Australia: The Land Where Time Began

A biography of the Australian continent 

Oceanic Ice Shelf Melting – the Effect of Basal Channels

A number of ice shelves in Antarctica and Greenland have been found to have basal channels, though it is not fully understood what their impact is on basal melting. In the study reported in this paper Millgate et al. used the general circulation model from the Massachusetts Institute of Technology (MIT) to investigate the effect of basal channels on the oceanic melt rate for an idealised ice shelf that resembles the floating tongue of the Petermann Glacier, Greenland. The formation of a single geostrophically balanced boundary current is prevented by the introduction of basal channels; the flow is instead diverted up the right-hand (Coriolis-favoured) side of each channel, with the return flow being in the opposite direction on the left-hand side. In agreement with previous studies, the mean basal melt rate decreases as the prescribed number of basal channels is increased. The subice flow was found to be a largely geostrophic horizontal circulation for a small number of channels that were relatively wide. The melt rate reduction is then found to be caused by a relative contribution increase of weakly melting channels crests and keels. The subice flow changes to a vertical overturning circulation for a large number of channels that are relatively narrow. The result of this change in circulation is a weaker sensitivity of melt rates to the size of the channels. The Rossby radius of deformation governs the transition between the 2 regimes. The important role basal channels play in regulating basal melting is explained by the results of this study.

0.59 mm per year was contributed to sea level rise between 1992 and 2011 from the ice sheets of Antarctica and Greenland (Sheppard et al., 2012). The acceleration and thinning of the ice streams has been found to be the largest contribution to this, which is believed to have resulted from oceanic melting of their floating Ice shelves (Sheppard e al., 2004; Holland et al., 2998a). It is implied by this that being able to predict the stability of the ice shelves and gain an understanding of the ocean-ice interactions at their base is an important step in having the ability to predict the stability of the ice sheets of Antarctica and Greenland.

The outlet glaciers of Greenland terminate in long narrow fjords as either a tidewater glacier, or a floating ice tongue which is less common. There is a surface layer of Polar Water that overlies warmer modified Atlantic Water within the fjords, and with Atlantic Water more generally being warmer in fjords that are further south (Straneo et al., 2012). Glaciers that feed fjords with warmer Atlantic Water have a tendency to terminate as tidewater glaciers without ice shelves. The retreat of the Greenland glaciers have been linked by various studies to the warming of subsurface waters (Holland et al., 2008a; Nick et al., 2009; Christoffersen et al., 2011).

In northern Greenland the Petermann Glacier is a major outlet glacier which drains about 6 % of the area of the Greenland Ice Sheet, and is1 of 4 glaciers in Greenland that are grounded deeper than 500 m below sea level (Falkner et al., 2011). Petermann Glacier terminated in an ice shelf that was 70 km long until 2010 and was confined to the Petermann Fjord (Rignot & Steffen, 2008; Johnson et al., 2011). Over the last century the ice shelf had been relatively stable in terms of ice volume and the extent of the ice shelf (Higgins, 1991; Falkner et al., 2011) that was potentially due to the modified Atlantic water layer that was cooler than the temperatures of other fjords that were at more southerly latitudes around Greenland (Straneo et al., 2012). An area of approximately 275 km² has been removed by 2 large calving events from the Petermann Ice Shelf since 2010, which reduced the ice shelf to a length of about 40 km (Falkner et al., 2011).

Around Greenland and Antarctica basal channels, channels that are carved into the base of several ice shelves, typically those that have a strong oceanic thermal driving, have been recorded. It has been found (Rignot & Steffen, 2008) that the floating tongue of the Petermann Glacier has pronounced channels that are aligned in the direction of the ice flow, while it has been revealed (Motyka et al., 2011) there was a large channel in the base of Jakobshavn Isbrae ice tongue prior to its retreat from 1998 onwards. In Antarctica channels have also been found in the base of the Pine Island Glacier (Payne et al., 2007; Mankoff et al., 2012; Vaughan et al., 2012; Dutrieux et al., 2013).

The presence of these channels has been found (Payne et al., 2007; Mankoff et al., 2012) to have had an impact on the oceanography within the subice-shelf cavity, which directed the meltwater from the inner cavity along these channels towards the ice front. It was suggested (Payne et al., 2997) that this would result in enhanced melting within the channels and channel deepening.  It has been shown (Dutrieux et al., 2013) that instead these channels have been carved by ocean melting near the grounding line, and then diminished downstream by melting at the keels between the channels. It was found (Payne et al., 2007) that the result of channelling of meltwater plumes was that enough residual heat reached the surface of the ocean at the ice front to cause the formation of small Polynyas. It was noted (Mankoff et al., 2012) that basal channels are common in ice shelves, though are prominent only in those which undergo intense basal melting.

Several mechanisms for the formation of basal channels have been proposed. It was found (Gladish et al., 2012) that undulations in the thickness of the ice at the grounding line are amplified by oceanic melting to form longitudinal channels, whilst channels failed to form that had a smooth (constant) thickness. It was suggested (Le Brocq et al., 2013) that the formation of ice shelf channels can result from subglacial water crossing the grounding lines in a channelized manner, which would entrain warmer ocean water, which then induced large melt rates that were localised, which for small basal channels are enhanced by oceanic melting. It was shown (Sergienko, 2013) that basal channels can appear spontaneously in the presence of lateral shear, even without undulations at the grounding line. At Petermann Glacier, however, Millgate et al. expect the lateral shear to be low, and it has been shown that undulations are high at the grounding line (Rignot & Steffen, 2008).

According to Millgate et al. the question of the overall importance of basal channels to the stability of ice shelves remains open. It is suggested (Rignot & Steffen, 2008) that mechanical weakness increases at the crest of channels, where the ice is thinnest. It was shown (Vaughan et al., 2012) that the settling of crests and keels towards hydrostatic equilibrium is responsible for the ice fracturing, which weakens it further.

Contrasting with this a coupled ice shelf-ocean plume model was formulated (Gladish et al., 2012) and found that basal channels actually increased the ice shelf stability by preventing the development of focussed high melt rates which melted through the ice shelf without any channels. Also, ice shelf melting decreased monotonically as the number of smaller channels increased. In the simple model (Gladish et al., 2012) with reduced ocean physics precluded an investigation of the physical mechanisms involved in this sensitivity. In this study Millgate et al. used the full 3D Massachusetts Institute of Technology general circulation model (MITgcm) to investigate further the impact basal channels have on ice-ocean interactions and thereby the stability of the ice shelves.

Conclusion

Millgate et al. used the MITgcm to assess the impact the basal channels have on the melting of ice and circulation within an ice shelf cavity. It was found by Millgate et al. that the flow of the surface layer of the ocean beneath the ice shelf was altered by the inclusion of the channels, which changed the focus and intensity of the melt. They also found that as the number of channels increases there is a decrease in the overall melt rate, which is in agreement with previous findings (Gladish, 2012).  Millgate et al. also found that this sensitivity is high for a small number of larger channels, but the sensitivity drops for a greater number of smaller channels. It was also found that a geostrophic flow circulates around the channels for larger channels. More “no flow” regions are added beneath keels and crests as channels are narrowed, though on the slopes the geostrophic flow remains unchanged. The result of this is a decrease in the mean flow of the ocean surface layer, and therefore basal melting.

Ageostrophic overturning circulation, that is much less sensitive to the width of the channel, replaces the horizontally sheared circulation which is not viable. The number of channels permitted before the circulation changes from a geostrophic circulation to an overturning circulation changes with varying viscosity changes, though at the viscosity chosen for this study this transition is governed by the deformation radius.

It was suggested (Dutrieux et al., 2013) that near the grounding line melt-enhanced features rapidly reach maximum surface expression before thinning towards the ice front. As a result of the static nature of the ice shelf Millgate et al. were not able to model this directly, it is suggested, however, by their modelled results that strong melting in the steeper section of the channel, near the grounding line, would promote rapid growth of the channel, though a widening and thinning of the channels would be promoted by melting further downstream on the channel keels, which is in agreement with a previous study (Dutrieux et al., 2013) and the channel profiles seen in the ASTER DEM image.

The mechanical stability of the ice shelf is increased by the presence of basal channels (Rignot & Steffen, 2008; Vaughan et al, 2012). The susceptibility of an ice shelf to basal melting is, however, also decreased by the addition of channels for 2 reasons. 1. The melting is distributed more evenly, moving away from predominantly beneath the boundary a current which is Coriolis-generated to over more of the ice shelf. 2. There is a decrease in the mean melt rate. A possible explanation as to why basal channels are observed in warm-water ice shelves in Greenland and Antarctica is this stabilising effect. Ice shelves with channels are more likely to persist, and a “survivor bias” then makes them more likely to be observed, if the channels stabilise ice shelves.

There are limitations to the model that was used in this study. There is a requirement for coupled models with an evolving ice shelf to test the full impact of the physical processes described in this paper. It is shown by observational data that there is a shallow sill that separates a deep basin within Petermann Fjord from Hall Basin (Johnson et al., 2011), though the bathymetry beneath Petermann Glacier is not known. The profile of the channels is also highly idealised, so Millgate et al. plan to model a realistic Petermann Ice Shelf domain with a more realistic bathymetry.  The use of a higher resolution model would allow modelling current subgrid scale processes and reduce the amount of parameterisation. Millgate et al. also say the model is limited by the lack of seasonal forcings, tides and winds which will be included in further model studies.

Sources & Further reading

1. Millgate, T., P. R. Holland, A. Jenkins and H. L. Johnson (2013). "The effect of basal channels on oceanic ice-shelf melting." Journal of Geophysical Research: Oceans 118(12): 6951-6964.