Sea Ice Prediction – Update To 2012

Update#7 

I agree with Grant Foster and dana1981 that  it is a bad idea to make claims about short trends because the error bars in mapping to longer term trends are going to be large.  I should have presented my reasoning for doing this more clearly.

In examing the Vinnikov et al 1999 paper, I did not explain that my use of their trend values to state that short term assessments have value for quantities which involve inertia (mass) such as heat and ice. If the sea ice area were to recover to its original area and thickness (for whatever reason), for example, it does not matter what its long term trend was.  The long term trend (if there is one) would be reset. I have made this point often with respect to ocean heat content (e.g. see). It also applies to sea ice (although area is only one part of it).

Trend analysis, as being interpreted by Grant Foster and dana1981  is fundamentally an assumption that there is an overarching control on the long time period. My perpsective is to see if there are change points which interrupts that trend (e.g. 2006). I look forward to seeing what they obtain for sea ice area and for insolation-weighted sea ice using the same approach they used for sea ice extent.

Update #6  Skeptical Science has joined with Grant Foster to dismiss the claim in my posts that the Vinnikov et al 1999 model prediction overpredicted Arctic sea ice loss. Indeed, they concude the opposite. The Skeptical Science post is Lessons from Past Predictions: Vinnikov on Arctic Sea Ice. They both critizise my post in that I visually extracted the trend and also cherrypicked the start of the time period.

The visual examination of a time series to look for patterns is a classic approach that has been used in developing atmospheric boundary layer forumla for use in models, as I learned from Hans Panofsky years ago.  The selection of a starting time (2006) was based on  this figure, and it does show a clear breakpoint. David Douglass has used this concept of breakpoints (although for different years) with respect to ocean heat content in his paper

D.H. Douglass, R.S. Knox, 2012: Ocean heat content and Earth’s radiation imbalance. II. Relation to climate shifts. Physics Letters A. http://dx.doi.org/10.1016/j.physleta.2012.02.027

The Skeptical Science weblog post concluded that “the Arctic sea ice decline is about 27 years ahead of the Vinnikov model predictions.”

I did make an error in my post by assuming that anomalies in sea ice extent and sea ice area would be the same. I am convinced from the analyses by Grant Foster and dana1981 that this is incorrect. I have always focused on sea ice area since this is the more accurate way to diagnose the albedo radiative feedback.

Since, Grant Foster and dana1981 at Skeptical Science have the statistical tools ready to analyze the data, to convincingly show that Arctic sea ice decline is as advanced as they state, that they

i) perform the same analysis for sea ice area that they have done for sea ice extent

and

ii) perform the analysis of insolation-weighted sea ice trends; e.g. see

Pielke Sr., R.A., G.E. Liston, and A. Robock, 2000: Insolation-weighted  assessment of Northern Hemisphere snow-cover and sea-ice variability.  J. Geophys. Res. Lett., 27, 3061-3064

and

Pielke Sr., R.A., G.E. Liston, W.L. Chapman, and D.A. Robinson, 2004:  Actual and insolation-weighted Northern Hemisphere snow cover and sea  ice — 1974-2002. Climate Dynamics, 22, 591-595 DOI10.1007/s00382-004-0401-5.

With these two other analyses, we will see if the Arctic sea ice is actually in the “death spiral” as they report based on their analysis.  They may be correct, but from the figure below using sea ice area from the Cyrosphere Today, it certainly looks like a hiatus in its demise since 2006.

Update #5   Some of the readers of my weblog are wondering who Tamino is.  The scientific debate on the sea ice issue, of course, is independent of who makes comments, but professional courtesy requires that an individual identify themselves. Tamino is Grant Foster of Tempo Analytics in Garland,  Maine [h/t Anthony Watts]. He has a peer-reviewed publication with Stefan Rahmstorf

Grant Foster and Stefan Rahmstorf 2011: Global temperature evolution 1979–2010. Environ. Res. Lett. 6 044022 doi:10.1088/1748-9326/6/4/044022.

and one wirth Mike Mann, James Annan, and Gavin Schmidt

Foster, G., J. D. Annan, G. A. Schmidt, and M. E. Mann (2008), Comment on ‘‘Heat capacity, time constant, and sensitivity of Earth’s climate system’’ J. Geophys. Res., 113, D15102, doi:10.1029/2007JD009373.

His expertise clearly is in statistics, and he wants to apply this skill to climate analysis.

He does have a quite good study published February 12 2012 titled

Pine Beetle Infestation and Fire Risk in the Black Hills

where he writes as part of his conclusion

Surely, excessive rhetoric about the urgent fire danger posed by pine beetle infestation, sometimes to the point of hysteria, does not serve the public interest.

[he should apply the same advice to his treatment of climate science]

His real world identity, and measured scientific approach, however, conflicts with his handling of his posts and those of his commenters on Tamino.  Tamino has a number of the same disparaging commenters that appear on Skeptical Science.  The failure to identify yourself when discussing a scientific issue shows a lack of professional courtesy.

Now, with respect to the sea ice discussion,  I have asked the question of Grant Foster

What criteria in the observations would have to occur, before you would reject the model predictions of sea ice coverage?

Provide a benchmark criteria to assess for the coming seasons. From my perspective, I view that the models are not refuted if the anomalies in sea ice areal coverage fall at close to or greater than the rate of, say, the order of 10o,000 square kilometers per decade in the next few years.

Update #4  April 26 2012   In the latest comment in the Tamino post Do the Math, I am critizied for using sea ice area and not sea ice extent (along with the now to be expected personal insults by a number of commenters on Tamino).   However, in my view since it is area, not extent that better maps with radiative feedback, the Cryosphere Today presentation is preferred. Perhaps Tamino, should show the Vinnikov et al prediction (if he can obtain it) for the years up to 2012 and beyond in terms of  sea ice area.

In terms of the Vinnikov et al plot, there clearly is an ambiquity as the Hadley and the GFDL results for sea ice extent are so different. To repeat, the use of anomalies provides a way to avoid the absolute value of sea ice extent or area, as both are quite dependent on the precise definition that is used to define them (as illustrated by the Hadley and GFDL analyses).   Tamino would have to show that the anomalies for the extent and are coverage were distinctly different, and then I would agree that the two should not be interpreted as having the same anomalies.

There is one valid point, however, that is raised in the comments. Is the break-point visible in 2006 statistically significant? We are pursuing this analysis and will report on this weblog when we have results.

Finally, if Tamino wants to have a constructive discussion, I recommend he also examine the magnitude of the insolation-weighted radiative feedback from sea ice area coverage, as we presented in our papers

Pielke Sr., R.A., G.E. Liston, and A. Robock, 2000: Insolation-weighted  assessment of Northern Hemisphere snow-cover and sea-ice variability.  J. Geophys. Res. Lett., 27, 3061-3064

and

Pielke Sr., R.A., G.E. Liston, W.L. Chapman, and D.A. Robinson, 2004:  Actual and insolation-weighted Northern Hemisphere snow cover and sea  ice — 1974-2002. Climate Dynamics, 22, 591-595 DOI10.1007/s00382-004-0401-5

Of course, this discussion of a break point becomes moot if the sea ice decline prior to 2006 returns.  However, until and unless that happens, Tamino (and a number of the commenters) are not, in my view, following the scientific method which is to seek to refute hypothses (i.e. in this case the Vinnikov et al or other model preditions) rather than defending the models. Tamino asks the question

Could it be that Roger Pielke is actually aware of that, but that he really doesn’t care about portraying sea ice changes correctly, he only cares about discrediting global warming science?”

This question is absurd and insulting, but it does show Tamino’s mindset.  I am not discussing global warming science at all in the sea ice post.  I am seeking to refute (test) the Vinnikov et al prediction. That is the scientific method.  If you are convinced and present evidence, as you have, that I have been unsucessful so far, that is an appropriate scientific debate. I can then counter with other information.  However, to misrepresent my views on climate science, when I specifically referred you to a summary article on my views;

Pielke Sr., R., K.  Beven, G. Brasseur, J. Calvert, M. Chahine, R. Dickerson, D.  Entekhabi, E. Foufoula-Georgiou, H. Gupta, V. Gupta, W. Krajewski, E.  Philip Krider, W. K.M. Lau, J. McDonnell,  W. Rossow,  J. Schaake, J.  Smith, S. Sorooshian,  and E. Wood, 2009: Climate change: The need to consider human forcings besides greenhouse gases.   Eos, Vol. 90, No. 45, 10 November 2009, 413. Copyright (2009) American   Geophysical Union

is dishonest.

Update #3  April  25 2012 The comments on Tamino on their post Do the Math raised the issue of why I chose to use the long-term analysis of near daily of anomalies of sea ice area rather than yearly averages of sea ice extent for my comparison with the Vinnikov et al result.  The answer is that there is more information in the anomalies. For example, see

The absolute values on the left side of the Vinnikov et al plots are also misleading, as clearly the real world annual average of sea ice extent shown in the above figure from The Cryosphere Today is not even close to the values plotted for the GFDL model or what are reported in their paper to be the Chapman and Walsh, or Parkinson et al observed data.  The reason for the disagreement is likely due to different definitions of sea ice extent. Showing anomalies from the current data on The Cryosphere Today is a way to focus just on the anomalies and their trend over time, and to avoid dealing with the absolute values themselves.

The comments on Tamino disparage the use of visual information to interpret the behavior of the sea ice.  Instead they focus on linear (or near-linear trend) quantitative analyses. However, if the system is not behaving in a near linear (monotonic) way (such as shown in the Levitus et al figure 1 right), their analysis will miss obvious changes in the behavior of the data.  Everyone, if they are being objective, will see a break in the slope of the anomaly plots since 2006.

Finally, reading the comments on Tamino illustrates the lack of professional courtesy that is the benchmark of science.  I am not presenting my comments on Tamino for that reason.  Most of the commenters also hide behind anonymity to insult rather than actually debate a scientific issue. The Tamino weblog approach illustrates the unfortunate state of scientific debate with respect to climate state. They are defending the model predictions, rather than follow the scientific method to determine if there are real world observstions that refute the models (which are, afterall, hypotheses). If all attempts fail to reject the model predictions, than it is accepted as a robust prediction and attribution tool. A simple question to Tamino would be what criteria in the observations would have to occur, before he would reject the model predictions of sea ice coverage?

Update #2 April 24 2012  There is a comment by “Ned” at Tamino that presents the plot of “annual mean sea ice extent” and show that it has decreased faster than the Vinnikov et al model results.  I disagree that this is the proper metric to show, and that anomalies provide the more appropriate comparison as this avoids the confusion as to which value is actually plotted in the Vinnikov et al paper with respect to the models.  Moreover, the plot of anomalies shows why using linear (or quasi-linear trends) is misleading. The sea ice data had a significant change in its trend in 2006. Tamino and his commenters appear unable or uninterested in actually constructively discussing this issue. They have also chosen to ignore the inconvenient behavior of the Antarctic sea ice coverage.

Update April 24 2012 – Tamino has a post titled

Do the Math

where he disagrees with my post. Unfortunately, he does not present an honest view of what I wrote. He writes

I refer the reader to our advice on “Defense Against the Dark Arts.” His misdirection is revealed by Step 3: look at more than they show you, and be especially wary of time spans that are too brief and areas that are too small.  In this case the “time spans that are too short” alarm is flashing red — not only has Pielke cherry-picked his starting point, he’s comparing a predicted long-term trend to an observed time span of far less than a decade.  That’s foolish of him, and misleading to his readers.

Tamino also writes

Note also that when you start “since 2006″ or later, the error bars on the estimated rates are rather large.  The trend for such short time spans is so uncertain, it really doesn’t give much information.

In other words, Tamino, for some reason claims “error bars” are too large which is absurd, as the sea ice coverage data analysis is quite robust, and 2006 was chosen as it is clearly a break-point in the sea ice data.

He estimates the trend not just from 2006 to the present, but from every starting year since 1999 to the present and has, for example, the claim that the trend from 2009 to the present is -2 million km squared per decade. This is obviously quite different from what is shown below from The Cryosphere Today below. Perhaps Tamino should take his own advice and “do the math“.

He then concludes with a motive

Could it be that Roger Pielke is actually aware of that, but that he really doesn’t care about portraying sea ice changes correctly, he only cares about discrediting global warming science

Tamino is quite disingenuous in his post.  I do not disagree that the Arctic sea ice has been decreasing. My post was to compare the sea ice anomaly trends that were presented in the Vinnikov et al paper to real world observations updated to 2012. The figure in the Vinnikov et al 1999 paper shows a rather monotonic (but increasing over time) decrease  in Arctic sea ice content with time.  Tamino ignores what is obvious in even a visual comparison between the Vinnikov et al plot and the real world observations that the decline has stopped, at least for now.

As to motive, I did not even discuss global warming in the post. His ad hominem end statement that I only care about “discrediting global warming science” is an example of him seeking to discredit anyone who introduces a counter viewpoint. If Tamino was actually honest in his post, he would present the same type of figure from the models as done by Vinnikov et al 1999.  In addition, he would report honestly on my view on climate science and global warming which we summarized in our paper

Pielke Sr., R., K.  Beven, G. Brasseur, J. Calvert, M. Chahine, R. Dickerson, D.  Entekhabi, E. Foufoula-Georgiou, H. Gupta, V. Gupta, W. Krajewski, E.  Philip Krider, W. K.M. Lau, J. McDonnell,  W. Rossow,  J. Schaake, J.  Smith, S. Sorooshian,  and E. Wood, 2009: Climate change: The need to consider human forcings besides greenhouse gases.   Eos, Vol. 90, No. 45, 10 November 2009, 413. Copyright (2009) American   Geophysical Union.

**********original post******************

In 2009 I posted the following

A Comment On A 1999 Paper “Global Warming And Northern Hemisphere Sea Ice Extent By Vinnikov Et Al

I have edited the post below using crossouts and underlining, to show that the models are even more out of sync that they were in 2009. There are also very recent discussions on the Arctic sea ice anomalies at WUWT and Real Climate.

Ten  Thirteen years ago, the following paper was published.

Vinnikov et al., 1999: Global warming and northern hemisphere sea ice extent. Science. 286, 1934-1937.

In this paper, there is a presentation of the model predictions of sea ice extent along with observations up to 1998.  This weblog introduces the subject of how well have the model predictions done.

Their abstract includes the statement (referring to the GFDL and Hadley global climate models)

“Both models used here project continued decreases in sea ice thickness and extent throughout the next century.”

In the conclusion to their paper, they write

“Both climate models realistically reproduce the observed annual trends in NH sea ice extent. This suggests that these models can be used with some confidence to predict future changes in sea ice extent in response to increasing greenhouse gases in the atmosphere. Both models predict continued substantial sea ice extent and thickness decreases in the next century.”

In their paper (in Table 1) they have model predictions (in units of linear trend in 106 square kilometers per decade) listed for the GFDL climate model from 1978-1998 of -0.34 (and -0.19 using a “smoothed model output“) and for the Hadley Centre climate model -0.18 (and -0.16 using a “smoothed model output“).

A value of -0.18 is the loss of sea ice area of 180000 square kilometers per decade, for example.

The first figure below is from the Vinnikov et al 1999 paper  with respect to the model predictions, while the second and third figures are the sea ice areal converage for the Northern Hemisphere up to the present (April 20 2012)  and the Antarctic sea ice areal anomaly from The Cryosphere Today.

Until later in 2007, the sea ice areal extent continued to decrease in a manner which, at least visually, is consistent with the Vinnikov et al 1999 predictions (although the actual values of areal coverage differ substantially between the observations and the predictions, perhaps as a result of their formulation to compute areal coverage).

However, since 2006, the reduction has stopped and even reversed. Perhaps this is a short term event and the reduction of sea ice extent will resume. Nonetheless, the reason for the turn around, even if short term, as well as the long term increase in Antarctic sea ice coverage, needs an explanation.  Moreover, this data provides a valuable climate metric to assess whether the multi-decadal global models do have predictive skill as concluded in the Vinnikov et al 2009 paper.

It has been claimed that most of the recent sea ice is thin and thus will melt quickly this spring. Perhaps so. However, in terms of the albedo feedback into the atmosphere that we discuss in our papers

Pielke Sr., R.A., G.E. Liston, and A. Robock, 2000: Insolation-weighted  assessment of Northern Hemisphere snow-cover and sea-ice variability.  J. Geophys. Res. Lett., 27, 3061-3064

and

Pielke Sr., R.A., G.E. Liston, W.L. Chapman, and D.A. Robinson, 2004:  Actual and insolation-weighted Northern Hemisphere snow cover and sea  ice — 1974-2002. Climate Dynamics, 22, 591-595 DOI10.1007/s00382-004-0401-5

the albedo feedback would be muted and could be even negative if the positive global sea ice converage continues.

Comments Off

Filed under Climate Change Metrics

Comments are closed.