East Antarctic Ice Sheet - Initiation and Instability

Australia: The Land Where Time Began

East Antarctic Ice Sheet - Initiation and Instability

Ice sheets have evolved over the past 50 million years on a continental scale in Antarctica (Kennett, 1977; Coxall et al., 2005; Kominz et al., 2008; Mudelsee, Bickert, Lear & Lohmann, 2014). Understanding of the past of the East Antarctic Ice Sheet (EAIS) behaviour and therefore the ability to evaluate its response to environmental change, which is ongoing, is limited by the shortage of ice proximal geological records (Naish et al. 2001; Naish et al., 2009; Cooper et al., 2009; Escutia et al., 2011). The EAIS terminates at the sea and is grounded within the Aurora subglacial basin, the catchment of which drains ice to the Sabrina Coast (the part of the coast of Wilkes Land, Antarctica, between Cape Waldron 115^o33’E and Cape Southard 122^o05’E) that is below sea level, which indicates it may be sensitive to perturbations of the climate (Fretwell et al., 2013; Golledge et al., 2015; DeConto & Pollard, 2016). In this paper Gulick et al. use marine geological and geophysical data from the continental shelf seawards of the Aurora subglacial basin to show that glaciers that terminate at the sea are present on the Sabrina Coast by the Early to Middle Eocene. It is implied by these findings that there had been a substantial volume of ice in the Aurora subglacial basin prior to the establishment of ice sheets on a continental scale about 34 Ma (Kennett, 1977; Coxall et al., 2005; Kominz et al., 2008; Mudelsee, Bickert, Lear & Lohmann, 2014). Subsequently there were at least times at which the ice advanced across and retreated from the continental shelf on the Sabrina Coast during the Oligocene and Miocene. It is indicated by the presence of tunnel valleys (Kehew, Piotrowski, & Jørgensen, 2012) that are associated with half of these glaciations that there were subpolar glacial systems rich in surface meltwater under climate conditions similar to those that are anticipated with continued anthropogenic warming (Golledge et al., 2015; DeConto & Pollard, 2016). Since the Late Miocene cooling (Herbert et al., 2016) has resulted in an expanded polar EAIS and the limited response of glaciers to warmth during the Pliocene in the catchment of the Aurora subglacial basin (Cook et al., 2013; Aitken et al., 2016; Rovere et al., 2014). It is indicated by geological records from the continental shelf of the Sabrina Coast that as well as ocean temperature, atmospheric temperature and meltwater derived from the surface influenced the ice mass balance under warmer climatic conditions than those prevailing at present. It is implied by the results of the study by Gulick et al. that a dynamic response by the EAIS with continuing anthropogenic warming will contribute to future sea level projections (Golledge et al., 2015; DeConto & Pollard, 2016; Aitken et al., 2016; Masson-Delmotte et al., 2013) that may be underestimated.

The response of the WAIS to anthropogenic warming and contribution to global sea level are the largest uncertainties in climate models because the formation, evolution and behaviour of the WAIS during past warm climates are not well understood (Golledge et al., 2015; DeConto & Pollard, 2016). It is indicated by oxygen isotopes, δ¹⁸O, obtained from deep sea foraminifera, that the Earth experienced the warmest conditions of the past 65 My during the Early Eocene, 53-51 Ma. About 15 My of cooling followed this warm period, with declining atmospheric CO₂, tectonic reorganisations, and the development of Antarctic ice sheets on a continental scale by the earliest Oligocene, 33.6 Ma (Kennett, 1977; Coxall et al., 2005; Kominz et al., 2008; Mudelsee, Bickert, Lear & Lohmann, 2014; Masson-Delmotte, 2013; Anagnostou et al., 2016; DeConto & Pollard, 2003). It is suggested by deep sea δ¹⁸O and far-field (far from Antarctica) sea level records that ice sheets advanced to and retreated from the continental selves of Antarctica responding to astronomically paced changes in solar insolation, as atmospheric CO₂ declined throughout the Oligocene and Miocene (Kominz et al., 2008; Mudelsee, Bickert, Lear & Lohmann, 2014; Anagnostou et al., 2016; Pӓlike et al., 2006; Liebrand et al., 2017). Also, it is suggested by these records that larger Antarctic ice sheets that had growth that was less pronounced and decay cycles that were less pronounced after the Middle Miocene, about 13.8 Ma (Kennett, 1977; Mudelsee, Bickert, Lear & Lohmann, 2014), at a time when global climate was cool and there was low CO₂ relative to the Eocene and Oligocene (Mudelsee, Bickert, Lear & Lohmann, 2014; Masson-Delmotte et al., 2013). While a good general framework for cryosphere development in the Cainozoic in Antarctica is provided by far-field records, little direct evidence is provided for the extent and location of ice or thermal conditions that are required to assess climate forcings and feedbacks involved in Antarctic cryosphere and global climate evolution; ice volume in the Northern Hemisphere in the Pliocene and Pleistocene also complicated these records (Kennett, 1977; Kominz et al., 2008; Mudelsee, Bickert, Lear & Lohmann, 2014).

The continental margin of East Antarctica and sediments of the Southern Ocean provide direct evidence of the evolution of the EAIS, which indicates that in the Late Eocene there was marine terminating ice (Scher, Bohaty, Smith & Munn, 2014; Carter, Riley, Hillenbrand & Rittner, 2017; Passchier et al., 2017) and glacial-interglacial cycles through the Pliocene (Naish et al., 2001; Naish et al., 2009; Cook et al., 2013). Existing ice-proximal records are, however, geographically limited and temporarily discontinuous, which makes regional comparisons difficult. Additional insight into the evolution of the EAIS is provided by recent ice sheet models. Catchments with deep subglacial topography that dip landwards, and surface meltwater, including the Aurora subglacial basin (ASB), are indicated by outputs to possibly be sensitive to perturbations of the climate, e.g. atmospheric and/or oceanic temperatures, atmospheric CO₂, sea level) (Fretwell et al., 2013; Golledge et al., 2015; DeConto & Pollard, 2016; Golledge et al., 2017). Outputs depend, however, on boundary conditions that are poorly constrained (Masson-Delmotte et al., 2013; Passchier et al., 2017; Golledge et al., 2017), feedbacks (Anagnostou et al., 2016), and mechanisms of retreat (DeConto & Pollard, 2016). Therefore there are uncertainties that remain in regards to the evolution of the EAIS that can be resolved by ice-proximal marine geological and geophysical data that are well dated (Kennett, 1977; DeConto & Pollard, 2003).

In order to improve predictions of response to warming and contribution to global sea level rise of the EAIS in the future, knowledge of the evolution of the EAIS in catchments that have large potential sea level contributions is critical. The Aurora subglacial basin, which is low lying and has been sculpted by glaciers, and contains ice that is the equivalent to 3-5 m of sea level rise (Fretwell et al., 2013; Aitken et al., 2016; Young et al., 2011), drains ice to the Sabrina Coast from the Gamburtsev Mountains via the Totten Glacier, which is experiencing the largest mass loss of East Antarctica (Li et al., 2016) and is influenced by warm subsurface ocean waters, down to deeper than of 400 m at its grounding line (Rintoul et al., 2106). Several basins that have been over-deepened comprise the Aurora Coast catchment (Aitken et al., 2016; Young et al., 2011), which suggests that regional outlet glaciers may be susceptible to progressive retreat (Herbert et al., 2016) and changing subglacial hydrology (Wright et al., 2012). Therefore, regional glacial dynamics and, ultimately contribution to sea level during a given warm interval depends both on catchment and glacier boundary conditions, e.g. subglacial topography, substrate, and the presence and volume of meltwater, coupled to forcings of atmosphere and ocean.

In this paper Gulick et al. present the first ice-proximal geophysical and geological records of the glacial evolution of the Arora subglacial basin. In order to document regional glacial development, ice dynamics, and the timing of major environmental transitions, Gulick et al. integrated seismic and sedimentary data from the continental shelf of the Sabrina Coast, at the Aurora subglacial basin outlet. During the rifting of Australia and Antarctica in the Late Cretaceous this margin was formed, and subsidence continued through the Palaeozoic (Escutia et al., 2011). The continental shelf of the Sabrina Coast is about 200 km wide, about 600 m deep, and has a landward slope. There are 3 distinct packages of sedimentary rocks that are bounded by basement, unconformities that are extensive regionally, and the sea floor, termed Megasequences I-III (MS-I, MS-II and MS-III) that were identified by Gulick et al. Recovery of date from near the top of MS-I and at the base of MS-III was allowed by glacial erosion of the sea floor.

MS-I, which is the deepest unit, overlies basement, and consists of a unit that is an approximately 620 m thick sequence that dips to seaward and is of discontinuous reflectors of low amplitude that increase the amplitude and lateral continuity upsection. There is no evidence within these strata of glacial erosion. Gulick et al. imaged 2 intervals of inclined stratal surfaces (clinoforms) on the middle shelf, which indicates times of high sediment flux to a continental margin that was not glaciated. Silty sands that are rich in mica were recovered 15-20 below the upper clinoforms by the jumbo piston core NPB14-02 JPC-55 (1.69 m). These marine sediments are indicated by terrestrial palynomorphs and benthic foraminifers to be of Late Palaeocene age, which confirm the pre-glacial seismic interpretation of MS-I.

Within MS-I above the upper clinoforms there is a series of reflectors that are of moderate to high amplitude and laterally variable. Piston core NBP14-02 JPC-54 that was recovered from this interval contains lonestones that were interpreted to be debris that had been ice rafted. It is indicated by terrestrial palynomorphs that these sediments date to the Early to Middle Eocene. It is indicated by reflectivity, which is laterally variable, with no chaotic facies, and with debris that had been ice rafted, with no evidence of cross-shelf glacial erosion, that glaciers that are marine terminating were present at the Sabrina Coast by the Middle Eocene, though grounded ice had not yet advanced across the shelf.

Episodes of enhanced sediment flux from the Arora subglacial basin are revealed by MS-I strata, followed by the arrival in the Early to Middle Eocene of glaciers that are marine-terminating to the Sabrina Coast. It is indicated by models and observations that the ice sheets of Antarctica nucleated in the higher elevations of the Gamburtsev Mountains, first reached the ocean near the Sabrina Coast and Prydz Bay (DeConto et al., 2003), which increased the flux of sediment to the Australian-Antarctic Gulf (Close, Stagg & O’Brien, 2007). There is a series of basins that are topographically constrained within the Arora subglacial basin that probably hosted ice volumes that were progressively larger (Aitken et al., 2016; Young et al., 2011) as ice expanded in the catchment. Gulick et al. speculate that following the Early Eocene climate optimum, 52-51 Ma, as regional and global temperatures cooled and the atmospheric concentrations of CO₂ declined (Masson-Delmotte et al., 2013; Anagnostou et al., 2016) the northern Arora subglacial basin highlands was breached by glacial ice (Young et al., 2011), which allowed glaciers that are marine-terminating to deliver ice-rafted debris to the shelf of the Sabrina Coast by the Early to Middle Eocene. This important finding indicates that the presence of a substantial East Antarctic volume of ice by the Early to Middle Eocene and the arrival of marine-terminating glaciers relatively early at the Sabrina Coast, compared with the arrivals in the Late Eocene in Prydz Bay and the Weddell Sea (Scher et al., 2014; Carter et al., 2017; Passchier et al., 2017). It is not clear if this early arrival is unique to the Sabrina Coast or if relevant data has yet to be recovered.

The first known evidence that has been preserved of grounded ice on the shelf of the Sabrina Coast is provided by the deepest regionally mappable surface that is roughly eroded about 13 m upsection from core JPC-54 which separates MS-I strata from MS-II strata. According to Gulick et al. MS-II has a thickness of up to 675 m with 10 additional erosive surfaces that truncate reflectors and have morphology that is rough and channels which indicate glacial erosion in an environment that is rich in meltwater (Cooper et al., 2009; Kehew et al., 2012; Right et al., 2012; Close, Stagg & O’Brien,, 2007). Gulick et al. say they observed strata that had parallel high-amplitude reflectivity and strata of varying thickness that are prograding which indicates conditions of open marine and intervals of high flux of sediments (Cooper et al., 2009; 7,8), respectively, between the 11 glacial advances and retreats from the Arora subglacial basin.

The Sabrina Coast MS-II reveals multiple erosive surfaces (2-6, 8 and 9) and U-shaped channels that are carved into the sediment strata, which differs from East Antarctic shelf sequences that have been imaged previously (Cooper et al., 2009; Escutia et al., 2011). It is indicated by their geometry and size that these channels are consistent with subglacial tunnel valleys that have been observed in subpolar glacial systems rich in surface meltwater (Kehew et al., 2012). According to Gulick et al. the channels that are most prominent are associated with surfaces 3-5, 8 and 9. The Overlying erosive surface that is an approximately 330 m thick sequence of strata that dip seawards that have no rough erosive surfaces, which indicates prolonged progradation of the continental shelf and a high flux of sediment in an open marine setting (Cooper et al., 2009; Escutia et al., 2011). MS-II (and in some places MS-I) strata is truncated by a regional unconformity that dips seawards. Gulick et al. recovered diatomites from the Late Miocene to earliest Pleistocene from the unconformity as well as immediately above it. Gulick et al. consider the Late Miocene, around 7- 5.5 Ma, to be the youngest possible age for the base of MS-III, because they may not have recovered sediments below the unconformity.

From the Early to Middle Eocene to the Late Miocene ice advanced across the continental shelf of the Sabrina Coast at least 11 times, at times when the average CO₂ concentrations in the atmosphere, global temperatures, and global sea levels were similar to, or higher than, those at present (Mudelsee et al., 2014; Masson-Delmotte et al., 2013; Anagnostou et al., 2016). The pacing of these glaciations is not known without additional age constraints, though it is indicated by far-field and ice-proximal records that the sensitivity of the cryosphere to insolation that is astronomically paced changed during the Oligocene-Miocene (Naish et al., 2001; Pӓlike et al., 2006; Liebrand et al., 2017). It is suggested by the scale of the tunnel valleys on the Sabrina Coast shelf, and similar channels within the catchment of the Arora subglacial basin, which is about 400 km from the grounding line of the present, that the regional subglacial hydrologic systems were fed during the Oligocene-Miocene glacial-interglacial cycles by large volumes of surface meltwater (DeConto & Pollard, 2016; Aitken et al., 2016). Therefore, meltwater that is surface derived may play an important part in the behaviour of the East Antarctic Ice Sheet (Wright et al., 2012), as is indicated by models (DeConto & Pollard, 2016). At the top of MIS-II the prograding sequence is similar to Middle to Late Miocene sequences in Wilkes Land and Prydz Bay, which reflects the transition from subpolar to polar glacial regimes (Cooper et al., 2009; Escutia et al., 2011).

MS-III consists of a veneer that is 0-110 m thick, above the regional unconformity, of sub-horizontal to landward-dipping strata that thicken towards the land, which indicate there was substantial glacial erosion of MS-II or lower regional sediment flux and the onset of ice loading by the Late Miocene (Escutia et al., 2011). There are no visible channels in the strata of MS-III, which suggests there was reduced regional meltwater influence and basal meltwater flux that was more diffuse (Kehew, Piotrowski & Jørgensen, 2012; Aitken et al., 2016; Young et al., 2011; Wright et al., 2012). Within MS-III strata, that are acoustically chaotic, are there high-amplitude reflectors indicating the presence of erosional surfaces in the tills of Miocene to Pleistocene age and an advance or retreat of an expanded EAIS (Aitken et al., 2016). There are open marine sediments but it is suggested by the lack of preservation the limited regional retreat of ice or interglacials that were shorter since the Late Miocene.

The Arora subglacial basin catchment was occupied by an expanded WAIS and occupied the Sabrina Coast continental shelf since the Late Miocene (Aitken et al., 2016), which was coincident with global climate, reorganisation of the carbon cycle and hydrologic cycle (Mudelsee et al., 2014; Herbert et al., 2016), expansion and reorganisation of the ice sheet on a continent-wide scale (Cooper et al., 2009; Cooper et al., 2009; Escutia et al., 2011; Herbert et al., 2016), intensification of the Antarctic Circumpolar Current, cooling of the Southern Ocean, and the development of the modern meridional thermal gradient (Kennett et al., 1977; Herbert et al., 2016). The amount of surface ablation was probably limited by atmospheric cooling, which resulted in the expansion of ice and reduced meltwater derived from the surface in the Arora subglacial catchment. A maximum grounding line retreat (Aitken et al., 2016) of about 150 km from its location at the present since the Late Miocene is suggested by thickness of MS-III and erosion patterns in the catchment of the Arora subglacial basin, though intermittently there were open ocean conditions on the shelf. Therefore, the Arora subglacial basin did not substantially contribute to the rise in sea level during the warmth in the Pliocene (Aitken et al., 2016; Rovere et al., 2014), which contrasts with the adjacent Wilkes subglacial basin (Cook et al., 2013).

The importance of atmospheric temperatures and meltwaters that are derived from the surface to the ice mass balance of Antarctica is revealed by the records from the Sabrina Coast. The ice-proximal record from the Sabrina Coast shelf confirms model predictions of the sensitivity of the region to climate in the long term (Golledge et al., 2015; DeConto & Pollard, 2016; Aitken et al., 2016; Young et al., 2011), though deeper, more continuous samplings of the sediments is necessary to assess the timing, magnitude and rates of evolution of the EAIS in the Arora subglacial basin. The potential for glaciers of the Arora Basin catchment to revert from being the extensive polar system of the past 7 Myr to the subpolar system of the Oligocene-Miocene, which was a time when there were average global temperatures and atmospheric CO₂ concentrations that were similar to those that are anticipated under current warming projections (Golledge et al., 2015; DeConto & Pollard, 2016; Masson-Delmotte et al., 2013) is critical for the global sea level rise scenarios of the future. The Totten Glacier is thinning more rapidly than any other outlet glacier in East Antarctica (DeConto & Pollard, 2016; Masson-Delmotte et al., 2013; Rintoul et al., 2016) due to thermal forcing of the ocean (Rintoul et al., 2016). It is suggested by the findings of Gulick et al. that ice in the Arora Basin Catchment has the potential to respond dramatically to anthropogenic climate forcing if surface meltwater production results from regional atmospheric warming.

Sources & Further reading