This is's Typepad Profile.
Join Typepad and start following's activity
Join Now!
Already a member? Sign In
Recent Activity
They must either converge in the end or 2013 has to slow rapidly.2013 will slow rapidly, because we will reach the end of the melt season. I think 2013 is going to run out of time. Looking at Jim's graph here, it looks to me as though 2013 is roughly 1 million square km behind 2012, and 2012 lost roughly 3 million square km from now until the end of the melt season. So, in order to catch up with 2012, 2013 would have to lose ice at a rate one-third faster than 2012. Your comparisons fall a little bit short of that. I'm not saying it is impossible - if the cloud clears and a dipole anomaly gets itself set-up ice melt and export could be very dramatic - but I reckon it is strongly odds-against at this point.
This is amazing work Rob. Thank you! From my reading of the Student's t-test, it suggests that the probability of breaking the 2012 record is less than 0.05%, which is perhaps over-stating things, but this is the best methodology for predicting September sea-ice extent/area that I have seen. Interestingly, your methodology suggests that the chance of this season coming in below 2007 is a bit more than 20%. One of the interesting things about some previous records is that when a year [say 2007] has set a new record, all subsequent years also break the previous record [2005]. Will this hold?
Toggle Commented Jul 2, 2013 on Problematic predictions 2 at Arctic Sea Ice
@Wayne - The variation in the year-to-year increases in CO2 have very little effect - if any - on the year-to-year variations in global temperature/Arctic sea ice. Carbon dioxide is actually a pretty weak greenhouse gas, and it's only the accumulation over many decades and centuries that is now having an effect on the climate. The year-to-year increases in carbon dioxide are affected by ENSO - this is why 1998 is the record year, as the strong El-Nino meant that the land and oceans did not absorb as much carbon dioxide as they usually do. The strong El-Nino is also the reason why it was such a warm year in the global temperature record, because during an El-Nino heat is redistributed around the tropical oceans in response to changing wind patterns. At the moment the ENSO phase is almost at La-Nina levels, the opposite of El-Nino, so for carbon dioxide levels to be increasing so quickly does indicate that human emissions are still increasing.
I am reminded of a post that Tamino made some time ago, where he showed that adding in information on the area/extent in June added very little predictive power for predicting the September minimum extent. What matters from here onwards is the same that has mattered in previous years: weather and the trend. We have seen the effect that the weather can have, in that it has delayed the melt in areas such as the Kara and Beaufort Seas. The effect of the trend is also there to see, in that the ice in the central Arctic is so thin that it has been pulverised by PAC-2013. We can't predict what the weather will do for the next couple of months - it will be interesting to watch and see - but we can be quite confident that, because of the trend, the minimum extent will be one of the lowest on record.
NeilTToday's AMSR2 image was showing significant open water (or almost open water), north of Svalbard and past the pole on the Russian side.I've been looking at the Uni Bremen AMSR2 maps and the MODIS images, and I'm beginning to wonder whether there will be any ice left on the Russian side of the Greenwich Meridian and the Date Line by the end of the melting season. The cyclone has really broken the ice up so there's lots of open water, and lots of ice edge for sun-warmed ocean waters to eat away at. On the other hand, we're only eight days away from the Summer Solstice now, and cloud from the cyclone is providing cover at the time of the year when the sun is at its strongest, so perhaps there won't be enough time to melt it all.
3.64 million square kilometres. I think one has to be very cautious about deviating from the excellent predictions of Tamino, which are for 3.97 +/- 0.9 million square kilometres. However, I have chosen to do so. My prediction is that there is nearly a 50% chance of a new record being set this year, with my 3.64 figure being slightly above last years 3.61. My reasons for this are: 1. Last year did not seem to be set up for a big melt. We thought the ice was thicker in the Beaufort Sea. There wasn't a consistent Arctic Dipole as in 2007. Yet the record was still smashed. 2. PIOMAS has been largely vindicated by the Cryosat observations. At some point the decline in thickness will accelerate the decline in extent - perhaps the cracking indicates that this year is that year? I've not gone too far below Tamino's forecast though. It's worth remembering that the NSIDC mean September sea ice extent has not yet set a new record two years in a row. And tomorrow will be six weeks from midsummer. We are already in the quarter of the year with most solar insulation into the Arctic, and there is still 13 million square kilometres of sea-ice reflecting that sunlight back to space. On Midsummer's day itself, it is likely that sea-ice extent will still be more than 10 million square kilometres of sea-ice. The final interesting point I would make is that following the 2007 minimum, none of the subsequent years have had a minimum extent above the previous minimum (of 2005). All have had less sea-ice than in 2005. If 2013 follows this pattern then we would expect that it would also be below 2007, if not necessarily below 2012.
R. Gates This was very interesting to read. Thanks. I haven't read as many papers on SSWs as I would like, but from what I am aware of, the common working hypothesis on SSW formation is that mid-latitude Rossby wave-breaking events are responsible for disturbing the polar vortex and creating a SSW event. This also fits with the observed disparity between SH and NH occurrence of SSWs, because Rossby wave-breaking occurs more often in the northern hemisphere, due to the storm tracks, Rocky Mountains, etc. This does lead to the somewhat unsatisfying situation that blocking events (where wave-breaking of the flow creates a block in the flow) can be seen both as a precursor to a SSW, and as a consequence of a SSW. What do you think of the mid-latitude Rossby wave-breaking hypothesis? SSW events occur roughly one year in two - I would guess that it would be interesting to look at composites of the meteorology in your source region for years with and without a SSW. It would also be interesting to look at these regions in seasonal forecast models, to see whether differences there are responsible for explaining the difference in predictability of SSW events. This is a long-winded way of asking for your email address.
ScottPerhaps I'm unaware of a environmentally sound process to extract methane hydrates, but this initially strikes me as the last thing we would want to do in the Arctic.I suppose the best thing you can say about extracting the methane hydrates to burn is that converting it straight into carbon dioxide is marginally less damaging to the climate than having it released as methane directly into the atmosphere. People have talked about the importance of not leaving this as a problem for governments to address. A positive development in this regard has been the Transition Towns movement, which is seeking to involve local people in creating a more resilient local economy. It would be a good use of my time to become more involved in this, I think.
Toggle Commented Apr 17, 2013 on Perception of the Arctic at Arctic Sea Ice
Neven, you said this:the jaw-dropping sight of an ice-free Arctic that was projected to occur somewhere towards the end of this century, could be witnessed towards the end of this decade (if not earlier).Here is what I fear will happen. All the Arctic sea ice will melt, relatively soon. There will be a lot of talk about how this is very important, and we must do something, but nothing will really change in people's day to day lives. Deniers will post that photo of that submarine and claim that nothing exceptional has happened, that there were no satellites when the Vikings were in Greenland, or the Romans were growing vines in northern Brittania, and so it's quite likely there was no Arctic ice then and this is no big deal. No Arctic sea ice will become the new normal, and, well, some days it will rain and other days it won't and it will generally be colder in winter than in summer and most people will see that nothing much has changed in their daily lives, except for food prices going up, etc, and although this huge momentous thing will be different - an ice-free Arctic for the first time in Millennia - everything will pretty much stay the same. There's a lot of very interesting rational analysis, but it turns out that humans simply aren't as rational as we thought. The problem is too big, and yet too incremental, for us to address it. Like the proverbial frog we are going to allow ourselves to be boiled alive (figuratively speaking). I mention Arctic sea ice melt and the response of most people is to shrug their shoulders because it doesn't affect how easy it is for them to rent a flat, or pay for the groceries, or forge a romantic relationship, or anything to do with their day-to-day concerns. Everyone is transfixed at how amazingly cold this March has been - the 14th coldest in the Central England Temperature record stretching back to 1659. Nobody really noticed when March last year was the 4th warmest in the same record. The ridicule heaped on Dr David Viner for that article in the Independent on snow now comes from mainstream people looking for an easy chuckle, whenever it happens to snow. Everyone ignores that it is still getting warmer. I tend to think that we have lost, or at least that we are losing very badly. Facts and data are not helping us when the other side is proving so very good at using the natural variability to hide the problem.
Toggle Commented Apr 16, 2013 on Perception of the Arctic at Arctic Sea Ice
Ice is absorbing almost as much heat as the atmosphere? What is going to happen when we lose a big chunk of that? Scary...Well, judging by the graphic alone, most of the heat will go into the Oceans instead, and it will make very little difference. Looking at it in more detail, I did some back-of-enveloped calculations on what would happen to SSTs in the Arctic once it was ice-free, and, because the specific heat capacity of water is a lot lower than the latent heat of fusion of water-ice, I'd naively expect SSTs in the Arctic to shoot up rapidly in that scenario. This has implications for delaying the re-freeze of sea-ice in the Autumn, supporting above freezing temperatures in the Arctic for a longer period of time, thereby extending the length of the melting season for the Greenland Ice Sheet, etc.
Toggle Commented Apr 10, 2013 on A new round of vids at Arctic Sea Ice
I've been somewhat sceptical of the PIOMAS data in the past, due to the relative lack of verifying data, but the comparison with Cryosat is impressive. I retract all my previous comments relating to the unreliability of the PIOMAS data. This means the volume declines shown by PIOMAS have to be taken at face value, rather than hedged with doubts about it being modelled. Following PIOMAS we can be confident that there will be a complete melt of the Arctic sea-ice in a melting season well before the end of this decade, and if the right conditions for melt occur it could happen this melt season. Looking at the wipneus extrapolation graphs, we could have an Arctic free of sea-ice for six months of the year by the end of this decade. I reckon there's a 1-in-5 chance of the complete melt-out happening this year. I wonder if they'll ever launch another satellite to look at Arctic sea ice? There might not be any left by the time the next one would be due to go up.
@crandles - There are two different situations. 1. For the IPCC experiments then the models are run with fixed 1860 conditions for a spin-up (ideally until they reach a steady equilibrium). Then you have a transient run started from the end of that spin-up point with historical forcings up to nearly present-day, and various different scenario forcings up to 2100. 2. For testing purposes - ie when the sea-ice modelling scientists are testing changes to their parametrisations - then you can't do runs like that. It would take too long. So if you want to test a change in something like the cloud scheme, that might affect the solar radiation that reaches the sea-ice in summer, you have to do so with a shorter run. Then you face a choice of doing this with fixed present-day conditions - where you have the best available observations of clouds, atmospheric circulation, etc - or with fixed conditions at 1860 - where most of the observations are missing. --- It isn't a simple problem to fix, because there are constraints on supercomputer time and on the observations available. With the benefit of hindsight, though, it does seem pretty obvious that the models should have a stable Arctic sea-ice bias for this reason.
Toggle Commented Jan 11, 2013 on The real AR5 bombshell at Arctic Sea Ice
If I remember the Stroeve paper correctly, it said that the models that got the past trend right, had a much lower SIE to start with. I'll look it up. Found that interesting at the time. This is to do with how the models are developed and tested. The [coupled] climate models are [mostly] not tested by being run in transient mode, but in a steady-state mode. So they are run with constant "present-day" boundary forcings and compared with present-day average observations. This means that they have a bias towards modelling sea-ice that is too stable, because the sea-ice has to survive for a 50-100 year test run with something like year 2000 CO2 levels. So, why not test the models in transient mode? A couple of reasons - firstly you want to use that test later on for attributing climate change to greenhouse gases, so you can't use a transient run for your testing. Secondly, errors in the ocean drift require a long (many hundreds of years, ideally as long as possible) spin-up of the model before the transient runs are started, and there isn't enough computing capacity to do that for every development test run. It would be better [in sea-ice terms] to do development runs in pre-industrial conditions, but then you have the problem of not having good observations to compare your model with. Good observations only existing for recent decades (and even then they have problems). A broader point about the IPCC - this was exactly the intention. If you go back to when the IPCC was established, it was because ad-hoc groups of scientists were saying more forceful things about the climate, and the powers-that-be wanted to tie the scientists into a process that would constrain them from speaking out. The IPCC makes any scientists who warns of more extreme risks look like a mad alarmist, and also provides false reassurance to most climate scientists that there work is being listened to by working through the IPCC channel.
Toggle Commented Jan 7, 2013 on The real AR5 bombshell at Arctic Sea Ice
GlennYou have to have some sympathy for the guys working in Sea Ice modelling. Still a newish field... need to model the terminal phase of the very thing they are modelling. On the other hand, this provides a massive opportunity for young scientists to make a name for themselves, so it would be a pretty exciting time to work on sea-ice modelling. Also, while we look to be quite close to an effectively complete melt-out in September, the sea-ice will still come back during the long polar night for many years to come. I don't know how much the Arctic would have to warm for the Arctic ocean to store enough heat to prevent sea-ice formation over the winter, but I would guess at a very large amount. Thus the focus of sea-ice modelling will switch pretty quickly to forecasting the timing of the opening and closing of the major shipping routes, rather than the annual minimum (which will in the relatively near future always be close to zero). This sort of transition was one of the reasons given for the mooted (but now aborted) merger between the British Antarctic Survey and some other NERC research institute (was it NOC?).
Toggle Commented Dec 11, 2012 on PIOMAS November 2012 at Arctic Sea Ice
@Ggelsrinc - Have you read the trilogy of books Forty Signs of Rain, Fifty Degrees Below, and Sixty Days and Counting by Kim Stanley Robinson? I think you'd find them interesting. I think most people would. As part of the narrative there are various large-scale geo-engineering efforts to repair the damage created by global warming, which are not too dissimilar to your own idea in terms of their scale and feasibility (ie the scale is gargantuan and the feasibility low). Theoretically, I think what you propose is possible, though there would be several challenging technical problems. What I would envisage would be large buoys which could be deployed to the Arctic during summer. If we assume that each buoy has a very powerful pump then one might be able to thicken a square kilometre of sea-ice per buoy. The technical challenges that would need to be overcome would be: (1) The pump. To cover a square kilometre it would have to be very powerful - farm irrigation systems can cover about one-third of a square kilometre with a "big gun" style sprinkler - so this is feasible. (2) The temperatures. You would be operating in temperatures of below 30C at times, maybe for four months of the year. This would make pumping large quantities of water very hard. (3) The energy. The buoys would need a dense, high output power source for the pump. Something nuclear, inevitably. I doubt that the systems designed for satellites would deliver enough power to operate (and heat) the pump, so you might need to use something derived from nuclear-powered submarines. This would be quite large (and very expensive). (4) The ice. One of the big risks to these buoys would be that they would be crushed by movements in the ice-pack, so they would have to be quite strong, particularly if they are nuclear powered. (5) The numbers. If we assume that we have a modest aim, to thicken just 1 million square kilometres of sea-ice, and that we have a very powerful pump as above, then we would require 1 million buoys. By comparison, the international ARGO project has 3594 floats/buoys deployed as of yesterday, so you would need nearly 280 of your large, nuclear-powered, cold-resistant, very high pressure pump buoys for each ARGO float. I think all these technical problems would be solvable, just about, but it would be really hard, and ruinously expensive. By way of comparison, each ARGO float costs about US$ 15,000. The annual cost of maintaining a fleet of one million ARGO floats would be about US$ 6.7bn. How many times more expensive would our nuclear-powered Arctic pump buoys be than ARGO floats? I think US$ 15 million would be very cheap for this. So the cost would be, very roughly, at least seven trillion United States Dollars. Every year. By way of comparison, one estimate I have seen is that World War Two cost the United States five trillion dollars (in present prices), in total. So, to conclude my lengthy and assumption-filled comment: I can imagine what you propose being possible, but it would be so difficult that there are very many things we should do first, and it's possible that removing CO2 from the atmosphere would be a cheaper way (of repairing the Arctic by reducing temperatures) with the benefit that once it was done, you wouldn't have to do it again.
Toggle Commented Oct 10, 2012 on Naive Predictions of 2013 Sea Ice at Arctic Sea Ice
Observed 2012 values in every case fell below the central predictionLarry, thanks for presenting these statistical predictions. I find it interesting that the 2012 values were below those predicted. This can mean one of two things: 1. Overall, the weather of the melting season was more favourable for sea-ice melt than on average, so that 2012 is below trend. 2. The trend is increasing at a faster rate than predicted by the Gompertz curve, so that a weather-neutral year like 2012 ends up below the estimated trend. I know there are papers in the literature that attempt to decompose the contribution to sea-ice melt from the weather, and I wonder whether the metrics that they use - based on wind patterns, etc - would help to determine what effect the weather had on the melt this year. The Great Arctic Cyclone is interesting to consider in this context, because it likely would not have had anywhere near as large an impact if it had occurred during the 1980s, because the thicker sea-ice at that time would have been less susceptible to being churned-up. Also, one might consider the very existence of such an anomalously strong Arctic Cyclone as itself a feedback resulting from the very low sea-ice extent. This tentatively leads me to suggest that #2 may be closer to the mark, but really I'm just guessing. I have one direct question for you. Is it possible to calculate the probability that PIOMAS volume in 2013 will be below 2,000 cubic km? Obviously, one can tell that it is somewhere in the range 5% - 45%, but it would be interesting to know where.
Also, if we're thinking about predictions, here is one worth keeping an eye on. After 2007 set a new record below 2005, all the subsequent years have also been below 2005. My prediction is that all years subsequent to 2012 will now be below 2007. Yes, next year might well be higher than this year, but I think it's doubtful it will fail to go below 2007.
Neven:You thought 2007 was big when it smashed the 2005 records? Well, 2012 is to 2007 what 2007 was to 2005. If my calculator doesn't deceive me, the 2007 record was a drop of 22.2% from the 2005 record. The 2012 record is a 23.6% drop from the 2007 record.I think you have made some sort of a mistake, or you have a mischievous calculator, or I will shortly be greatly embarrassed. [Extent] 2007/2005 = 4.30/5.57 = 77.2% ie 2007 was a drop of 22.8% on 2005. [Extent] 2012/2007 = 3.61/4.30 = 84.0% ie 2012 was a drop of 16% on 2007. [Area] 2007/2005 = 2.78/4.03 = 69.0% ie 2007 was a drop of 31% on 2005. [Area] 2012/2007 = 2.11/2.78 = 75.9% ie 2012 was a drop of 24.1% on 2007. Have I done something wrong? 2012 was still a big decline, and what was most notable is that this decline happened without what we've thought of as the optimal conditions for sea-ice melt. I'd also be interested to know if this year has increased the linear rate of decline - I'd guess that it has, but I don't have the old September graph saved anywhere to check.
The big problem with models is that few are designed with an open water arctic in mind. They talk about a warmer arctic, ignoring the water side of the equation. There's a caveat in that I haven't dived into the actual way the models are built, but I seriously doubt a "wet" arctic could be modeled, since there's no precedent or easy comparison.Can I recommend that you take some time to read about how global climate models are constructed? A good place to start might be the introduction. The way climate models are set up they would automatically account for the change in the Arctic, as it warms up and the sea-ice disappears in the model simulation. Of course the models are not perfect, so their response to this change could be wrong for some reason - for example errors in the way the model represents cloud could lead to the model producing too much/little cloud over the newly wet Arctic - but the model would at least be having a try. Just as an example, as part of CMIP5, the same models that are used for 21st century simulations are also used for paleo-climate simulations, where the boundary conditions are altered to those of the last glacial maximum, so one can see how well the models recreate those conditions.
Oops! There's a sentence above that should read: " random fluctuations would have become self-reinforcing..."
It IS implausible, given that the Arctic has begun absorbing MUCH more energy in the past few years due to the onset of significant melt-out in the Arctic basin during the period of maximum insolation (late May-beginning August).The trends for June are in the lowest half of the trends for each month. See: Compared to September: I don't think that the ice-albedo feedback has started properly, given the much lower trends in June. In any case, we know that local feedbacks cannot be overwhelmingly strong, because otherwise Arctic sea-ice extent would never have been all that stable, as random fluctuations wouldn't have become self-reinforcing, leading either to a complete lack of sea-ice or a new ice-age. The most important factor is external forcing to the Arctic basin, and I'd have thought that some fluctuation in the external forcing could lead to an apparent, temporary, slowdown in the trend. I'm not saying it's likely, just that it's a possibility. I'd give it a probability of roughly 10-25%. Before this melt season I'd have given it a probability of 25-50%, but seeing as how the massive melt happened this year despite unfavourable atmospheric conditions, it's now looking less likely. I'm just providing a realistic scenario to try and keep the hyperbole under control.
It looks like models have a hard time keeping up with the rapid changes of the Northern Hemisphere cryosphere that are unfolding as we speak.Isn't there a bit of a danger that we're making too simplistic a comparison between the model projections and recent observations, in a similar way to that made between recent global surface temperature observations and model projections by the other side? If you look at the individual model projections in detail, rather than the multi-model mean, you can see that the models do project periods of rapid sea-ice decline, followed by temporary stabilisation - an inverse of the pattern seen with increases in global temperature. This is not surprising, because we have a situation where you have natural variability imposed on a trend. Sometimes the natural variability will act to exaggerate the trend, and sometimes it will act to hide it - as we are recently seeing with global surface temperatures. It's not implausible that the Arctic will see a period with relatively modest further declines in September sea-ice extent, if some facet of natural variability (the Atlantic Multi-decadal Oscillation, perhaps?) acts in a contrary sense to the trend.
What's going on north of 85 between about 60E and 165E anyway?It's melt Jim, but not as we know it. The point about extent is that it is the best observed measure of sea-ice. If there were thirty years of satellite ice thickness observations then I'm sure it would be used more often.
If Arctic sea ice continues melting at this rate, the question arises whether there is any sense in trying to optimize models.It does currently look very likely that the September sea ice will all be gone before the models are reasonably competent at predicting its disappearance. However, it is still worth improving the models because: (a) It will become increasingly important to forecast the timing of the seasonal Arctic sea-ice melt, as more trade uses the North-West passage and Northern sea route. Being able to do this a couple of weeks in advance could be very valuable. (b) Related to this, the big ice-albedo feedback kicks in with sea-ice declines in the months of May, June and July, and not September. These months give the models longer to catch-up. (c) Knowledge gained by modelling the Arctic sea-ice will be useful when modelling the Antarctic sea-ice, and possibly also the Greenland and Antarctic ice sheets, as well as other myriad smaller ice shelves. All that said, it is much easier to improve a model when you have detailed and accurate observational data to compare it against, otherwise you are flying blind when trying to work out why the model is failing to produce the right answers.
Someone, somewhere, on a blog, said something about the inevitability of the return of sea-ice each winter, which set me thinking about heat storage in the Arctic Ocean and the ice-albedo feedback. So I had a look at the monthly mean extent trend plots from NSIDC, such as for May, which is the month with the lowest rate of decline. June also has a low rate of decline, in the slow half of the year, and these are two months near the summer solstice, in late June, when the potential for the ice-albedo feedback is at its greatest. What this suggests to me is that the decline of Arctic sea-ice is only just getting started, and an ice-free date in September in the near-future is just the end of the beginning, rather than the beginning of the end. When the sea-ice decline in May and June starts to take hold, then we'll really see the ice-albedo feedback at work, and there will be all manner of rapid changes.