constrained by ice-thickness evidence from nunataks suggests a contribution of about 6 m (Mackintosh et al., 2011) to about 14 m (Peltier, 2004), which is about half the contribution made by the LIS.
The largest, and possibly the most intensely studied meltwater reservoir from the last deglaciation, the proglacial Lake Agassiz-Ojibway (aka Lake Agassiz), cannot be easily converted into volumes of freshwater. Though the southern shorelines of Lake Agassiz have been well mapped, it has proven more difficult to constrain the margins of the ice to the north, and the authors1 say this is not a trivial issue as the margins bordering the ice tend to be the deepest. It is typical for 1o shifts from the favoured ice margin position to be allowed by the most detailed reconstructions (Leverington, Mann & Teller, 2002) that are available, which when it was translated into values they deviated from the favoured volumes for the lake by about 45-200 %.
Abrupt climate change and sea level jumps
Throughout the last deglaciation meltwater pulses are prominent features, though attention has overwhelmingly been focused on the 10o-101-metre-scale phenomena that were associated with the collapses of major ice sheets up to the transition from the Pleistocene to the Holocene. The early Holocene 10o-101-metre-scale cousins are said by the authors1 to be of particular interest as they have been connected to relatively well-constrained sources of freshwater, in particular, Lake Agassiz.
The North Atlantic deep water formation is widely believed (Alley, 2007), resulting from the high density of cold surface waters that are relatively salty, to play a pivotal role in global oceanic circulation. Heat transport to the high northern latitudes is carried out by the resulting Atlantic meridional overturning circulation (AMOC). There is compelling evidence that the AMOC can be reduced by large fluxes of freshwater into the North Atlantic Ocean and thereby act as potential triggers for abrupt cooling events (Alley, 2007). It has been shown by model studies at varying levels of complexity (Wiersma et al., 2006; LeGrande & Schmidt, 2008) that in this context volumes of freshwater are a critical variable and this is where sophisticated sea level studies are able to play a prominent role. RSL reconstructions at high resolution, that are defined by the authors1 as those that attain decimetre-scale time resolution, are increasingly capable of quantifying the magnitudes of volumes of freshwater. The Early Holocene enables the study of climate changes that are abrupt under interglacial conditions which could make them more suitable analogues for the future than their counterparts from the glacial period, unlike the earlier stages of the last deglaciation.
According to the authors1 currently the only instance of major abrupt climate change for which the full chain of events, from a source of meltwater that has been well mapped to a climate response that has been dated exceptionally well, can be tied together is the 8.2 kyr cooling event, that is characterised in the Northern Hemisphere by cooling with an annual mean temperature decrease of 3.3 ± 1.1oC in Greenland (Kobashi, 2007). When geochronological data due to an anomalous marine 14C reservoir effect in the area of Hudson Bay (Barber et al., 1999) were revised it became conceivable that there may be a triggering mechanism that was associated with the final draining of Lake Agassiz, and a coeval jump of sea level, defined by the authors1 as "an abrupt annual to decadal-scale sea-level rise" in the Mississippi Delta (Tornqvist et al., 2004), though in view of the limited temporal resolution these inferences were tentative. The draining of Lake Agassiz, centred on 8.47 ky BP (Barber et al., 1999), and the abrupt cooling that was observed in the ice cores from Greenland (Thomas et al., 2007), was the basis for a process that occurred in 2 stages (Teller, Leverington & Mann, 2002), the final drawdown taking place in the second drainage phase, the critical mass was then provided to trigger the 8.2 ky BP climate response. This idea is supported by an increasing body of evidence, including the deep-sea records from the North Atlantic (Ellison, Chapman & Hall, 2006), that have identified 2 spikes in the freshening and cooling of the ocean surface, and terrestrial speleothem records indicating monsoonal activity changes in Asia and South America (Cheng et al., 2009). The authors1 say it is necessary to stress that the 'precursor' to the 8.2 ky BP event proper is not observed in many palaeoclimate records (Daley et al., 2011), including the Greenland ice cores. Sea level jumps have been identified in the Rhine-Meuse Delta (Hijma & Cohen, 2010), that appear to record the first and second stages of drainage, and the Mississippi Delta (Daley et al., 2011) that records only the second stage of drainage events of Lake Agassiz, the first stage being initiated 8.54-8.38 ky BP, 2σ error, and the second stage being initiated 8.31-8.18 ky BP, 2σ error.
Based on GIA model calculations that had previously been published (31), after correction, mainly for gravitational and elastic effects, the RSL record from the Rhine-Meuse Delta indicates there was a jump in sea level of 3.0 ± 1.2 m (Hijma & Cohen, 2010), 1σ error, that corresponds to a drainage volume of about 6-15 x 1014 m3, though the record from the Mississippi Delta implies a jump of 1.5 ± 0.7 m (Li et al., 2012), that has been interpreted conservatively in this paper as 1σ error, with a drainage volume of about 3-8 x 1014 m3 for the final stage. The estimate (Leverington, Mann & Teller, 2002) of the volume of Lake Agassiz during the final episode, 1.63 x 1014 m3, is exceeded by even the latter range, Tornqvist et al. suggest either the lake may have been larger than believed and/or there was a significant contribution from the LIS, such as by the rapid collapse of an ice sheet saddle (Gregoire, Payne & Valdes, 2012). According to the authors1 reconstructions of the LIS sea level contributions have inferred rates that were relatively low around this time, and also suggest a contribution from the AIS at around this time cannot be ruled out.
Fluxes, as well as volumes of freshwater, are considered of particular relevance within the context of the sensitivity of the AMOC. It has been concluded from the results of hydraulic modelling (Clark, 2004) that 1 or more peaks of freshwater fluxes of about 4-9 Sv (1 Sverdrup = 1 x 106 m3/sec, over 6 months) occurred in the final stages of drainage of Lake Agassiz. This is much higher than predicted as high-end estimates for the melting of GIS for the 21st century, 0.07 Sv being inferred from the GIS component in the sea level rise scenario proposed in (3). Catastrophic freshwater fluxes from earlier in the Holocene have been discussed elsewhere (Smith et al., 2011; Teller, Leverington & Mann, 2002), the authors1 suggesting they were probably too small to leave a detectable sea level signature.
At about 9.2 ky BP (34) a cooling event has been identified that was of lower amplitude and duration, though with a comparable spatial extent. It has been suggested by a recent study that an amount of water that is vastly smaller, 4 x 1012m3, about 1 Sv if it occurred in a year, meaning it is beyond the resolution of sea level reconstructions, could possibly have triggered this abrupt cooling (Yu et al., 2010). The triggering outburst has been hypothesised to have been the culmination of a period of enhanced freshwater discharge that was responsible for preconditioning the North Atlantic Ocean in a similar fashion to that of the 8.2 kyr BP event (Fleitmann et al., 2008).
An enigmatic meltwater pulse has been inferred at about 7.6 ka BP, that is potentially much larger, that further complicates the problem of freshwater forcing volumes and their possible climatic responses. The accumulation of high-resolution RSL data obtained from Fennoscandia (Yu et al., 2007), as well as reconstructions of LIS retreat (Carlson et al., 2008) has led to this phenomenon receiving considerable interest in recent years. The proposed eustatic sea level (Yu et al., 2007) rise of about 4.5 m cannot be detected in detailed RSL records from northwest Europe (Van de Plassche et al., 2010), that was relatively nearby. There is no known evidence of a climatic response in the North Atlantic region at this interval of time, it remains a conundrum that needs to be resolved.
|Author: M.H.Monroe Email: firstname.lastname@example.org Sources & Further reading|