## Watt about the ensemble of models?

Robert G. Brown (who posts as rgbatduke) has a new “sticky” post at Watts Up With That (WUWT) called the ensemble of models is completely meaningless, statistically.

What Robert G. Brown says in this post is

by forming a mean and standard deviation over model projections and then using the mean as a “most likely” projection and the variance as representative of the range of the error, one is treating the differences between the models as if they are uncorrelated random variates causing > deviation around a true mean!.

What I take this to indicate is that if you have a set of different models each with a most-likely projection, you can’t simply add them all to get a mean (which you interpret as the most likely result) and then use the variance to determine the error in the model estimates. Indeed, I would agree with this. This is not the correct way in which to determine the most likely result and in which to determine the likely range.

What would be the correct procedure? Well, as far as I understand it you would normally consider a single model. That single model will have various calculations associated with the known physics and chemistry, and various parameters that have a range of uncertainty. A single run of the model will produce a single result for a given set of parameters. If the parameters are uncertain (as they almost certainly are) then one would normally run the model many times with different choices of parameters. One can then combine the different results to get a mean and to get the range of likely results. Of course, some parameter values are more likely than others, so some model results are more likely than others and so when combining the models results, this has to be taken into account.

When I look at the typical figure from the IPCC’s AR5 leaked report – which is shown below – I see 4 different models, each with a range which indicates the uncertainty in each model’s predictions. I don’t know for certain that they’ve done the mean and error calculations correctly, but this does seem consistent with what I would expect.

Comparison of 4 different model scenarios with observed temperature anomaly data (from IPCCs leaked AR5 report)

I don’t know if all climate model predictions published in the literature have done their error analysis correctly (it is something that scientists do get wrong at times) but I’m unaware of an example where they’ve simply added together the mean results for a large ensemble of different models. I have seen this somewhere though. Where was it? Oh yes, it was in a WUWT post by Roy Spencer called Climate modelling epic FAIL – Spencer: the day of reckoning has arrived. In this post Roy includes the figure below that appears to be simply an ensemble of 73 model runs which have been averaged to get a mean and in which it is implied that the range of the different results gives an indication of the error.

Results from 73 different models compared with RSS and UAH mid-troposphere measurements (credit : Roy Spencer, WUWT).

I think I agree with Robert G. Brown that doing what Roy Spencer has done so as to compare measurements with model results is an “abuse of statistics“. It’s good to see WUWT including posts that criticise bad practice on their own pages. To be honest, I would quite like to see more of this.

This entry was posted in Anthony Watts, Climate change, Global warming, Roy Spencer, Watts Up With That and tagged , , , , , , , . Bookmark the permalink.

### 60 Responses to Watt about the ensemble of models?

1. BBD says:

The figure from AR5 shows the composite model estimates from all four previous Assessment Reports: FAR, SAR, TAR (first, second, third) and AR4. IIRC 21 models were used in AR4.

2. I did wonder if I was going to get caught out on this one. I looked it up briefly and thought that maybe it was a single model, but didn’t read very deeply.

Okay, so maybe that figure is produced in a way that isn’t what I thought. Would be interesting to know how they combined the models and how they get the range. One would hope that some kind of proper error analysis was performed when making the composite, rather than simply adding the means from each model and then using the range of the means to determine that range of the composite. I’ll have to do some reading to see if I can work it out.

3. BBD says:

There’s some insight into the IPCC process here, but I have a haunting feeling there is another, more specific AR4 WG1 reference that is eluding me.

4. Okay, so it seems that they do calculate the “mean of the ensemble” which according to their text, outperforms individual best-estimates. Hence my use of the AR5 figure as an example of doing what I thought was the correct approach was wrong (got to put my hand up to that I guess). What’s unclear is how they get the range/error of the multiple-models. It does say

By sampling modelling uncertainties, ensembles of AOGCMs should provide an improved basis for probabilistic projections compared with ensembles of a single model sampling only uncertainty in the initial state

so it seems that calculating the range is not simply from the variance in the individual model means, but is done by considering the modelling uncertainties.

I guess it also depends on how the multiple different models vary. Can they be perceived as models with the same physics and chemistry (as one might hope) that survey different regions of parameter space. Even so, simply taking the average of all the model means still seems a little simplistic as the mean of one model may be more likely than the mean of another. If their calculation of the mean of the ensemble and the range actually includes the uncertainties in the parameters used in each model, it may still effectively be consistent with the approach I had thought was most suitable. If not, then maybe it is a little simplistic but if the range is done in a manner consistent with proper error analysis, it may still be an appropriate approach. I don’t actually know the answer to this, but – in my experience – the chance that a large group of experienced scientists have made some fundamental error in how they combine different model results is probably quite low. Doesn’t mean that it can’t have happened, simply that it is unlikely.

5. Both single model and multi-model ensembles are used. The figure you showed, likely shows 4 different multi-model ensembles. The IPCC likes multi-model ensembles as they summarize a lot of information in just one figure.

A single model ensemble will likely underestimate the uncertainty in the climate projections. Even if you vary model parameters (knowing over which range is extremely difficult), there are still processes not included in the model or simplifications of processes that lead to too little variability.

Even a multi-model ensemble is expected to underestimate the uncertainty. For example, most models have a very similar modelling of convection (mass flux scheme), which does not take into account that also vorticity is transported upwards. None of the models has ground water hydrology with lateral flows. And it looks like all models are not that good in predicting the Arctic sea ice.

You should not interpret the mean and the variability of a multi-model ensemble without thinking. I think I saw a comment by this WUWT guest poster and he is a physicist (like I am/was myself). In physics you have well defined concepts, “simple” laboratory experiments and consequently crisp answers. Answers about which you do not have to think any more; although one of the most common errors of students is that they really do not think any more and do not notice that they made an error somewhere and answer is beyond anything.

Dealing with the complexity of reality you have to find the best concepts, you have to simplify the computation, but you cannot simplify the reality itself. Thus you never know whether you have overlooked something and you should carefully interpret the answer.

Single- and multi-model ensembles are also used in weather prediction to get an idea about the uncertainty. The advantage here is that you can validate the prediction and its uncertainty every day. Consequently, in weather prediction it is possible to numerically calibrate the prediction and its uncertainties.

6. Victor, thanks. That’s an interesting comment. I very much agree with your statement that one shouldn’t interpret the mean and variability of a multi-model ensemble without thinking. It does seem as though many forget that climate modelling is essentially modelling a very complex system and that analogies with simply laboratory experiments are not always that appropriate. As you say, the range of the models (or the error) is an estimate of the uncertainties, but the models may well have missed something that could influence the result.

It is a great pity that much of the criticism of this modelling relies on very simplistic comparisons between model results and observations, over what are essentially short timescales. It’s unfortunate hat there isn’t an attempt to appreciate that these models are trying to represent a very complex system.

7. Sou says:

The draft from AR5 is not very good in that it starts off at a single data observed temperature rather than being aligned with the observed trend. Tamino wrote something about that. Here it is:

http://tamino.wordpress.com/2012/12/20/fake-skeptic-draws-fake-picture-of-global-temperature/

There is a good reason for keeping the IPCC drafts confidential, because they tend to get misused and in any case are subject to change.

8. Skeptikal says:

I did wonder if I was going to get caught out on this one.

I did have a giggle reading this post.

Hence my use of the AR5 figure as an example of doing what I thought was the correct approach was wrong (got to put my hand up to that I guess).

At least you are willing to admit to making a mistake… gives you more credibility than someone who tries to talk their way around a mistake.

Here’s something for you to consider, look at Roy’s graph… it shows the individual model outputs as well as the multi-model mean. Look at the inconsistencies between the models. There’s only two things consistent in all the models… they all predict a warming world and they’re all inconsistent with observations.

9. BBD says:

Skeptikal

At least you are willing to admit to making a mistake… gives you more credibility than someone who tries to talk their way around a mistake.

10. Yes, I’ve read Tamino’s stuff about this and generally agree. There was every chance that the final report would have produced a more appropriate representation. Now that the draft figure is out there, any changes will be regarded as cheating.

11. Skeptikal says:

It is a great pity that much of the criticism of this modelling relies on very simplistic comparisons between model results and observations, over what are essentially short timescales.

If the models can’t produce anything meaningful on a short timescale, then longer timescales will only exaggerate the errors present in the short timescale.

You say; “simplistic comparisons between model results and observations”. These models are predicting a future temperature…. what else can you compare it to besides actual temperatures?

12. BBD says:

Wotts

I don’t actually know the answer to this, but – in my experience – the chance that a large group of experienced scientists have made some fundamental error in how they combine different model results is probably quite low.

That’s uncannily close to the way I view these things: experts know what they are doing. The clue being in the name etc. Clearly neither of us has what it takes to make it as a leading light in the “sceptic” blogosphere.

13. Skeptikal says:

When the world starts warming again. 😉

14. BBD says:

If the models can’t produce anything meaningful on a short timescale, then longer timescales will only exaggerate the errors present in the short timescale.

Nope. Wrong again.

15. BBD says:

You mean when OHC 0 – 2000m has decreased for at least a decade.

You and those pesky basics!

16. Skeptical, always good to admit one’s mistakes. Having said that it is still not clear that the AR5 figure is quite as simple as the figure presented by Roy, for example. The error analysis may be more complicated than simply looking at the variation in the means of the models used to create the ensemble.

There are two issues I have with Roy’s figure. One is that there aren’t any errors. Simply comparing with the model means isn’t really correct. The second is that his satellite data is an average of the RSS and UAH data. The trends in these two datasets differ (I believe – have meant to check but haven’t done so) by a factor of 3. The RSS trend is 0.09oC per decade while the UAH is 0.03oC per decade (the errors are about 0.028oC per decade). Clearly the UAH data does not compare at all well with the models. This, however, isn’t as true for the RSS data. Now, I don’t know which of these datasets is correct (if either) but it seems that the discrepancy may not be as great as Roy is indicating in his figure.

17. Skeptikal says:

No… I mean when the global thermometer starts moving again. You know, the thermometer which the IPCC predicts will rise… but which those pesky observations show are not rising.

18. @Skeptikal. Our weather prediction for two weeks in advance is very bad, but I can predict that this Christmas it will be colder.

That is sufficient to disproof your simplistic argument.

19. BBD says:

Skeptikal, you clearly don’t understand the basics.

Most of the energy accumulating in the climate system as a result of radiative imbalance caused by an increase in the amount of CO2 is in the oceans. Variability in the rate of atmospheric warming is neither unexpected nor likely to affect the long term (multi-decadal to centennial) trend.

The troposphere is not the climate system. Basic # 1!

OHC is increasing as expected: OHC 0 – 2000m. The red curve is the three month average. Look at it closely.

Natural (?) variability in ocean heat uptake affects the rate of surface warming. You cannot run around pretending that “AGW is falsified” on the basis of natural (?) variability in OHU. That would be very silly.

20. Skeptikal says:

There are two issues I have with Roy’s figure. One is that there aren’t any errors. Simply comparing with the model means isn’t really correct.

Just looking at the graph… you can see a clear and widening gap between the multi-model mean and the UAH/RSS composite. Error bars wont fix this. The global temperature needs to pop up and fairly soon or the gap is going to reach a size where the models become untenable.

21. Skeptikal says:

Variability in the rate of atmospheric warming is neither unexpected nor likely to affect the long term (multi-decadal to centennial) trend.

You really haven’t got a clue. If you can’t get things right on a decadal timeframe, what hope have you got for getting things right on a centennial timeframe?

“Trust me, the world is going to turn into a pressure cooker in a hundred years because my model says so” just doesn’t cut it in the real world. When the models can accurately predict on a decadal timescale, then and only then will the models be taken seriously.

22. BBD says:

The global temperature needs to pop up and fairly soon or the gap is going to reach a size where the models become untenable.

You mean like this?

It’s not the models that will “become untenable”, it’s all known paleoclimate behaviour and a serious chunk of the laws of physics.

Like Wotts, I’m not a betting man, but if I had to lay my wager, I’d bet on consistency in the laws of physics. I’d bet that modern climate will behave as climate has behaved before. I’d bet that the “sceptics” are wrong.

23. BBD says:

You really haven’t got a clue. If you can’t get things right on a decadal timeframe, what hope have you got for getting things right on a centennial timeframe?

If you don’t understand that natural variability hasn’t just stopped then you will remain confused by what you are seeing. If you accept that it continues and will continue, you will easily be able to reconcile observations with the projected long-term warming trends under various forcing scenarios.

Going on as you do just emphasises what is missing from your understanding of the topic.

24. Sou says:

Only by the 8% dismissives. And if it wasn’t changed the disinformers would tout it as an “error”. There is no “win” with deniers. They twist whatever the facts are (as seen in this thread).

The next IPCC report may well have some errors, but going by past reports, the number of errors is likely to be miniscule compared to the amount of information they contain. The main problem is that of compromise and conservatism. It overcomes compromise to a large degree by reporting as much as is practicable. The conservatism is built into science itself and ordinarily wouldn’t be a bad thing. But it can lead to a false sense of security.

25. BBD says:

When the models can accurately predict on a decadal timescale, then and only then will the models be taken seriously.

This is a standard contrarian misrepresentation. It richly deserves a smack-down, and who better to deliver it than that notorious model-sceptic, James Hansen?

TH: A lot of these metrics that we develop come from computer models. How should people treat the kind of info that comes from computer climate models?

Hansen: I think you would have to treat it with a great deal of skepticism. Because if computer models were in fact the principal basis for our concern, then you have to admit that there are still substantial uncertainties as to whether we have all the physics in there, and how accurate we have it. But, in fact, that’s not the principal basis for our concern. It’s the Earth’s history-how the Earth responded in the past to changes in boundary conditions, such as atmospheric composition. Climate models are helpful in interpreting that data, but they’re not the primary source of our understanding.

TH: Do you think that gets misinterpreted in the media?

Hansen: Oh, yeah, that’s intentional. The contrarians, the deniers who prefer to continue business as usual, easily recognize that the computer models are our weak point. So they jump all over them and they try to make the people, the public, believe that that’s the source of our knowledge. But, in fact, it’s supplementary. It’s not the basic source of knowledge. We know, for example, from looking at the Earth’s history, that the last time the planet was two degrees Celsius warmer, sea level was 25 meters higher.

And we have a lot of different examples in the Earth’s history of how climate has changed as the atmospheric composition has changed. So it’s misleading to claim that the climate models are the primary basis of understanding.

26. Skeptikal says:

Oh look… a smack-down from a climate activist who’s been arrested how many times?

27. Skeptikal says:

Our weather prediction for two weeks in advance is very bad, but I can predict that this Christmas it will be colder.

Colder than what?

28. Skeptikal says:

Now that the draft figure is out there, any changes will be regarded as cheating.

I wouldn’t think so. It’s better to have errors detected in a ‘draft’ than in the final report. Considering that the IPCC is basically funded by taxpayers around the world, an open and transparent process would be better for their image… as would a final report free of errors.

29. Whatever.

Colder than today, colder than this month, colder than in summer.

And before you ask it: in Europe or the USA.

30. What I meant by short timescale was two-fold. One was that trying to model the complex variability in the climate (rather than the local weather) on short timescales timescales is difficult. For example, we know that 90% of the excess energy is going into the oceans, while only 10% is heating the land and atmosphere. A small variation in the amount going into the oceans can have a big effect on the land and atmosphere temperature. That energy hasn’t disappeared, it has simply gone somewhere else and will presumably manifest itself in a different way and probably return to heat the land and atmosphere at a later time. So, a small “error” in how our model treats ocean heating can have a big effect on the land and atmosphere on short-timescales. One average, however, these short-term variations should wash out and the models may well do better. The other issue is that the errors in the temperature anomaly is large if we consider short timescales and so doing comparisons with models can be misleading if one only considers short timescales as the uncertainties are large and comparing mean values can be misleading.

31. Skeptikal says:

Well Victor, it looks like you’ve worked out the seasons… well done!

32. Indeed, I do hope that they produce something this is as error free as possible and that represents the science as accurately and honestly as possible. If I remember correctly, the link that Sou included in her comment is to a post by Tamino arguing that they’ve lined up the observations with the models at a point when the anomaly was particularly high (rather than using a mean value, for example). This has made the comparison look particularly poor when, in fact, it’s not nearly as bad as the figure appears to show. One could therefore argue that they should correct this in the final draft, and I hope they do. However, if the final figure makes the comparison look better than the figure that was leaked, I imagine many will be throwing around accusations of cheating.

33. Skeptikal says:

One average, however, these short-term variations should wash out and the models may well do better.

You’re putting a lot of faith in a model having long term ability to predict when short term ability isn’t really there.

The other issue is that the errors in the temperature anomaly is large if we consider short timescales…

The warming which created all this fuss can be considered ‘short timescale’, yet you don’t seem concerned about possible errors in that.

34. BBD says:

Gentlemen, we have a troll.

35. Yes you finally got it. And seasons are a change in the radiative balance, just like global warming.

Just because you cannot accurately predict the short-term details, be it the weather in two weeks or the climate in two years, does not mean that you cannot long-term changes. On the contrary, long-term changes are much easier as short-term changes. The longer the period, the more data you have (less measurement noise, weather and climate variability) and the stronger the signal is.

36. No, not that much faith in the models. As BBD has been trying to point out, global warming is about more than simply an increase in global surface temperature. It is an increase in energy going into the climate system. This is measured (both directly by satellites and indirectly by looking at surface temperatures and ocean heat content). It is clear that this is happening (i.e., the total energy is increasing). What’s more (and this isn’t stressed enough) is that climate models (unless I’m mistaken) actually estimate this energy imbalance correctly. So, they get the energy imbalance right but don’t always correctly determine where this energy goes. However, to remove this imbalance will require an increase in surface temperature. So my reason for being reasonably confident that surface temperatures will rise is not because the models predict it, but because there is no other known way (given the laws of physics) to remove this energy imbalance without this happening (unless we can somehow suddenly remove 100ppm of CO2 from the atmosphere).

37. > the thermometer which the IPCC predicts will rise

Citation needed.

38. Skeptikal says:

However, to remove this imbalance will require an increase in surface temperature.

There it is… something we agree on. What doesn’t make sense to me (and why I’m sceptical) is that temperatures have stalled. Considering that CO2 levels are still increasing, this shouldn’t happen. Logic tells me that there has to be a temperature imbalance between the ocean and the atmosphere for the ocean to accept heat… this means that the ocean can never take more heat than a percentage of what is available in the imbalance. Therefore, ocean absorbing heat could slow, but not stop, surface warming. Can you see that?

39. Skeptikal says:

40. Marco says:

Somebody who wants to have some fun can ask John Christy (whose graph Roy Spencer posted) how the RSS and UAH data set compare. I have a hint: they differ by a factor 3 in trend. You can also ask him to add the NOAA analysis of TMT tropical trend; I have a hint for that, too: its trend is 4 times larger than the UAH trend. Finally, ask him scientific justification of chosing rcp8.5, the most extreme scenario considered in any climate projections.

I’m not going to do it, because I am not interested in the misdirection that will follow with absolute certainty (if not outright abuse for daring to question John Christy or Roy Spencer). But “Skeptical” could try and show us his moniker actually is a reasonable description of his mindset…
(yeah, yeah, I know, I’m delusional for even suggesting this).

41. I don’t why you think the latter. Somebody else is welcome to correct me, but the energy enters as radiation. Some is absorbed in the atmosphere (infrared and UV primarily I think). The rest makes it to the surface. So when that radiation hits the surface it is absorbed or reflected. From what I’ve seen, the oceans only reflect 5% of the incoming radiation, so 95% goes in and must be absorbed in some form of energy or another. In a simple sense, one could argue that given that 70% of the surface is ocean, one would expect 0.95 x 70% of the incoming excess radiation to be absorbed by the ocean and 30% by the land and atmosphere. That’s very crude and ignores any conduction that can transfer energy from the atmosphere to the ocean and vice versa. Given that the atmosphere is likely to be warmer than the oceans, I might expect there to be a tendency for energy to go from the atmosphere to the ocean. So, getting to 90% going into the oceans seems reasonable.

As far as there being a plateau is concerned, I think one has to be a little careful. The uncertainty in the temperature anomaly since 2000 is large, so we can’t actually say with certainty that it has stopped rising. However, the evidence suggests that the rate has slowed compared to some longer term average. However, I’m still unclear why this is necessarily such an issue. If one considers the process, then there is expected to be some long-term trend (due to the average rate of increase driven by the energy balance) and shorter term variations. Some variations (say driven by an ENSO event) could push the surface temperatures above what they would be had they only been responding to the long-term warming trend. Hence one would expect the trend to appear slower until the long-term warming trend “catches up” with the warming due to this shorter term variation.

I guess what I’m saying is that it’s not that ocean warming has stopped surface warming, it’s that some event (ENSO for example) has enhanced the surface warming on a short timescale and (given that this will reduce the ocean heat content) presumably, at the same time, enhanced the rate of energy going into the ocean. Together this produces what appears to be a reduction in the rate of surface warming, but doesn’t mean that global warming isn’t happening or that the surface won’t start warming again once the ocean heat content has recovered and the underlying warming trend hasn’t “caught up” to the warming driven by the short-term variation.

I’m trying to write this quickly as I have to do something else before heading home, so happy for any misconceptions or errors to be corrected.

42. Indeed, I’ve heard this from a number of sources. Averaging the RSS and UAH data gives an average between the two and makes the fit look very poor despite the RSS trend being quite close to some of the model trends (I hadn’t heard about the NOAA analysis of TMT trend). Also using rcp8.5 means the models have the most extreme forcing. Overall, the choices made in producing that figure have been to make the comparison look as discrepant as possible. Like you, I’m not that keen to actually ask John Christy to explain why he chose to do things the way he did.

43. BBD says:

@ Skeptikal

What doesn’t make sense to me (and why I’m sceptical) is that temperatures have stalled. Considering that CO2 levels are still increasing, this shouldn’t happen.

I repeat… Basic # 1: The troposphere ≠ the climate system!

But that is *exactly* how you are treating it: “temperatures have stalled… I’m a sceptic”. This despite the fact that the climate system is mostly ocean and the measured increase in OHC 0 – 2000m is there for all to see (NODC OHC link already provided upthread).

There is a short, useful overview of this topic at RealClimate with links to relevant new studies, notably Balmaseda et al. (2013).

Logic tells me that there has to be a temperature imbalance between the ocean and the atmosphere for the ocean to accept heat…

Ah. More confusion over the basics.

Basic # 2: A warming atmosphere does not heat the ocean. That is mainly done by sunlight (DSW). Instead, it slows down the rate at which the ocean cools.

So OHC increases.

Remember that the decadal average for *atmospheric* temperature since 2000 is the highest in the instrumental record.

44. climateprediction says:

Wottsupwiththatblog, I don’t know what you mean when you express concern about “simplistic” comparisons between models and observations, but here’s a comparison which speaks to your other concern with respect to short timescales.

Wottsupwiththatblog, I’d be interested in your assessment of the accuracy of these specific models versus observations over this long timescale. Is this comparison too simplistic?

You seek sympathy for the modelers because the models are so complex. You can pretty much duplicate the mean of the recent global temperature forecasts generated by these complex models by extending the 1977-2007 trendline with a ruler and a piece of paper. Complexity isn’t the problem.

There is a refusal to recognize the obvious, that half the warming during the 1977-2007 was the result of the positive ocean cycle. Extending this trend as if it were all CO2 driven makes no sense. The current observed flat temperatures reflect the negative ocean cycle and based on history, temperatures will remain flat for several more years. If you understand this you can make what has been and continues to be an accurate global temperature prediction without a computer.

Contrary to your “gut feel” mentioned in another post, many if not most of the climate scientists recognize there are significant impacts of ocean cycles on multidecadal global temperatures but they chose to ignore them because they “don’t know how to model them”. I think the real reason is that they don’t like the lower temperature forecasts which result.

And as long as they get sympathy from people like yourself instead of criticism , then their over-forecasts of global warming are likely to continue. This may sound harsh, but Is there any chance that you prefer erroneous forecasts that are far too high ? Your comments about the Hansen forecasts versus observations may be telling.

45. dana1981 says:

There are other issues with Spencer’s graph as well. For example, he doesn’t seem to have any baseline – it looks like he just matches the data and model mean around 1979.

More importantly, the possibility (probability) that the discrepancy is in large part due to a cool bias in the data is not even mentioned. The likelihood of this bias is illustrated in the large discrepancy between the RSS and UAH trends.

Finally, I’m not sure why we should particularly care about the tropical mid-troposphere. Even if the model simulations of tropical TMT warming aren’t great…so what?

46. BBD says:

Is this comparison too simplistic?

GISS model II runs from 1988? Simplistic… oh yes. Meaningless? Getting on for. Deliberately and transparently chosen for rhetorical purposes? Absolutely.

Apart from being laughably crude by modern standards, GISS model II had a sensitivity of 4.2C IIRC.

I’ll leave Wotts to respond to the rest of your comment, but from what I can see the quality of argument doesn’t improve.

47. climateprediction says:

This is a climate model. We haven’t even reached 30 years yet which is often used as climate length. What do you think of the accuracy? You perhaps aren’t a serious commenter.

48. climateprediction, let me answer the second half of your comment. I’ve never seen the figure before, so need some context to respond to that part.

You say,

There is a refusal to recognize the obvious, that half the warming during the 1977-2007 was the result of the positive ocean cycle. Extending this trend as if it were all CO2 driven makes no sense. The current observed flat temperatures reflect the negative ocean cycle and based on history, temperatures will remain flat for several more years.

Here’s the issue I have with that statement (and we may have discussed this before). Ocean cycles do not generate energy, they simply move energy around. It is indeed probably the case that during the period 1977-2007, ocean cycles played a big role in heating the surface. However, the surface temperature today is considerably higher than it was in 1977. The ocean heat content is considerably higher than it was in 1977. If we’ve just entered the negative ocean cycle, why isn’t everything returning to what they were in 1977. Why is the total greater and likely to remain greater. Ocean cycles cannot do that. An ocean cycle cannot act to maintain a high surface temperature. The only way it can remain high is if something in the atmosphere has changed the equilibrium temperature so that the surface temperature needs to be higher today than in 1977 so as to reach equilibrium. Just because we can associate various signals in the data with ocean cycles, does not mean that ocean cycles are what is causing the energy in the system to rise.

That’s why I’m reluctant to criticise the modellers too much. As far as I’m aware they are getting the energetics correct (within the uncertainties) but may not be quite getting where these energy is going at the moment. This is much better than those who invoke ocean cycles, but can’t explain why there is a long-term increase in the total energy in the climate system.

49. Yes, there are clearly issue with mid-troposphere measurements given that there seem to be a wide range of different trends in the different datasets.

I presume from the last bit of your comment you’re suggesting that we shouldn’t be fixating on regions that are only absorbing a small fraction of the excess energy. I tend to agree. It’s clear that small changes in the rate at which the oceans absorb energy can have a big effect on these other regions and so it is not that surprising that there will be times when the models do not match observations all that well.

50. BBD says:

You perhaps aren’t a serious commenter.

?

You are the one trying to make argumentative capital out of an obsolete study using a primitive and obsolete model.

I’m just pointing out the nonsense as it appears.

51. climateprediction says:

I don’t disagree with any of the physics in your response. But you seem to be missing my point. As you say, when energy is added to the system there is a new equilibrium point. But I’m saying that at any given time the system may not be at the equilibrium point. Temperatures, for example, may be pushed from the equilibrium point by forces which don’t add energy to the system. Ocean cycles are an example. They sometimes push temperatures higher than the equilibrium point and lower lower than the equilibrium point. ENSO is an obvious example. Global temperatures are often moved by large amounts – sometimes by 0.5C in a few months. Two El Ninos without an intervening La Nina over a 3 year period can push temperatures, quite far above the equilibrium point. But as you say, the equilibrium point doesn’t change.

So over a period of time, while ENSO is pushing temperatures up and down, the energy level of the system will change because of added greenhouse gases. So if La Ninas exactly offset the effect of the El Ninos over a given period, then the temperatures will return to equilibrium at the end of the period, but it will be a higher temperature than at the start because of the greenhouse gases.

Let’s put some numbers to 2 consecutive 10 year periods. Let’s start with an anomaly of 0.0C. Let’s assume there are only 2 significant factors affecting temperatures, greenhouse gases and Ocean Cycles. Let’s assume that the greenhouse gases increase steadily throughout the 20 years such that the equilibrium temperature grows at the rate of 0.1C every 10 years. So the equilibrium anomaly in year 20 is 0.2C.

Now let’s assume that the first 10 years is a period where there are more El Ninos than La Ninas and the El Ninos are stronger. Lets also assume that the PDO is significantly positive and overall the Ocean Cycle is in its warm phase throughout the period which pushes temperatures up by 0.1C. So the anomaly at the end of year 10 is 0.2C. Half is due to the greenhouse gas increase and half is due to the Ocean Cycle.

Then let’s assume that during the second 10 year period the Ocean Cycle reverses and provides a cooling effect which exactly offsets the earlier Ocean Cycle warming effect. This would provide a cooling effect of 0.1C over the second 10 year period which offsets the greenhouse gas warming effect.

Over the 20 year period, the Ocean Cycle would have had no net effect on the anomaly. It would have had no net effect on the system energy.

At the end of the 20 year period, the anomaly would be 0.2C, the same as at the end of the 10 year mark. The rise over the 20 year period would be totally due to the greenhouse gases.

But if a forecaster were to be unaware of the Ocean Cycle Effect or doubt its strength or otherwise choose to ignore it perhaps because of modeling difficulties, they might look at the anomaly of 0.2C in the 10th year and proceed on the basis that it was all due to greenhouse gases. They then might project the anomaly in year 20 to be 0.4C based upon a continuation of the perceived greenhouse gas effect.

Thus in this example the forecaster would have over-estimated the anomaly by a factor of 2. They would not have predicted the flat temperatures of the second 10 year period.

I don’t think the above scenario breaks any laws of physics. Do you?

.

52. wotts
Given what’s happened in this thread and the pattern, I predict this entire blog will be going down the drain in a mere matter of days to a couple of weeks, if the same thing persists

53. What’s happened in this thread? There’s been a bit of the snark that I’ve been trying to avoid, but I didn’t expect to avoid it altogether. Or, are you referring to something else?

54. Marco says:

Probably referring to too much factual discussion… 😉

55. climateprediction, yes I agree with most of what you’re saying. Let me add some caveats though. In one of your paragraphs you say

Now let’s assume that the first 10 years is a period where there are more El Ninos than La Ninas and the El Ninos are stronger. Lets also assume that the PDO is significantly positive and overall the Ocean Cycle is in its warm phase throughout the period which pushes temperatures up by 0.1C. So the anomaly at the end of year 10 is 0.2C. Half is due to the greenhouse gas increase and half is due to the Ocean Cycle.

The scenario you present is that GHGs are increasing the equilibrium temperature smoothly at the rate of 0.1oC per decade and La Ninas and El Ninos can produce variations also of 0.1oC per decade.

Let’s imagine that during the first 10 year period, the El Ninos have acted to increase surface temperatures by 0.1oC per decade before the end of the 10 year interval. Well, at that stage the GHG forcing disappears because the surface temperature is above or at equilibrium. It will either remain flat or cool slightly until the GHGs have increased so that the equilibrium temperature is above the actual temperature.

So, that’s where I slightly disagree with your analogy. The GHG forcing produces a change in equilibrium temperature. Ocean currents act to heat or cool the surface, without changing the equilibrium temperature. Therefore, if an El Nino has increased the surface temperature so that it is closer to the current equilibrium temperature, the GHG forcing becomes less significant and the increase in surface temperature slows until the GHG forcing increases again. I guess what I’m saying is that the influence of GHGs and El Ninos/La Ninas don’t necessarily add in a given period so, again, I think what we’re seeing at the moment can be explained as being consistent with GHG forcing together with other variations that can push the surface temperatures closer to equilibrium and hence slow the rate of increase for some period of time.

You really haven’t got a clue. If you can’t get things right on a decadal timeframe, what hope have you got for getting things right on a centennial timeframe?

If you can’t predict how many sixes you wil get in ten rolls of a dice how can you possibly predict how many you will get in 100 rolls of a dice?

57. Oh no, it is not the snark. (please continue whoever’s doing it). I am talking about the people contributing comments on this thread.

58. Based on encounters with them in the past or based on what they’re saying on this thread?

59. acckkii says:

Reblogged this on acckkii.