Hot Topic has devoted a lot of posts to events in the Arctic over the last northern hemisphere summer. The loss of sea ice was dramatic – there was 25% less ice in September than the previous record, set in 2005. The little graph to the left shows just far off the trend line last year’s September area really was. And as I posted yesterday, recent studies suggest that the Arctic is primed for more significant losses in the near future. If the reduction in summer sea ice continues, there are some pretty major implications for the climate of the northern hemisphere and for our modelling of the global climate, and it’s those things that I want to consider in this post. Please note: I am not a climate scientist, and there are a lot of ifs and handwaves in this argument, but bear with me…
What really got me going was the news from the American Geophysical Union conference in December that half of the volume of sea ice had melted between 2004 and 2007. That’s a lot of ice, and it set me thinking about the “heat budget” for the Arctic. Assume for a moment that the Arctic climate is stable. Every summer, the ice melts back to roughly the same area, and every winter it refreezes to the same, larger, area. The freeze/thaw cycle is determined by the seasonal changes in incoming solar radiation and winter radiation of heat out to space (and quite a few other things, like albedo, but let’s keep it simple). Heat gained in summer is matched by heat lost to space and to the atmosphere over winter. The key thing to keep in mind is that the winter heat losses are essentially fixed.
Now lets add some heat. We’ll ignore changes in the relationship of the earth to the sun, the Milankovitch cycles, because they operate over much longer time scales than we’re interested in here. There are several ways to get heat into the Arctic. We can increase the amount of greenhouse gases in the atmosphere: this will reduce the heat lost to space, especially in winter. We can increase the amount of warm air and water vapour moved into the Arctic by the atmosphere through weather patterns, and we can increase the amount of heat carried into the Arctic by the oceans, especially the North Atlantic. There are feedbacks, too, that speed up the process, such as the “albedo flip” when white snow and ice is replaced by dark, heat absorbing water. If more heat is going in than can be lost each winter, then the Arctic will accumulate heat. Where does this heat go? It melts ice. All of these factors can be seen at work in the Arctic today.
To change water to ice, it has to lose a lot of energy – heat. Wikipedia explains enthalpy of fusion, also known as latent heat of fusion:
The heat of fusion can be observed if you measure the temperature of water as it freezes. If you plunge a closed container of room temperature water into a very cold environment (say −20 °C), you will see the temperature fall steadily until it drops just below the freezing point (0 °C). The temperature then rebounds and holds steady while the water crystallizes. Once completely frozen, the temperature will fall steadily again.
The reverse is also true. To melt the ice, you have to put heat in. You don’t see much temperature change until the ice is all gone. I’m not going to attempt any back-of-the-envelope calculations based on estimates of ice volume losses and the heat of fusion of water (333.55kJ/kg), but for the Arctic to lose as much ice as it has in the last ten years, the overall heat budget must be strongly positive.
The most aggressive predictions for the loss of the remaining summer sea ice are those made by Wieslaw Maslowski of the US Navy – a researcher with access to sea ice data from nuclear submarines patrolling in the Arctic ocean. Maslowski’s ice model suggested that the ice could all be gone by 2013, based on data up to 2004. In other words, prior to the huge losses of 2005 and 2007. “You can argue that maybe our projection of 2013 is already too conservative,” he told the BBC. Even if he is wrong in detail, other workers in the field have been rapidly revising their estimates for the loss of summer sea ice, bringing them nearer and nearer to Maslowski. And the consequences for what we think we know about the likely future state of the global climate are profound.
Once the last multi-year ice is gone, the Arctic will be much more like the Antarctic, where the sea ice melts away every summer, and then reforms every winter. The difference is the heat accumulation – which will continue. All the positive feedbacks will remain positive, and the Atlantic won’t suddenly stop shipping heat north of Spitzbergen because the ice has gone. The heat that’s been going into the latent heat of fusion will go into raising the temperature of the Arctic Ocean. How fast that might happen, I don’t know enough to speculate, but the end state could look like the Paleocene-Eocene Thermal Maximum (PETM) when the summer ocean temperatures could have been as high as 23C. Not much chance of winter ice forming in a sub-tropical Arctic…
When the summer sea ice is reduced to a few little chunks clinging to northern Greenland and the Canadian archipelago (within ten years?) the autumn freeze-up will start progressively later as heat continues to accumulate, and the spring melt will begin earlier. As the Arctic Ocean warms, the climate of the entire northern hemisphere will change – and in ways that no-one can project, because none of the models get the current ice loss – and therefore Arctic heat budget – right. The current crop of climate models aren’t very good at modelling the PETM Arctic – and they’re not doing a good job of keeping up with the current rate of change. This means that IPCC projections of likely climate change are all old news. Until the models can produce the changes we’re seeing, and then progress them into a future where the Arctic is seasonally – and eventually annually – ice free, they can’t give us good information about the sort of global and regional climate states we might see over the next 100 years.
There is a real danger that events are overtaking the science. A rapidly warming Arctic will bring significant – and rapid – changes to the northern hemisphere climate, and the potential for some very nasty feedbacks (melting permafrost releasing methane, methane hydrates boiling up from the sea bed) to kick in. We will not be immune down south. The impacts of an increase in rate of sea level rise will affect us all, and the economic dislocation of rapid change could be huge.
I am aware that this will be interpreted as “alarmist” by sceptics, and those who prefer to take what they feel to be a “moderate” view of the impacts of climate change. If there are holes in my argument, tell me. Perhaps the amount of heat going north is really not that great, and the ice will bounce back towards the long term trend line, and will still be around in 50 years. I hope it is – I hope I lose my bet with Stoat. But I don’t think I will. At the very least, the scenario I’ve described here needs to be considered as a “worst case” and evaluated scientifically and strategically. And there needs to be a lot of urgent work on the modelling of oceanic and atmospheric heat transport into the Arctic Ocean. Really urgent work, because I can’t see how any amount of emissions reductions are going to stop this happening.
Since you won’t publish your own back of the envelope maths, here’s mine
Latent heat of fusion H2O=333.55kJ/kg
specific heat capacity H2O(liquid at 25C)= 4.184J/cm-3/1K = 4.184kJ/kg/1K
333.55/4.184 = 79.72
So the energy required to melt 1kg of ice is equal to the energy required to lift the temperature of a litre of water by about 80C.
Is that right so far?
If so, seasonal melting and freezing of arctic ice must have a significant moderating effect on high latitude NH extremes.
Your figures look good, but the interesting thing is that the heat loss side of the budget is essentially fixed (barring volcanoes) – determined by the length of winter. The seasonal cycle is “cooling” enough to cope with some heat coming in from the North Atlantic and Pacific, but over the recent past, extra heat has been going into melting permanent, multi-year ice – the stuff left at summer minimum. (Gobally, heat has also been going into melting ice in Greenland, Antarctica and mountain glaciers).
When the multi-year ice is gone, the excess heat arriving in the Arctic will heat the ocean – and as your calculation shows, the water temperature will rise, making it harder for the winter ice to freeze. The accumulated heat has to be lost before the re-freeze can begin, so the onset of freeze is delayed. Hence my conjecture that once the multi year ice is gone, the winter ice is in deep trouble. There is a tipping point of no return, and I suspect we’ve passed it.
Bugger.
Bugger indeed.
Science is by nature a conservative beast, with the collation of data, writing up of results, peer review, acceptance into journals, and publication often glacially slow…no pun intended. Indeed, as with glaciers, in the wake of an overheated Washington lobbyists, politicians, and vested interests, our understanding of the impact of climate change on the human condition seems to have retreated, evaporating through obstinate disbelief.
Two texts that have influence my thinking (and unashamedly, my writing) in this regard have stood the test of time well, I think:
1. ‘The Younger Dryas period has caused much debate, since it challenged the previously-held idea that climate could only change very gradually. It had been thought that the thermal inertia of the ice sheets was so large that significant advances or retreats could only happen over long periods of time. The Younger Dryas demonstrated unambiguously that change can be abrupt. Climate appears sometimes to respond in a manner similar to earthquakes where stress and strain builds up over years, leading to sudden abrupt changes, rather than slow incremental changes.’
(Ref: In Confronting Natural Disaster: Engaging the Past to Understand the Future, G. Bawden and R. Reycraft, editors, pp. 75-98. University of New Mexico Press, Albuquerque, 2000.)
2. ‘The Holocene abrupt climate changes, hemispheric and global in extent, were of lesser magnitude than those characteristic of the Pleistocene, but they profoundly disrupted late hunter-gatherer, pastoral, and agriculture-based societies within various environments and at various levels of socioeconomic hierarchization, centralization, and regional command…. after a thousand-year period of post-Pleistocene climate amelioration (Alley 2000; Peteet 2000). The effects of this climate change on humans–and possibly Hordeum and Triticum populations as well (Rossignol-Strick 1999)–were radical. Hunting and gathering bands were forced to adapt to rapid drying and cooling of niches where wild plants and animals had formerly provided abundant subsistence (Bar-Yosef and Belfer-Cohen 1992; Moore and Hillman 1992).’
(Ref: Beyond the Younger Dryas: Collapse as Adaptation to Abrupt Climate Change in Ancient West Asia and the Eastern Mediterranean by Harvey Weiss)
My point being that there’s been considerable research on the impact of abrupt climate change on a younger Homo sapiens who had not yet learned to be maladaptive to rapidly changing circumstance. It will no doubt prove to be an interesting research project for some future student to examine how Homo urbanis and his siblings, Homo technologi and Homo consumption, coped in the first decades of the twenty-first century.
As Joseph Campbell pointed out, references to a world wide flood appear in every major mythology including, of course, Biblical mythology.
Hmm, maybe Noah was a myth, but I’m beginning to understand how he felt.
Gareth, I’ve never seen the details, but it’s my understanding that all of the model results show the Arctic winter ice persisting for a long time. Given the relative stability of the Arctic Ocean, the upper layer of the water would have to start out as pretty warm water to avoid dropping below freezing during the long night. Of course as soon as even a very thin layer of ice forms it acts to conserve the heat (even while allowing the atmosphere above it to cool much more).
Steve, I don’t doubt there will some winter ice for a considerable period of time after the summer ice is gone, but if we assume the “excess” heat being transported into the Arctic continues at the same rate as today (and, really, the main question here is about rates), then as soon as multi-year ice is gone, more of it can go into warming up the ocean. To me, this means that eventually (how long?) the winter ice will go too – unless we can reverse the warming trend. That would seem to involve not just cutting emissions but actively removing carbon from the atmosphere.
The GCMs are currently of little use from a policy perspective because until they get the Arctic right, they cannot get future climate states right. Forgive me, then, for being sceptical (hah!) about their projections for the persistence of winter ice…
Per Maslowski, his RCM got it right because it has the resolution to pick up the increasing warmth from ocean currents. If that’s representative of the problems the GCMs are faced with, we’ll be in for a bit of a wait since the computer capacity needed to do that kind of resolution in a GCM is years off.
The written material I’ve seen on Maslowski’s presentation included nothing about what happens to the sea ice post-2013, but I assume we’ll be seeing a publication shortly.
All of that said, IIRC there’s paleo evidence of winter sea ice persisting into quite warm times. I wasn’t trying to ascribe a lot of significance to this, BTW, since a short period of skinning over isn’t going to do much to slow down the melting of the Greenland ice sheet. The other recent Arctic results (sorry, can’t find the cite) showing that the troposphere is warming in winter darkness are especially worrisome since that would appear to put the GIS on a much shorter track to melting at high altiudes and for much more of the year.
I just found this paper (“An Ice-Free Arctic? Opportunities for Computational Science”) discussing the future of Arctic sea ice modeling (although I haven’t read it yet).
Speaking of ice sheets, see this from GRL this week:
Neglecting ice-atmosphere interactions underestimates ice sheet melt in millennial-scale deglaciation simulations
Abstract: Dynamic and thermodynamic interactions between the atmosphere and underlying ice sheets are generally not represented in the traditional one-way boundary condition forcing used to drive ice sheet models. This shortcoming is investigated through a series of idealized millennial-scale deglaciation simulations designed to isolate the mechanisms regulating the deglaciation timescale of the Laurentide ice sheet. Sensitivity experiments indicate that the conventional use of one-way (non-interactive) atmospheric forcing fields leads to an unrealistically insensitive melt response in the ice sheet model even when atmospheric carbon dioxide is set to modern preindustrial levels and Earth’s angle of obliquity is set to its early Holocene value. A more realistic deglaciation timescale is obtained only through the application of a new two-way (interactive) asynchronous ice-atmosphere coupling scheme and a seasonal ice albedo parameterization that accounts for the observed darkening of ice in the moist summertime ablation zone.
Also new from GRL relating to ice shelf collapse:
Channelized bottom melting and stability of floating ice shelves
Abstract: The floating ice shelf in front of Petermann Glacier, in northwest Greenland, experiences massive bottom melting that removes 80% of its ice before calving into the Arctic Ocean. Detailed surveys of the ice shelf reveal the presence of 1–2 km wide, 200–400 m deep, sub-ice shelf channels, aligned with the flow direction and spaced by 5 km. We attribute their formation to the bottom melting of ice from warm ocean waters underneath. Drilling at the center of one of channel, only 8 m above sea level, confirms the presence of ice-shelf melt water in the channel. These deep incisions in ice-shelf thickness imply a vulnerability to mechanical break up and climate warming of ice shelves that has not been considered previously.
Here’s the winter warming aArctic troposphere paper (actually RC’s discussion of it) I couldn’t locate above.
Thanks Steve, some very interesting stuff there – especially the Tremblay et al paper (An Ice-free Arctic?). A quick skim suggests it provides an extremely useful overview of what’s going on… I shall read it properly forthwith.
Thanks for the site and discussion, which I’ve just discovered. Here’s a post I made yesterday at The Oil Drum, where CC is discussed on occasion. I’m just an amatuer, but I’ve been looking closely at the declining ice mass (area x thickness) for several years now, and I’m puzzled why I don’t find more discussion of the latent heat of fusion. Here’s my comment in response to a question about why the loss of Arctic sea ice would affect the THC:
It’a a complex system, so there are multiple possible effects. But the main one is that as the ice sheet shrinks, the albedo declines, and solar radiation that would have been reflected by white ice (perhaps 90%) is instead absorbed by blue water (again, about 90%). So it’s a postive feedback that rapidly accelerates. This is why I expect to see an ice-free arctic within the next 3 years – perhaps even this year. That allows the arctic in general to warm very rapidly, which of course is already happening. That puts additional pressure on Greenland’s ice sheet. As that melts (again, already happening) massive amounts of fresh water are infused into the North Atlantic. This dilutes the salt water, making it less dense, and less apt to sink. It is the sinking of dense, saline water in the far north that drives the system. So anything that affects sinking, could slow or stop the circulation.
Not mentioned nearly as often is the latent heat of fusion. It takes 160x(!) more energy to melt a given volume of ice than it does to raise the temp of the same volume of ice by one degree C. So once that ice is no longer there to be melted, that 160x energy is free to go about warming the arctic waters and surrounding air, rather than breaking the bonds that hold ice in its solid form. This seems huge to me. Not only will there be less energy reflected by the ice that is no longer there, but that same energy will be doing vastly more heating of the environment when the amount previously absorbed by the phase change from ice to water is reduced.
Heat Numbers for H20
Ice to ice: 0.50 cal/g-C
(half a calorie to raise a gram of ice by 1C)
Ice to water: 80 cal/g
(80 calories to melt a gram of ice to water)
Water to water: 1.0 cal/g-C
(1 calorie to raise a gram of water by 1C)
Water to steam: 540 cal/g
(well, at least we’ll have to wait awhile to boil the seas)
Steam to steam: 0.48 cal/g-C (whew)
So for every gram of ice that melts, the next equivalent amount of incoming solar energy can heat 80 grams of water by 1 degree C. We’ve already lost about 80% of the ice mass of the Arctic. Not only the shrinking area matters – what used to average 10 ft thick now averages about 3 (or less) feet. IMHO, it’s the combination of albedo reduction and the latent heat effect that is causing the Arctic to warm so much more rapidly than the globe as a whole, and I believe plays a key role in abrupt climate change. It’s like a heat cancer. And Greenland is next in its sights.
http://www.theoildrum.com/node/3928#comment-337650
Hi Dan,
There are some things that can act to cool the Arctic. A persistent freshening of the upper layers of the ocean could encourage winter ice formation, and there are signs this is happening (the freshening, that is). You could see a situation where there’s virtually no ice in summer, but a very rapid freeze up in autumn/fall to quite large areas.
This summer is going to be very interesting, as I’ve recently suggested…