JAMES OVERLAND >> Good morning, everybody! And welcome to the people that are on the Web as well. I'm James Overland. I'm a Research Oceanographer here at NOAA in Seattle, Pacific Marine Environmental Laboratory. The theme I want to go over this morning is how do you interpret when you have a large extreme event that you've never seen before. And it's not an easy question to answer and goes back to Enlightenment times without any really definitive answer on that. So part of these...working in the Arctic for the last 40 years, but then in the last 15 years, every year or two, we see what I call "surprises", unexpected large magnitudes of change. Probably the major one was in summer of 2007. Here's a map looking down at the North Pole. Greenland here. And the white is how much sea ice we had in summer of 2007. And that compares to the yellow line, which is the previous average value. So there was 40% less of the sea ice in that year compared to the average. And on the right-hand side we have a plot of how much sea ice was in the Arctic, September every year. If I stand over here and just show the curve through 2007, we saw this enormous drop...one year drop in 2007. Even in mid-summer 2007 we didn't really know that something like this was came. And so we were sitting down here with a lot less sea ice and a whole lot more open water. And the question was, "What's going to happen next?" If you think all of these changes are random, you might think, well, it's going to return more to the long-term downwards change. But in sea ice lore, when you have more open water rather than the sea ice, the white sea ice reflecting the sunlight, you absorb more sunlight into the upper ocean. And so, you're adding a null amount of heat into the Arctic. And the question was, "Have we added enough heat in here that it's changed the whole system?" So once you started adding more heat, then it would be hard to get rid of and we would continue whatever causes the start to happen, after you had your minimum amount of heat here, would you have a runaway loss. And that's what they call a "tipping point". However, you got down here, you had different physics related to adding more heat. In fact, what happened, we did go back to the long-term trend down. So we didn't really have a tipping point, in this case. And in fact, this would be called regressing back to the mean. When you have a real outlier, the standard story is it will return more to the long-term average. Another example happened a couple of years later. All the snow on Greenland all melted out one July, hadn't happened before. During the early part of the decade, here, we had several years in a row where Greenland ice melted a lot faster than it had been melting. But in the last few years, it slowed down compared to those rapid rates. So again, it's a question of a real extreme we never saw before but a general change back to the long-term rate. So those are types of surprises. But in this, we have a real long-term change in what's happening with sea ice. And again, we're looking at maps and through models and satellite data, people were able to calculate how old different pieces of regions of ice was. In fact, back in the 80s, you had a lot of old, thick ice that was moving around in the Arctic more than 7 years. So it was 3 or 4 meters thick. And most of the Arctic was made up of this really old, thick sea ice. And why that was important is for the climate, that was a real big flywheel. Doesn't matter what the winds are doing all that, this big enormous chunk of ice was pretty stable, didn't vary a whole lot from year to year. But then, what we've seen, especially in the last 15 years, is most of that old sea ice is melted out or moved out into the Atlantic and melting. And now most of the ice is just a year or two old. So rather than this big climate flywheel, we now have thinner ice that will move around faster and respond quicker to changes in winds and temperatures. So another way of measuring that, a friend of mine works on, from satellites, you can get a direct measure of that. The ice that froze in a certain fall is called first-year ice. And it has more salt in it than ice that's been around for multiple years. And the satellites can tell the difference. They can't tell the ultimate age of the ice but they can tell whether it was first-year ice or older than first-year ice. So these show the map for every January, starting in the year 2000. And the red are the amount of this old, thick multiyear ice. And this is for January, so the blue is now open water. It's actually this thin, first-year ice. So, back before 2007, we had mostly multiyear ice. And then in the winter after that, 2007 loss, most of that 40% open water froze and became first-year ice. So when we were down here in winter 2008, after 2007 summer, we had mostly first-year ice and a little bit of multiyear ice. And then in the following years after that, the first-year ice that stuck around is now second and third year ice. And so the actual amount of multiyear goes up because we now have summer 8-year old ice but we have 2 or 3-year-old ice. And then again after 2012, we had a real drop in the multiyear but then we came back with 2 and 3-year sea ice. And then this year we had another major drop. And so overall we've lost 60% of that old, thick multiyear ice compared to 20 years ago. And we've lost the climate flywheel from the Arctic. So one thing that happened is 2007, 2012, we had an extreme but we went back to the long-term average. And that's the general case for meteorologists. If we have an outlier case, it'll probably turn back more to average values. So that's certainly one thing that can happen. Another thing that can happen is we pass certain thresholds. North of Bering Strait, in the old days, walrus would, in the summer, would drift around on individual ice floes. But in the last decade, we've had more and more cases where there's no sea ice north of Bering Strait and they've had to come out on land. And as long as we continue with global warming, this is going to be primarily the case. A third thing is we can have some irreversibility going on. Permafrost is made up of frozen soil and ice. And if you melt that permafrost, the ice part runs away as water and there's no way you can put Humpty Dumpty back together again. The water part is gone and so, you know, the permafrost that had been there for tens of thousands of years is irreversibly lost. The other thing that can happen is with less sea ice and more open water, you can have more waves. So we're getting more erosion on the north coast of Alaska and that's an irreversible process, at least on decadal kinds of scales. So that's all introduction. What I want to get to is what we've seen in terms of Arctic temperatures in the last year. And so this is a temperature map. We're looking down at the North Pole. Here's Alaska. Here's Greenland. And we can contour areas that are all the same temperature. And so in the summer part of the Arctic near the North Pole and around there, we have 6-degree average temperatures. Averaged over January through March last winter. And this is nearly double the previous record for Arctic temperatures. So we're not talking some, initially of some process, well, we've just increased a little bit compared to where we were before and the same process is going on. The first thing you think about is, wow, this is really out there, something special going on. And we look into it and I'll show later, there was a change in the winds around the outside of the Arctic that brought warm air in. So it wasn't an internal process in the center of the Arctic, like reflection or radiation, it was primarily this outside source. So we say, well, this is a real outlier but it was caused by some random weather events, we'll never see it again. Well, this last fall, here's the same extreme temperatures going on and they had roughly the same cause from changing the winds coming into the Arctic. And now we have 2 examples of real outliers. So is it random or have things changed? Certainly from a meteorological point of view you can have 2 years that look identical. But, you know, the fact that we have such large numbers is still something to really try and assimilate. And not only did we have the warm temperatures, but I showed you the minimum sea ice in summer months, and we've had records on that. The red bars show that every month during the year had less than average sea ice area around. And not only that but if you look what was the rank where most of the winter and spring and again in the fall, this last year had the minimum amount of sea ice for that particular month, compared to all history. So there's a lot going on on the sea ice that's tied in with the temperature increases. So, you have to bear with me a little bit of geeky meteorology. We also want to look at the wind patterns. And meteorologists tend to look at the differences in pressure that caused the winds. And so if we're at the surface, you can look at a level surface or sea level, you look at differences in pressure and they cause the winds. But if you go up in the atmosphere, rather than looking at differences in pressure on a constant level surface, you turn that back around and you pick a constant pressure surface and then you look at how high that is and how it slopes. And so those are the types of weather maps that meteorologists look at. This is a map of the heights of one of those pressure surfaces. In regions where it's warm, air is less dense. And so a constant pressure level, the distance between the high and the ground will be higher in warmer areas generally, and lower in lower areas. So around mid-latitudes, you know, you have higher heights. The Arctic is colder; you tend to have lower heights. And so the difference between those 2 regions causes the steepness of the difference to be strong. And so the winds follow the lines of constant height. And if the lines are bunched, you have stronger winds. And also notice that there's some irregularities here. It's not just a uniform blob of cold temperatures. So the other thing about this jet stream or lower atmosphere polar vortex, is it can change from a real tight pattern with one cold spot to more of a wavy pattern where the flow is unstable or chaotic on a rotating earth. And so you can have a more wavy pattern. In the strong pattern, the jet stream is across northern Canada in winter. But if you have the wavy pattern, it can reach further down into the US and further down into eastern Asia. One of the things we're looking at is the difference between the more uniform pattern and the wavy pattern. Part of the reason for the warm Arctic is that last 2 years we've been more in a wavy pattern. So now for your final on this. Here is this height pattern for last winter. So the winds follow the constant colors. If you look at the light purple, you see 2 separate pieces of the polar vortex. It's split into 2 pieces. And so if you follow the light purple, the winds are coming up over Alaska into the Arctic. And they're flowing from around Iceland all the way towards the North Pole and they're bringing warm air with them. And so this last fall, if we look at the average from October to December, we have a very similar split vortex with not quite as wavy a pattern over North America. The other thing that was going on this fall is, I noted before, was record minimum sea ice. And so if we look at the sea ice map for November, the purple is where it used to be and the white is where it was now. So north of Norway here we have a whole lot of open water. And north of Alaska we have a whole lot of open water compared to what it used to be. And so...what's happening is this warm air is coming in from the south and 2 things are happening. One is the warm air is helping the ice, or keeping the ice from growing in the fall. This should be all ice covered by November but it's really retarding the growth of the ice as you're bringing warm air in the both of these regions. And also the winds are holding the ice back. But once you started having these regions of open water rather than ice, the warm air is flowing over a warmer surface and so the air can get further into the Arctic and reach the North Pole and places like that. So we have this new feedback that the warm air coming from the south is holding back the ice formation in the fall. That's allowing the air to go further and then the warmer air is less dense so it can change the height field and continue that split polar vortex longer than it might otherwise happen. So people say well, we should have this kind of feedback. They've been saying it for 15 years or so. But I think this last fall is a very clear case that feedback between the ice, the weather, and the winds, has kicked in. Now another question that comes up is do all these changes in the Arctic, do they affect the weather down where we live? And it's a real controversial question and part of what we're working on. One of the issues might be how can global warming and a warm Arctic affect cold temperatures down over the US? It's a complex issue. So one of the things we've done is, on the left picture here, when the eastern US is really cold, often times the western US and the ocean off of the West Coast is warmer. So warmer air has higher heights. So you tend to, if you map the jet stream, you tend to have this more wavy part, where the winds are coming up over Alaska bringing warm air on the east side of high pressure, which is called a "ridge", you're bringing cold air from the Arctic down into the central US and that eventually moves over the East Coast. This is a long-term pattern, there's nothing real Arctic-related to this type of pattern. But now, with no ice north of Alaska, things are a little different--that normally the northern end of this ridge feature goes over Alaska. But now we have an ice-free area that's warmer over north of Alaska and that can help lock in more of that wavy pattern. And the northern extent of the higher heights or pressure reach well into the central Arctic in early December and so you have a lot higher latitude for the start of bringing cold air down. This might be one first real example where the Arctic did play a role in the weather on the East Coast. There were, during early December, there were record cold temperatures on the East Coast. So where are we now? We had a really warm winter last year. And then we started off with a warm fall, this winter. But this winter is progressing just like last winter as well. And here's the map for January of the warmer temperatures. You can see there are warm colors throughout the Arctic. On the right here is a map that shows how much ice there is in the Arctic in every month, for every year. The blue line represents the amount of ice that's in the Arctic every fall month this year. And you can see that the blue line is below every other year. So we're continuing less ice this year than ever before. One of the differences is while the waviness of the jet stream is really strong over the Atlantic side, it's a little more variable on the US side. So we've been having cold and warm events throughout this winter. So it hasn't been all cold. In summing up, the Arctic is certainly moving to uncharted territory. Not only did we have extreme temperature but we just about doubled the previous ones. We're pretty much out of old, thick ice. So the ice is variable in speed and it's thinner so there can be a lot more interaction between the ice, the ocean, and the weather. So these are places we've never been before. And so what's next, how do we look at that? If you look at sort of one process at a time, you say, well, according to statistics, when you have an extreme that future cases will rather than stay out there, will return more towards the long-term conditions, sea ice is tending to do that. 2007, 2012 and when they run models, you can't really make the models have a tipping point and explode to completely go away. You need the general long-term increase in global warming to keep that going. So that's a possibility. If you think in terms of the wavy or not-wavy jet stream, the last 2 winters had the wavy jet stream. The change between these patterns is rather random. So pretty soon, we've got to eventually end up with a year with more of the straight wind pattern which would knock out the really warm Arctic. And it's not something happening in the Arctic that's the main cause of the warm temperature. It's the new connection between the Arctic and the mid-latitude winds. On the other hand, we have 2 years where we've had the warm Arctic that when we have the wavy pattern, we've ended up with a new extreme temperatures. We expect the wavy pattern and the non-wavy pattern to oscillate back and forth in different years. But it sure looks like when we have the wavy pattern with the winds that are warm coming to the south and retarding the ice and then the open water allowing the temperatures to increase, we certainly have 2 good examples of that. So we have additional information now that perhaps every time we get a wavy pattern we're going to be more likely to have the warm temperatures. So that's an example of Bayesian statistics where we have a certain condition and then we have new information and we need to look at the new information rather than just saying, oh, it's all random. So in summary, the Arctic is not only changed, but in the last year it's really taken a clown leap to places it had never been before by a lot different. And part of the speculation then is things tend to be random. We do think that there's a new example of the feedback between the weather, the winds, the temperatures, and the loss of sea ice. So thank you very much. We can take some questions. [APPLAUSE] Chris. CHRIS SABINE >> Thanks. That was a wonderful talk, Jim. Thanks. So what happened last September? It seemed like we were on-track to be exceptionally low ice, then all of a sudden you broke like in your rankings, you went from 1, 1, 1, 1 all the way up to 5 and then now you're back down to 1s again. What happened there? OVERLAND >> Yeah, good question. So, whether you have low sea ice in September depends on 2 things. One is, the amount of ice you start with. And we started with a lot less ice in May. So we said well, if you just have average conditions for the winds and the weather during the summer, you're probably going to have an extreme. Rather than the high pressure that we've seen for the last 5 years in the summer Arctic, we went back to the kind of weather patterns that we used to 20 years ago in the Arctic summer where we had more low pressure and storms. So, the weather itself during the summer was not conducive to a low sea ice in September. And when you added the 2 together, that's why we hit the third or fourth minimum case. It was a minimum case based on a starting condition. But it was countered by the weather in the summer. NICK BOND >> You mentioned what is indeed a controversial topic about the potential for feedbacks of the open water on the circulation itself, and as you say, you know, how wavy the pattern is. And of course there's lot of ways for the atmosphere to have those kinds of circulation. Are people, um...What are they doing in terms of kind of dynamically linking...trying to link the loss of sea ice to more likely asymmetric patterns like that. In particular, is anybody doing things where they take the actual atmospheric conditions and like, imposing sea ice and seeing if it's different? OVERLAND >> Yeah. Important but difficult question. Two weeks ago I was at a workshop on that very question and there were 100 people that showed up. So, because it impacts where people live, there's a lot of interest in that. The problem is that you can come up with ideas for how the loss of ice has warmer temperatures and increases these height pressure fields...and impact. But there's a whole lot of chaos going on and different processes both in the Arctic and then when you get down to mid-latitudes--you know, we have tropics and we have the Blob, warm temperatures in the north Pacific, and the Arctic, all potentially feeding into that. So people are looking at it as an important problem, but it's an extremely difficult problem to tease out because there's so much chaos going on in the actual atmosphere. And there are a lot of model studies to try and do what you say. They take out the ice and run the cases again. Different models give different answers. And the same model, if you run it over and over, give different answers as well, due to just the background chaos. So where the field is, is we don't think the whole East Coast is going to move to colder winters. You can't say the climatology is really affected, but you can find case studies such as here, and you can find case studies for eastern Asia where you can show the linkage from the ice sea through 3 or 4 different processes, ending up with cold temperatures. CHIDONG ZHANG >> Jim, you showed that 6-degree warm temperature anomaly from Global Reanalysis - are there any in situ observations or satellite observations that can give that results, give you similar amplitude of warm temperature? OVERLAND >> When we published this, that was the main question we got. So there are now 4 different re-analyses, as you know, out there. So we compared the different re-analyses that the re-analyses take the individual observations and try and make fields of them but they do it in slightly different ways. And so, there was about a 1-degree difference between in the re-analyses there. And also go back and look at individual stations in northern Scandinavia and northern Canada and Barrow, Alaska that map real well on to those. Those observations are in the re-analyses as well so we're pretty confident that those overall patterns and the magnitudes are real. Certainly within a degree or two and that puts them way out of the bounds from what we've seen before. WEI YONG>> So you showed a map of November ice cover and there is a big open water in the GINS Sea--Greenland Sea, Norwegian Sea, Iceland Sea--that area. So being an oceanographer, I keep wondering, do you think the oceanic heat transport into that region might also play a role besides the atmospheric forcing? That's one question. The other question is--You showed this satellite ice thickness time series in the past 20-ish years. If you looked at the thick ice, there actually seems to my eye, there is a regime shift after 2007. In other words, in sea ice thickness, there might actually be a tipping point. Comparing to ice extent which tends to fluctuate a lot so I was wondering if you have any thoughts on that? OVERLAND >> I think that's a good point. And if you look at the curve from all the years, we recover slightly because we're getting some 2 and 3-year old ice after we had mostly first-year ice. That's not a real recovery back to where it was with 7 or 8-year old ice. That's a really important point is we don't necessarily see it in the ice extent but we may see a bit of a tipping point in the ice thickness. The other point that you're bringing is, the variability of the inflow of warm water in the ocean helps keep that ice back north of Norway and that varies. So that helped the case here but the real "why did it happen quickly in one year" had to do with bringing the warm air and not only the warm air but the air was moister and that radiates more radiation energy and keeps that there. So there's some effect there but I think the bigger one is we had some warm temperatures off the northeast coast of North America. One of the questions is how did the sea surface temperatures also help setup the long wave pattern. And there's a lot of chaos in those patterns but the way I like to think about it, if you have the warm sea temperatures or different sea ice, it's like rolling loaded dice that you'll have more of a tendency for the wavier pattern. People think that the North Atlantic temperatures help set this flow into the Arctic and that was due to...you have this wavier pattern, but the wavier pattern can move west or east and that affects locally what's going on. You can't ignore the oceanography in the North Atlantic or the North Pacific. NED COKELET >> So Jim, I always look forward to your talks on the Arctic. I never look forward to the message you bring though. [CHUCKLE] Not your fault. Split polar vortices - presumably you can look back through re-analyses in the past and look at those. Have they happened in the past? Are these more extreme than what we've seen in the past? That sort of question. OVERLAND >> In fact, they are a fairly common feature in winter. It's just that this one was stronger than that. When you go back looking about these warm intrusions of warm air into the Arctic, they happen every decade or two decades or so. In fact, I was looking over the whole winter but there are several papers that talked about not only over time is that warm air coming in but it's coming in in a couple of really strong storms. You see in the paper that, in fact last week, that the North Pole is above freezing in the first week in February. Those have been seen in the past but not as clearly as this. The magnitudes of the temperatures have never been seen. Yeah? ALAN MEARNS >> Jim, what's happening to the glaciers, like in Baffin Island? Are they breaking off ice islands? I guess I'm wondering what's happening to those chunks of ice that are coming off the land. [INDISTINCT COMMENT] OVERLAND >> [CHUCKLE] Good question. I don't know for sure off of Baffin Island. There's a long-term downward trend in loss of glaciers throughout North America. And those regions are certainly involved so... We have cruise ships going through the Northwest Passage, and there's a lot of channels between the islands north of there, and so it's fairly dangerous if you have some ice up in those islands even though it looks like the Northwest Passage is clear. You can have some bigger icebergs drifting down in there. I'm partly finishing up on Ned's plate. MAN >> Be afraid! [AUDIENCE GROANS AND LAUGHS] OVERLAND >> Part of the story is we keep having these surprises. Those don't really show up well in the computer climate models. And the computer models end up being slower. Muyin and I work at looking at these climate models and we've written up the Arctic should be 4 degrees warmer by 2040. Well, we've had 2 years where already the real world is saying we have a winter that's on average already 4 degrees. So presumably we're not going to go back but it's these interactions that I'm talking about between the weather, the winds, the temperature, and the sea ice, that are not really involved very much in these models and are showing that they can really amplify where we are. So we're less solid about what's the 3 or 4-year prediction because we have these interactions. On the other hand, the modelers that do sea ice are really looking at that steady downward trend and saying well maybe things aren't as bad as they might be. But I think it's these surprises that make me really nervous about that. And if we're able to slow down the emissions and limit the global temperatures to 2-degree warming, the Arctic's still going to be 4 or 5-degree warming by mid-century. So that's completely going to change the whole Arctic--what you can do economically, what animals are there, how the ecosystem is structured. So, even if we're somewhat stabilized and can live with a 2-degree warming here, the Arctic is going to be way out of what we've ever known in the Arctic. So that's the bad news. The good news is, hey, there's lots of cool data to look at. JOHN JANSEN >> Permafrost, with, you know, Greenland snow cap, and of course, the glaciers and everything. Is freshwater input or is that new freshwater circulation patterns a factor at all in setting a new baseline for oceanic currents or sea surface temperatures and whatnot, either locally or regionally? OVERLAND >> Uh, yes! I'm involved with some global assessments of the Arctic and one of the main chapters is on freshwater and...that the whole freshwater cycle, the rain runoff, the freshwater in the oceans is all increasing and speeding up. The models are projecting that the Arctic should have more net rainfall than it used to, simply because you're bringing more moisture into the Arctic. Yeah, that's a part of it. Ok, thank you very much! [APPLAUSE]