Holocene thinning of Darwin and Hatherton glaciers, Antarctica, and implications for grounding-line retreat in the Ross Sea

Abstract

Abstract. Chronologies of glacier deposits in the Transantarctic Mountains provide important constraints on grounding-line retreat during the last deglaciation in the Ross Sea. However, between Beardmore Glacier and Ross Island – a distance of some 600 km – the existing chronologies are generally sparse and far from the modern grounding line, leaving the past dynamics of this vast region largely unconstrained. We present exposure ages of glacial deposits at three locations alongside the Darwin–Hatherton Glacier System – including within 10 km of the modern grounding line – that record several hundred meters of Late Pleistocene to Early Holocene thickening relative to present. As the ice sheet grounding line in the Ross Sea retreated, Hatherton Glacier thinned steadily from about 9 until about 3 ka. Our data are equivocal about the maximum thickness and Mid-Holocene to Early Holocene history at the mouth of Darwin Glacier, allowing for two conflicting deglaciation scenarios: (1) ∼500 m of thinning from 9 to 3 ka, similar to Hatherton Glacier, or (2) ∼950 m of thinning, with a rapid pulse of ∼600 m thinning at around 5 ka. We test these two scenarios using a 1.5-dimensional flowband model, forced by ice thickness changes at the mouth of Darwin Glacier and evaluated by fit to the chronology of deposits at Hatherton Glacier. The constraints from Hatherton Glacier are consistent with the interpretation that the mouth of Darwin Glacier thinned steadily by ∼500 m from 9 to 3 ka. Rapid pulses of thinning at the mouth of Darwin Glacier are ruled out by the data at Hatherton Glacier. This contrasts with some of the available records from the mouths of other outlet glaciers in the Transantarctic Mountains, many of which thinned by hundreds of meters over roughly a 1000-year period in the Early Holocene. The deglaciation histories of Darwin and Hatherton glaciers are best matched by a steady decrease in catchment area through the Holocene, suggesting that Byrd and/or Mulock glaciers may have captured roughly half of the catchment area of Darwin and Hatherton glaciers during the last deglaciation. An ensemble of three-dimensional ice sheet model simulations suggest that Darwin and Hatherton glaciers are strongly buttressed by convergent flow with ice from neighboring Byrd and Mulock glaciers, and by lateral drag past Minna Bluff, which could have led to a pattern of retreat distinct from other glaciers throughout the Transantarctic Mountains.

Stability of the Antarctic Ice Sheet during the pre-industrial Holocene

“…Despite the overall picture of ice sheet stability during the Late Holocene, empirical evidence and ice sheet modelling indicate that the ice volume was still changing in some regions of Antarctica. In the east of the Ross Sea sector, >400 m of ice sheet thinning is recorded between 3,200 and 1,900 years ago 53 , and in the west of the sector, 190 m of thinning is recorded between 6,700 and 300 years ago as well as 25 m of thinning in the past 200 years 54,133 . In East Antarctica, 70 m of thinning and episodes of increased glacier melting are recorded from 4,000 years ago in eastern Dronning Maud Land 27,37 , ice unloading accelerated from 2,700 years ago at Larsemann Hills 48 and glacial discharge increased from 4,500 years ago in Prydz Bay and from 1,700 years ago in Adélie Land-George V Land 134,135 .…”

Inland thinning of Byrd Glacier, Antarctica, during Ross Ice Shelf formation

“…Our data-model comparison demonstrates that inland thinning extends 100s km from perturbations at the grounding line and ice shelf front and is linked to temporal changes in the basal sliding regime. F I G U R E 8 Surface exposure age data and generalised interpretation trends from this study and three other glaciers along the TAM (Hillebrand et al, 2021;King et al, 2020;Spector et al, 2017) showing local variation but broad agreement in constraining an early Holocene pulse of abrupt thinning. Data from Reedy (Todd et al, 2010) and Shackleton (Spector et al, 2017) glaciers included without generalised trends due to limited data.…”

“…Samples labelled with grey text indicate samples interpreted to contain inherited cosmogenic nuclide inventory due to previous exposure, (d) satellite image (Bindschadler et al, 2008) showing location of a, b, and c. Dashed grey line in B is location of Lonewolf Escarpment. [Color figure can be viewed at wileyonlinelibrary.com] and Pennsylvania State University 3-D Ice Sheet Model (Pollard & Deconto, 2012), we extract the modelled ice sheet thinning and grounding line retreat histories from two ensembles of the last deglaciation for the Lonewolf Nunataks (Hillebrand et al, 2021;Lowry et al, 2019Lowry et al, , 2020. These ensembles are chosen because they cover the necessary geographic domain to investigate the relationship between regional grounding line dynamics and ice thinning at Byrd Model ensemble A (Lowry et al, 2019(Lowry et al, , 2020) is run at 10-km horizontal resolution for the Ross Sea catchment only.…”

Cosmogenic-nuclide data from Antarctic nunataks can constrain past ice sheet instabilities

“…Many of these sites are adjacent to glaciers in the Transantarctic Mountains that are, in reality, major conduits of ice from the East Antarctic Ice Sheet into the Ross Sea but are not large enough to be resolved in the 40 km model. These include data from Taylor Glacier as mentioned above (Balco et al, 2014), Reedy Glacier (Todd et al, 2010), and Hatherton Glacier (Hillebrand et al, 2021). These sites could be used for model-data comparison if the model resolution was increased sufficiently to correctly resolve ice flow in these regions, perhaps by embedding a nested model domain in the low-resolution 5 Ma model runs.…”