Support the Arctic Sea Ice Forum and Blog

Author Topic: The 2014 Melting Season  (Read 1789765 times)

Frivolousz21

  • Nilas ice
  • Posts: 1915
  • Live in Belleville, IL..15 miles SE of St. Louis.
    • View Profile
  • Liked: 598
  • Likes Given: 7
Re: The 2014 Melting Season
« Reply #3250 on: August 24, 2014, 04:21:58 AM »
I think melt pond formation has everything to do with the June cliff.

I got a nickname for all my guns
a Desert Eagle that I call Big Pun
a two shot that I call Tupac
and a dirty pistol that love to crew hop
my TEC 9 Imma call T-Pain
my 3-8 snub Imma call Lil Wayne
machine gun named Missy so loud
it go e-e-e-e-ow e-e-e-e-e-e-blaow

Neven

  • Administrator
  • First-year ice
  • Posts: 9470
    • View Profile
    • Arctic Sea Ice Blog
  • Liked: 1333
  • Likes Given: 617
Re: The 2014 Melting Season
« Reply #3251 on: August 24, 2014, 10:35:03 AM »
ASI 2014 update 8 is up on the ASIB: neck and neck.

Not much to say really, from the conclusion:

Quote
With just 2-4 weeks left in this melting season, we wait and see where the minimum will end up. With 2014 relatively high, but steadily declining, and 2013 stalling around this time, the race is still somewhat exciting. Unless 2014 completely stalls over the next two weeks, which of course, is also possible.

But more important than this rearguard action, are the implications for ice pack health. As we have seen, modeled volume is now higher than it has been for a couple of years, and with a substantial part of the ice on the Pacific side of the Arctic surviving, multi-year ice will see another rebound. Last year's volume rebound had been all but wiped out at the start of this melting season, but maybe the ice pack gets some more flesh on the bones after this year of in situ melting.
The enemy is within
Don't confuse me with him

E. Smith

ChrisReynolds

  • Nilas ice
  • Posts: 1764
    • View Profile
    • Dosbat
  • Liked: 20
  • Likes Given: 9
Re: The 2014 Melting Season
« Reply #3252 on: August 24, 2014, 11:18:50 AM »
The fact that June sea ice extent losses have a stronger upward trend than the June area losses is interesting.  But extent in June is affected by regions like Hudson, Baffin etc., so I don't think this disproves that increased melt ponding plays an important role in the June Cliff of CT area.

I also thought it sounded plausible that regions like Hudson and Baffin are the reason why this most interesting data shows such a distinct trend. That could as well explain the seemingly low correlation with the September minimum. The problem is just that June area loss in these regions also seems to be on the rise. Either way, I don't think this can explained properly without taking a closer on the numbers of each individual region.

Regional analysis already done. I had been hoping to blog this weekend but I think I'll need more time to think about my results.

I've just done an analysis from 1990 to 2013 - producing the area loss contribution to overall area loss from 1 June to 30 June as a function of grid box concentration on 1 June (NSIDC gridded data). The results are correct but are a bit weird. However for the years with strong June losses (indicative of cliffs) 2007, and 2010 to 2013, were those losses due to melt pond formation I would expect that concentrations above 0.9 (well away from the ice edge) to play a consistent large role in the June loss - there is no such consistent pattern for those years. This suggests to me that what is going on is regional and complex, not one factor - melt ponds.

Regression of surface air temperature over the pack does imply a reasonably strong relationship on June area losses.

Michael,

Arctic Ocean compactness (area/extent), has been falling at a rate of -0.0024 /yr, R2=0.345 (statistically significant), from around 0.8 in the 1970s to around 0.7 in the last few years. This implies a decline in concentration - the pack has been becoming more dispersed. EDIT - That's data for 1 Sept.

Steven,

The image you post does suggest a strong role for melt ponding. But check out the plot I've attached. The Arctic Ocean plays a strong role in 2007 (and other years), but for the 2010 to 2013 period the picture is less clear. For 2012 June losses in the Arctic Ocean* only accounts for about half of the total June losses. Of the remaining half about 1/6 is from the Pacific (Bering, Okhotsk, and 2/6 from the American, mainly Hudson, but also Baffin (Gulf of St Lawrence is included in that region but is not a major player at that time).

*Arctic Ocean is classed as Beaufort, round to Barents, Greenland, Central and CAA.

Friv,
I thought like that too, then I looked at the data, and I saw less reason for certainty.
« Last Edit: August 24, 2014, 11:36:16 AM by ChrisReynolds »

ChrisReynolds

  • Nilas ice
  • Posts: 1764
    • View Profile
    • Dosbat
  • Liked: 20
  • Likes Given: 9
Re: The 2014 Melting Season
« Reply #3253 on: August 24, 2014, 02:52:34 PM »
Maybe if I re do everything, and instead of using June 1 to June 30 use the specific period of rapid anomaly falls in 2007, and 2010 to 2013.

It will take some time...  :(

Jim Hunt

  • First-year ice
  • Posts: 6268
  • Don't Vote NatC or PopCon, Save Lives!
    • View Profile
    • The Arctic sea ice Great White Con
  • Liked: 893
  • Likes Given: 87
Re: The 2014 Melting Season
« Reply #3254 on: August 24, 2014, 05:13:57 PM »
Last year's volume rebound had been all but wiped out at the start of this melting season, but maybe the ice pack gets some more flesh on the bones after this year of in situ melting.

As I suggested on the blog, I'm reserving judgment until after all the cyclones have done their work and the refreeze is well under way. A wave borne battering is already on its way to the edge of the "ice pack". For further information please see:

"David Cameron Practices for Big Wave Surfing Contest"
"The most revolutionary thing one can do always is to proclaim loudly what is happening" - Rosa Luxemburg

ghoti

  • Grease ice
  • Posts: 767
    • View Profile
  • Liked: 12
  • Likes Given: 15
Re: The 2014 Melting Season
« Reply #3255 on: August 24, 2014, 09:46:12 PM »
The SWERUS second leg has run into swells near Wrangel sufficiently large to prevent them from doing sediment cores and CTD sampling. I don't know how big this makes the waves but maybe you should consider Wrangel Island beaches for your surfing :P

http://www.swerus-c3.geo.su.se/index.php/martins-blog-leg2/285-from-calm-to-rough-weather

viddaloo

  • Nilas ice
  • Posts: 1302
  • Hardanger Sometimes
    • View Profile
  • Liked: 1
  • Likes Given: 0
Re: The 2014 Melting Season
« Reply #3256 on: August 25, 2014, 12:32:30 AM »
This graph will have another data point in a month or so. We'll see how much that changes its dark prognosis. Currently it says 2017.
[]

OSweetMrMath

  • Frazil ice
  • Posts: 137
    • View Profile
  • Liked: 0
  • Likes Given: 0
Re: The 2014 Melting Season
« Reply #3257 on: August 25, 2014, 12:55:31 AM »
Yes, but what is the justification for the cubic fit? We would expect any polynomial fit to exceed 100% in the short term. If you're looking for curves to fit, a logistic or Gompertz function is more physically realistic, but I suspect the data supports a wide range of curves and parameters, which limits the usefulness of any particular fit for forecasting.

viddaloo

  • Nilas ice
  • Posts: 1302
  • Hardanger Sometimes
    • View Profile
  • Liked: 1
  • Likes Given: 0
Re: The 2014 Melting Season
« Reply #3258 on: August 25, 2014, 01:09:16 AM »
Thanks, OSweetMrMath, that's always an interesting discussion, so let's exchange some views and definitions. To my view, working with this stuff I realise *all* statistics, all graphs, are lying, simplifying and/or guessing in some way. This graph, however, seems to fit the data tight:

Quote
Linear regression was the first type of regression analysis to be studied rigorously, and to be used extensively in practical applications. This is because models which depend linearly on their unknown parameters are easier to fit than models which are non-linearly related to their parameters and because the statistical properties of the resulting estimators are easier to determine.

What if a certain development just isn't linear? After all, it seems preferred — according to Wikipedia — historically simply because it's easier to calculate. Well, adding a trendline in Google Chart was also quite easy, so there :) If it better predicts the future, then fine. But we'll see how much it changes next month.
[]

cesium62

  • Frazil ice
  • Posts: 330
    • View Profile
  • Liked: 33
  • Likes Given: 5
Re: The 2014 Melting Season
« Reply #3259 on: August 25, 2014, 01:24:34 AM »
Yes, but what is the justification for the cubic fit? We would expect any polynomial fit to exceed 100% in the short term. If you're looking for curves to fit, a logistic or Gompertz function is more physically realistic, but I suspect the data supports a wide range of curves and parameters, which limits the usefulness of any particular fit for forecasting.

Why would a sigmoid type function be more realistic physically?  While I agree that the curve will flatten when it hits 100%, I don't see any reason for the curve to be continuous.

And now I'm curious...  if we restrict ourselves to functions of the form y = (ax + b)^k, with 'k' an integer, for what value of 'k' would we get the best fit?

plinius

  • Frazil ice
  • Posts: 403
    • View Profile
  • Liked: 0
  • Likes Given: 0
Re: The 2014 Melting Season
« Reply #3260 on: August 25, 2014, 02:08:39 AM »
I suppose sigmoid curves satisfy the eye of the beholder, because they avoid the Achilles-and-the-turtle problem... ;-)

But concerning fits - I believe that a curve should have a physical justification. So, not any polynomial is easily justified. Linear is quite ok, since on those "short" timescales, we can at least pretend the forcing to be somewhat linear. If you assume some feedback, I would not want to go beyond quadratic without a proper justification (by the way, the PIOMAS volume vs. time is fit near perfectly by a quadratic function + some seasonal signal).

@June area/extent loss: did you consider discarding the Hudson bay? Shift of its melt-out could be partially responsible.

viddaloo

  • Nilas ice
  • Posts: 1302
  • Hardanger Sometimes
    • View Profile
  • Liked: 1
  • Likes Given: 0
Re: The 2014 Melting Season
« Reply #3261 on: August 25, 2014, 02:55:46 AM »
Hi, plinius, just to be precise, this isn't volume over time, but melt percentage over time. The two curves will be quite different, because the melt volume over time probably won't be able to steadily grow, as ocean warming and lack of refreeze will limit the amount of ice there is left to melt. Smaller start (April) volume will however allow for 85, 90 and 100% meltdowns, as we can see from the graph.

As to the curve not landing smoothly and softly on the 100% mark, I don't see any reason why it should? I think it will go with a bang when it finally goes, and to a large extent surprise «everyone». Further, if we were to change the title of the graph and the properties we are measuring, into eg. Arctic ocean warming, then there's no reason the graph couldn't continue «through the ceiling», past 100% (I know, it sounds ridiculous), in terms of warming the CAB further and also expanding in each direction the period of Summer & Autumn when the ocean is a) ice–free and b) excessively warm. (Those would be entirely different numbers, of course, but one has to reflect upon what is happening out there in the Natural world, where there's actually no such thing as a STOP sign at 100% meltdown.)
[]

Frivolousz21

  • Nilas ice
  • Posts: 1915
  • Live in Belleville, IL..15 miles SE of St. Louis.
    • View Profile
  • Liked: 598
  • Likes Given: 7
Re: The 2014 Melting Season
« Reply #3262 on: August 25, 2014, 02:59:16 AM »
The ESS has fallen completely apart.  The bremen update today will be nasty.  Not sure it can melt out completely with the weather getting colder but it's going to be close.  The water is doing work.

The wind and water is doing work.



http://ocean.dmi.dk/arctic/satellite/plots/satanom.arc.d-00.png
I got a nickname for all my guns
a Desert Eagle that I call Big Pun
a two shot that I call Tupac
and a dirty pistol that love to crew hop
my TEC 9 Imma call T-Pain
my 3-8 snub Imma call Lil Wayne
machine gun named Missy so loud
it go e-e-e-e-ow e-e-e-e-e-e-blaow

Frivolousz21

  • Nilas ice
  • Posts: 1915
  • Live in Belleville, IL..15 miles SE of St. Louis.
    • View Profile
  • Liked: 598
  • Likes Given: 7
Re: The 2014 Melting Season
« Reply #3263 on: August 25, 2014, 03:58:25 AM »

The arctic basin itself has a lot of open water.  At least 250K or more of the extent is in the Canadian Arpichelago versus 2007, 2008,2010, 2011 and 2012.

The weather may prevent a major melt out now in the ESS region.  The vast majority of the melt has been from the bottom.  With the water and wind packing a punch we might see another 500-600K drop extent wise.  We will see.

I got a nickname for all my guns
a Desert Eagle that I call Big Pun
a two shot that I call Tupac
and a dirty pistol that love to crew hop
my TEC 9 Imma call T-Pain
my 3-8 snub Imma call Lil Wayne
machine gun named Missy so loud
it go e-e-e-e-ow e-e-e-e-e-e-blaow

pearscot

  • Frazil ice
  • Posts: 385
    • View Profile
  • Liked: 190
  • Likes Given: 65
Re: The 2014 Melting Season
« Reply #3264 on: August 25, 2014, 05:04:26 AM »
Damn Firv, that's a good graphic. I've been waiting all week to see one of those images. Looks like the central ice pack is in good shape but the rest if in tatters.  Although this year is far above the others, the bite has really grown significantly this month. I wonder if we have a brutal rest of the melting season with an El Nino next year if the nice will be in the hurtlocker. Either way, this year is pretty interesting to look at. Wish I knew more about the ice though...pictures only show so much.
pls!

Michael Hauber

  • Nilas ice
  • Posts: 1114
    • View Profile
  • Liked: 168
  • Likes Given: 16
Re: The 2014 Melting Season
« Reply #3265 on: August 25, 2014, 05:14:37 AM »
The extension of ice into the Barents sea is unusual.  Within the satellite record 1989 and 2003 are the only other years with ice extending beyond the islands on the north side.  There is currently more ice in the Barents sea as at this date than all but one year between 1980 and 2000, and ice has been expanding in this region for most of August.  (blown by wind - I'm quite sure its still melting)

At the same time the Laptev bite is clearly bigger than any other year.  The ice edge hasn't retreated as the northern most point in 2013 - which is a litter further around towards the Atlantic, but it looks like we are in with a fair chance of exceeding the northernmost point in 2013, which I'm pretty sure was the northern most point achieved in the satellite record.  Although the 7 day forecast currently shows winds reversing over the Laptev bite for a few days to blow the ice away from the pole.

So record low ice in one region, and near record high ice in an almost adjacent region.  Definitely in part due to the wind tending to blow ice from the Laptev sea around towards the Barents.

I do wonder if the high ice in the Barents will have any consequence for the northern most parts of the THC, or for the amount of warm water entering the Arctic from this current.  Perhaps more fresh water in this region will slow the sinking of heavy salty surface waters and slow the gulf stream down?
Climate change:  Prepare for the worst, hope for the best, expect the middle.

OSweetMrMath

  • Frazil ice
  • Posts: 137
    • View Profile
  • Liked: 0
  • Likes Given: 0
Re: The 2014 Melting Season
« Reply #3266 on: August 25, 2014, 05:51:17 AM »
Viddaloo,

In general, the problem of extrapolation is much harder than the problem of fitting a curve to be used inside the data. The problem is that you can fit the existing data as closely as you want, but you really want to fit the future data. Without other constraints on the problem, better fits on the historical data are likely to be worse fits on the future data. (With no other restrictions, it's possible to find a curve which goes through every existing data point and any value you choose for next year.)

One place to start with restrictions is with the mathematical properties of the data. The percentage ice melt is between 0 and 100 by definition, so I'm suspicious of prediction curves which don't stay in those bounds. You could say that the curve is the cubic until 2017 and then constant 100% after that. This avoids the bounds problem, but this solution is more ignoring the problem than solving it.

You run into the same problem in the other direction. The data is more or less flat through the 80s, and as a first guess I would expect it to have been flat during earlier time periods, so if we had the data for the 60s and 70s, the melt percentage would have consistently between 40% and 50% during that time period. But the cubic fit turns hard toward zero just to the left of the graph. The first problem is that you have to cut the curve off in the past, just like you have to cut it off in the future.

The second problem is that polynomials generally do not handle flat data very well. Assume I'm correct, and generate 10 or 20 years of fake, more or less constant, data on the early end of the data set, and try fitting the cubic again. Or go the other direction, and throw out the first 10 or 15 years of the data and fit the cubic to that. What you are likely to find is that the amount of data at the left end of the graph controls how fast the curve rises at the right end of the graph.

This could lead to the conclusion that the prediction for when 100% ice loss will occur is more dependent on how much data you include at the start than it is on the actual data points at the end, which should make you wonder if this makes sense.

I suggested the logistic or Gompertz curve because they are mathematically consistent with the past and what we expect in the future, since the data starts flat and could be expected to rise to nearly 100% every year. Even then, the parameter estimates are not likely to be very precise, which means that instead of predicting a particular year for a melt out, you should be trying to predict a likely range of years in which the first 100% melt could occur.

Another approach, as plinius suggested, is to use a linear fit. The argument is that nearly any function is approximately linear over short time scales. So use a time range where the data appears to be linear, assume it will continue to be approximately linear in the near future, and run with it.

This has two problems. First, the assumption of linearity is only valid for short time ranges. In particular, it may not hold as far as you are trying to predict. Second, since the data as a whole is not linear, you probably don't want to use the full data set. But then deciding where to cut the data off is not obvious, and as you use less data, the precision of the parameter estimates goes down. Effectively, you could end up with a variety of straight lines which appear to go through the data reasonably well, but result in an unreasonably large range of predictions for the date of 100% melt.

In conclusion, fitting curves to data for purposes of extrapolation is hard. Just because you can add a trend line to a Google Chart, it doesn't mean you should take it seriously.

viddaloo

  • Nilas ice
  • Posts: 1302
  • Hardanger Sometimes
    • View Profile
  • Liked: 1
  • Likes Given: 0
Re: The 2014 Melting Season
« Reply #3267 on: August 25, 2014, 06:36:40 AM »
OSweetMrMath — can I call you «»? — I stand corrected! Of course it's a poor mathematical model, but for very short periods from now until Zero Ice Day, it may suffice. As I mentioned in my last comment, the shift of the whole graph to heat potential or heat content in the Arctic ocean would probably be better, as that would allow for the graph to continue through Zero Ice Day to the other side.

BTW, some preliminary runs of the graphing script for other months than September show that August is likely to be the month we 'celebrate' Zero Ice Day. At least it will be ice–free the same year as September, or perhaps the year after. (I've got 2017.95 for August, and it's hard for August to arrive that late in the year!) July and October will follow suit in 2024 and 2025, respectively. That's no Arctic sea ice at the end of October, the 31st, not close to already for 8 years ice–free September.

PS: As an admitted newbie who just plays around with the new Google Graph toy, I'm hoping you and/or others can post some better predictions based on more reliable math. I'm skeptical of the linear attempts I've seen, though. Seems values are way too far away from the trendline, plus prediction is poor for the coming years. If a trendline can't reasonably predict the next 5 years or so, It's not much of a trendline, is it?
[]

Bob Wallace

  • Young ice
  • Posts: 3855
    • View Profile
  • Liked: 41
  • Likes Given: 5
Re: The 2014 Melting Season
« Reply #3268 on: August 25, 2014, 08:03:04 AM »
You're familiar with this?


viddaloo

  • Nilas ice
  • Posts: 1302
  • Hardanger Sometimes
    • View Profile
  • Liked: 1
  • Likes Given: 0
Re: The 2014 Melting Season
« Reply #3269 on: August 25, 2014, 08:21:40 AM »
You're familiar with this?

Yes, seen it before, it's very good, very convincing, even for novices. It's the way things are going, based on that registered data. Yet, I suspect the function is pretty similar to the one Google Graph drew up for me? Wipneus, can you provide us with the function? Great work, BTW!
[]

ktonine

  • Frazil ice
  • Posts: 363
    • View Profile
  • Liked: 0
  • Likes Given: 0
Re: The 2014 Melting Season
« Reply #3270 on: August 25, 2014, 09:20:10 AM »
Quote
OSweetMrMath writes: "One place to start with restrictions is with the mathematical properties of the data. The percentage ice melt is between 0 and 100 by definition, so I'm suspicious of prediction curves which don't stay in those bounds."

I think this is, at least in a physical sense, incorrect.  Ice is simply a solid form of H2O.  What we're measuring is how much H2O will have a temperature above/below it's freezing point.  The process doesn't stop once the H2O melts - it continues to increase in temperature with continued warming.

I've never seen an estimate of the average temperature of *all* the water in the arctic.  A time series of this metric may shed some light on the actual function we're seeking to describe.  A function that goes above 100% or below 0% is merely showing that temperature changes don't stop just because a phase transition has occurred.

jdallen

  • Young ice
  • Posts: 3410
    • View Profile
  • Liked: 650
  • Likes Given: 244
Re: The 2014 Melting Season
« Reply #3271 on: August 25, 2014, 10:07:07 AM »
Quote
OSweetMrMath writes: "One place to start with restrictions is with the mathematical properties of the data. The percentage ice melt is between 0 and 100 by definition, so I'm suspicious of prediction curves which don't stay in those bounds."

I think this is, at least in a physical sense, incorrect.  Ice is simply a solid form of H2O.  What we're measuring is how much H2O will have a temperature above/below it's freezing point.  The process doesn't stop once the H2O melts - it continues to increase in temperature with continued warming.

I've never seen an estimate of the average temperature of *all* the water in the arctic.  A time series of this metric may shed some light on the actual function we're seeking to describe.  A function that goes above 100% or below 0% is merely showing that temperature changes don't stop just because a phase transition has occurred.

We have spent a lot of time here speculating on the measurement of ice, and from that, trying to ascertain how soon it will disappear.

I say what we *should* be doing is measuring heat, and how it is moving around in the artic environment.  Some of the discussion I've seen here and elsewhere examining and attempting to measure total volume melt is heading in the right direction.

Albedo, phase change, re radiation, transport via current, etc, comprise some of our key variables.  Insolation, in narrow limits is a constant. Save for its effect on albedo, ice is more a spectator than actor in the drama.  We need to improve our skill at understanding the energy flow.  Do that, and we may better determine what the ice will do.
This space for Rent.

Andreas T

  • Nilas ice
  • Posts: 1149
    • View Profile
  • Liked: 18
  • Likes Given: 4
Re: The 2014 Melting Season
« Reply #3272 on: August 25, 2014, 10:54:35 AM »
What is missing from these discussions of the shape of a curve is that there has to be a physical mechanism underlying it. What is the mechanism which drives this expectation of a certain type of curve? Without that it just becomes a beauty contest of likes and dislikes.
How different predictions can be is seen here http://www.realclimate.org/index.php/archives/2011/09/the-unnoticed-melt/
paper here http://www.see.ed.ac.uk/~shs/Climate%20change/Climate%20model%20results/Tietsche%20ice%20recovery.pdf

The point this makes is that a reduction in sea ice  does not necessarily produce subsequent reductions in sea ice.


......

Albedo, phase change, re radiation, transport via current, etc, comprise some of our key variables.  Insolation, in narrow limits is a constant. Save for its effect on albedo, ice is more a spectator than actor in the drama.  We need to improve our skill at understanding the energy flow.  Do that, and we may better determine what the ice will do.


I agree with jd, the difficulty is to fathom how these processes which have brought the ice to its present state are responding to that changed state.
I disagree with ice being more spectator than actor. Its reduction of heat loss in winter is what makes Tietsches model rebound so quickly. Its effect on clouds is hard to guess at, could open water produce and maintain clouds to keep itself from refreezing? Those  clouds would surely produce precipitation, what does that (snow) do to affect subsequent developments?

This discussion probably belongs into "the slow transition" thread.
« Last Edit: August 25, 2014, 11:30:19 AM by Andreas T »

cesium62

  • Frazil ice
  • Posts: 330
    • View Profile
  • Liked: 33
  • Likes Given: 5
Re: The 2014 Melting Season
« Reply #3273 on: August 25, 2014, 11:03:20 AM »
You're familiar with this?



As Chris Reynolds points out (paraphrased) in The Long Transition thread, if you draw an exponential curve through the central arctic september minimum ice volume, that curve will hit zero well after the curve you get for the full arctic.  So you need to explain why the full arctic uses one model, but the northern-most subset of the full arctic uses a different model.

ChrisReynolds

  • Nilas ice
  • Posts: 1764
    • View Profile
    • Dosbat
  • Liked: 20
  • Likes Given: 9
Re: The 2014 Melting Season
« Reply #3274 on: August 25, 2014, 12:38:15 PM »
JD Allen,

Part of the PIOMAS data is ocean temperature.
http://psc.apl.washington.edu/zhang/IDAO/data_piomas.html
Otemp - ocean temperature of the first ten levels. Off the top of my head I can't say how deep this is. But using the large 'thermal mass' of water, and given the possibility of re-phrasing volume loss as energy gain, it should be possible to look at ice/ocean energy gain (at least as deep as ten levels) in PIOMAS.

Actually that only takes us up to 2004 - one of the indices they've not updated beyond the 2004 study.

ChrisReynolds

  • Nilas ice
  • Posts: 1764
    • View Profile
    • Dosbat
  • Liked: 20
  • Likes Given: 9
Re: The 2014 Melting Season
« Reply #3275 on: August 25, 2014, 12:43:12 PM »
You're familiar with this?



As Chris Reynolds points out (paraphrased) in The Long Transition thread, if you draw an exponential curve through the central arctic september minimum ice volume, that curve will hit zero well after the curve you get for the full arctic.  So you need to explain why the full arctic uses one model, but the northern-most subset of the full arctic uses a different model.

The model being used on this thread is wrong. It is based on the decline of volume primarily due to loss of multi-year ice, that won't continue.

Gray-Wolf

  • Grease ice
  • Posts: 948
    • View Profile
  • Liked: 131
  • Likes Given: 458
Re: The 2014 Melting Season
« Reply #3276 on: August 25, 2014, 12:58:43 PM »
Hi Chris !

It's whether 'extreme melt seasons' will continue to occur though?

After 2007 we looked back at past 'perfect melt storm' synoptics and found them to come around every 10 to 20yrs with the two before 2007 having only ten years spacings.

How do you see the current ice dealing with a return of the 'perfect melt storm' synoptic with high export and high melt rates throughout the season?

My greatest concern is the recurrence of a 'perfect melt storm ' year and that , should the past two events prior to 07' serve as a guide, could be 2017?
KOYAANISQATSI

ko.yaa.nis.katsi (from the Hopi language), n. 1. crazy life. 2. life in turmoil. 3. life disintegrating. 4. life out of balance. 5. a state of life that calls for another way of living.
 
VIRESCIT VULNERE VIRTUS

crandles

  • Young ice
  • Posts: 3379
    • View Profile
  • Liked: 239
  • Likes Given: 81
Re: The 2014 Melting Season
« Reply #3277 on: August 25, 2014, 02:01:44 PM »

As Chris Reynolds points out (paraphrased) in The Long Transition thread, if you draw an exponential curve through the central arctic september minimum ice volume, that curve will hit zero well after the curve you get for the full arctic.  So you need to explain why the full arctic uses one model, but the northern-most subset of the full arctic uses a different model.

I think it reasonable to suggest that outer seas will fully melt out earlier than Central Arctic Basin. So you are likely to get different dates using the same type of model. This does leave the question of whether

a) date using CAB is nearer correct while using all Arctic is a blend of dates of earlier seas melt out and later CAB melt out. (Perhaps to get the right date you have to use the last tiny portion of the CAB?)

or

b) Once the outer seas all fully melt out then waves get more access to CAB and the rate of melt accelerates. In this case date using all Arctic might be nearer correct.

Therefore these different dates do not as much of a problem as different dates when using volume compared with area or extent. Gompertz shaped curves could solve this. If however you are expecting continued volume reduction acceleration then why the area/extent curves predict so late looks to be a problem.

greatdying2

  • Grease ice
  • Posts: 574
    • View Profile
  • Liked: 0
  • Likes Given: 0
Re: The 2014 Melting Season
« Reply #3278 on: August 25, 2014, 03:36:43 PM »

As Chris Reynolds points out (paraphrased) in The Long Transition thread, if you draw an exponential curve through the central arctic september minimum ice volume, that curve will hit zero well after the curve you get for the full arctic.  So you need to explain why the full arctic uses one model, but the northern-most subset of the full arctic uses a different model.

I think it reasonable to suggest that outer seas will fully melt out earlier than Central Arctic Basin. So you are likely to get different dates using the same type of model. This does leave the question of whether

a) date using CAB is nearer correct while using all Arctic is a blend of dates of earlier seas melt out and later CAB melt out. (Perhaps to get the right date you have to use the last tiny portion of the CAB?)

or

b) Once the outer seas all fully melt out then waves get more access to CAB and the rate of melt accelerates. In this case date using all Arctic might be nearer correct.

Therefore these different dates do not as much of a problem as different dates when using volume compared with area or extent. Gompertz shaped curves could solve this. If however you are expecting continued volume reduction acceleration then why the area/extent curves predict so late looks to be a problem.

Clearly it is (b). Ice melts mainly from the edges. If you look at higher latitudes, of course it is melting slowly... until the edge reaches it, at which time melting accelerates.

If you arbitrarily defined a region, say "CAB core", as being 88-90 N, what would the curve look like? Would the trend be a good predictor of when 88-90 eventually will melt out? Of course not. Once the ice edge reaches that area, it will tank.
The Permian–Triassic extinction event, a.k.a. the Great Dying, occurred about 250 million years ago and is the most severe known extinction event. Up to 96% of all marine species and 70% of terrestrial vertebrate species became extinct; it is also the only known mass extinction of insects.

crandles

  • Young ice
  • Posts: 3379
    • View Profile
  • Liked: 239
  • Likes Given: 81
Re: The 2014 Melting Season
« Reply #3279 on: August 25, 2014, 04:53:11 PM »

As Chris Reynolds points out (paraphrased) in The Long Transition thread, if you draw an exponential curve through the central arctic september minimum ice volume, that curve will hit zero well after the curve you get for the full arctic.  So you need to explain why the full arctic uses one model, but the northern-most subset of the full arctic uses a different model.

I think it reasonable to suggest that outer seas will fully melt out earlier than Central Arctic Basin. So you are likely to get different dates using the same type of model. This does leave the question of whether

a) date using CAB is nearer correct while using all Arctic is a blend of dates of earlier seas melt out and later CAB melt out. (Perhaps to get the right date you have to use the last tiny portion of the CAB?)

or

b) Once the outer seas all fully melt out then waves get more access to CAB and the rate of melt accelerates. In this case date using all Arctic might be nearer correct.

Therefore these different dates do not as much of a problem as different dates when using volume compared with area or extent. Gompertz shaped curves could solve this. If however you are expecting continued volume reduction acceleration then why the area/extent curves predict so late looks to be a problem.

Clearly it is (b). Ice melts mainly from the edges. If you look at higher latitudes, of course it is melting slowly... until the edge reaches it, at which time melting accelerates.

If you arbitrarily defined a region, say "CAB core", as being 88-90 N, what would the curve look like? Would the trend be a good predictor of when 88-90 eventually will melt out? Of course not. Once the ice edge reaches that area, it will tank.

I don't think it is so clear at all. As Chris Reynolds points out, the decline in volume has been a lot of MYI reduction. That cannot continue to accelerate.

OTOH there are frequently occurring thickness that are thinning and approaching thickness where we could expect a large extra area to all melt out.

The models suggest Gompertz shaped curve in approaching ice free conditions. This seems much more credible after 2013 and 2014. After 2012 and before 2013 melt season, it seemed that there was so little time for the Gompertz shape to make an appearance. Much more space and hints of it now. Thus I now think it foolish to be too certain of a rapid transition.

Too much certainty of a rapid transition was probably always a bad idea - just gives conservatives' ammunition to say these environmentalists are always crying wolf.

BornFromTheVoid

  • Nilas ice
  • Posts: 1339
    • View Profile
  • Liked: 679
  • Likes Given: 299
Re: The 2014 Melting Season
« Reply #3280 on: August 25, 2014, 05:18:50 PM »
We've now dropped below 26 out of the 35 previous minima record (using the 5 day NSIDC average).
 
We've just 2002, 2005 and the 2007-2013 group left. 2002 we should pass with relative ease, the rest are uncertain for now.

I recently joined the twitter thing, where I post more analysis, pics and animations: @Icy_Samuel

Lord M Vader

  • Nilas ice
  • Posts: 1406
    • View Profile
  • Liked: 60
  • Likes Given: 39
Re: The 2014 Melting Season
« Reply #3281 on: August 25, 2014, 05:50:43 PM »
BFTV & Co: if we are looking at JAXA numbers we find that we already have surpassed 2002 minima and now only have 2005 and 2007-2013 ahead. I think the minima of 2005 should be passed in a couple of days or so given the conditions of the ice in ESS, Beaufort and Kara Sea.

One word about NSIDC numbers: they should be too high and downgraded. The reason for this is the area in CAA at Amundsens route that is ice free on both Wipneus and Bremens ice maps.

In addition, we are now having a quite solid predicition where the minima will be. From 2005-2013 there was a drop in SIE in the range from 0,15-0,73 Mn km2 from Aug 25 until minima giving a range from 4,60-5,18 Mn km2. Most of these years saw a drop of at least 0,5 Mn km2. Given the current conditions I don't think this is an unfair number. This should mean that we approximatively will equal 2013 in SIE.

BornFromTheVoid

  • Nilas ice
  • Posts: 1339
    • View Profile
  • Liked: 679
  • Likes Given: 299
Re: The 2014 Melting Season
« Reply #3282 on: August 25, 2014, 05:56:37 PM »
As long as the NSIDC methods are consistent over time, then they don't need downgrading, as you say LMV. Where you might have issue is comparing the newer AMSR-2 data with the older SMMR or SMMI data, primarily due to the large differences in spatial resolution.

When comparing over long time periods, consistency is key! If you want the most accurate data for now, then MASIE is probably the way to go.
I recently joined the twitter thing, where I post more analysis, pics and animations: @Icy_Samuel

Lord M Vader

  • Nilas ice
  • Posts: 1406
    • View Profile
  • Liked: 60
  • Likes Given: 39
Re: The 2014 Melting Season
« Reply #3283 on: August 25, 2014, 06:28:38 PM »
BFTV. yes, consistency is key!! :) Especially if one wnat to do long-term statistics  8) But I do think NSIDC are filtering obvious errors away when they are doing their final analysis.

Where do I find MASIE?


BornFromTheVoid

  • Nilas ice
  • Posts: 1339
    • View Profile
  • Liked: 679
  • Likes Given: 299
Re: The 2014 Melting Season
« Reply #3284 on: August 25, 2014, 06:36:34 PM »
All the MASIE stuff is here http://nsidc.org/data/masie/
I recently joined the twitter thing, where I post more analysis, pics and animations: @Icy_Samuel

TeaPotty

  • Frazil ice
  • Posts: 322
    • View Profile
  • Liked: 79
  • Likes Given: 121
Re: The 2014 Melting Season
« Reply #3285 on: August 25, 2014, 06:39:23 PM »
As long as the NSIDC methods are consistent over time, then they don't need downgrading.

In programming we say "Garbage in, garbage out." Consistency has no relation to bad data, and it definitely does not improve the bad data in any way. The NSIDC data has been proven to be faulty and overly conservative on many occasions.

ktonine

  • Frazil ice
  • Posts: 363
    • View Profile
  • Liked: 0
  • Likes Given: 0
Re: The 2014 Melting Season
« Reply #3286 on: August 25, 2014, 06:52:43 PM »
BFTV says:"As long as the NSIDC methods are consistent over time, then they don't need downgrading..."

This is a relatively common canard that I think is accepted too easily.  It relies on several assumptions whose truth is not clear.

1) That average ice concentration has not significantly changed.
2) That average ice thickness has not significantly changed
3) That average ice temperature has not significantly changed
4) That average ice geographic distribution has not significantly changed. (?)

Of these, only #4 seems to be relatively constant. 

Thus, is it informative to say that year xxxx has equivalent SIE to year yyyy if any of these parameters has significantly changed.  Is a pack with 5Mk^2 at an average 80% concentration the same as one with and average 50%?  Or large differences in average thickness or ice temperature?

Even if the SIE numbers came in identical to (for instance) 2005 would anyone here want to make the claim that the ice is the same now as then?  If not, then what exactly is extent telling us?

We spend a lot of time looking at it, but it's really a crude first order approximation of the ice state.

BornFromTheVoid

  • Nilas ice
  • Posts: 1339
    • View Profile
  • Liked: 679
  • Likes Given: 299
Re: The 2014 Melting Season
« Reply #3287 on: August 25, 2014, 07:00:51 PM »
As long as the NSIDC methods are consistent over time, then they don't need downgrading.

In programming we say "Garbage in, garbage out." Consistency has no relation to bad data, and it definitely does not improve the bad data in any way. The NSIDC data has been proven to be faulty and overly conservative on many occasions.

It's not so much bad data or errors, it's just a feature inherent to the lower resolutions sensors. While it's great being able to get more accurate data with modern sensors, we still need to be able to put this into context.

BFTV says:"As long as the NSIDC methods are consistent over time, then they don't need downgrading..."

This is a relatively common canard that I think is accepted too easily.  It relies on several assumptions whose truth is not clear.

1) That average ice concentration has not significantly changed.
2) That average ice thickness has not significantly changed
3) That average ice temperature has not significantly changed
4) That average ice geographic distribution has not significantly changed. (?)

Of these, only #4 seems to be relatively constant. 

Thus, is it informative to say that year xxxx has equivalent SIE to year yyyy if any of these parameters has significantly changed.  Is a pack with 5Mk^2 at an average 80% concentration the same as one with and average 50%?  Or large differences in average thickness or ice temperature?

Even if the SIE numbers came in identical to (for instance) 2005 would anyone here want to make the claim that the ice is the same now as then?  If not, then what exactly is extent telling us?

We spend a lot of time looking at it, but it's really a crude first order approximation of the ice state.

Well, no measure of extent is going to inform you of the quality or thickness of the ice. Extent is simply informing us of the extent, or the grid cells with 15% or greater ice concentration. It certainly isn't a good measure of the "health" of the ice. For that we need other metrics.
I recently joined the twitter thing, where I post more analysis, pics and animations: @Icy_Samuel

deep octopus

  • Grease ice
  • Posts: 559
    • View Profile
  • Liked: 6
  • Likes Given: 17
Re: The 2014 Melting Season
« Reply #3288 on: August 25, 2014, 07:51:20 PM »
This is interesting. JAXA is indicating that the northwest passage is about to open up via the Parry Channel. Guess we'll have to stay tuned.

Lord M Vader

  • Nilas ice
  • Posts: 1406
    • View Profile
  • Liked: 60
  • Likes Given: 39
Re: The 2014 Melting Season
« Reply #3289 on: August 25, 2014, 07:57:32 PM »
Two weeks to go before there will only be very small differences in SIE. What do we've to look at until that?

Bremen SIE-map and my analysis of plausible outcome with areas 1-6 but first a look at the weather forecast. Latest GFS 12z run indicates cyclonic dominance. There are warm air at 850 hPa that tries to enter the Arctic basin but probably won't reach any further than the shoreline. Initially, we'll have a cyclone located in the Laptev bite which is forecasted to move further to the North American coast but not giving too strong winds. However, these winds maybe will be strong enough to rip through the ice...

Area 1: this is the one I think will have the best opportunity to melt out completely. It depends on how thick the remaining ice is.

Area 2: I would be extremely surprised if this area manage to melt out. I would take some strong winds followed by relatively warm air to have such a scenario. I do think we'll either see a polynya be established followed by refreezing or the ice to be dispersed and very fragmentated until refreezing occurrs.

Area 3: This small area of ice have been extremely resistant, much more than I had expected it to be. However, given the cyclonic weather that is forecasted, at least some of this ice should vanish until the melting season ends.

Area 4: little or some of this ice in Amundsens Route should melt away but I don't think we'll see the Northwest Passage open up completely this year. Even if the ice most south go away, refreezing in the northern part , e.g Parrys Strait, will prevent the NP to open.

Area 5: this area won't melt out but maybe a polynya may form. The cyclone that was present here a few days ago did a good job and the ice concentration was mainly only about 50% so if another cyclone appears here it would be very interesting!

Area 6: some people here have already noticed it as interesting. I agree, but my main question is whether this ice will melt  more after we've passed the minimum...

What are your ideas about these areas?

sincerely, LMV

plinius

  • Frazil ice
  • Posts: 403
    • View Profile
  • Liked: 0
  • Likes Given: 0
Re: The 2014 Melting Season
« Reply #3290 on: August 25, 2014, 08:09:54 PM »
Hi viddaloo,

of course I know that it's percentage of ice melt, and I fully agree that there is no reason, why the graph should asymptotically approach the zero, instead of a hard crash. Only thing I was worried about was the not fully founded use of a higher order polynomial. So, the thing you would have to convince me about is why this should have an order larger than 2, in your case 3 (which is responsible for the aggressive prediction for an ice-free arctic - again I also assume this will happen, but cannot see why we should expect it before 2018 or so.).


Hi, plinius, just to be precise, this isn't volume over time, but melt percentage over time. The two curves will be quite different, because the melt volume over time probably won't be able to steadily grow, as ocean warming and lack of refreeze will limit the amount of ice there is left to melt. Smaller start (April) volume will however allow for 85, 90 and 100% meltdowns, as we can see from the graph.

As to the curve not landing smoothly and softly on the 100% mark, I don't see any reason why it should? I think it will go with a bang when it finally goes, and to a large extent surprise «everyone». Further, if we were to change the title of the graph and the properties we are measuring, into eg. Arctic ocean warming, then there's no reason the graph couldn't continue «through the ceiling», past 100% (I know, it sounds ridiculous), in terms of warming the CAB further and also expanding in each direction the period of Summer & Autumn when the ocean is a) ice–free and b) excessively warm. (Those would be entirely different numbers, of course, but one has to reflect upon what is happening out there in the Natural world, where there's actually no such thing as a STOP sign at 100% meltdown.)

plinius

  • Frazil ice
  • Posts: 403
    • View Profile
  • Liked: 0
  • Likes Given: 0
Re: The 2014 Melting Season
« Reply #3291 on: August 25, 2014, 08:20:07 PM »
@SweetMaths: I think my point was slightly different. Of course it boils down to the limitations of Taylor series, but I rather wanted to point to that we have some clues about the physics, and hence have some constraints on the fitting function. Correct thing would be to work that in and doing a nice Bayesian fit with proper hyperparametrisation, but that's overkill here.

@ChrisReynolds: I agree to most points, though I am not sure, why we should have reached the end of multi-year-ice burnout. In all cases there was something left and the reason why it did not happen the past two years was not lack of ice, but lack of Fram transport that lead to accumulation.
A simpler question would be in my eyes: Would the arctic be ice free if there was no multi-year-ice right now? And also - how much memory effects can the underlying ocean provide - your hypothesis is that we see the intertial behaviour of the past decade, because there was multi-year-ice melting, providing a longer timescale. But what if the ocean carries a major part of the heat (or also weakening of the low salinity layer)?

ChrisReynolds

  • Nilas ice
  • Posts: 1764
    • View Profile
    • Dosbat
  • Liked: 20
  • Likes Given: 9
Re: The 2014 Melting Season
« Reply #3292 on: August 25, 2014, 09:07:11 PM »

As Chris Reynolds points out (paraphrased) in The Long Transition thread, if you draw an exponential curve through the central arctic september minimum ice volume, that curve will hit zero well after the curve you get for the full arctic.  So you need to explain why the full arctic uses one model, but the northern-most subset of the full arctic uses a different model.

I think it reasonable to suggest that outer seas will fully melt out earlier than Central Arctic Basin. So you are likely to get different dates using the same type of model. This does leave the question of whether

a) date using CAB is nearer correct while using all Arctic is a blend of dates of earlier seas melt out and later CAB melt out. (Perhaps to get the right date you have to use the last tiny portion of the CAB?)

or

b) Once the outer seas all fully melt out then waves get more access to CAB and the rate of melt accelerates. In this case date using all Arctic might be nearer correct.

Therefore these different dates do not as much of a problem as different dates when using volume compared with area or extent. Gompertz shaped curves could solve this. If however you are expecting continued volume reduction acceleration then why the area/extent curves predict so late looks to be a problem.

Clearly it is (b). Ice melts mainly from the edges. If you look at higher latitudes, of course it is melting slowly... until the edge reaches it, at which time melting accelerates.

If you arbitrarily defined a region, say "CAB core", as being 88-90 N, what would the curve look like? Would the trend be a good predictor of when 88-90 eventually will melt out? Of course not. Once the ice edge reaches that area, it will tank.

But then first year ice will grow back vigourously and make up the losses - but the proper place for that argument is on the slow transition thread.

viddaloo

  • Nilas ice
  • Posts: 1302
  • Hardanger Sometimes
    • View Profile
  • Liked: 1
  • Likes Given: 0
Re: The 2014 Melting Season
« Reply #3293 on: August 25, 2014, 09:27:02 PM »
So, the thing you would have to convince me about is why this should have an order larger than 2, in your case 3 (which is responsible for the aggressive prediction for an ice-free arctic - again I also assume this will happen, but cannot see why we should expect it before 2018 or so.).

Hi, plinius! Well, the answer is probably as prosaic and boring as it is amateurish? :) Fact is, I was looking at the Google Developer pages for ways to graph these 1979—2013 melt data, and these lines were there in a graph code example:

Quote
type: 'polynomial',
degree: 3,

As they made for such a tight fit to the data points, I went with that. For 2018, most months, even September, are predicted to crash post–2018 when 2014 data points are added, even with a Degree 3 'aggressive' graph:

August crash is postponed from 2018 to 2022, July crash from 2020 to 2024, September.... We'll see!
« Last Edit: August 25, 2014, 10:00:09 PM by viddaloo »
[]

Greenbelt

  • Frazil ice
  • Posts: 167
    • View Profile
  • Liked: 39
  • Likes Given: 24
Re: The 2014 Melting Season
« Reply #3294 on: August 25, 2014, 11:23:27 PM »
According to the 12utc deterministic Euro, post-tropical storm Christobal heading toward Iceland with the power of a big winter storm.

Going to be an interesting week up there with volcano and now storm watching.

http://download.ecmwf.org/data/web248/get_legacy_plot-web248-20140825190516-28308-13609.gif

Michael Hauber

  • Nilas ice
  • Posts: 1114
    • View Profile
  • Liked: 168
  • Likes Given: 16
Re: The 2014 Melting Season
« Reply #3295 on: August 25, 2014, 11:32:41 PM »
Will be interesting to see the impact of the current low pressure system on the ESS as it moves in that direction.  The ice there is starting to look thin in that area with the floe size small enough I have to zoom into 250m resolution to see them.  Although at that zoom they still look significant so I think they'd need to be a bit thinner to be really vulnerable to flash melting. 
Climate change:  Prepare for the worst, hope for the best, expect the middle.

OSweetMrMath

  • Frazil ice
  • Posts: 137
    • View Profile
  • Liked: 0
  • Likes Given: 0
Re: The 2014 Melting Season
« Reply #3296 on: August 26, 2014, 02:37:01 AM »
You're familiar with this?



This is the graph that brought home the danger of arbitrary curve fitting for me. I believe the points on the graph are the average ice volume for the month of September from PIOMAS for each year through 2013, and the curve is an exponential fit to those points, so something like volume = a - k exp(c year) where a, k, and c are constants determined by the fit. Maybe Wipneus can chime in with the exact formula.

As a fit for the data, it's really good. But we want to use it to predict when the sea ice volume will go to zero, and it runs into problems for this purpose. The first thing to notice is that although most points fall inside the confidence interval, the data point for 2013 is above the confidence interval. The fact that the last actual value is off the curve should lead to doubt about the curve's ability to predict the future.

We don't have the 2014 value yet, but we can expect it to be close to the 2013 value based on the melt so far this year. So it looks like the 2014 value will be far outside the confidence interval for the predicted value. Looking ahead, the curve predicts that the 2015 value will be less than 2000 cubic kilometers. I could be wrong, but that seems very unlikely to me.

In addition, the 2012 version of the curve (without the 2013 data point) was an even better fit to all of the existing data, but the predicted value for 2013 was much too low. The prediction curve was even lower before the 2013 data point pulled it back up.

What this means is that the 2013 curve could not accurately predict the 2014 value, and the 2012 curve could not accurately predict the 2013 value. It does not have reliability for one year predictions, and yet people want to use it to predict when the ice volume will go to zero, 4 or 5 years in the future.

This curve's claim to fame is that the 2010 data predicted the 2011 and 2012 data points, but to me that just points up the danger of arbitrary curve fitting. Just because a particular curve works for some period in the future, that does not mean it will work for other periods.

At this moment a quadratic fit to this data gives a better fit than the exponential fit, but again there's no reason to assume that will continue to be true. A simple linear fit, while not great, is not unreasonable. The linear fit is also somewhat more stable than the other fits, with new data having a relatively small impact on the trend line.

On the basis of the data, the linear fit is likely to overestimate how long the ice will last. This is a personal preference, but I prefer this to the other alternative. There are two possible statements: This linear fit predicts the Arctic will be ice free by 2040, but there's a good chance it will be earlier than that.  Or: This (exponential or quadratic) fit predicts the Arctic will be ice free by 2015, no 2017, no 2020, no, but real soon, I assure you.

I try to cautious in the predictions I make, so I try to be cautious with the regression functions I use.

Michael Hauber

  • Nilas ice
  • Posts: 1114
    • View Profile
  • Liked: 168
  • Likes Given: 16
Re: The 2014 Melting Season
« Reply #3297 on: August 26, 2014, 03:48:19 AM »
I believe projecting a polynomial fit onto PIOMAS data was behind projections of an ice free Arctic as early as 2013 (but more likely later) which were made in 2007.  BBC

So the projection worked from 2007 to 2012, but seems to now be definitely going wrong.
Climate change:  Prepare for the worst, hope for the best, expect the middle.

Steve Bloom

  • New ice
  • Posts: 20
    • View Profile
  • Liked: 0
  • Likes Given: 0
Re: The 2014 Melting Season
« Reply #3298 on: August 26, 2014, 04:33:25 AM »
No, Michael.  Maslowski's model-based projection is for 2016 +/- 3 years, so don't get excited just yet.  Remember also that similar comments were made after the 2009 season.   

viddaloo

  • Nilas ice
  • Posts: 1302
  • Hardanger Sometimes
    • View Profile
  • Liked: 1
  • Likes Given: 0
Re: The 2014 Melting Season
« Reply #3299 on: August 26, 2014, 05:07:05 AM »
I try to [be] cautious in the predictions I make, so I try to be cautious with the regression functions I use.

But is it fair to say that giving key decision makers in our society a false impression that the Arctic Sea Ice will last till 2040 — when you know this is most likely way too late — has any semblance of being cautious? Some people would perhaps rather call that irresponsible.

The way I interpret the Ice Volume data from PIOMAS, we may be looking at an ice–free Arctic all year round just ten years after your «cautious» 2040 prediction:
[]