New modelling and the Devensian limit

The modelled extent of the Celtic Ice Sheet,with basal thermal regime assumptions.  The red line shows the commonly assumed Devensian limit, which has now been shown to be incorrect in many areas.  The orange areas show where enhanced erosive capacity probably occurred.  Note that there was no great potential for erosion in St George's Channel and the Celtic Sea approaches.

This new paper (Patton et al, 2016) has just been published online.  It's very influential, and will be much cited!  It's a big and complex paper, and I have yet to fully digest its contents, but there are a few things that spring out on an initial reading.    First, the British and Irish Ice Sheet (BIIS) is referred to here as the Celtic Ice Sheet (CIS).  Second, the maximum Devensian ice extent is deemed to be about 22,700 years ago -- that's rather earlier than many other papers have suggested.  Third, the authors support the conclusions of other authors that the behaviour of the Devensian Irish Sea Glacier was highly erratic and dynamic.


This is what they say:


The sensitivity of this region to temperature and precipitation fluctuations (Fig. 6B–C) supports the contention that this ice sheet was highly dynamic, and led to a complex geomorphological palimpsest with multiple advance and retreat cycles (Greenwood and Clark, 2009, Clark et al., 2012 and Hughes et al., 2014). In this context, model output supports and resonates with recent evidence of expansive marine-based limits found on Porcupine Bank, west of Ireland (Peters et al., 2015), and at the Celtic Sea shelf break south of Ireland (Praeg et al., 2015). However, our model results presented here indicate that these far-field limits likely relate to short-lived surge-phases of ice stream activity, associated with enhanced orographic precipitation input to the Irish ice sheet, rather than long-term stable ice-sheet limits.

 Given that the modelling work here covers a vast area and is portrayed on small scale map illustrations, we cannot read too much precision into the ice limits for Southern Great Britain, but what interests me is that when the new evidence from Dartmoor, the Scilly Isles and the South-west is brought into the frame, the Devensian Irish Sea Glacier is shown on nearly all of the models as filling the Severn Estuary, the Bristol Channel and the Celtic Sea approaches. The ice is even shown pushing well into Somerset -- and in some scenarios as far east as Salisbury Plain.  Bear in mind that this modelling is for the Devensian -- I am quite convinced that when similar modelling is done for the Anglian Glaciation, the ice will be shown as being even more extensive.......

Glacial transport of bluestones, anyone?
 

-------------------------------------
Henry Patton, Alun Hubbard, Karin Andreassen,  Monica Winsborrow,  Arjen P. Stroeven. 2016.
The build-up, configuration, and dynamical sensitivity of the Eurasian ice-sheet complex to Late Weichselian climatic and oceanic forcing.  Quaternary Science Reviews, Volume 153, 1 December 2016, Pages 97–121

http://dx.doi.org/10.1016/j.quascirev.2016.10.009
http://www.sciencedirect.com/science/article/pii/S0277379116304498

Note:   Late Weichselian = Late Devensian

-----------------

Abstract

The Eurasian ice-sheet complex (EISC) was the third largest ice mass during the Last Glacial Maximum (LGM), after the Antarctic and North American ice sheets. Despite its global significance, a comprehensive account of its evolution from independent nucleation centres to its maximum extent is conspicuously lacking. Here, a first-order, thermomechanical model, robustly constrained by empirical evidence, is used to investigate the dynamics of the EISC throughout its build-up to its maximum configuration. The ice flow model is coupled to a reference climate and applied at 10 km spatial resolution across a domain that includes the three main spreading centres of the Celtic, Fennoscandian and Barents Sea ice sheets. The model is forced with the NGRIP palaeo-isotope curve from 37 ka BP onwards and model skill is assessed against collated flowsets, marginal moraines, exposure ages and relative sea-level history. The evolution of the EISC to its LGM configuration was complex and asynchronous; the western, maritime margins of the Fennoscandian and Celtic ice sheets responded rapidly and advanced across their continental shelves by 29 ka BP, yet the maximum aerial extent (5.48 × 10⁶ km²) and volume (7.18 × 10⁶ km³) of the ice complex was attained some 6 ka later at c. 22.7 ka BP. This maximum stand was short-lived as the North Sea and Atlantic margins were already in retreat whilst eastern margins were still advancing up until c. 20 ka BP. High rates of basal erosion are modelled beneath ice streams and outlet glaciers draining the Celtic and Fennoscandian ice sheets with extensive preservation elsewhere due to frozen subglacial conditions, including much of the Barents and Kara seas. Here, and elsewhere across the Norwegian shelf and North Sea, high pressure subglacial conditions would have promoted localised gas hydrate formation.

------------------

Highlights

Maximum ice sheet coverage was 5.5 × 106 km2 ∼22.7 ka BP.

The EISC grew to 7.2 × 106 km3, equivalent to c. 17 m of global sea level lowering.

Maximum ice extension was asynchronous - 2–5 ka later east of the main ice divide.

Subglacial erosion patterns reveal potential for widespread landscape preservation.

The optimal reconstruction reveals a moderately thick ice complex with nunataks.

----------------


Conclusions

In this study we use a higher-order, thermomechanical ice sheet model to reconstruct the build-up of the Last Glacial Maximum Eurasian ice-sheet complex (37–19 ka BP), as well as its sensitivity to a variety of key glaciological and climatological parameter configurations. Boundary condition data for each semi-independent ice centre are forced separately across three contrasting zones of the ice-sheet complex, reflecting the disparate oceanographic and climatological regimes of the northern Eurasian and High Arctic domains.

The optimal experiment presented indicates rapid growth of the ice-sheet complex, with margins first present along much of the western Eurasian shelf break by 29 ka BP. The Celtic, Fennoscandian and Barents Sea ice sheets continue to expand south and east under further temperature cooling, reaching a peak extent and volume c. 22.7 ka BP (5.48 × 106 km2 and 7.18 × 106 km3 respectively).

Required climate and oceanographic forcings differ significantly across the three ice sheet sub-domains, ranging from maritime conditions across Britain and Ireland to a polar desert regime in Siberia. Minimum temperature suppressions required to drive extensive glaciation decrease northwards, indicating an apparent insensitivity to climate cooling across Arctic regions. Considerable precipitation gradients that simulate rain-shadow effects are also required to keep ice within known empirical margins in eastern sectors. A heterogeneous sensitivity to calving losses is applied across the domain to simulate differences in sub-surface ocean temperature and the buttressing effects of perennial sea ice/ice shelves.

Maximum LGM margins are not contemporaneous, with major ice-divide migrations forcing a relatively late incursion into eastern sectors c. 20–21 ka BP compared to c. 23–25 ka BP along western margins. Although rain-shadow effects amplify this asynchrony, the most compelling driver for this behaviour is a pronounced difference in topography either side of the major nucleation centres.

Flowsets previously reported in the literature and attributed with an LGM age closely align with general predictions of ice flow direction across Fennoscandia and the southern Barents Sea. However, the relative chronologies of some flow packages appear to be closely associated with the timing of large magnitude ice-divide migrations as the ice complex migrated eastwards. Ice flow from central sectors of the BSIS and northern Fennoscandia therefore dominated only during the latter stages of the LGM.

A relatively high enhancement factor of deformation ice flow is needed to force low-aspect growth of the ice complex, particularly for the Barents Sea ice sheet. Although the resulting ice cover is generally thinner than under conventional parameter values, comparison of ice profiles with cosmogenic-exposure age transects along the margins of Svalbard and Norway, as well as central Fennoscandia, reveal the optimum experiment to be a good fit, and an improvement on previous modelled reconstructions.

Crustal deformation imposed by the EISC at 21 ka BP matches closely to the broad patterns of observed uplift across the northern Barents Sea and Fennoscandia. During the LGM, maximum depression by the ice sheet was c. 290 m east of Svalbard and in northern Sweden. More limited ice cover based over the Scottish Highlands led to a maximum depression of c. 125 m.

Subglacial properties of the model reconstructions reveal that basal erosion by the EISC was widespread at the LGM, particularly beneath the major ice streams that drained the Fennoscandian and Celtic ice sheets. Conversely, cold-based conditions dominated across the Barents Sea and upland regions of Fennoscandia. Such conditions across the hydrocarbon-rich continental shelf of the Eurasian Arctic would have created a stable environment conducive for the widespread growth of sub-marine gas hydrate as well as paraglacial permafrost.