Another month has passed and so here is the updated Arctic sea ice volume graph as calculated by the Pan-Arctic Ice Ocean Modeling and Assimilation System (PIOMAS) at the Polar Science Center:
Last month I wrote:
The 2012 trend line isn't quite hugging the 2011 trend line as much as the latter hugged the 2010 trend line, keeping a polite distance. It's still 1069 and 1537 km3 below 2011 and 2010 respectively. Hopefully the trend line starts pulling up a bit.
Unfortunately, it seems the trend line didn't pull up and has stayed in position, still 1099 and 1136 km3 below 2011 and 2010 respectively.
Here is Wipneus' version of the same graph, for the first time with a calculated "expected" value for 2013 (dotted lines), based on the same date values of 1979-2011 and an exponential trend. A caveat from Wipneus: "Note that the statistical error bars are quite large."
The trend line on the PIOMAS anomaly graph is now well within the 2 STD zone:
Mind you, this is anomaly from the downward linear trend, not from some baseline. 2010 and 2011 showed the same behaviour, so let's see whether the current trend line will reach the 1 STD zone again.
Time now for the PICT thickness graph. I have used my crude method of dividing PIOMAS volume numbers by
Cryosphere Today area numbers to calculate the ice pack's average
thickness. Again, this is just an indication that allows us to compare
with previous years:
Average thickness for December 1st (in m):
- 2005: 1.46
- 2006: 1.54
- 2007: 1.35
- 2008: 1.32
- 2009: 1.32
- 2010: 1.18
- 2011: 1.14
- 2012: 1.06
With CT sea ice area returning to 'normal' (see graph), meaning it's not freakishly below other years, and PIOMAS volume not returning to 'normal' (in other words: 2010 and 2011 levels), thickness will automatically be low. Maybe the rapid expansion of the ice pack in the Bering Sea is having to do with that.
The thickness graph from the Polar Science Center shows the same:
That was the last PIOMAS report for this year. Next year: new trend line, new colour, new data, new hints, new melting season.
And as Wipneus shows, for 2013 zero is within the error range for the volume minimum... (although it takes 2 or 3 more years for zero to become likely)
Posted by: AmbiValent | December 12, 2012 at 23:22
Looking at ARC ice speed and drift, the ice is moving, meaning that it is being fractured. It takes a long time to conduct that heat out of ice. Surface water was more subject to mixing in 2012 so there is more salt in the formed ice. Also, there was more heat than normal in the near surface waters that will tend to keep the ice warmer than normal. Thus, the ice there is, is weak.
The ice we see today, is likely to melt rapidly next summer. I expect a powerful Arctic cyclone next summer that will facilitate melt.
I still expect max ice in 2013 @ just under 14 million km^2, and that the Arctic will approach sea ice free condition by the end of the melt season.
Posted by: Aaron Lewis | December 13, 2012 at 01:26
I believe annual global insolation is relatively constant and so should be the average ice melting potential that enters the Arctic each summer. Any heat held back in one year is still there to add to next year's flux. That melt potential for at least the last ten years exceeds winter ice formation by at least 750 km3. If only a linear decline, the ice will inevitably be gone by September in only a few years and by June in just a few more. To me it appears that Wipneus's exponential decline appears an even better fit and that he is the Nate Silver of ice melts.
Posted by: Ron Mignery | December 13, 2012 at 01:37
I noticed an event in Wipneus' graph of exponential trend broken out by month.
The trendlines for July & November are crossing right about now, as June and January did in 2011. This sort of shift in the seasonal pattern is is notable and is clear in the data from the last few years. If the trends continue, the relative seasonal change will be quite large.
I certainly think the 0 intercept for June is more plausible than for November. I hope I'm right about that, because zero volume in November by 2019 is hard to fathom. Of course, the same can be said about the entire picture conveyed by the graph -_-
Posted by: Ethan O'Connor | December 13, 2012 at 02:02
There is a new report out from the US National Academy of Sciences on Artic Sea Ice Loss that might be of interest.
"Seasonal-to-Decadal Predictions of Arctic Sea Ice: Challenges and Strategies"
With registration, one can download the pdf.
http://www.nap.edu/catalog.php?record_id=13515&utm_medium=etmail&utm_source=The%20National%20Academies%20Press&utm_campaign=NAP+mail+new+12.11.12+B&utm_content=&utm_term=
Posted by: Apocalypse4Real | December 13, 2012 at 03:34
I have a lot of banners of the Arctic from space I've been using on the freeways here in California:
http://freewayblogger.blogspot.com/2012/12/one-thousand-signs-plus.html
If anyone has any good text to go with them, please send it along. ("Now disposable for your convenience" has been up over a week in a spot passed by 325,000 cars per day)
Don't know if this changes anyone's attitude or behavior much, but it's a hell of a lot of fun.
Posted by: Freewayblogger.blogspot.com | December 13, 2012 at 05:18
It is from 1979-2012, that is why the "expected" december values are not there.
Posted by: Wipneus | December 13, 2012 at 12:41
Wipneus
An open question. Does PIOMAS gridded volume information? Is it possible to do the sort of analysis you are doing but with regional data so one could look at the trends and projections region by region?
With the likely minimum for next year removing nearly 40% of the remaining volume, the impact on Area/Extent could be huge. It would be interesting to try and separate the volume of ise in the bastion north of Greenland/Ellesmere from the rest of the Arctic. Might give a better indication of just how much Area/Extent are likely to get hit in the rest of the Arctic.
Posted by: Glenn Tamblyn | December 13, 2012 at 21:53
Chris Reynolds did look at doing regions per NSIDC region mask. We haven't got the volume calculation perfect which might be due to area calculations or something else. Also haven't got any 2012 data.
It would be nice if google could get all the information necessary to put it in an app for further analysis. I don't know whether they intend that to become publicly accessable. Beta testers may get an early look and experiment opportunities but probably have to agree not to publicly share information and need to be comfortable writing javascript and/or python.
Posted by: crandles | December 14, 2012 at 00:01
Glenn:
That is entirely possible using the data files on PIOMAS page: http://psc.apl.washington.edu/zhang/IDAO/data_piomas.html As crandles notes teh needed data are only available to 2011.
Volume per gridpoint can be calculated by thickness*length*width
thickness from files heff.H*, length and width from the file grid.dat.pop, which also gives you longitude/latitude of each grid point.
I have R-code that does this. If anyone is interested I can publish it.
Whether such analysis will lead to more insight I do not know. For me one of the striking features of the (all arctic) volume losses is that the signal (the decline) stands so much out of the fluctuations: the story is plain to see and denying is clearly ridiculous.
Regional volume losses will have a lower signal-to-noise ratio.
Posted by: Wipneus | December 14, 2012 at 11:44
What happens after the ice is gone? In broadest terms to my understanding, heat enters the Arctic in summer and exits into space in the winter. Currently about 1000 km3 of ice melting potential melts old sea ice instead of radiating to space. Presumably as the summer ice disappears, the Arctic must warm up until winter radiation matches summer influx. How much warmer is that? Is there still winter ice in that scenario? What then happens with Greenland, permafrost, shelf ice, etc.?
Posted by: Ron Mignery | December 14, 2012 at 15:46
Ron, your calculus is way off!
If you add up the numbers from the Greenland Records thread, we are this year above 1400 cubic km's in extra melt compared with last year. If you add the (sofar unknown) melt from mountain glaciers, we may exceed 1500 cubic km's this year.
When all the snow is gone, sea ice melt picks up. When the sea ice is gone, the tidewater glaciers may start to disappear in earnest. Then you only have the mountain glaciers as a buffer, before Greenland plunges into the ocean. Forget about a local sunshine-albedo feedback. Since we are talking big-scale advection here (both in the oceans and in the atmosphere), you may start considering how to spend some 334 KJ/kg of lost ice in your "Arctic oven".
Posted by: P-maker | December 14, 2012 at 16:11
P-maker
Sorry, I meant when the summer sea ice is gone. This year's melt was, for now at least, an outlier, and 1000 km3 seemed to me a better argument for sea ice only. The melting from glaciers will continue for many decades after the summer sea ice is gone and so will not add to the radiation deficit. I was asking about the consequence of the deficit in the years between the loss of summer sea ice and the loss of the glaciers.
Posted by: Ron Mignery | December 14, 2012 at 16:40
Rubbish!
This year' melt is NOT an outlier. It's the most recent year of an increasing series of melt years!
I am not sure about the "many decades" you mention. Tidewater glaciers and mountain glaciers have a radiative balance as well.
We are already losing tidewater glaciers and mountain glaciers at an alarming speed, so there are no "years between".
Posted by: P-maker | December 14, 2012 at 17:09
Still not frozen up north of Svalbard on 14 Dec 2012 according to today's data from DMI.
Below I consolidated their two views by overlaying the sea surface temperature anomalies above 1C (oranges) over the sea surface temperatures themselves (blues).
Posted by: A-Team | December 14, 2012 at 17:26
The exact number does not matter as regards my question. To restate, how much warmer does the Arctic need to get to radiate an additional 1000 km3 or 1400 km3 of ice melting potential during the winter? What does the warmer Arctic then look like? Dates and numbers please;)
Posted by: Ron Mignery | December 14, 2012 at 17:32
Average of last 10 year minimums have fallen 890km^3 per year and the trend seems to be upwards. PIOMAS last data at 30 Nov was 1100 km^3. 2007 and 2010 were outliers for high volume loss, and 2012 isn't. 2012 may however have been an outlier for extent and area loss.
So that is qualified support for P-maker's not an outlier. But 1000km^3 remains a sensible number.
I think Ron Mignery's statements are pretty reasonable. It was clear to me he meant sea ice. Also I think it is clear that 'years between' referred to years after loss of summer sea ice and before all glaciers are lost. There was no reference to tidewater or mountain glaciers. I believe there will be many decades of that. How long do you expect Petermann Glacier to last?
If you are going to pick on statements partly by misunderstanding what is being said, how about we pick on "before Greenland plunges into the ocean" and call that ridiculous.
Posted by: crandles | December 14, 2012 at 17:35
I think there is an enormous difference between ice surface at -30C in winter and water at -1C for the amount of heat that is lost to space. It is partly going to depend on whether a thick fog develops. I think convection will prevent a real pea soup dense fog. So I think there will still be ice in winter for a long time but I am afraid I am not up to putting numbers and dates on this yet.
Posted by: crandles | December 14, 2012 at 17:39
I'd run the numbers using 1210 km2 as an average over 3 years and came to the figure of .7C temp rise for the top meter of the Arctic ocean (after my math was corrected).
I'd think an an additional .7C/yr would cause much later freeze-up, much foggier/cloudier conditions and an end to seasonal ice in the not to distant future.
Terry
Posted by: Twemoran | December 14, 2012 at 18:28
Ron (and a comment for Twemoran),
Increased cloud is something that bothers me. While there have been warmer periods in the Arctic during the last million years, these periods weren't associated with the very high CO2 levels we have at present.
Terry, with regards your comment at Dosbat, more than the issue of inundation of the East Siberian Sea, I suspect it is CO2 and the year round impact on downwelling infra red that could prove to make what comes different from warm events in the last million years.
As I've recently posted here, the warming in Autumn is massive, and it's due to open ocean venting much of the heat gained during the summer. This negative feedback is already in play. Yet we still see substantial losses of ice volume from the sea ice. I had been very critical of talk about an imminent sea ice free state based on trend extrapolation. I'm still critical of trend extrapolation. But I now expect a sea ice free state in late summer this decade - due to behaviour of the ice since 2010 and the apparent failure of autumn heat loss to stem the rate of volume loss.
Crandles is correct that there is a massive difference between -30degC and -1degC - the physics make this a rather trite observation, but it's important to keep in mind. So this is another negative feedback. A third major negative feedback is the saturation of IR bands as water vapour increases - you get the best warming effect from increasing water vapour early in the process.
So I think we will see substantial warming once the Arctic transitions, after which the summer ice free period will expand adding to the regional warming. But people shouldn't get carried away, there will be limits, unless...
Unless cloud radiative forcing steps in to take the system to a whole new level of warming - but even that has limits, they're just much warmer.
Posted by: Chris Reynolds | December 14, 2012 at 18:49
Perhaps the increased fog due to increasing Arctic ocean temperature, then in time leads to dense clouds especially in winter.
And perhaps that is how the equable climate came to be during the Eocene.
Posted by: Sam | December 14, 2012 at 18:59
Sam, I'm in total agreement.
Posted by: Chris Reynolds | December 14, 2012 at 19:08
Would fog in winter in the Arctic really matter? With no sunshine, isn't it just a black body situation where fog would radiate as readily as water at the same temperature below it?
Posted by: Ron Mignery | December 14, 2012 at 20:19
Lads,
In IMHO dense fog is closely related to advection and pollution in London in the '50ies. You may also experience dense fogs over the banks off New Foundland - which is also closely related to advection of warm humid air over a cold surface.
If you foresee dense fogs over the Arctic Ocean, you need to make the case, that all this space will be dominated by a shallow layer of fresh and cold surface water throughout the entire winter season.
As I have tried to explain several times, the dominant feature of this autumn has been the persistent katabatic winds off the North Slope of Alaska, off the Greenland ice sheet, off the Siberian North Slope and off Antarctica. Thus, as long as you have off-slope winds, you will have up-welling near the coast and no fog (except for sea smoke, which is a completely different kettle of fish).
I therefore kindly urge you to consider carefully, whether fog or low-lying clouds really makes any difference, when it comes to the Arctic radiation balance in the winter season.
Posted by: P-maker | December 14, 2012 at 20:29
Entirely speculation but sea smoke is more what I envisage:
With little convection the top of the 'fog' would get cold while the bottom is warmed by the oceans creating convection to move warm moist air upwards while the cold top of 'fog' falls and warms but there is more evaporation than that air can absorb creating sea smoke which rises as more cold air falls to replace it.
The rising moist air is likely to rain. If rained out, the process of creating a fog then starts again.
So there is upward movement of heat by latent heat transfers as it rains and by convection.
This loss of heat probably isn't as quick and efficient as radiation through clear sky, but winter lasts a long time so a lot of heat has to be built up to avoid this heat running out. So I would tend to suggest quite a number of years where the winter ice gets thinner and is lost slightly earlier before enough heat is built up to last the winter. As I say just speculation.
Posted by: crandles | December 14, 2012 at 21:13
Ron, P-Maker,
I think people are too fixated on fog. On the issue of fog, it will still reduce IR because the nature of forming fog over ocean is that the ocean is warmer than the air above. Fog in an inversion forms because the ground is colder than the air, so cools air in contact below dew point, a different process.
Clouds matter because they back radiate infra red, so they act as a blanket, reducing radiation to space. Because droplet size is of the same order as the wavelength of infra-red concerned, a layer of stratus acts effectively as a blackbody, stopping the direct radiation of IR to space. A warmer atmosphere also matters because it holds more water vapour, and the enhanced greenhouse effect that ensues further backradiates IR radiation.
Low lying clouds do make a difference. There is no getting around it. Katabatic winds from Greenland are a regional issue and have little impact over the Siberian coast. During MIS11 when Greenland melted entirely the warming effect in distant (>3000km) NE Siberia was of the order of 0.3 deg C. The cooling effect of katabatic winds from Greenland on the Siberian permafrost and ocean sediment clathrate warming will be similarly inconsequential.
Sam mentioned the 'equable climate' issue. Which I discussed in my most recent blog post. Theoretically (Abbot & Tzipperman "Sea ice, high-latitude convection, and equable climates") at doubled CO2 with an increase of atmospheric heat transport from the current 100W/m^2 to 140W/m^2 it is feasible that the Arctic could be kept sea ice free year round. Back off from that admittedly extreme case (it's a very large increase in atm heat flux) and it is apparent that at doubled CO2 a staggeringly small amount of sea ice could be all that is left through the Arctic winter. In such a scenario, where the refreeze is severely retarded, Siberia could easily be kept warm by enhanced atmospheric heat transport and increased clouds.
Posted by: Chris Reynolds | December 14, 2012 at 21:17
I believe one of the icebreakers late this summer was commenting on the dense fogs encountered, so dense they were being picked up as ice by the satellites.
I'd assume anywhere you have open water and an inversion layer, fog would result (think LA smog).
Ron - I think a dense fog/cloud cover would act as a very efficient blanket separating the warmth below from the cold above, which would be where the black body radiation would be emanating from.
P-maker - Not sure of Alaska, but katabatic winds are so prevalent in Greenland in the fall they even have their own name. The Piteraq winds are particularly fierce & in Oct & Nov they're often noticeably blowing things around far out in the Greenland Sea.
The addition of .7C across the whole of the Arctic Ocean because of the release of the latent heat of fusion has to initiate a lot of evaporation since phase change is the only efficient way to remove the heat.
With this water vapor blanket in place things can stay warm far later in the season allowing far less time for the next season's FYI to thicken. - A vicious positive feedback that occurs whenever we don't have enough ice to continue the melt/freeze cycle uninterrupted.
Terry
Posted by: Twemoran | December 14, 2012 at 21:25
Radiation heat loss from Earth to space is a function of the temperature at the top of the atmosphere.
Raising the temperature at sea level from -30 to -1 C means that less water vapor condenses out, and therefore there is more latent heat in the atmosphere. If that water vapor were to condense, there would be a lot of heat to radiate (through a dry atmosphere) to the upper atmosphere, and then off into space. Water vapor in the atmosphere “blocks” radiation from the surface. Thus, when there is moisture in the atmosphere, the top of the atmosphere is colder, and less heat is radiated into space.
I think we also must expect warm spring and summer winds off of the continents, onto the relatively cold Arctic waters resulting in seasonal fogs. The winds will also facilitate overturn, so the seas can warm more deeply in the summer.
Ice clouds clearly raise the albedo. However, I am not convinced that fog has that much effect on the total heat budget of a local system. I think fog tends to convert visible light into latent heat in the atmosphere. While fog reflects more visible light back, less infrared from the surface is transmitted to the upper atmosphere through the fog, and the net heat budget changes by less than what is seen by the eye looking at the Blue Marble.
Open water in the Arctic is a powerful warming feedback. In the summer, it absorbs heat. In the winter, it allows a blanket of water vapor to keep heat near the surface.
Posted by: Aaron Lewis | December 14, 2012 at 22:02
Terry
Are you then suggesting that open water at -1degC with fog above it would radiate less than ice at -30degC?
Posted by: Ron Mignery | December 14, 2012 at 22:05
Gents,
this heated discussion is taking us around basic physics, which none of us apparently can grasp in full detail.
In an attempt to summarize, I would like to state the following:
Dense fog over open Arctic waters will "insulate" against further heat loss, thus preserving heat in the Arctic Ocean during the winter half year.
Preserving heat in the Arctic ocean is a positive feedback mechanism. Thus to answer your question Ron, yes -1 deg C water would preserve heat in the Arctic, whereas -30 deg C sea ice would lose heat to the atmosphere during winter.
In addition to this particular winter feedback mechanism, there would be plenty of other positive feedback mechanisms working troughout the remaining part of the year.
Positive feedbacks working together would lead to accelerating ice loss, as I was trying to demonstrate earlier.
In order to evaporate H2O from the surface, you would need to have a temperature gradient from the water to the atmosphere above. If you have fog above, you will typically have condensation and NO evaporation.
Posted by: P-maker | December 14, 2012 at 22:35
For those interested in current and future Greenland and global climate change, the IPCC AR5 draft report was leaked today.
Chapter 5 is on the cryosphere. Chapter 11-12 include short and long term climate change assumptions/conclusions.
Apologies for the source, but the skeptic, Alek Rawls, leaked it.
http://www.stopgreensuicide.com/
Posted by: Apocalypse4Real | December 15, 2012 at 01:29
Open Arctic water will become Atlantic water. The fresh water lens will collapse because of the lack of a protective cap of ice melting in late Summer. The Warmer Atlantic water rises with the widespread mixing and the result is a dense Winter fog caused by 6-10 degree surface water.
As soon as the Arctic is ice free most of the Summer it is ice free all year.
Posted by: Jim Williams | December 15, 2012 at 04:06
In western Washington we have fog and low clouds that last for a week or more sometimes if no storms happen to come along to clear the fog. During periods when solar radiation is not great enough to clear the fog out(November 10 to February 1) under a typical scenario the temperature falls and ground fog under clear skies. Over the next few days the fog thickens and and lifts slightly and temperatures warm close to the ground below the fog which is now just a low stratus cloud. Typically, if this lasts for a week the bottom of the clouds is now about 200 meters above ground level and drizzle and light rain may fall with temperatures that may vary less than 1 degree C. over the course of a day or several days. I suspect that in the Arctic with ice free conditions and no storms to clear fog and low clouds, conditions would develop similarly and very little heat would be lost to space under those conditions.
Posted by: VaughnA | December 15, 2012 at 05:20
Ron
A comment before answering.
We'd have to go back a long way to find the average winter ice temperatures at -30 C according to the North of 80 chart.
That said I do think that thick fog/heavy clouds would drop outgoing radiation at -1 C to lower rates than -30 C under clear skies.
Global temps weren't too much higher than where we stand when palm trees, alligators and turtles were flourishing through winter months with no solar radiation. The only way that I can imagine this being possible without invoking lots of geothermal activity (for which there is no evidence), is if fog/clouds really are capable of retaining the heat through the long Arctic night.
Last year we saw the result of heat retaining fog/clouds & it melted a lot of ice. P-maker makes the point that as humidity rises evaporation (and sublimation)slow down considerably & I have to agree with Jim's
"As soon as the Arctic is ice free most of the Summer it is ice free all year"
although I think the latent heat of fusion released when there's no ice to melt may play a bigger role than he envisions.
A4R
Reading the new IPCC Report is going to eat up lots of spare time. Thanks so much for crawling through the muck to find the 1 jewel.
Terry
BTW - I've had to go through the verification process on my last two posts - Type-pad must be in an overly zealous mode.
Posted by: Twemoran | December 15, 2012 at 06:50
Ron Mignery said :
That is a very good question, and unfortunately not so easy to determine.
Allow me to give it a shot.
Let me start with some basic physics, take on some known feedbacks, and see how far we get :
For starters, a loss of 1000 Gton ice every year suggests a annual deficit of 3.3 * 10^20 Joule in the energy balance in the Arctic.
Over 10 Million km^2 of Arctic ocean, that is something like 1 W/m^2 (365/24/7) power imbalance.
Incidentally, this is comparable to greenhouse gas forcing over the Arctic, although that may be a coincidence (or maybe not).
Either way, if the Arctic were a black-body, without any feedbacks, and without any additional heat input from lower latitudes, then, to radiate 1 W/m^2 extra away to space at the Top Of Atmosphere (TOA), the Arctic TOA would need to warm up about 0.25 C (Stefan Bolzmann).
Then there are known feedbacks :
First, due to GHGs and water-vapor in the Arctic atmosphere, the SURFACE needs to warm up MORE than 0.25 C to cause the same 1 W/m^2 extra radiation to space (think climate sensitivity factor). This is conservatively a factor of 2 or so, so the Arctic surface would need to warm up (365/24/7) at least +0.5 C above it's current temperature to eliminate the 1000 Gton/year annual ice volume loss.
Second, this +0.5 C warming would have to be ON TOP OF the already warming Arctic. Lower latitudes are bringing in more heat due to global warming, at a rate of something like 0.15 C/decade. This warming rate is doubled by Arctic amplification, resulting in the observed Arctic annualized warming trend of about +0.3 C/decade.
Incidentally, it seems that the +0.5 C necessary to stop the 1000 Gton ice loss/year is about equivalent to some 6 years of Arctic (+0.3/decade) warming. So, it is as if that ice volume loss trend is running 6 years behind schedule, which tells something about the time frame over which we may see this rate of ice volume decline change significantly (or not).
So don't expect Arctic sea ice volume decline to suddenly reverse within the next 6 years at least.
In summary, If we just look at basic physics and known feedbacks that would apply anywhere on the planet, then the Arctic would need to warm up at least 0.5 C above it's current trend to stop the 1000 Gton annual ice loss trend, and the timeframe over which this can happen is longer than 6 years.
Next, I'll try to incorporate actual Arctic-specific feedbacks, and you will see that the outlook for Arctic sea ice in summer will get A LOT worse...
Posted by: Rob Dekker | December 15, 2012 at 09:13
I estimated before that, in order to stop the 1000 Gton ice loss/year volume trend, the Arctic surface would need to warm up at least +0.5 C above current trend.
Here, I included a TOA-SURFACE temperature feedback factor 2, which would be a conservative estimate anywhere on the planet (even without ice).
What I did NOT include yet is an estimate of what the effect will be on ice volume for a +0.5 C warmer Arctic (necessary to increase TOA radiation to prevent 1000 Gton/year ice melt).
After all, a warmer Arctic will itself cause a reduction of ice volume.
And the key question is that IF a warming of +0.5 C causes close to, or even more than, 1000 Gton ice loss per year, then radiation losses cannot compensate the 1000 Gton ice loss we wanted to save in the first place. In other words, the Arctic will be unstable, and ice volume will simple collapse.
So how much ice loss does a +0.5 C warmer Arctic cause ?
Here, let us conservatively assume that this +0.5 C temperature increase is spread evenly over all seasons.
Then, how much less ice volume will grow in the Arctic winter ?
Well, average temperature in the Arctic winter is something like -20 C, so +0.5 C represents a 2.5 % increase in temperature.
Basic ice growth physics tell us that ice thickness (volume) will between 1.25 % and 2.5 % less at the end of the freezing season (depending on how much ice dynamics are present).
PIOMAS tells us that ice volume at the max (in spring) is about 22,000 Gton, so 1.25-2.5% means that 275-550 Gton less ice will be available at the start of the melting season.
Ough. That's already half of the 1000 Gton that would lead to an instable Arctic, meaning there is a 1.5-2X positive feedback factor right there from winter warming alone..
Now summer. That's when albedo feedback kicks in.
Suppose that ice is 1.25-2.5% thinner at the start of the melting season. This suggests that more ice area will melt out. But since insolation varies quickly over the melting season, the amount of albedo-induced heat absorption depends heavily on WHEN the ice margin is WHEN (at which latitude). So, summer effect is very difficult to calculate, but here is a first-order ballpark estimate of the influence of summer :
If thickness reduced equally by 1.25%-2.5%, then we can expect AT LEAST an ice area reducion of 1.25-2.5%.
During the melting season the Arctic ocean surface receives about 2.5 GJ/m^2 insolation. With an albedo change of 0.5 between ice and open water suggests that over the melting season on average 1.25 GJ/m^2 extra energy is absorbed into open water that used to be ice. If the 10Mkm^2 ice area reduced by 1.25-2.5 %, we may thus expect another this causes an additional extra heat absorption of 1.5-3 *10^20 J, which represents an additional summer melt of 470-940 Gton.
Combined with the winter melt, the +0.5 C that we needed to increase radiation losses to space to prevent 1000 Gton ice loss, may thus by itself cause a total 745 - 1490 Gton ice loss.
These back-of-the-envelope calculations suggest that the Arctic may be close to (or already past) the point where it is not possible to reverse the PIOMAS trend of 1000 Gton ice loss/year any more.
IOW : The Arctic may indeed be in an unstoppable death spiral at this point.
Posted by: Rob Dekker | December 15, 2012 at 10:53
Rob, I really envy you for your capacity of making those calculations. I wouldn't know where to start.
Posted by: Neven | December 15, 2012 at 12:22
Wow, those are scary calculations. Even your minimum of 745 Gtons would mean multiplying your 0.5C by 4 to give a 2C rise which would have serious effects on Greenland if nothing else.
I am impressed by the calculations. I also wonder if they need constructive criticism.
I also think your 2 for feedback factors is very conservative and 2.5 would still be conservative. Multiply Stefan Bolzmann 0.25C by 2.5 gives 0.625C.
So far most of the warming has been in winter so if we continue that, we need a 1.25C warming in winter. If average is -20 degrees below freezing (-21.5C looks more like it off that 80N graph though perhaps that doesn't last a full 6 months) then we have lost 6.25% of the amount below zero.
Rob then states this would give rise to 687 to 1375 Gtons less ice at maximum depending on amount of ice dynamics.
I am afraid I am getting a little lost at this point:
Ignoring effects of leads. At maximum this may allow 6.25%? more heat to flow out through the 6.25% less insulating ice to keep the temperature warmer by the amount needed. But that is at maximum. Earlier in the freezing season there is going to be no ice instead of some ice for quite a while and even after ice starts to form in the new regime it will be a lot less than 6.25% down on previous levels. So that 6.25% less ice looks like it could have quite a bit more effect than the calculations Rob has presented suggest.
I could be completely misunderstanding.
Posted by: crandles | December 15, 2012 at 13:43
Rob,
Not sure if I agree, although after a long day (coincidentally) calibrating industrial ovens on site I'm not very sharp.
Firstly the issue of using Stefan Boltzman to calculate the emission. Due to the T^4 term SB is highly non linear, so the difference is dependent on the initial temperature from which you increase by o.5degC.
SB is J = S * T^4.
Where S is SB's constant.
If I work out J for 5degC steps in temperature, then work out J for each step with a 0.5K increase I get the following differences in J (j/m/S).
Temp (K) , Difference (over 0.5degC or K)
243 , -1.632195064
248 , -1.734927242
253 , -1.841882153
258 , -1.953144847
263 , -2.068800373
268 , -2.188933781
273 , -2.313630122
278 , -2.442974446
Freezing (0degC) is about 273K, I don't do Fahrenheit. ;)
Again, I may be tired and losing the plot, but I don't follow this:
"...it seems that the +0.5 C necessary to stop the 1000 Gton ice loss/year is about equivalent to some 6 years of Arctic (+0.3/decade) warming."
0.3degC/decade -> 0.03degC per year, 0.5degC/0.03 = 16.667 years.
Or have I got the wrong end of the stick?
PS -
In models the Open Water Formation Efficiency is highly non linear, for a given loss of thickness, the amount of open water produced at the end of the melt season increases non linearly as the ice thins. See figure 2 of Holland et al, "Future abrupt reductions in the summer Arctic sea ice." PDF.
Posted by: Chris Reynolds | December 15, 2012 at 18:55
Rob
Thank you for understanding and answering my question with numbers and dates even!
Interesting that GHG effects in the Arctic could account for all the decline without no increase in convection from lower latitudes required. Am I interpreting that correctly?
So temperatures should rise on average by 0.5degC ignoring positive feedbacks. This will decrease the difference with lower latitudes and increase the difference with the GIS. I think I've heard that winds in the Arctic more-or-less cycle around the pole. What fraction of polar air then passes over the GIS in a summer? Would the GIS extract all the increased heat with increased melting or will some be left over to actually raise polar temps?
If polar temps do rise, what will be the quantitative effects on lower latitude temps?
Just asking...
Posted by: Ron Mignery | December 15, 2012 at 19:13
Jim Willams suggested that another consequence of the loss of summer sea ice will be extinguishing of the Arctic halocline that currently protects surface waters from much warmer deeper waters (6-10 degrees! really?) Would the Arctic then warm by 6-10degC? If so that certainly dwarfs a 0.5degC rise.
Posted by: Ron Mignery | December 15, 2012 at 19:40
Thanks guys, for constructive feedback on my answer to Ron, on how much the Arctic needs to warm up to compensate for 1000 Gton annual ice loss.
First some answers to your comments and questions, and then some notes at the end.
crandles said :
I'm sorry crandles, I should have been more clear on this step.
Ice growth (thickness) in winter theoretically (if ice sits still, and there is no snow, nor ocean heat flux) follows a SQRT function of temperature-below-freezing * time. Inb that ideal case, if temperature below freezing is 2.5 % less, then we can expect ice thickness to be about 1.25 % less at the end of the freezing season.
But if ice moves around a lot in winter, then most freezing (heat transfer from ocean to atmosphere) may happen through polynias rather than through the ice. Also, if snow cover is the determining insulation factor, then that SQRT ice growth function becomes more linear with winter temperature changes. Actually, if there is some (change in) ocean heat flux, due to influx of warmer (Atlantic or Pacific) water, then ice growth can be severely inhibited, and may even show bottom melt while atmospheric temperatures are well below freezing.
So it may be (much) worse than I estimated, but the 687 Gton loss (1.25 % of max volume) does seem to be a lower bound for stopping the 1000 Gton/year volume trend.
Chris Reynolds said
You are absolutely right, Chris. Thanks for correcting my mistake.
The 16 year timeframe is kind of scary though. Seems to suggest that it will take a long time (or a completely different Arctic) before the energy deficit (resulting in an annual 1000 Gton/year ice loss) in the Arctic would slow down.
Regarding Stefan Boltzmann, and the table you provided ; I assumed that TOA effective radiative temperature over the Arctic is 255 K (as it is on average over the planet). At that point, we can assume about 3.7 W/m^2/K response. In other words, a 1 C increase at TOA will radiate about 3.7 W extra to space. And as I think your table shows, a 0.5 C increase will radiate about 1.9 W/m^2 to space. And the 0.25 C increase at TOA that I deduced will thus radiate the 1 W/m^2 power needed to avoid the 1000 Gton/year ice loss. That all sounds about right, no ?
Ron Mignery said :
Yes and no. Sorry for being so vague, but I still think that the 1 W/m^2 energy deficit that matches both the 1000 Gton/year ice loss and at the same time matches Anthropogenic GHG forcing over the Arctic, may simply be a coincidence.
After all, that 1000 Gton/year loss did not come out of the blue. It is the result of an increased amplitude of the diurnal cycle of volume loss/gain. Compared to the 80's, a whopping 3000 Gton extra ice freezes each year now during the winter, but since the summer knocks out an extra 4000 Gton, the annual loss amounts to the 1000 Gton/year that you started your question with. So it is very clear that negative feedbacks (increase ice growth in fall/winter due to thinner ice) are fighting positive feedbacks (albedo effect in summer) and it is clear that the positive feedbacks are winning, and actually gaining ground.
What this means for
I recall a study that suggests that if the Arctic summer is essentially ice free, that this will result in a 10 C increase in winter temperatures, even without any influence from the halocline. Please allow me until Monday to dig up the reference.
So it may not even be necessary to demolish the halocline for such a change in Arctic temperatures, but Jim is certainly correct that below the halocline there is a vast amount of heat stored, and if that heat gets released to the surface it will be game over for Arctic ocean sea ice.
Posted by: Rob Dekker | December 16, 2012 at 10:21
Rob,
I'm still not 100% convinced, I need to think about it. My firm shuts down over Christmas so I might try to do a large spreadsheet then.
One problem I didn't post on is with regards the assumption of 255k as global average. If we assume 10degC/km lapse rate, 500mb in summer is around 5.7km, in winter about 5km (if I recall correctly). There's still a lot of weather at 500mb, and lapse rate makes it colder than 255K. So I _suspect_ the effective emission temperature in the Arctic is lower than 255K, which would make the temperature increase needed to radiate 1W more a bit larger. It's tricky trying to work out an efficient way to cut through the problem.
Sorry if I always seem to be picking fault, I am a pedantic old git. To be fair I do it with my own work, the current ratio of blogged to investigated is about 1/5. Maybe I should blog on my dead end work.
PS 1
I have serious doubts about the abyssal water being 6 to 10degC warmer than the surface. It seems to me that if it were it would be of low enough density (despite higher salinity) to overturn.
PS 2
With regards the 1 watt and anthro forcing. Current net anthro forcing is around 1.6W/m^2. Francis & Hunter 2007, PDF, indicates decadal trend in downwelling infra-red of between + 5 and 9 W/m^2 per decade over Arctic seas. Against this the CO2 forcing will be lost. These increases are mainly due to clouds and water vapour.
To quote from their conclusion:
The role for CO2 forcing is hidden amongst these changes, it's only in model studies that its role can be extricated. As with its role in global changes and the global atmosphere; remove CO2 and the planet freezes as water precipitates out with cooling. CO2 is the framework upon which the whole greenhouse effect hangs.
Posted by: Chris Reynolds | December 16, 2012 at 11:09
Temperature & Salinity profiles of Arctic Ocean.
Posted by: Chris Reynolds | December 16, 2012 at 11:25
Terry, I'll agree that tropical atmosphere will have its role to play in Arctic warming, but the analyses were completely ignoring the effect of ocean currents.
Chris, Historical subsurface temperatures are certainly a bit less than 6 degrees, but recent incursions have been 6 degrees and greater. (4 degrees is already quite enough...) What do you think is going on north of Europe anyway? I think the thing to watch over the next few years is salinity in the North Atlantic far more than temperature in the Extreme Arctic.
Also, I think it's the Gulf Stream which will provide the real shocker when it comes to the GIS -- but so far that is just a hunch.
Posted by: Jim Williams | December 16, 2012 at 19:39
Jim,
Graph of density/temperature/salinity here.
Mercator Salinity here.
Take typical Arctic/Atlantic boundary region salinity of about 33psu. The graph is in per mil, but around 32 I think the two scales are in reasonable agreement. From Mercator the lowest salinity in bulk of Arctic is about 32 psu.
Let's say the intruding Atlantic water is about 6 degC, as you say. And the Arctic about 0 degC, to lessen the density difference in your favour.
Using the graph both bodies of water will have near equal density. So why is the Atlantic water hiding beneath the Arctic? Historical temperatures are a lot less than 6degC. Do you have a reference for this 6 degC water?
North of Europe?
Do you mean the persistent warm tongue (atmosphere) over Barents? Otherwise not sure what you mean.
I think the Gulf Stream is overstated. Stopping ocean overturning around Greenland won't affect UK's climate- it's not a pipe, the spin of the Earth draws wind & water over the Atlantic, overturning circulation isn't important in that respect. I agree with Carl Wunsch - when the THC stopped in the past it was winds that changed stopping it, those same wind shifts drove coincident changes in European climate- not the THC itself.
It's worth noting that the summer high pressure pattern over Greenland should act against Northwards flow.
Posted by: Chris Reynolds | December 16, 2012 at 20:19
Since plain water reaches maximum density at 4degC and does not get lighter than water at 0degC until it reaches 8degC http://en.wikipedia.org/wiki/Properties_of_water, won't there always be a layer of cold water above the warmer waters below as long as Arctic temps stay below 4degC? How thick would that layer be? Would it be thicker than the Arctic Mixed layer http://en.wikipedia.org/wiki/Mixed_layer?
Posted by: Ron Mignery | December 16, 2012 at 21:42
Here is the University of East Anglia press release link to the new research by Renfrew that impacts this conversation:
'Missing' polar weather systems could impact climate predictions"
The focus is on polar hurricanes and their impact on heat transfer in the Arctic.
Renfrew comments: ""We have shown that adding polar storms into computer-generated models of the ocean results in significant changes in ocean circulation - including an increase in heat travelling north in the Atlantic Ocean and more overturning in the Sub-polar seas."
Condron adds: " "By simulating polar lows, we find that the area of the ocean that becomes denser and sinks each year increases and causes the amount of heat being transported towards Europe to intensify.
"The fact that climate models are not simulating these storms is a real problem because these models will incorrectly predict how much heat is being moved northward towards the poles. This will make it very difficult to reliably predict how the climate of Europe and North America will change in the near-future."
http://www.uea.ac.uk/mac/comm/media/press/2012/December/polar-storms-climate-change
Posted by: Apocalypse4Real | December 17, 2012 at 02:11
I'm not impressed with the overturning scenario either, mostly because exactly the opposite seems to be happening. What I find interesting about the Gulf Stream is the general poleward trend of the Western Boundary Currents as temperatures rise. Poleward for the Gulf Stream could become West of Greenland, and there have already been a few brief instances where the Stream has connected directly to the Labrador Current. Notice that right now the North Wall is unusually far North and the Stream is sort of pooling just south of Greenland. There's no deepwater channel to the west, but in many ways the Gulf Stream can be treated as a surface current.
If the Gulf Stream were to switch West of Greenland I guess it would have an effect in Europe, but nothing compared to the effect upon Greenland.
Posted by: Jim Williams | December 17, 2012 at 13:57
Rob
You observed that the Arctic summer now melts about 4000 Gton more ice than it did in the 80's and that Arctic GHG effects could account for about 1000 Gton of it. Is it then correct to conclude that an additional 3000 Gton of melt is now imported to the Arctic yearly from lower latitudes compared to the 80's?
Posted by: Ron Mignery | December 17, 2012 at 15:45
Jim Williams wrote,
In order to do so the Earth should stop turning and then return in the opposite direction.
A performance which looks rather out of scope for matka Earth.
Posted by: Kris | December 17, 2012 at 17:31
Ron,
Arctic temperatures have risen more than lower latitudes so there is less of a gradient which would tend to imply lower heat import into Arctic. However if the jet streams are meandering more there might be more heat being pumped into the arctic.
I would suggest that a sizable proportion of the 4000 Gton extra melt is due to albedo effect causing more solar energy to be absorbed. So I doubt extra heat import from lower latitudes accounts for an extra 3000 Gtons of melt.
Posted by: crandles | December 17, 2012 at 17:57
There is an article from Science from 2011 on the temperature of the waters entering the Arctic Ocean via the Fram Strait. The temperature has increased from a long term average of 3.4C to an average of 5.2C measured since 1890.
The article argues that even without increased mixing of the water layers, the heat flux will increase by 40%, which in time will impact the temperature of the upper water layers - and the ice.
Article:
Enhanced Modern Heat Transfer to the Arctic by Warm Atlantic Water,
28 JANUARY 2011 VOL 331 SCIENCE, 450-453
Abstract:
Instrumental air and AW (Atlantic Water) temperatures in the Arctic during the 20th century and beyond display quasi-synchronous multidecadal oscillations that make isolation of the industrial warming trend difficult (3, 21). Basinwide observations since the 1980s detected multiyear events of AW spreading in the Arctic Ocean that featured both a strong warming and an increased inflow to the Arctic (7, 27, 28). Although we cannot quantify from our data the variability of previous AW inflow to the Arctic by volume, our temperature data series and the above observational link suggest that the modern warm AW inflow (averaged over two to three decades) is anomalous and unique in the past 2000 years and not just the latest in a series of natural multidecadal oscillations. Both effects—a temperature rise as well as a volume transport increase—introduce a larger heat input into the Arctic Ocean. Although there is no direct contact of the AAWL with the ocean surface in the Arctic, such an increased heat input has far-reaching consequences. The strong AW warming event in the Arctic Ocean in the 1990s caused a shoaling of the AW core and an enhanced heat flux to the surface (29), concurrent with decreasing sea ice (4). Recent oceanographic data from the Laptev Sea continental margin indicate the impact of warm AW-related water masses on the shallow (<50 m) shelf (30), a feature not observed before in a >80- year time series. The data also provide evidence for a significant heat flux to the overlying shelf waters (30). Even without any modification of the vertical heat transfer processes, the enhanced temperature contrast between the AW and the surface sea water freezing point (increased from ~5 to 7 K as identified here) leads to an increase in the vertical heat flux of ~40%. Any positive feedback mechanism will magnify the effect of this flux increase on the ice cover. Complementing the strong feedback between ice and atmospheric temperatures (1), warming of the AW layer, unprecedented in the past 2000 years, is most likely another key element in the transition toward a future ice-free Arctic Ocean.
It would seem that the Arctic ice is fighting both the warming from above plus very unfavorable longterm fluctuations in ocean currents.
Posted by: John Christensen | December 17, 2012 at 18:03
I am probably writing rubbish again. Water temperature differences may well be rising rather than falling and I am not sure if most of the heat transfer is ocean rather than atmosphere, there being a lot more heat capacity in water but water flows are slow compared to atmosphere.
Posted by: crandles | December 17, 2012 at 18:03
John Christensen,
"Enhanced Modern Heat Transfer to the Arctic by Warm Atlantic Water,"
Paywall free copy.
http://instaar.colorado.edu/~marchitt/reprints/spielhagenscience11.pdf
I did post a lengthy post on the other thread where I explained how Atlantic Water is a sideshow. The main drivers are ice atmospheric changes and ice dynamics - but even though the comment was saved and displayed the system has lost it.
Posted by: Chris Reynolds | December 17, 2012 at 20:34
Crandles re albedo
You may very well be right about the importance of albedo in the 4000 Gton increase in annual Arctic melting seen since the 80's if my calculations are anywhere near correct.
The paper cited by Lou Grinzo two years ago http://www.atmos-chem-phys.net/10/777/2010/acp-10-777-2010.pdf modeled the consequence of total Arctic sea ice loss from the average of 1983 to 2007 values. They reported a 19.7 Wm-2 gain for typical sky overcasts, 10.2 Wm-2 gain for full overcast, and 34.9 Wm-2 for clear skies.
Rob calculated an equivalence of a 1 Wm-2 gain over the entire Arctic with 1000 Gtons of ice melt. Eyeballing the sea ice extent graphs suggests to me about a 20% decline in summer solstice sea ice extent. By my very crude calculus, this represents 20% of 19.7 Wm-2 or about +4 Wm-2, enough to melt 4000 Gtons of ice. Increased albedo thus seems sufficient to account for all of the increased summer melt seen since the 80's, never mind the 1000 Gton increase from GHG mentioned by Rob or the massive influx of warmer Atlantic waters mentioned by John.
If current trends have 16 years of momentum as Rob suggests then all Arctic ice that ever sees the sun will soon be gone. Wipneus's extrapolations show June ice free by 2021. When that happens, the Arctic will see between 10,000 and 40,000 Gtons of ice melting potential in search of a target from albedo increase alone.
Please tell me I'm crazy.
Posted by: Ron Mignery | December 17, 2012 at 21:15
Ron,
It's not just the change in albedo between open ocean and sea ice.
Here's a graphic from Perovitch & Poshenski:
http://farm9.staticflickr.com/8452/8030738470_b90a73494c.jpg
It shows sea ice albedo changes during the season for First Year and Multi Year sea ice. To quote from a post on my blog: Over a square meter covered by MYI they find that by September cumulative solar heat input amounts to around 900MJ/M^2, however for a square meter of FYI the gain over the same period, with the same solar input is almost 1200MJ/M^2. That's almost 1/3 more energy gain.
I think this should be paywall free:
Perovich & Polashenski, 2012, "Albedo evolution of seasonal Arctic sea ice."
http://www.agu.org/journals/gl/gl1208/2012GL051432/2012GL051432.pdf
Posted by: Chris Reynolds | December 17, 2012 at 22:02
I meant decreased albedo...
Posted by: Ron Mignery | December 17, 2012 at 22:51
Can we attribute a Gtons of extra freeze to less insulating ice to the reduction in maximum (or average?) winter ice volume fall from circa 32000 km^3 to circa 22000 km^3?
Presumably the effect of a fall from 22000 km^3 to 17000 km^3 is more than the effect of the fall from 27000 to 22000. In what manner does this effect grow?
What is the pattern of the albedo effect growth? Does it continue to outpace the extra freeze effect?
Is it possible to estimate when the earliest and latest the extra freeze effect could catch the albedo effect (that at some stage presumably starts leveling off after uncovering most water for most of the summer)?
Or is it extra radiation from higher temperatures that stops albedo effect running away with the process with extra freeze from less insulation being a minor player?
Posted by: crandles | December 17, 2012 at 23:39
So here is my worst-case nightmare take on the near future based on the responses here to my questions: By 2021 summer sea ice has virtually disappeared. Albedo decrease results in 20,000 Gtons per year of additional melt on the only ice left, principally the GIS. The heat that used to melt the sea ice is also available to melt another 17,000 Gtons per year. Any of this heat diverted to warming the Arctic generally will result in more positive feedbacks from permafrost thawing and destruction of the ocean halocline/thermocline resulting in even more GIS melting. With the GIS approaching a 2% loss of its mass per year, 7 meters of sea level rise will occur in 50 years or 1 meter in 7 years. By 2030 massive coastal flooding will be commonplace and every location unable to withstand a 7 meter rise will be a financial write-off. My house at 10 meters will be ok but my friend's house at 3 meters will not. Maybe now is the time to sell...
Posted by: Ron Mignery | December 18, 2012 at 03:03
@ Ron: There will probably be many things that will make the rise far slower then what you are saying, there are 2 articles I came across that do back up what you are saying to a point.
http://e360.yale.edu/content/feature.msp?id=2115
http://e360.yale.edu/feature/how_high_will_seas_rise_get_ready_for_seven_feet/2230/
In both cases the wording was carefully put in terms of conservative. You can bet your bottom dollar that they're best guesses are far far worse. They are also basically admitting that what is going on in the ice fields are not holding to traditionally to behavior because there are forces that are exerting changes on those glaciers that were not witnessed pre 1950's. The problem that scientists are having now is that the changes are so rapid it is almost impossible to discover what new forces and by how much are bring about the recent changes.
Also from what I can infer is that most of the concentration has been looking at heat and possible melt from the ocean waters. Has any one been thinking about how much faster the glaciers will disintegrate into oceans as you start adding SLR to the picture.
Let us say that you add 1 metre by say 2050. That would mean that ice sheets that are currently grounded could then become floating sheets and as in the case of Greenland the ice that is attached to the sides of fjords could then be broken off. On top of the the ocean then enters far farther upstream. All those stresses means that in both the of both GIS and AWIS you would not have to melt the ice at the source at all, just get it into the ocean currents and let them do the work somewhere else.
As for the land based ice, it is then pushed into the vacated areas that are then kept vacated because the ocean carries it away as soon as it arrives, raising SL that much faster.
The point I am trying to make is that all you need to do is get enough SLR from heat to break up the front wall of the ice before other forces such as pressure and kinetic energy take over. That is what I see different from land locked glaciers and sea going glaciers. Mechanical energy becomes a far more important factor once you get beyond a critical point. We saw that in the Arctic once the wall in Frams Straight broke up. It changed from a stopping point to an expressway for ice.
Do not get me wrong. Heat is still very important, but we do tend to ignore the importance of mechanical energy and its impact on ice.
Posted by: LRC | December 18, 2012 at 05:15
Ron Mignery said :
It seems that you are mixing up volume loss due to an uncompensated forcing (1 W at TOA) versus an annual RATE of ice loss (1000 Gton/year). The two are not a priori related at all. In fact, much depends on the summer and winter 'feedback' processes, which on an interannual basis also affect each other and create 'delayed' responses. And then there is subsurface ocean heat flux, which has effects on this process on decadal timeframes.
Thus you cannot so easily directly attribute 1000 Gton/year to 1 W/m^2 GHGs. At all.
To give an idea of what the RATE change of GHG effect over the Arctic since the 80s, consider that CO2 increased 2ppm/year since the 80's. Considering that 280 ppm constitutes a 'doubling' of CO2 since pre-industrial times, which should create 3.7 W/m^2 forcing, the 2 ppm/year probably increased TOA forcing over the Arctic by something like 3.7*2/280= 26mW/m^2/year.
Using the calculation in previous post (1 W/m^2 causes 1000 Gton ice melt) if feedbacks are fast in the Arctic, then the direct GHG effect of 26mW/m^2/year could not melt out more than about 26 Gton/year, but that is about as much as we can attribute to direct GHG effect on the rate of volume loss.
The rest will be lower-latitude heat influx and Arctic feedbacks. Which apparently make up the majority of the ice volume losses.
Posted by: Rob Dekker | December 18, 2012 at 09:52
Ron Mignery,
After spending way to much time debunking self-proclaimed skeptics claims of Arctic sea ice recovery, and batling strawmen arguments like cloud cover (such as Dr. Curry on the climatedialogue.org site) your proposition of "worst case scenarios" feels 'refrehing' in a way. It made me realize that we should have been talking about worst-case scenarios, and what happens after summer Arctic sea ice is gone, in the first place.
Still, I think there is little evidence that your worst-case scenario would materialize.
For starters, remember that if there is no ice in the Arctic in summer left over to melt, then the excess heat will FIRST be absorbed by the Arctic ocean water itself, BEFORE the atmosphere will have enough heat to melt GIS in any substantial way.
How much will the Arctic warm up in summer if there is no ice ?
Let's do a ballpark estimate for the scenario where Wipneus' extrapolations to an ice free June by 2021 materialize (and they may, if the feedback balance does not change significantly). In that case half of the Arctic summer insolation will then be absorbed by water instead of ice. Accumulated Arctic summer insolation 'at-the-surface' (through cloud cover and all) is something like 2.5 GJ/m^2 (80 W/m^2 annualized), which is 1.25 GJ after June. Since a switch of albedo from old snow/ice to water is about 0.5, we can expect some 630 MJ extra heat be absorbed after June. This will warm the 'mixing' layer (upper 20 meters) of the Arctic ocean, which could thus warm up to about (630[MJ/m^2]/4125[kJ/meter]/20[meter] = 7.6 C.
That is significant, but remember that none of that heat melts the GIS.
Also remember that in fall/winter, the Arctic will give up all (or at least most of) that 630 MJ/m^2 heat again to space (simply by staying at 0 C while previously it would be deep below freezing), and during that time GIS will be well below freezing already.
So I think that there is no scientific evidence for your worst-case scenario.
Posted by: Rob Dekker | December 18, 2012 at 10:33
On the other hand, Arctic ocean water peaking at an average 7.6 C in summer (and presumaby much higher near the coast) could be considered absolutely insane and worse than anyone even dared to predict.
I wonder why we did not see more of such 'worst-case' scenarios worked out in scientific literature, especially since this is where we are heading if the feedback in-balance does not change.
Posted by: Rob Dekker | December 18, 2012 at 10:49
Rob
Thank you for your reassurances. I appreciate how simplistic it is to imagine all the excess heat attacking the GIS as the coldest remaining target in the neighborhood. Firstly it may not be in the neighborhood at all. If a summer high sits on Greenland, would that isolate the GIS from the circling Arctic winds? Also if those now warmer, wetter winds could climb up 2km to the GIS would they not cool adiabatically below freezing and drop their moisture as snow and tend to lower sea levels? These are not rhetorical questions; I have no idea and would appreciate expert responses.
I also understand that glaciers and ice sheets are not the same thing and that disappearance of all the glaciers of Greenland would have little affect on sea levels. Also the GIS would have to thin 10 meters per year to lose 2% of its mass per year and that might be improbable fluid dynamically.
If the Arctic ocean water does warm to 7.6degC as you suggest, would that not destroy the clines and allow the warm waters below to enter the mixed layer. If so, would not the surface waters stay warmer than 0degC for the winter until all that legacy heat is vented? Would that dump more snow on the surrounding land areas possibly including the GIS?
Posted by: Ron Mignery | December 18, 2012 at 16:00
I meant 40 meters, not 10.
Posted by: Ron Mignery | December 18, 2012 at 16:39
Ron & Rob
Before you get carried away on the wings of pure speculation, it may be timely to consider one of the major potential feedbacks from all that melt you suggest.
In the most recent version of the SWIPA report, which was published here recently:
http://amap.no/swipa/SWIPAOverviewReport.pdf
it is stated: “All the main sources of freshwater entering the Arctic Ocean are increasing – river discharge, rain/snow, and melting glaciers, ice caps, and the Greenland Ice Sheet. Recent calculations estimate that an extra 7700 km3 of freshwater – equivalent to one metre of water over the entire land surface of Australia – has been added to the Arctic Ocean in recent years.”
We all know that neither Australia nor the Arctic is flat as pancake. We also know that all that freshwater will most likely stay above the halocline under most circumstances mixing into a depth of about 20 meters one way or the other.
To give you just one example of added freshwater, I will describe the outflow this year from the Sermeg Kujalleg ice-stream east of Ilulissat (Jakobshavn Isbræ in the old days). During the months of September-December this year, a steady river of icy meltwater has “left the building” (as seen on http://sermitsiaq.ag/icecam/ ). This constant flow of 0 deg C meltwater has spread out over the Disko Bay according to daily SST-maps from DMI. Roughly estimated, we are talking about 50 x 100 x 0.02 km = 100 cubic km in one “meltwater pulse” (best seen as a cold SST anomaly in the eastern half of the Disko Bay since the end of August at this site: http://ocean.dmi.dk/arctic/disko.uk.php ). As long as this meltwater keeps entering the Disko Bay, we should not expect to see sea-ice cover in this area. But, as soon as it stops, we may expect to see a rare freeze over (rare in this century at least). It will also require a cold atmosphere, which is able to cool about 20 meters of freshwater above the halocline through radiative cooling.
If you for a moment try to imagine what 20,000 – 40,000 cubic km of freshwater would mean wrt. refreezing of sea ice, you would also need to consider geography. If late season river discharge, snow melt and sea ice melt no longer contributes (as may be the case in the Barents Sea in recent years), sea ice may be of only historical interest. On the contrary, if you have a constant discharge of fresh water in the start of the freezing season, you may expect to see occasional sea ice form in the future near to these “outlets”. Thus, discussing the radiation balance over the Arctic as a whole, may be of only academic interest. You need to look at it as a complex 3D puzzle, which is very difficult to model in extenso.
It is also evident that discussing sea ice and glacier ice separately on this blog will become a more and more futile exercise, as we move towards a situation, where the GrIS is an increasingly important source of fresh meltwater.
Posted by: P-maker | December 18, 2012 at 20:59
P-maker,
I find this an intriguing vision. It parallels my view that we are confronted with a maze-like development in space and time. Very complex. Not easy to fit into simple equations.
The question is 'how do we cope with consequences' while they cannot be specified.
Posted by: Werther | December 18, 2012 at 22:59
P-maker wrote:
Actually, Disko Bay always has been "frozen over" in Winter till the winter 2007/2008.
The bay remained ice cover free in the Winter 2008/2009, 2009/2010 and 2010/2011. Last winter, from February 2012 on the bay had been frozen solid again.
Bottom line, till further notice an open Disko Bay in the entire Winter period is rather an exception, thus a rarity. :-)
Posted by: Kris | December 19, 2012 at 01:55
Rob Dekker: a late reaction, I am trying to follow your calculations, but loosing you quite early.
When you say:
You say that you ignore heat input from lower latitudes (probably atmospheric). How can that be?
As OLR I take 150 W/m^2 ( ERBE's measurements in Jan 1986, from a graph in Pierrehumerts Planetary Climat's book).
Heat input from below I have to estimate, growing 1 meter of ice in 90 days gives 43 W/m^2, heat conduction of 1 meter pure ice
@20 degC temperature difference is 46 W/m^2.
So either way the atmospheric heat input from lower latitudes seems to be in the order of 100 W/m^2, larger than any other factor. Possible changes in this heat transport will need to be considered when calculating the effects of surface heating.
With the 150 W/m^2 you can calculate a equivalent black-body temperature of 226.8 K. 151 W/m2 gives a delta T of 0.38 degC.
Is this the TOA temperature that you are talking about?
Now this TOA temperature corresponds is somewhere in the higher troposphere. If the height is known, the surface temperature can be worked back normally assuming an adiabatic lapse rate of something 7-10 degrees/km, probably higher. The exception is in arctic winter, where the temperature in the lowest km's of the atmosphere rises, an inversion. That inversions causes that the temperatures close to the surface and higher in the atmosphere are decoupled: change in one will just change the height of the inversion layer.
Another effect from the inversion, is that the heat transport becomes purely radiative, no convection. That makes the heat transport even more sensitive to the presence of low clouds, clouds block radiation but not the convective transport that normally dominates.
I summary these are my doubts:
1. You cannot make an arctic heat balance without considering the heat transport from lower lats;
2. As long as there is an winter temperature inversion, you will have to forget surface temperatures derived from the radiative-convective normal. For instance clouds will play a larger role.
Posted by: Wipneus | December 19, 2012 at 09:55
Kris,
it all depends on how you define "the new normal" ;-).
Three out of the last four years certainly were totally ice free, but I actually based my assumption on news received about a decade ago, that the greenlanders had to cancel their national championships in dog sledge racing on the Disko Bay due to unstable ice conditions. Please remember that it is also a question of the ice quality for the locals...
Posted by: P-maker | December 19, 2012 at 10:32
Hi Wipneus,
Thanks for your constructive response. This is very helpful, and made me realize that I was not very clear in many of the steps I made in my post.
Sorry for being unclear. In all my calculations, I used deltas over the 'current' situation.
After all, the goal was to see how much (delta) the Arctic needs to warm up to eliminate the (delta) 1000 Gton/year ice loss.
For simplicity, I assumed that the heat input from lower latitudes remains unchanged.
Yes indeed. Chris Reynolds already suggested that I overestimated TOA temperature over the Arctic, and you quantify that very nicely. Thanks ! So 0.38 C delta T at TOA it is (for radiating 1 W/m^2 more).
You were trying to determine the surface temperature here from the TOA temperature, and yes, that is difficult and involves thinking about inversion layers, and water vapor in winter and clouds etc. But if you think in deltas over the current situation, it's much easier. For starters, you already KNOW the surface temperature of the current situation. It's the winter surface temperature (from NCEP/NCAR or so).
With deltas (derivatives), the only thing we need to estimate is how much will the surface have to warm up to cause an increase of 0.38 C at TOA, which should be a simple factor and not a complex equation, since with delta calculations the atmospheric situation (clouds, water vapor etc) remains the same. I used a standard multiplier 2, hinting at climate sensitivity factor for a standard atmosphere, simply out of laziness. With TOA delta-T 0.38 C, that would mean 0.76 C surface temp increase is necessary to radiate 1 W/m^2 more to space.
Now where it gets interesting is what happens next. You mention :
If temperature goes up 0.76 C this obviously leads to less ice growth, which leads to less ice volume in spring, and that's where the rest of my calculations go, which finally conclude that maybe the Arctic is currently very close to being 'unstable' in the sense that it cannot compensate for the rate of 1000 Gton/yr that PIOMAS is suggesting.
What do you think ?
Posted by: Rob Dekker | December 19, 2012 at 11:15
Wipneus,
"So either way the atmospheric heat input from lower latitudes seems to be in the order of 100 W/m^2, larger than any other factor."
That's the figure found in the literature, for example Smedsrud, 2008, "Recent and future changes of the Arctic sea-ice cover" Who find around 100W/m^2 from NCEP/NCAR. Rising from between 90 and 95 in 1950, to a peak of around 105 around 1980, then falling to 2006's 100W/m^2.
Lapse rate in the winter Arctic atmosphere will be about 10degC/km, dry adiabat. In summer slightly lower due to increased atmospheric humidity.
Posted by: Chris Reynolds | December 19, 2012 at 18:37
This is not a correct description. In standard atmosphere the delta-T's are similar. The main effect of feedbacks is to raise/lower the temperatures simultaneously. Here is a sketch of a earth-normal situation:
Two temperature-height profiles for the atmosphere are drawn. These lines have a slope, called lapse rate, that follows from the physical gas constants and humidity. In the no-feedback case the lapse rate is assumed constant and delta-T is the same at all heights. Humidity feedback will lower the lapse rate (in degC/km), causing the surface temperature change to be less than the high atmosphere delta-T.
This is a negative feedback, the same change in humidity will cause another effect: an increase at the height at which the outgoing radiation is "radiated". Since higher, thus colder, means less energy is radiated implying a positive feedback.
The constant lapse rate is maintained when the heat rates change: try to increase the temperature from below will make the atmosphere unstable, increase the convective heat flux until lapse rate is restored. Try to heat the upper atmosphere will slow the convective heat transfer until the lower atmosphere is heated enough to restore the lapse rate. This illustrates as well that the constant lapse rate requires a sufficiently large net heat flow from surface upward.
In the case when there is an inversion, I think it will look something like this:
Now there is no convective coupling between the surface- and higher atmosphere temperatures. The delta-T in the higher atmosphere will depend on the radiated energies from the surface and the inversion layer, but also any changes in the heat flux coming from lower latitudes.
Posted by: Wipneus | December 20, 2012 at 13:39
Thank you Wipneus. That is a very informative post on lapse rates and inversion layers.
I'm trying to see how your post applies in the context of delta-T analysis, to determine how much the Arctic should warm up to eliminate the ongoing 1000 Gton/year volume loss.
Do I understand you correctly that you point out that heat loss to space may not be a simple factor, since it heavily depends on wether an inversion layer is present or not ?
And that therefore we should also consider that an increase in Arctic temperature reduces the heat flux from lower latitude ?
If so, I understand you point, and indeed it is likely that there are two paths for the Arctic to loose heat when it is warming up : one is to space, and one is to lower latitudes. This means that the 0.76 C increase mentioned before is probably an upper bound, and may be smaller if we consider the second path of loosing heat.
Is that a correct summary of your point ? Or was it something else ?
Also, let me note that the purpose of my back-of-the-envelope was not to exactly determine the warming necessary considering all physical properties. It was simply to determine a ballpark number for the warming necessary to lose an extra 1 W/m^2 out of the Arctic.
The interesting part (as far as I am concerned) is what happens next ? As far as my calculations show, the increased heating does cause additional ice thinning, and at this point the ice in the Arctic seems to be so thin already that any more thinning causes significant ice area reductions in the summer, which will increase heat absorption due to albedo change, and when the year is over we caused more ice loss than the 1 W/m^2 we were trying to eliminate. In other words, the Arctic may be unstable at this point, and will continue to lose more ice each year until it starts to heat up significantly.
Can you please comment on that reasoning ?
Posted by: Rob Dekker | December 21, 2012 at 07:15
Here are two interesting plots which seem to be relevant in this discussion :
Neven
1) Here is NCEP/NCAP reanalysis of Arctic (70N-90N) surface temperatures over the past decades, and
2) Here is OLR (Outgoing Longwave Radiation) at TOA over the same period.
There two graphs suggest that over the past 3 decades (the satellite era), annual average Arctic surface temperatures increased from -11.5 C to about -9 C (increase of 2.5 C) while OLR increased from 191 to 196 (increase of 5 W/m^2).
Thus this seems to suggest that it takes the Arctic about 0.5 C warming to shed 1 W/m^2 to space, which is right in line with our estimates.
Also, interesting to note that my estimate of 0.3 C/decade Arctic warming seems to be a severe underestimate. The 2.5 C warming over 3 decades (0.83 C/decade) suggests that the Arctic is warming about 5 times faster than the rest of the globe.
Posted by: Rob Dekker | December 21, 2012 at 07:58
Oops, I forgot to put closing >'s in the href HTML tags. I hope the links still work.
[Fixed the links, N.]
Posted by: Rob Dekker | December 21, 2012 at 08:00
Rob:
At the moment I am not worried about your numbers. It would not surprise me if they are basically correct.
I am concerned that the uncertainties caused by ignoring changes in a significant heat flux and the application of global factors to a situation that is in many ways different than the rest of the world.
About heat from lower latitudes. Heat fluxes tend to be linear with temperature differences. If the average temp. difference with the source of this heat is guessed to be -20 degC, a delta-T will cause -5% delta-heatflux, or about -5 W/m^2.
On the other hand I hear that some people think less difference in temperature will cause weakening jet-streams, causing heat exchange to increase, probably not in a linear way. So I'm not even sure about the sign of this effect.
So my best uncertainty estimate could be +/-5 W/m^2, clearly having the potential to change your numbers quite a bit in any direction.
Similar thoughts about the situation in the lower atmosphere , except the ignored delta's that I can think of (inversion thickness, low clouds) will cause increase in delta-T over your estimates.
So yes to most of your questions, except the sign of the change in heat flux from lower latitudes.
Posted by: Wipneus | December 21, 2012 at 09:45
Wipneus,
A .5C delta T on -20 degC appears to me to be 2.5% delta heat flux (not 5%).
That is presumably dealing with atmospheric heat flow and if we aren't sure of the sign maybe we will be lucky and the effects cancel out?
What of oceanic heat flow? Water out of arctic from deep water at 4C is not affected and temps of water in may increase but if we are dealing with a short period presumably not much change. If half? of the 100W/m^2 is by ocean then presumably the 2.5% atmospheric changes may also need to be halved again? (Still above 1W/m^2 but at least closer.)
In addition to your:
"the application of global factors to a situation that is in many ways different than the rest of the world"
What about seasonality? Does the assumption of .5C warmer all year round make sense? There has been larger temperature warming in winter and none in summer. With no ice in summer, then summer may well warm up but I would still expect growing temperature difference in Oct, Nov Dec, and falling temperature difference in Jan Feb Mar. Maybe this makes little difference but the ice is going to be a lot thinner in Oct Nov Dec and allow more heat to escape through much thinner ice.
The reason I am trying to push this: Could the calculations be trying to rid the heat balance through one effect - temperature change when there are more changes than just this. e.g. a growing negative feedback of much thinner autumn and winter ice. So far albedo has been beating insulation but the insulation effect is going to grow stronger with thinner ice. The albedo effect will also grow stronger as ice retreats more quickly but it presumably starts to level off at some point. I am not sure I can see how the question can be answered without showing how the major feedbacks, i.e. albedo and insulation, change.
Posted by: crandles | December 21, 2012 at 13:16
A .5C delta T on -20 degC appears to me to be 2.5% delta heat flux (not 5%).
read: a 1 deg delta-T causes 5% change in heat flux.
What of oceanic heat flow?
Bottom of the ice will have a more-or-less constant temperature equal to sea ice freezing point, -2 degC IIRC. No influence of ocean temperature there. Increased ocean temperatures will of course cause the ice cover to be thinner (not to mention prevent), which looks to be among the most important factors.
What about seasonality?
The original question specified "in winter". I kept to that, this is already difficult enough.
What is interesting what happens to the heat once the summer ice (Aug-Sep) is gone. If it disappears to space, winter ice volume decline will slow down.
The reason I am trying to push this: Could the calculations be trying to rid the heat balance through one effect - temperature change when there are more changes than just this.
I see what you mean. But it is the heat balance on the surface that you are most interested in. A heat balance over the whole (bottom to top) arctic atmosphere looks difficult to me. So I wonder if limiting to the lower layers would be feasible.
Posted by: Wipneus | December 21, 2012 at 14:02
When I wrote "What of oceanic heat flow?" I was still talking about heat flow from lower latitudes and suggesting 50W/m^2 by ocean and by atmosphere unless someone knows better than an arbitrary 50:50 split. Ocean 50W/m^2 growing slowly while atmosphere could reduce due to smaller differential but increase due to wandering jetstream. I don't think these changes are likely to swamp the ballpark analysis but you are correct to point out that they might.
You indicated you were concerned about 'application of global factors to a situation that is in many ways different than the rest of the world' in the calculation. I though it was natural to add to this that the assumption of .5C warmer all year round also looks suspect to me.
Yes this does add complexity to an already complex analysis which may not be wanted but if the problem requires the analysis, then ignoring it doesn't make the problem solved unless it can be shown to have little effect.
This post was just meant as clarification. If I am not helping, perhaps I should interfere less.
Posted by: crandles | December 21, 2012 at 15:09
Wipneus said
You certainly make a good point. It is very well possible that a delta-T in the Arctic may cause more heat loss to lower latitudes than to space.
Also, I agree that oceanic heat flux is most likely a very important factor (if not the cause) of the continuing ice thinning, especially for thinning of multi-year ice.
However, I am not convinced that the 'uncertainty' or 'variability' in either of the two atmospheric heat sinks is a determining factor in the inter-annual ice-volume loss. In fact, if we look at PIOMAS' anomaly chart
http://psc.apl.washington.edu/wordpress/wp-content/uploads/schweiger/ice_volume/BPIOMASIceVolumeAnomalyCurrentV2.png?%3C?php%20echo%20time%28%29%20?
it seems that the variability in ice volume, superimposed on the relentless downtrend is small, and does not seem to be able to stop or even affect the 1000 Gton (1 W/m^2) annual loss, regardless of which path (to space or to lower latitude) is dominant.
Posted by: Rob Dekker | December 22, 2012 at 09:07
Rob: about those global factors applied to the arctic.
Figure 9.1 from ipcc's AR4 shows modelled temperature profile trends over the 1890-1999:
Compared with the lower latitudes, most of the arctic warming is near the surface. In fact surface delta-T could be five times or more than delta-T at 300-400 mbar where most of the TOA OLR will take place.
I think this has everything to do with the inversion layer and low clouds.
Posted by: Wipneus | December 22, 2012 at 13:16
Ouch!
.25C times 5 instead of 2 gives 1.25C which then becomes 42 years of 0.03C per annum forcing.
Is this why we haven't seen worst case scenarios in the literature? Or is there any good reason(s) to think there has to be major errors in this?
Posted by: crandles | December 23, 2012 at 01:28
Crandles, Wipneus,
Warming is indeed very much stronger at the surface level than at higher altitudes, and maybe even a factor 5 at Wipneus suggests.
The same finding is also confirmed by Screen and Simmonds 2010, who present a very good overview of what may be happening in the Arctic atmosphere :
http://sciences.blogs.liberation.fr/files/arctique-ann%C3%A9es-2000-tures.pdf
Incidentally, they also conclude that Changes in cloud cover, in contrast, have not contributed strongly to recent warming. and suggest that the main cause of Arctic atmospheric warming is reduced ice cover (albedo changes cause more heat absorbed in summer, which needs to be released later (in fall and winter)).
Also, the altitude-dependent warming that Wipneus reports decidedly does NOT result in a factor 5 'sensitivity' factor over the Arctic. As I showed in my December 21, 2012 at 07:58 post (Neven, thanks for fixing the links), NCEP/NCAR data suggests that the increase of 2.5 C surface temperature over the past 3 decades caused a 5 W/m^2 increase in OLR at TOA over the Arctic (70-90N).
This means the Arctic atmosphere seems to support a sensitivity of 0.5 C/1 W/m^2, which is a factor 2 over non-feedback (black-body) response, and right in line with the guesstimate that started this conversation.
The reason that the uneven warming at different altitude does not cause a factor 5 in delta-T to OLR relation seems also addressed in Screen and Simmonds 2010. The find a significant increase in lower altitude water vapor (after all the most powerful greenhouse gas), especially in fall, which makes sense (due to a warmer fall due to summer sea ice losses). So maybe the inversion layer and special Arctic atmosphere condition do not significantly change the big picture. The factor 2 found over the Arctic is not unlike response factors found in non-Arctic atmospheres.
So prior arguments still stand, I think. The Arctic probably needs to warm up about 0.5 C to lose the additional 1 W/m^2 needed to eliminate the relentlessly consistent downtrend in volume (some 1000 Gton/year now).
Also, I wonder if you guys have any comments on my main finding, that the Arctic may at this point be unstable, and needs to warm up very significantly (probably my means of ice-free summers) before ice volume will level off.
Posted by: Rob Dekker | December 24, 2012 at 08:56
crandles said
I agree, Chris. The Arctic summer cannot warm up as long as the Arctic is covered in ice. Which suggests that Arctic winter needs to compensate for the entire year, and thus wam up about 1 C to compensate for the downtrend in ice volume (1000 Gton/year). Which makes prospects of 'recovery' (of volume, let alone extent) even more unlikely.
Posted by: Rob Dekker | December 24, 2012 at 10:30
Rob wrote:
Speaking about unlikely phenomenae, according to IARC-JAXA the SIE is going down since a couple of days ....
Posted by: Kris | December 24, 2012 at 11:10
IJIS December 23 2012;
Is heading south, lowest measured on this date: 11.507.969 km2
http://www.ijis.iarc.uaf.edu/en/home/seaice_extent.htm
Merry Christmas and Happy New Year.
Posted by: Espen Olsen | December 24, 2012 at 12:38
>"I agree, Chris. The Arctic summer cannot warm up as long as the Arctic is covered in ice. Which suggests that Arctic winter needs to compensate for the entire year"
I thought we were considering warming effect after last of 3300 Km^3 gone in summer so that further 1000 Km^3 loss per year seemed unlikely.
I think I have already said the analysis seems directed at how much warming would compensate when there may be two compensating effects - temperature increase and extra loss of heat in autumn and winter.
If you are considering the possibility of stopping ice loss before loss of summer ice then it definitely becomes a matter of considering extra heat loss in autumn and winter. This will only happen if the ice insulation feedback rapidly strengthens to catch up with albedo effect.
Albedo has been winning so far but that does not mean that the nature of the growth pattern cannot be such that the ice insulation effect doesn't start to grow more strongly and catch the albedo effect. With just 3300Km^3 ice left the time to do this seems short, but how do we rule it out? Characterizing the nature of the growth would seem a good start, if possible. Unfortunately I don't think I am managing to get my head around your sqrt of temp below freezing * time formula.
It might also be worth pointing out winter temps now seem about 5C above 1958 to 2002 average for about five months (Nov to Mar). This is substantially more than equivalent to .5C for whole year. So an extra 1.2C for five months looks like about 1.2/5*32years= ~8years effect.
Expecting the 8 years to shorten to 3300/1000 = 3.3 years might not be impossible considering growth in the ice insulation effect. However we also have to cope with continued GHG and further albedo feedback effects so I cannot see that happening. There just isn't the time unless you are really very, very optimistic about the growth of ice insulation effect while confidently expecting not much growth in albedo effect.
Any thoughts on the 8 years of effect calculated above versus the 16 years of forcing? Does this give an indication of the margin of error we are working with or indicate my calcs are foolishly deluded or perhaps give a hint of the ice insulation effect being strong or .... ?
Merry Christmas and Happy New Year.
Posted by: crandles | December 24, 2012 at 13:24
Kris wrote
Speaking about unlikely phenomenae, according to IARC-JAXA the SIE is going down since a couple of days ....
And today, at least, is at a record low of 11507969 km2, just below the prior record for this date set in 2010.
Posted by: Donald | December 24, 2012 at 15:00
Sorry for the late reply, guys. I enjoyed some long overdue family Christmas time.
I hope that everyone enjoyed the holidays, and very best wishes for you and your family for 2013.
Unfortunately, outlook for Arctic sea ice for 2013 and beyond does not look too bright.
First a clarification :
crandles said :
The 16 year forcing above was based on the assumption that the Arctic is warming 2X the global rate. The NCEP/NCAR data suggests that the Arctic actually warms about 5X the global rate over the past 30 years, which brings the 'lag' of ice melt behind Arctic warming to about 6.5 years. Your 8 years 'lag' is based on the 50 year record.
Either way, it seems that Arctic ice volume melt 'lags' behind Arctic warming, by some 6-8 years, simply because of energy imbalance. So that is the timeframe over which we may not expect a halt to the 1000 Gton/year ice loss.
I do not see any evidence at all that ice insulation negative feedback is getting stronger. If anything, ice volume loss rate has DOUBLED (instead of decreased) since the volume losses finally started to cause some extent losses around 2007.
This suggests that albedo effect is stronger than reduced ice insulation effects in fall/winter, and is getting stronger the more ice extent is lost...
You are right Chris; when summer ice is gone, there obviously is no more ice to loose in summer. However, the issue is the energy balance, which won't go away after summer ice is lost.
The equivalent of 1000 Gton ice per year is some 3.3 * 10^20 Joule energy deficit. Over some 10 million km^2, that is a deficit of 33 MJ/m^2. So if a square meter of ice is gone in summer, this energy (33 MJ) will go to heating Arctic ocean water. That is enough energy to warm 20 meters (mixing layer) of water by some 0.4 C. This suggests that maybe an ice-free Arctic summer will force the Arctic water to warm up so much that increased heat loss to space will prevent further energy deficits, and thus stabilize the Arctic.
However, if one square meter of Arctic ocean becomes ice-free during summer, albedo effect kicks in and it is very strong : Multiple studies suggest that a square meter of ice-free Arctic ocean will absorb around 20 W/m^2 annualized more heat as compared to an ice-covered square meter. That means that some 630MJ extra heat is absorbed for that ice free square meter.
That is a factor 20 stronger than the heat loss to space that we can expect, which suggests that a more ice-free Arctic summer will accelerate melting (and thus cause earlier ice-free conditions).
If the heat absorbed is 20x the heat lost, then it is hard to imagine how the heat loss each year can be confined to 1000 Gton/year equivalent (3.3*10^20 J).
This suggests that the real changes in the Arctic are just starting....
Posted by: Rob Dekker | December 28, 2012 at 11:03
Thanks Rob.
>>Expecting the 8 years to shorten to 3300/1000 = 3.3 years might not be impossible considering growth in the ice insulation effect.
>I do not see any evidence at all that ice insulation negative feedback is getting stronger.
Probably different terms:
The feedback probably isn't getting stronger but the effect of that feedback is getting larger: If you calculate the ice volume increase from minimum to the maximum, this looks fairly flat to start with but then turns upwards in the last few years. If I assume the minimum ice volume declines at a steady rate then I would expect ice volume increase to continue getting larger at increasing rate until the minimum ice volume hits zero.
If you have an upper bound on this that is only 5% of the albedo effect that would be very useful to show that albedo will continue to dominate insulation effect. Where does that come from?
So far the upturn in ice volume loss during summer seems larger then the upturn in winter ice volume increase by a factor of more like 2 than 20. This would suggest very rapid increase in effect of albedo feedback is to arrive shortly for the effects to top out with albedo effect being 20 times stronger than ice insulation effect. That is entirely possible and would be expected with the rapid loss of extent over just a few years of continuing to have a positive minimum ice volume.
That makes a good case for balancing out the heat budget with just temperature increases and ignoring the ice insulation effect.
It also further hints the next few years are highly unstable and it will only be after effect of albedo feedback is rapidly reducing (when there is little sea ice while the sun is high in the sky) that temperature increase can start to make headway towards balancing the heat budget. Until then albedo will look like a runaway effect - a very scary prospect.
Posted by: crandles | December 28, 2012 at 12:21
crandles said :
Sorry Chris. I realize that I was not very clear in my post.
The albedo effect referred to the total increase in heat absorption during an ice free summer. The current heat deficit that leads to the current 1000 Gton/year rate loss is only 5 % of that. That is the factor 20 I reported, but I admit that with that I am kind of mixing absolute heat loss/absorption with current 'rate' of heat loss deficit (rate of ice volume loss). Still, it spells out what is still to come.
The point I wanted to make is that ice insulation effects in fall/winter do not seem to be able to keep up with summer albedo changes, despite the fact that ice is thinning big-time and fall/winter temps are up. Evidence for this assertion lies in the increase in rate of volume loss, which was some 400 Gton/year before extent started to decline significantly, and close to 1000 Gton/year loss afterwards.
So, there does not seem to be much (if any) evidence that fall/winter ice thinning causes the Arctic to warm up fast enough to get rid of the increased (albedo change) heat absorption caused by reduced summer sea ice extent. In fact, my calculations above suggest that ice is so thin now that any increase in temperature will start to reduce ice extent very significantly, leading to more albedo-change heat absorption in summer than the heat loss to space that the temperature change originated. In other words, the Arctic may indeed be close to being unstable now, and the prospects for stabilization or recovery of ice volume are very dim.
We may thus loose more ice (and increase summer Arctic ocean temperatures) until we reached an ice-free Arctic summer. To quantify that state, an Arctic ice free summer will absorb about 20 W/m^2 more heat than current, which would have to be shedded to space. That means that the Arctic would need to warm up around 10 C, leading to warm summers (close to 8 C as in earlier estimates) and warmer winters (-10 C average or so). In fact, multiple scientific papers indeed predict an Arctic winter that is around 10 C warmer than today once the Arctic is ice free. Such warm winters will result in very thin winter ice that melts away very early in spring, and thus is not unimaginable, physically speaking.
My point exactly.
Posted by: Rob Dekker | December 29, 2012 at 11:29
"The point I wanted to make is that ice insulation effects in fall/winter do not seem to be able to keep up with summer albedo changes, despite the fact that ice is thinning big-time and fall/winter temps are up. Evidence for this assertion lies in the increase in rate of volume loss, which was some 400 Gton/year before extent started to decline significantly, and close to 1000 Gton/year loss afterwards.
So, there does not seem to be much (if any) evidence that fall/winter ice thinning causes the Arctic to warm up fast enough to get rid of the increased (albedo change) heat absorption caused by reduced summer sea ice extent."
The point I am trying to make is that that is all ***so far***.
On this, the past may not necessarily be a guide to the future especially when the heat flow difference between 30cm of ice and 50cm of ice is much greater than the difference between 50cm of ice and 70cm of ice.
Of course past albedo feedback effect is not a good guide to future albedo feedback effect either. Especially so when extent might start to decline more rapidly as we approach ice free conditions.
I do expect albedo effect to stay ahead of ice insulation effect. If it didn't, further decline in ice would then be reduced to GHG forcing and be very slow. This seems unlikely as it would be very difficult to then explain past warm arctic climates.
While I expect albedo effect to stay ahead, the challenge is to show that has to be the case rather than a rather arm wavy 'it has so far'.
Posted by: crandles | December 29, 2012 at 15:41
Rob
"So if a square meter of ice is gone in summer, this energy (33 MJ) will go to heating Arctic ocean water. That is enough energy to warm 20 meters (mixing layer) of water by some 0.4 C."
You lost me somewhere.
If a cubic meter of ice is melted the latent heat released would raise the temperature of a cubic meter of water by 80C, or raise the 20 meter mixing layer beneath to 80C / 20 = 4C above it's present temperature. (assuming it's a nice even ice cube)
If we are losing 1,000 km3 of ice per year this energy would heat the mixing layer beneath to 4C, or by spreading it across the whole Arctic Basin 4C / ~14.06 =~ 0.2845C
A repeat of 2012's melt would give .765 x 0.02845C =~ 0.218C
A repeat of 2010's melt would give 2,465 x 0.02845C =~0.701C
A repeat of 2007's melt would give 2.535 x 0.02845C =~0.721C
This should be seen as an annual addition to all the other feedbacks and forcings that the Arctic will experience. The latent heat of fusion has always been offset by the winter freeze but that ends when we run out of ice.
With the decrease in albedo, the increase in cloud cover and an increase in GHG I have to believe that an ice free summer isn't a stable condition & that a virtual perennial ice free Arctic will follow close on it's heels.
I'm in total agreement with your closing sentence.
"This suggests that the real changes in the Arctic are just starting...."
Terry
Posted by: Twemoran | December 29, 2012 at 20:15
>"This should be seen as an annual addition to all the other feedbacks and forcings that the Arctic will experience. The latent heat of fusion has always been offset by the winter freeze but that ends when we run out of ice."
You lost me. Surely the winter loses more heat when we run out of ice rather than this heat loss disappearing.
Posted by: crandles | December 29, 2012 at 21:14
crandles
We may lose more heat during winter than at present, but this is heat that otherwise wouldn't be there to lose.
As long as some ice survives there is no release of sensible heat from the ice melt as it is simply carried over to the subsequent freeze cycle. Temperatures north of 80 stay flat as do Arctic water temperatures while the energy is used to melt ice.
When there is no ice to melt the energy presents itself as sensible heat and water temperatures rise.
Much of the heat will be transferred to the atmosphere & some portion of this will escape through the clouds & GHG to be radiated off to space during winter but the fact is that this heat simply would not have existed had there been ice to melt.
I don't have any idea what portion of this heat will end up radiated away but I'd expect some will stay in the water & that Arctic inversion will increase due to fog and low clouds so that some heat will be left for the next year to build upon.
Whatever portion that remains will be added to each following year with a cumulative build up culminating in a year round ice free Arctic.
I think that at present some heat from the previous summer is retained and not radiated away. Even with less winter ice cover I don't see this changing, primarily because of the increased fog and low clouds I'd expect over open water.
I probably should have included expected radiation losses in the above post but I was only trying to address the additional sensible heat that will be released as opposed to other positive and negative forcings that an ice free summer would cause.
Terry
Posted by: Twemoran | December 29, 2012 at 23:03
crandles said :
I tried really hard to do exactly that in my December 15, 2012 at 10:53 post.
We even had a discussion about the 687 to 1375 Gtons less ice at maximum from these calculations, where I explained that this range was based on linear versus SQRT ice growth assumptions in winter.
I readily admit that the calculations in that post are crude and back-of-the-envelope, but with all due respect, Chris, rather than dismissing them as "far out" and "arm wavy", it would be more helpful if you point out where I go wrong in reasoning, assumptions and math in that post on the balance between positive and negative feedbacks in the Arctic at this point.
Posted by: Rob Dekker | December 30, 2012 at 07:33
Terry said
It seems that you did not lose me at all, Terry. the only reason I obtained 0.4 C and you got 0.2845 C is because I used 10 M km^2 for Arctic basin area, while you used the more accurate 14.06 M km^2 number.
Posted by: Rob Dekker | December 30, 2012 at 07:36
Hello everybody,
I am reading this blog for some months with increasing interest and learned a lot. Now I have the feeling, that the natural sciences are at its limits with the ab-initio models and there is a need for a engineering science description of the complex feedback systems - especially for the arctic, since signals of change are clearest there.
Could you suggest me some good engineering science type of papers about qualitative and quantitative feedback models for the arctis?
There should be a way to describe the dynamics of the system moving from one stable point (pre-industrial clima) to 400 ppm CO2 stable clima (e.g. like mid-pliocene) with usual models decribing the dynamics of non-linear systems between 2 stable points using some empirical data. If it could model/fit piomas, it would give some hope that it could be used to decribe the change. After that step, the economics point of view could be addressed and finaly there is a chance for broad democratic legitimation of necessary possible decisions. I think, the natural science are still very important to understand the system, but they are confusing with details and limitted validity by first principle methodes.
Are you aware of some literature describing the dynamics of the artic in such a way? Peer reviewed articles would be best to confince some poeple here to start with some work.
With best regards,
Jens
Posted by: SATire | December 30, 2012 at 12:40