Thursday 5 July 2018, 11:30am–11:30am

This is a presentation by Dr Rob Massom (Australian Antarctic Division and ACE-CRC) on the causes of ice shelf disintegration. For well over 30 years, Rob has studied the sea ice and snow physics in both hemispheres. Rob is the co-leader of the ACE-CRC/AAD Sea Ice Processes and Change group. In this seminar, he reports on work he did in collaboration with Ted Scambos (NASA), Phil Reid (BOM), Vernon Squire (University of Otago) and Sharon Stammerjohn (University of Colorado).


Understanding the causes of catastrophic ice-shelf disintegrations on the Antarctic Peninsula since 1995 is crucial to improving coupled models of the Antarctic Ice Sheet, and predicting its future state and contribution to sea-level rise. A climate-related causal factor overlooked to date is regional sea ice loss. For the disintegrations events observed, i.e. Larsen A (1995), Larsen B (2002) and Wilkins (2008 and 2009), a change to greater seasonal absence of a protective sea ice ‘buffer’ enabled increased flexure of vulnerable outer ice-shelf margins by ocean swells that likely weakened them to the point of calving. This outer-margin calving triggered wider-scale disintegration of ice shelves weakened by multiple factors in preceding years; key prerequisites were extensive flooding and outer-margin fracturing. Wave-induced flexure is particularly effective in outer ice-shelf regions thinned by bottom crevassing. Based on combined analysis of satellite and ocean-wave data with ice shelf-sea ice-wave model outputs, this study highlights the need for ice sheet models to account for sea ice (change). The results underline the importance of both understanding the cross-cryosphere linkages and improving predictions of future sea ice coverage and conditions around Antarctica (underpinned by ocean and atmospheric change and variability).

The seminar will be held in the theatrette of the AAD (ground floor). All welcome!

AAD seminar team

Sea ice, ocean swell, and ice shelf disintegration

Video transcript

Barbara: Welcome everybody! Thank you so much for showing up for yet another exciting seminar. That reminds me, we won't have a seminar next week, and probably not the week after. So just watch out for emails; you'll get the invitation for probably, at the moment, for the 16 August.

Today's speaker is [Dr] Rob Massom, another well-known, happy face in the holy halls of science. Rob has been studying sea ice for no less than 35 years, which makes him an expert. He is currently the co-leader of the Sea Ice Process and Change program at the ACE-CRC. Way back in 1989, Rob did his PhD which was on data that he collected from drift buoys that were drifting around in the Weddell Sea, and Rob analysed the annual sea ice cycle based on these data. This was followed by a postdoc at the NASA Goddard Space Flight Centre, and in 1992 he finally made it to Tassie.

One of his many interests is the interaction between the Antarctic ice sheet and sea ice and that is what he will be talking about today. Thank you very much!

Rob: Thank you very much, Barb, for all those kind words. Thank you every one for coming to listen to me. Today I am going to talk about Surf's Up! It's about what happens - or what we think happens - to ice shelves when sea ice diminishes, as sea ice buffer around the ice shelves disappears I just like to start off by acknowledging my wonderful co-authors of this particular study; over there is Ted Scambos, who is an ice shelf and an ice sheet expert from the National Snow and Ice Data Centre from the University of Colorado. Next door to him is Luke Bennetts, who is a mathematician from the University of Adelaide who is an expert on sea ice and ice sheet - ocean interactions and modeling. And this character here who looks more like Amundson than a modern scientist is Professor Vernon Squire, who is also an expert on wave - ice interaction from the University of Otago. Close to home is Phil Reid from the Bureau of Meteorology, and also Sharon Stammerjohn from the University of Colorado. Both Phil and Sharon are experts on sea ice change and variability particularly change in duration which is a key factor when we are talking about ice shelves.

First of all, this is the structure of the talk; I'll say a little bit about why ice shelves are important in the scheme of things. Then we'll focus on five catastrophic disintegration events that occurred since 1995 involving the Wilkins, the Larsen A and the Larsen B ice shelves all of which are on the Antarctic Peninsula and we identified common factors that are involved in all of those, including disappearance of sea ice.

In a way this study is sort of breaking new ground - and I put there old ice as well - but it's breaking new ground in the sense that it's introducing sea ice change as a trigger mechanism that is involved in the disintegrations and that involves both pack ice, moving pack ice, and also stationary fast ice which is locked onto the periphery of these ice shelves. This has resulted in, culminated in a recently published article in Nature, about two weeks ago. I have to say that this is the culmination of 15 years of work. So thanks to my co-authors for helping me get this across the line.

Finally we'll wind up with a slide about what the future holds in terms of ice shelves and sea ice.

How, where and why Antarctica is losing mass is a very high priority and it's a major concern. Also two weeks ago, a there was a study in Nature by Shepherd et al. which had combined satellite measurements over the past 20 to 30 years, and they've shown - as in this plot here - that there has been a great acceleration in the loss of grounded ice from Antarctica to contribute to sea level rise into the ocean. They've shown that almost 2,700 billion tonnes of ice were lost since 1992 and that's the equivalent of about 8 mm of sea level rise. But more importantly, if you like, this loss is accelerating and 50% of the loss has occurred in the past 5 years. So this is looking quite serious.

The ice sheet contribution - I'm not an ice sheet scientist, I am very much a sea ice scientist - but the IPCC and all my colleagues tell me that the ice sheet contribution is the greatest uncertainty in predictions of future sea level rise. This is a quote from the IPCC down here. They say in the last report, the assessment report in 2013, that "due to a low level of scientific understanding there is low confidence in attributing the causes of the observed mass loss from the Antarctic Ice Sheet over the past two decades". Clearly we need to better understand what is causing this loss, what is driving it.

So what is at stake? The Antarctic Ice Sheet - as I am sure you know - holds roughly 90 % of all the ice on Earth and this is equivalent to - all melted in one go which isn't likely - but it's the equivalent of about 57metres of sea level rise. Importantly, a vast number of people live within 1 metre of mean high water level around the globe. If sea level rises again, you've got increased risk of storm surges affecting coastal margins.

So Ice shelf "health" - that's a hard one to say - is a key factor in all of this. Ice shelves and floating outlet glaciers fringe approximately three quarters of the grounded ice boundary of the whole of Antarctica. They play a key role; the role is to buttress or constrain the dynamic contribution, the flow of grounded ice down from the continent into the ocean past what's called the grounding line, which is this area over there. This is a very critical role in the sense that they constrain the amount of ice that comes down off the ice sheet and contributes to sea level rise. They do this by providing back pressure against outflow glaciers that are cascading down off the continent.

So if you remove this buttressing capacity, it's one of the ways and means of increasing the contribution of the ice sheet to sea level rise and a great deal of work has been done in the past decade or so looking at using satellite imagery in particular or satellite data to look at increased ice shelf thinning around the continent. You can see the red colours here show effectively where ice shelves have thinned due to incursions of warmer ocean waters on their bases. This map is from a study by Paolo et al. in 2015. Also there is a risk that the grounding line will retreat inwards and this could accelerate the amount of ice that comes off the continent. A lot of this work has been done in West Antarctica in the Pine Island glacier area. I just draw your attention to this position analysis from the ACE-CRC. It's well worth having a look at; it has recently been published by Gwyther et al. and it's "The Antarctic Ice Sheet and Sea Level"; it's a publication that pulls together all of that information.

It's critical, there is a critical need to better understand and indeed quantify the amount of melt, the patterns of melt, and the change that is occurring on the surface around Antarctica, but also how warm waters get underneath and interact with ice shelves to increase basal melt there.

A second way that more ice can come off the continent into the ocean is by disintegration of ice shelves, large scale disintegration and that's the focus of the talk here, indeed the study that was published. I show a sequence of images here that are showing the Larsen B; this is in late January 2002 just before it broke up and I draw your attention to melt ponds on the surface there, and also the fact that there is no sea ice in the ocean in the northwestern Weddell Sea just offshore from the ice shelf. Over just a very short period of time, in fact two weeks, the whole thing disintegrated. You would have seen these images; it was quite stunning the way this happened so rapidly. Over 3000 square kilometres were lost in just over 2 weeks and this was a feature that had been in place - based upon sedimentary measurements underneath the ice shelf once it disappeared - been in place for 11,500 years. So something that had been in place for a long period of time disintegrated literally overnight.

When ice shelves disintegrate, they don't directly contribute to sea level rise. The ice shelf is already floating. But in the case of this particular disintegration, the glacier flowing into it accelerated dramatically. There was an 8-fold acceleration of glaciers in this region here in the year following the disintegration of the ice shelf. This is based upon three studies, one of which is shown here, by Rignot et al. 2004, based again on satellite data analysis. That outlet glacier was also lowered by 38 metres over that time period.

Here we are going to concentrate on three ice shelves and five disintegrations events; it's the Wilkins ice shelf, which is here over on the south western part of the peninsula, and the Larsen A and Larsen B which are on the north eastern part of the peninsula. In each case, hundreds to thousands of square kilometres of ice disintegrated very, very rapidly in just a few days or even weeks. This was extraordinary, this abruptness and rapidity and the scale of the break-up was quite extraordinary, and marked a major departure from the typical calving scenario that occurs around Antarctica on decadal time scales, the kind of big ice bergs that break off naturally. Once these ice shelves broke up, there was again based on the Shepherd study, there was a five-fold increase in grounded ice loss from the Antarctic Peninsula since 1992. So the disintegration of these ice shelves played a big role in that. This is really the most special part of our study; understanding the causes of these disintegrations is crucial to improve the modeling of the ice shelf - ice sheet - ocean and atmosphere and here we add sea ice because we feel that this is a key part of the system as well, depending upon which ice shelves you are talking about. It is important, it is necessary to predict the fate of the remaining ice shelves in future and the fate of the ice sheet indeed as well, to reduce current large uncertainties in future estimates of the ice sheet contribution to sea level rise.

These disintegrations to date have been attributed to a couple of key things. Strong regional warming - as you are probably aware - the Antarctic Peninsula has been one of th strongest warming regions on Earth since measurements began in the 1940s. Where you get the warming, you get changes in material strength of the ice sheet, the ice shelves, and surface melt and brine infiltration as well. In addition to this, you get increased ponding on the surface as shown in this image of the Larsen B just before it broke up in January 2002. Now this increased ponding, as studies have shown, the water while it sits on the surface can actually infiltrate into existing crevasses and cracks or discontinuities within the ice shelf. The pressure of the water can force down and increase and open those cracks even more and compromise the structural integrity of the ice shelf. That has been called hydro-fracture. In addition to this warming, there are inherent glaciological factors involved as well. Outer margin fracture, the outer region here by bending stresses. This occurs where ice shelves melt and you get a redistribution of the stresses and that puts additional strain on the ice shelves and results in more cracking. You also get structural weaknesses which can occur inland where a glacier is cascading down from the ice sheet in contact with the bedrock or mountains and those rifts propagate down through time to affect the structural integrity of the ice shelf. Those two factors, strong regional warming and glaciological factors, appear to reduce the structural integrity or destabilize an ice shelf but they don't quite explain what we are seeing in terms of the break-up, that abrupt transition from a quasi-stable state to wide-scale disintegration. So what we are suggesting is could regional change in sea ice be an additional factor? And that is the story behind our paper.

In fact, all of the disintegrations to date - that's the Larsen A, the Larsen, and the Wilkins Ice Shelf over here - they have occurred adjacent to hot spots of decreased annual sea ice duration. All the reds in this map show where the sea ice annual duration has decreased. It's a trend and that is using satellite passive microwave data or ice concentration data. Where these disintegrations have occurred, the ice season, the amount of year when ice covers the ocean offshore is approximately three months shorter than it used to be. That is suggesting that these ice shelves have been exposed to open ocean and conditions more than they were in the past.

Every time I look at this I... I'm colour blind so you have to bear with me. I am trying to work through it as best as I can. Anyway, what these plots are showing is a time series of satellite derived ice concentration in boxes which are immediately offshore from the ice shelves that broke up. This is Larsen and this is Wilkins over here. What it's showing is this at the top here is 100 percent sea ice, so that is fully consolidated sea ice around the ice shelves. Down at the bottom is the opposite; it's open ocean around ice shelves. In both cases, you can see something of a change from more compact sea ice or more sea ice around the ice shelves to greater frequency of occurrence of open water periods around each ice shelf, particularly in the Wilkins but also in the Larsen here. The change seemed to occur roughly in 1991 for the Larsen and 1989 for the Wilkins. Looking at the annual values of these, we've shown that the Larsen B and A were exposed mainly - or more exposed - over a period from November to March, while the Wilkins was February to May and even June which is outside the melt season. That suggests to us there is something appearing here which is in addition to melt; hence, the sea ice loss. Previous studies have shown that this is in line with a study by Arigoney-Neto et al. in 2014 where they pooled together all the information on the Wilkins. They have shown that for the period 1947 to 1990, that indeed this Wilkins area didn't suffer any break-ups at all. But break-ups suddenly occurred once the sea ice started diminishing more and more.

What this plot is showing down here is particularly extreme event, an atmospheric event that occurred in January. The break-up occurred in January to March 2002 of Larsen B. This was related to an extreme exposure event, a wide open ocean corridor that existed from mid-November 2001 to mid-March 2002. This has been attributed to persistent northerly and north-westerly winds. I was stuck on an ice breaker with the Americans on the other side of the peninsula in the lead-up to this event. That's a different story but it produced a study there which showed the effect of compaction of sea ice and warm conditions on that side of the peninsula. What happened on the other side of the peninsula is that all these northerly winds swept the ice away and that is shown in the red colours in that bottom map. So it was total exposure of that ice shelf region.

So the first common factor is loss of sea ice; the second common factor appears to be this calving of outer - we call them sliver ice bergs; long, thin bergs - that occurred just off the margins of the ice shelves just prior to them disintegrating. You can see them here from the Larsen B and soon afterwards this mélange plume, the whole thing just disintegrated as a plume into the northwestern Weddell Sea. It's also the case with the Wilkins in 2008 and 2009, and also the Larsen A. So this let us to think are these two common factors - this lack of sea ice and this outer margin calving - are they somehow linked?

This was the basis of our paper and we put together this conceptual model. I am not expecting you to have a look at this and take it all in but what it is showing is on the left hand side that there are these ice shelf factors or glaciological factors that are part and parcel of these disintegrations but the ocean is playing a key role. The atmosphere is obviously playing a key role in terms of warming and the winds sweeping ice away. But what is new to this study is the sea ice factors over here. We are suggesting that sea ice loss somehow precipitated increased attack of these outer margins of the ice shelves, very vulnerable margins where with the waves and ocean swells and these triggered the disintegration of the ice shelves having first calved the sliver bergs.

Again I am not expecting you to take all of this in but the question is how does swell damage ice shelves in the absence of sea ice? Luckily, as I said we had some first rate modelling capability through Vernon in particular and Luke as well, of course. They looked at two aspects of this one of which -number 1 there - is the effect that sea ice has on in terms of damping or attenuating pressure waves that are coming towards the ice shelves. So we were able to look at the effects of sea ice and we fed into the models some information that we derived from the literature and also from satellite data. The second modelling is to do with the strain of the ice shelf itself due to the waves. We had a look at this strain with and without the sea ice cover around.

So what was the effect of this sea ice removal on the swell-induced strain? We were able to look - based upon wave model data - the peak periods we were looking at were in the vicinity of about 3 to 12 seconds for Larsen, and the Wilkins is 3 to 16 seconds, and significant wave height is about 3 to 7 metres and the directions were largely on-shelf when these disintegrations occurred. So with the sea ice buffer, there was strong attenuation of waves in these kind of ranges. That was shown by our model results and has also been shown by previous studies. The maximum strain that is due to this on the ice shelf itself were of the order of 10-8 to 10 -7. What these plots here are showing is here is a maximum strain at 10-5. The blue line is where there is no sea ice, and this area over here is including sea ice that is a sea ice zone which is about 80 km wide, and this is 250 km wide. So having sea ice makes a big difference to the amount of energy that can interact with the outer margins. When you take sea ice away, these strains increase significantly to about 10 -6. But for an 80 m thick ice shelf they increase to about 10 -5 and these kind of strains - based upon previous studies gain by Vernon Squire and colleagues way back into the 1970s - are sufficient to enlarge outer margin rifts and crevasses due to their fexural cycling and fatigue caused by these swells interacting with the outer margins. So 80 metres thick ice shelves seems very thin but studies on the Larsen C have shown that basal crevassing in the outer margin of the ice shelves quite dramatically affect the coherent thickness. Based upon those observations we suggest that the Larsen B and the Larsen A could be affected by the same processes whereby the basal crevasses decreased the thickness by approximately two thirds.

Another important thing is that the model showed us the plot over here; it's the maximum strain location in kilometres, from the outer margin in to the ice shelf and this shows the observed upper range of the waves, this is wave period down here. What this is showing us is that the maximum flexural strain occurred not deep into the ice shelf because all of the energy is dissipated by the thickness of the ice shelf itself. But it was concentrated on the very outer margins, and this is exactly where the sliver berg calving was occurring. It is also coincident with outer zones of increased margin fracturing. This is a satellite based image of the Larsen B before it broke up - that's the Larsen B there - and you can see this whole network of very important in this case crevasses that occurred towards the outer margins and indeed increased in length and number after January 2000 just before the ice shelf break-up. This is a similar feature in the Wilkins; this is the northern margin of the Wilkins and you can see the same structures which are barely holding together the ice shelf. Down in this area here, this is more protected by Latay Island from damaging swells and there has been less loss from that region just there.

There is further evidence of a relationship between sea ice change/variability and break-ups, in this case the Wilkins. As I said the break-ups of the Wilkins which were both catastrophic occurred not only in summer but also in winter and autumn. They coincide with increased exposure and low sea ice concentrations that occurred offshore. This is showing the 2008/2009, this is the seasonal mean sea ice concentration anomalies for each of the seasons again challenging my colour blindness but I had to write summer and autumn in there so I get it right. There were earlier break-ups here as well which coincide with these dips in sea ice coverage or sea ice anomalies around and certainly the case in 2008/2009 also.

We're on the home straight now! This is showing the Wilkins Ice Shelf; not only pack ice is the factor but an additional factor is land-fast or fast ice that is attached to the outer margins. You can see it in 2008 and 2009. A disintegration in 2008 and 2009 only occurred when this fast ice broke up and diminished. You can see here in 2008/2009 when there was fast ice, no break-up; when the fast ice broke up, the collapse quickly followed and it's even more dramatic in 2009. We think that fast ice plays an important role and acts as an additional buffer to destructive storm related swells. I should say we are looking at swells that come from distant storms across the global ocean. They also act as a mechanical bonding agent/buttress. They act as a glue if you like on the outer margins and minimise calving and also stabilise the ice shelves.

There is evidence from the Mertz Glacier Tongue of strong mechanical coupling between fast ice and ice shelves from a previous study. What you can see in this image here - it's a satellite synthetic aperture radar image from RadarSat - you can see that there is rifting in the glacier tongue there but this rifting was propagating through the fast ice which is attached to it and moving with it. That is suggesting that there is some mechanical coupling there and also we showed that the Mertz Glacier Tongue which was advancing at approximately 1 kilometre for the year - was advancing en masse with fast ice attached to it; this is shown in the relative displacement. We looked at the relative motion of the glacier and the fast ice using satellite imagery. This is just to show that on this side where there is fast ice, there is less calving, small scale calving, of ice bergs and on this side where there is a polynya there is significant calving of small scale ice bergs there.

This is showing that - based on additional work by Alex Fraser - this is a plot showing sea ice concentrations from that box region offshore from the Wilkins Ice Shelf. This is consolidated sea ice and open ocean just there; complete exposure down here. This is MODIS fast ice anomaly here; that is zero here, and negative and positive. What we are seeing here is that where there is extensive pack ice, there is also extensive fast ice; the pack ice protects the fast ice. In the years when there was break-up of the ice shelves, 2008/2009, there was persistent lack of pack ice offshore and this led to less fast ice and then break-ups. So pack ice effectively - this based on work that has been done over the years by Langhorne et al. and other people - shows that pack ice plays a key role in buttressing fast ice from swell break-up, something that anyone that studies fast ice, - penguin scientists like Barb or anyone else - understands this relationship. There is an intimate relationship; change in pack ice can affect fast ice, it affects - we suggest - ice sheet margin stability.

So, the synthesis, pulling all this together. We are thinking that in the cases we looked at, the Wilkins and the Larsen A and the Larsen B, a change in sea ice cover led to enhanced margin interaction of the ice shelves with swells. This over time cumulatively led to outer margin weakening to the point where these bergs, the sliver berg calvings, occurred. In effect, this removed key stone blocks from if you like an arch, relatively stable "compressive arch". This has been seen in satellite analysis that showed the strain rate trajectory is transvers when they are in place and the compressive arches are in place, then the ice shelf outer margin is relatively stable. But if you remove some blocks from that outer margin, it can lead to structural failure. Again basal crevassing would increase the likelihood of swell induced outer margin failure.

An earlier study by Doake et al. in 1998 predicted - and this is a modelling study - that outer margin failure, perhaps just a few kilometres inland, would trigger spontaneous chain reaction collapse of weakend ice shelves. This will be driven by stored potential energy of the system. A number of other studies have shown that this is water driving into cracks, hydrofracture. Once you set the thing in motion - and McAyeel and others have shown that this is ice blocks toppling like dominoes and moving outwards, and also localised tsunami-genesis where this toppling occurs that can generate significant waves which then feed back into the system to create further break-up. And another study by Banwell et al. has shown that chain-reaction drainage of the surface lakes can be another part of this chain reaction. The plume expansion - this whole plume of ice bergs - can extend out into the ocean if there is no sea ice buttress there. In the case of the Wilkins, this was unconstrained by the presence of fast ice. Also in the case of the Wilkins, fast ice plays a key role in reconstituting the ice shelf. So not always do the plumes blow out and remove all of the material. Some of it sits in place and is frozen together by fast ice. There are other studies that have shown there is a possibility of very long period waves affecting ice shelves. These are unaffected by sea ice presence or absence. Another possibility is that tidal effects can be playing a role but these are thought to be less significant and mainly occurring near the grounding line.

To wrap up: there are four common, essential prerequisites for disintegration in the cases that we have looked at, we think. The first is reduced coverage of sea ice and prolonged exposure to ocean swells. The second is extensive surface flooding and hyrofracture where the water goes down into crevasses. The third is extensive outer margin fracturing, and the fourth just before these disintegrations occur is that calving of sliver bergs. What we are suggesting is that - again in the cases that we looked at - while flooding and hydrofracture, which is the number 2 case there, are the drivers of catastrophic disintegration; it's the linkages between reduced sea ice, extensive margin fracturing and sliver berg calving which play a key role in not only preconditioning or destabelising ice shelves but also doing that to the point that they can trigger rapid disintegration of shelves weakened by multiple factors. Sea ice loss increases the odds of disintegration.

What does the future hold? IPCC and others have predicted that there is the rising temperature in the future in Antarctica. This could lead to weakening of ice shelves - that is remaining ice shelves, there is predicted sea ice loss and also more extreme wave states around the Southern Ocean. This could increase the likelihood of future ice shelf disintegrations. However, not all remaining ice shelves are likely to respond in the same way to such increased exposure. It also depends upon their glaciological characteristics, physical setting, and the degree and nature of flooding. This is a study [by Bell et al.] This image is of the Nansen Ice Shelf - it's by Bell et al. and came out a year or so ago; it shows that melt is not always destructive; sometimes - in this case - there are rivers or channels which move melt water off the surface into the ocean rather than the melt water sitting in ponds and creating fractures. Cold climate ice shelves with frequent open water offshore are apparently stable, we believe; examples being the Fimbul and Ross Ice Shelf. They can withstand wave flexure and this has been shown by a study by Bromirski et al. on the Ross Ice Shelf without sudden retreat IF they are not flooded and hydrofractured and the structural integrity is not compromised. But will this situation change in the future?

So final thoughts. The focus today has been largely on important ice shelf and ocean interactions regarding thinning. This is a very nice figure from the report by Gwyther et al. What we are suggesting is that sea ice change and variability could be an additional player to bring into this whole model regarding ice shelf stability and disintegration. This is in addition to the role of sea ice in modulating basal melt. A number of studies recently by Gwyther et al. and Silvano et al. and others have shown that where you get strong polynya activity, this puts a plume of salinity or saline water down into the water column which effectively cuts off incursions of warm water to lead to basal melt. This figure again, our study and other studies underline the highly coupled and complex nature of the ice shelf system which around Antarctica is undergoing really rapid change. It suggests that ice sheet models should include sea ice and waves as something of a step to forecasting the fate of the remaining shelves and reducing uncertainty in sea level rise. There are many unknowns; we need more understanding and modelling and observations on the linkages and the nature of these effects - mechanical or otherwise - and couplings. I guess the take away message here - I have to think hard about this - is sea ice change affects sea level rise. Sea ice is important on quite a number of levels; it's important for biology, the climate, biogeochemistry, etc. But this is another reason to better understand what is happening in sea ice, with sea ice around Antarctica. What is driving the observed trends, the patterns of change and variability? This is towards improved modelling and prediction and future coverage, and here we can add coastal exposure, too.

So follow up work in progress. We are developing through satellite remote sensing - Phil and myself and others - a Coastal Exposure Index. We are putting together a review paper on two-way linkages between sea ice and ice sheet margins and ice bergs, and that's both in the Arctic and Antarctic. The idea is to try and raise the profile of cross-cryosphere interactions/ linkages and identify gaps, and encourage inclusion in models, adopting a more holistic approach. With Alex and others, we'll be looking at case studies, regional, again of cross-cryosphere interactions and change involving pack ice, ice bergs, floating margins and fast ice.

I just like to wind up by acknowleding my co-authors and Wendy Piper and Nisha Harris who dealt very nicely - and I am very grateful - dealt nicely with the media around this about two weeks ago. This contributes to a WCRP CliC initiative and if anyone is interested, there is a Nature paper which is quite dense but there is also a conversation article which is hopefully a little bit more approachable if anyone is interested in following this up. And I always put this up and Go, the roos!

[end transcript]