I received a number of emails about the newly published Guemas et al (2013) paper titled “Retrospective prediction of the global warming slowdown in the past decade”. It’s paywalled. The abstract is here. It reads:
Despite a sustained production of anthropogenic greenhouse gases, the Earth’s mean near-surface temperature paused its rise during the 2000–2010 period1. To explain such a pause, an increase in ocean heat uptake below the superficial ocean layer2, 3 has been proposed to overcompensate for the Earth’s heat storage. Contributions have also been suggested from the deep prolonged solar minimum4, the stratospheric water vapour5, the stratospheric6 and tropospheric aerosols7. However, a robust attribution of this warming slowdown has not been achievable up to now. Here we show successful retrospective predictions of this warming slowdown up to 5 years ahead, the analysis of which allows us to attribute the onset of this slowdown to an increase in ocean heat uptake. Sensitivity experiments accounting only for the external radiative forcings do not reproduce the slowdown. The top-of-atmosphere net energy input remained in the [0.5–1] W m−2 interval during the past decade, which is successfully captured by our predictions. Most of this excess energy was absorbed in the top 700 m of the ocean at the onset of the warming pause, 65% of it in the tropical Pacific and Atlantic oceans. Our results hence point at the key role of the ocean heat uptake in the recent warming slowdown. The ability to predict retrospectively this slowdown not only strengthens our confidence in the robustness of our climate models, but also enhances the socio-economic relevance of operational decadal climate predictions.
Not too surprisingly ClimateProgress has a post New Study: When You Account For The Oceans, Global Warming Continues Apace about the paper.
The abstract suggests that the tropical Pacific and Atlantic Oceans are responsible for 65% of warming of global ocean heat content for the depths of 0-700 meters since 2000. However, the much-adjusted NODC ocean heat content data for the tropical Pacific (Figure 1) shows a decline in ocean heat content since 2000, and the ocean heat content for the Atlantic (Figure 2) has been flat since 2005.
Figure 1
###########
Figure 2
The abstract also mentions a new-found ability to predict slowdowns in warming. But the warming of tropical Pacific ocean heat content is dependent on the 3-year La Niña events of 1954-57, 1973-76 and 1998-01 and on the freakish 1995/96 La Niña, Figure 3. And the warming of sea surface temperatures for the Atlantic, Indian and West Pacific oceans, Figure 4, depends on strong El Niño events.
Figure 3
###########
Figure 4
CLOSING
Can Guemas et al (2013) can predict 3-year La Niñas and freakish La Niñas like the one in 1995/96? Can they predict strong El Niño events, like those in 1986/87/88, 1997/97 1997/98 and 2009/10? Both are unlikely—the specialized ENSO forecast models have difficulty projecting beyond the springtime predictability barrier every year.
FURTHER READING
For further information about the problems with ocean heat content data, refer to the post Is Ocean Heat Content Data All It’s Stacked Up to Be?
And for further information about the natural warming of the global oceans, see “The Manmade Global Warming Challenge.”
News Flash: A new era in clairvoyant prediction is upon us. What used to be back alley business has now bloomed into Wall Street penthouse business suites thanks to recent research proving the accuracy of Retrospective Prediction. The Enquirer has now overtaken the New York Times in prominence and believability. Our future is safe because of the science behind the most important discovery since penicillin.
These authors must be so proud, so proud.
@richardscourtenay Hmm, so my comment is ridiculous! Interesting! I’ll say no more.
@Gary Young I’ll attempt to respond as your response is at least reasonable.
You say
Being able to tell you where you’ve been when you know where you’ve been is a really small accomplishment — absolutely required if you hope to predict where you’re going, but nothing to brag about.
Well, this seems to be the point that I think most are missing. If you apply this process properly then you use your model without “knowing where you’ve been”. Imagine I have a model that I think can predict the future evolution of something (let’s say surface temperatures) using past data. The year is 2013 and I have data from 1950 till today. If I only consider data from 1950 till 1990 and then use my model to predict what will happen between 1990 and 2013, this will give me an indication of how well my model works. The point is, if this is done properly, the model doesn’t know what happened between 1990 and 2013. I don’t see why this is in any way an unreasonable thing to do. I also don’t quite see why the term “retrospective prediction” also isn’t a reasonable way to describe such a process. I’m not specifically commenting on this paper or whether or not the authors applied this correctly or whether or not there results suggest their model works well. I’m simply commenting that it seems odd that so many comments are mockingly dismissive of something that is fairly common practice in many areas of science.
It’s not whether an explanation is ad hoc that matters, or whether it is intended to bolster a “retrospective prediction.” What matters is whether it is well-supported by the evidence. If the evidence does not support oceanic warmth hiding, then that’s what counts. Don’t pounce on the explanation for being ad hoc, because ad hoc is no vice.
If their model now ‘works’ then predicting future events should be a doddle . . . so where are the REAL predictions?
It’s all academic really, because according ot the BBC News this morning the increasing Co2 in the atmosphere is effecting the wind, so it will make air travel far TOO expensive for the ordinary person to even contemplate flying anywhere in the near future.
Presumably a side effect will be the impact on wind turbines too! (probably why they never reach their stated maximum output – air too thick with CO2).
Another problem with “retrospective prediction” is that it allows the authors to smuggle in the term ‘prediction’. But they are using models and models cannot predict. They should use a term along the lines of “retrospective tuning.”
Bob Tisdale says:
April 9, 2013 at 6:08 am
> Hmmm. Lots of people jumping on the term “retrospective prediction”.
> Seems to me it’s the same thing as a hindcast.
Ah yes, that was the term I was trying to recall. People haven’t squawked about that much.
To Paul Homewood,Richard Verney and others:
About 90% of back radiation does go into the ocean.To correctly discuss this heat transfer problem one must first formulate the surface energy balance,for which it is sufficient to assume that the LWR is absorbed at the interface.See, for example,Mills “Basic Heat and Mass Transfer” Eq.9.54. When properly evaluated it will show that the effect of the LWR is to reduce the heat transfer(loss) from the water to the interface—thereby “heating” the ocean.
“Retrospective prediction”
yeah, let’s see them try to get away with that at the blackjack tables in Vegas.
But really, I knew I was going to get that hand, so you should let me double my bet.
How do the graphs at this site square with those presented by Bob Tisdale above?
/Mr Lynn
I have a question I need help with:
The paper states that at the top of the atmosphere there is a net energy imbalance of +0.5 to +1 W/msq. If this is true, and accurate, then this implies to me that whatever redistributions of heat may be happening within the system, (oceans etc), the system itself is failing to emit as much energy as it gains. Which in turn implies that by worrying about how this translates into observed temperature, or how heat is distributed in the oceans and atmosphere, we are missing the point. The system overall is still gaining energy, which surely can’t be sustainable without unpleasant consequences?
izen says:
“The value of this paper is the finding that physics that researchers use to explain how thermal energy is distributed in the climate can ‘retrospectively predict’ the actual observed data. That helps validate the physics used to analyse the climate.”
No, it doesn’t.
At best, it provides a hypothesis as to how a part of the climate system might work. Validating that hypothesis requires testing that hypothesis. That is accomplished by making rigorous predictions that prove out. Actual predictions. You know, of the future.
This new “ocean heat” prophecy of the global warming religious movement is simply a reboot of the failed surface temperature portent. For the last twenty five years, we have had hindcast based hypotheses about the effect of CO2 on surface temperature thrown at us by people who were claiming that those hypotheses “validate the physics used to analyze the climate”.
“We have these models”, we were told, “and they accurately hindcast the surface temperatures – but only when we include the awesome power of the God Compound, CO2. This proves that our eschatology is the One Truth.”
And we kept hearing that crap, even as the actual predictions made by the methods of those hindcast-based hypotheses began to fail. Miserably.
Doomsday was foretold. Great Trials and Tribulations! Repent!!! Purchase your Climate Indulgences before Judgement Day! But it did not happen.
A decade of no warming was shrugged off, and the High Priests of the Church of the IPCC told us that the augurs had assured them that under no circumstances would Lord CO2 forsake his promise of warming the believers for so long as 15 years. As 15 years of no warming neared, Reverend Santer reread the entrails, and pushed out the Holy Hiatus to 17 years. Now that many surface temp datasets (despite the fiddlings of the faithful) are but a few months shy of 17 years of no warming, Pope Kevin I has placed the Second Coming of the Warmth “20 years or more” into the future.
Meanwhile, the climate equivalent of the Great Disappointment is beginning to be interpreted by some warmists in exactly the way that the Millerites dealt with their crisis of faith – the predicted Second Coming of the Heat did happen, just as we predicted! It just happened on an invisible plane! His Warminess is there! Just hidden from the sight of the mortals, down in the third decimal place of the Great Deep. The true believer KNOWS this, because:
“We have these models and they accurately
hindcast‘retrospectively predict’ thesurface temperaturesocean heat content – but only when we include the awesome power of the God Particle, CO2. This proves that our eschatology is the One Truth.”Freaking witchdoctors. Learn how science works. Hint: this ad hoc gnostic bullshit ain’t it.
Models are hypothesis. Models don’t validate anything. Models get validated. Sometimes. Other times, they fail. Make some actual predictions with your “new and improved” models (Now, with Ocean Heat Accounted For!), and phone us back in thirty or forty years to tell us how well ya did. And do it on your own dime, thanks.
A “retrospective prediction” only makes sense when you divide the data you have in a training set that you use for training your model and a validation set that MUST NOT be used for training the model but only for validating whether your model has any skill.
I don’t know whether they did this and am not inclined to buy the paper. If they didn’t, the validation of their model is still missing and therefore it should be considered conjecture.
Retrospective prediction. Imagine if they had worked “teleconnection” in there too.
I keep coming back to the bouys. In order for the heat to get to the 700-2000m layer, it has to pass through the 0-700m layer first, and the bouys would see it. The bouys just haven’t seen it…
FrankK says:
April 9, 2013 at 12:19 am
“What is probably meant by “retrospective prediction” is actually “calibration” or “validation”. ”
No, calibration and validation are two separate things. We know they calibrated – you always do when training a model – ; what we don’t know is whether they made any effort at validation; and we also don’t know if any such effort, if undertaken, was honest (think double blind trials – did they REALLY not use their validation data during the training).
Hobbit says:
April 9, 2013 at 7:40 am
“I have a question I need help with:
The paper states that at the top of the atmosphere there is a net energy imbalance of +0.5 to +1 W/msq. If this is true, and accurate, ”
Do you think they measured that? Why do you think this? How do you think could they have measured it?
Usually these imbalance claims come, as amazing as it sounds, directly out of climate models – yes, climate scientists have no shame, and think their audience is as dumb as a box of bricks.
You know, they make a computer say this – there’s an imbalance. From there the next step is to get funding.
If izen or any other warmist can prove me wrong by showing the measurements that show such an imbalance, please bring it on.
So… after this analysis the climate models will be adjusted to take into account heat going in to the ocean. Good. And if the air temp goes down then the models can be adjusted to increase the heat going into the ocean. And if the air temp starts to climb again the models can be adjusted to account for that too. But if we just stay on this plateau we make our adjustment and stick with it. I get it. And if the ocean heat data conflicts in anyway some other adjusted metric can always be waved at to justify the process. Sea level rise or fall for example. It’s almost magical how that ocean sops up or releases the extra heat to support the theory. You can’t argue with magic. Especially retrospective magic.
By the way, this has encouraged me to now know that my bracket prediction of Kansas winning the NCAA tournament was actually robust. No more busted brackets for me. I won’t go into all the details, but it’s clear that with some proper retrospective adjustments to scoring and defensive statistics Kansas actually won the tournament. Too bad for you Louisville.
Funny how people ignores (willfully) the role of the sun.
To the left of centre:
My post at April 9, 2013 at 4:31 am explained why your post at April 9, 2013 at 3:43 am is ridiculous, pseudoscientific nonsense.
You have replied to that at April 9, 2013 at 6:34 am without referencing my post and by saying in total
I understand why you say nothing in response to my explanation. And everybody who reads that explanation will understand, too.
I write to thank you for providing me with this opportunity to draw attention to my explanation of your ridiculous nonsense.
Richard
RoHa’s Retrospective Prediction Company.
Retrospective predictions GUARANTEED 100% accurate !!!
Not so fast. Mann can’t even get retro-active preditions right. He “predicts” no medieval warm period, which is pretty damn robust.
@Hobbit Exactly. If the measured energy imbalance is correct, then we are gaining something like 0.5 Joules per square metre every second. If correct, then this energy has to be going somewhere (a fundamental property). It can heat the oceans, melt polar ice, increase surface temperatures. If the oceans appear not to be gaining heat, if the ice appears not to be melting and if the surface temperatures appear not to be rising then either our measurement of the energy imbalance is wrong, or our understanding of where this energy is going is wrong. I can’t really see an alternative.
Of course there are many who argue that the ice is indeed melting, the oceans are indeed gaining energy and we can’t yet tell if the surface temperatures have risen since the mid 1990s because of the natural variations in the temperature anomaly data. If so, then maybe we already know where this energy is going.
JJ says:
April 9, 2013 at 7:40 am
Hear!! Hear!! Well put!!
“To explain such a pause, an increase in ocean heat uptake below the superficial ocean layer 2, 3 has been proposed to overcompensate for the Earth’s heat storage.” However, Bob has shown that the temperatures have not increased. If the heat uptake has been traced to the oceans and the temperatures have not increased, where did the heat go? I suggest the invisible elephant in the room is chemical storage of heat in the ocean. The climate analysis world has disregarded the effect of endothermic reactions on the heat content of the oceans. There are several chemical processes including the production of hydrocarbon compounds, calcium carbonate, hydration of calcium carbonate and hydration of other chemicals including methane that remove heat in the form of thermal energy and convert it to chemical energy. Chemical energy stored in the ocean is removed from the active climate system.
DirkH,
Thanks for the response. I think I recall Dr Spencer saying in his book that they have satellites measuring the energy imbalance at the top of the atmosphere, but I may be wrong. Is that not the case? Naturally if this is an output from computer models, as well as an input, then there is a problem with their approach…!
HINDSIGHT: The Journal of the American Retrospective Prediction Society
Coming in the next issue:
* Late season snows predicted for Northeast USA in Spring 2013 by novel weather model.
* A financial analysis method that would have predicted the banking crisis of 2008 as early as 2006.
* A new model of thalidomide’s mechanism predicts widespread birth defects if given to large numbers of pregnant women.
* Re-analysis of archived British intelligence documents shows Hitler actually was preparing to invade Poland in 1939.
I will be impressed when they can make a
projectionprediction about future standstills and get it right because right now it’s a case of we have a robust climate model because we can now predict the past. Climastrology.