The skillful predictions of climate science

Smith et al (2007): 0.3°C in 10 years

In 2007, a team of climate scientists from the UK Met Office led by Doug Smith wrote a paper “Improved Surface Temperature Prediction for the Coming Decade from a Global Climate Model”, published in the journal Science.  Although published in 2007, the paper made predictions for the decade 2004-2014. (Presumably the work was started around 2004 and it took some time for the paper to be published). The paper made claims about the “skill” of the model, for example “Having established the predictive skill of DePreSys…

The Smith et al paper made the following specific predictions:

  • There would be 0.3°C warming over the decade 2004-2014
  • At least half of the years after 2009 would be warmer than the record year of 1998.

Note that at that time, 2007, the warmest year was thought to be 1998; subsequent adjustments to the method made 2005 warmer than 1998.

The predictions were spread far and wide. They were included in a Met Office Press release, and a glossy brochure on “Informing Government policy into the future”, with the almost obligatory scaremongering background pictures of black clouds and people wearing facemasks.  Vicky Pope gave a talk on these predictions, saying that “these are very strong statements about what will happen over the next 10 years.”
And of course the faithful media reported the story without questioning it.

These predictions have turned out to be wrong.  We are almost into 2014 and there has been no warming at all since 2004. Of the years since 2009,  none of them have broken the record of 1998 according to HADCRUT3 data. Using HADCRUT4, 2010 is warmer by a meaningless 0.01°C (that’s one tenth of the error estimate). 2011 and 2012 were cooler and it’s now clear that 2013 will be cooler also.

The warming prediction was for  0.30° ± 0.21°C [5 to 95% confidence interval (CI)], so unless we get some significant warming over the next few months it looks as though the observations will be outside the CI of the model.

COP19, Warsaw 2013: pause “expected”

The Met Office had a stand at the recent COP19 climate conference in Warsaw.
Here are two interesting pictures from this (HT Leo Hickman)

Apparently a slowdown in warming is “expected”. Although the Smith et al 2007 paper did say that natural variation would partially offset warming over the next few years, but then went on to make the incorrect predictions above, there was no mention of such an expectation in the Met Office press release or the talk by Vicky Pope (in fact she said “over the next ten years we are expecting to see quite significant changes occurring”).

The second graph is interesting since the axes are unlabelled, failing the first rule of elementary graph-plotting. I asked on twitter what might be plotted against what. Thanks to Gerry Morrow for suggesting that it showed trust in climate science against time. Presumably it is supposed to show the number of times per century that a pause of a certain length is likely to occur, according to climate models.  The text talks of a ten year pause, but the current pause is now more like 15 years, expected only twice a century if the models are correct.

Smith et al (2013): skillful models

A new paper, Smith et al (2013), “Real-time multi-model decadal climate predictions”, has been published in Climate Dynamics. This one has 22 authors, reflecting the fact that it involves many more climate models. The tone is somewhat more cautious than the previous one; for example it says that “We stress that the forecast is experimental”, but it also claims that the models “have undergone rigorous evaluation and individually have been evaluated for forecast skill”.

The paper muddles up predictions and hindcasts, something the Met Office has been criticised for before, for example by claiming that “forecasts of the year 2011 agree well with observations”. Reading the paper, it is not clear when these ‘forecasts’ were carried out.

There is no headline figure of a warming prediction over the next decade, but here is the graph showing the predictions in graphical form:

The uninitialised forecasts seem to start about 0.3C higher than present, while the initialised ones (red curve) show a rise of about 0.3C in about three years. One forecast from the Reading group (green curve) has for some reason been singled out, and shows a more gradual rise (Ed Hawkins tells me that’s because it is a statistical forecast rather than a GCM-based one).

But what is most remarkable about Smith et al 2013 is that there is no attempt to assess the accuracy or otherwise of the previous paper Smith et al 2007, even though that earlier paper is cited, and despite the statement in the Introduction that “Assessing discrepancies between forecasts and subsequent observations can reveal weaknesses in initialization strategies, model simulations of internal variability, model responses to external forcing, and uncertainties in future forcing factors, all of which are invaluable for improving future forecasts”.

Here’s a reminder of the famous one-minute summary of how science works, from a guy called Richard Feynman.

Update: Ed Hawkins says there is a paper in preparation on examining the prediction biases in a version of DePreSys.

45 thoughts on “The skillful predictions of climate science

  1. Hi Paul,

    keep up the good work. Amongst all the noise of the climate debate this is the stuff that shoud be right at the centre of the debate and clearly presented whenever vacuous claims of 95% certainty and the like are trotted out. After all what is the scientific method apart from checking actuals against predictions.

    The alarmist crowd go out of their way to avoid making testable predictions of less than a lifetime and also have huge ranges so they can hardly miss…yet they still do. I dread to think what the debate would be like if they had randomly got some stuff right.

    When they drop their guard and do predict easily checkable stuff like this they should be openly called out on it (hopefully in the MSM) not allowed to seamlessly move on to the next prediction. I work in the horse racing industry and this sort of stuff is standard fare for charlatan tipsters.

    This one, “at least half the years…” plus Jones “15 years (and Santer’s 17 years) of pause will invalidate the models” are easily understandable to the layman. Do you know David Rose, can he pick up on this one?

    As for presenting an unlabelled graph…ye Gods. That’s the Smug Elite for you.

  2. This is a classic statement from the abstract Smith et al. (2013) :

    However, in the absence of volcanic eruptions, global temperature is predicted to continue to rise, with each year from 2013 onwards having a 50 % chance of exceeding the current observed record.

    If you toss a coin it has a 50% chance of being heads or tails. Likewise 2015 say has a 50% chance of being warmer (or colder) than 2014.

    Therefore you also expect 3 times every century to get 5 successive warming years in a row !

    They are really hedging their bets on the validity of GCM models !

  3. Quote: Note that at that time, 2007, the warmest year was thought to be 1998; subsequent adjustments to the method made 2005 warmer than 1998.

    Uh! What does this mean? Some unspecified method is adjusted and global temperatures change? The warmest year was “thought” to be 1998? Does this mean that real world temperature data is simply “thought” into existence?

    Is “Climate Science” based on fractional reserve lending? If so can we expect hyperinflated temperatures as the IPCCs version of QE leaks into the real temperature record?

    The parallels between fiat currencies, banking and AGW are striking.

  4. BillyL,

    those tropical storm forecasts deserve a headline post of their own, particularly after the attempts to make political capital out of the suffering in the Phillipines (different area I know).

    I know Monty Python are reforming, they could use some of the material in there.

    “Recent studies have shown that dynamical models have considerable skill predicting the number of tropical storms – for example successfully predicting the change from the exceptionally active season of 2005 to the below-normal activity of the 2006 season”

    So they got something right 9 years ago. Closely followed by:

    “Last year the Met Office forecast was for 10 tropical storms and an ACE index of 90 with a 70% probability range of 7 to 13 storms and an ACE index of 28-152, respectively. In the event, the number of storms was 17 and the ACE index was 123”

    So as usual a barn door to hit….and missed comfortably. What skill eh?

    And what is it with 70% levels for such a huge range, that just screams “we don’t really know” to me…which the results would seem to show. What is the point of it all? Does anyone act on these forecasts? Anyone? If so are they happy customers?

    Do you know where to find the actuals for 2013 as there is only 1 week to go to see how well they did? Why don’t they show a rolling graph updating as they occur?

  5. That clip of Feynman is priceless. And perfect. Had to share it on my Facebook page to explain to my friends why I keep posting links debunking the posers predicting climatocalypse. Hope they can understand the “essence of science.”

  6. There’s an interesting new paper just out looking at the skill of decadal climate model forecasts. It’s by a postdoc Emma Suckling and Lenny Smith, a maths / time series / economics expert. They seem to be saying that simple empirical models (like basically just predicting the future based on the past) are more accurate than full-scale climate GCMs.

  7. Thanks Billy. Well if I’m reading it correctly, it looks like the Storm prediction isn’t bad at 14 (13 actual) it’s the Hurricane one which is spectacularly wrong this time.

    The forecast itself is totally preposterous though: 70% chance of between 4-14 hurricanes. What possible use is that?! In the history (back to 2004) the Met Office itself shows they’ve ALL been between 3-14. It’s like saying there’s a 70% chance of your newborn eventually growing to between 4 foot and 8 foot. You could do better blindfold playing Pin the Tail on the Donkey. Why don’t their superiors laugh them out of court when they present that useless nonsense (give me the job I’ll do it!).

    So this year they’ve projected an above average 9 and the result is an off the scale low 2 (and both of those only scraped in as the lowest category 1). With ACE they predicted above average130 (massive range again of 76-184) and it was a new low of 33. That’s not just missing a barn door, that’s missing the whole town. We’ve had an extremely rare event and they couldn’t even get the sign right.

    And yet all we get bombarded with is that non-existent Global Warming is causing more extreme weather. WTF! How do they get away with it?

    In the words of Mr Feynmann, “If it disagrees with experiment, it’s wrong”. It’s not hard is it.

    Paul, well done for blogging. Do your best to get all this out there so that the man in the street understands it.

    (all assuming I’m not misreading the stats post-vino, quite possible)

  8. “Curve-matching” is scorned by the professional statisticians and modellers alike. The reason is that knowledge and intellect are supposed to be superior to assuming that the near future will behave like the recent past. So you are a doofus if you expect tomorrow to be somewhat like today, because, as we are told by the IPCC, tomorrow’s physics is Special, and is not informed by the physics of yesterday.

    Looks to me like the curve-matching, the past-is-the-key-to-the-present doofuses should get Professorships and the peer-reviewed professors should get their college education costs refunded.

  9. Hi Paul

    I’m rather bemused by the negative slant you’ve taken on this area of research (eg. calling Smith et al (2013) “ridiculous”). You’ve shown the classic Feynman clip with his very wise words, but actually Doug Smith and colleagues are doing exactly what Feynman is advocating, which is to make testable predictions. They should be applauded for this.

    It’s a shame that you didn’t explain decadal forecasting very clearly. The aim is to forecast changes over the next 10 years arising from all causes, including natural internal variability as well as any long-term forcing. If you are going to cite Feynman saying (correctly) that “If [a proposed physical law] disagrees with experiment, it’s wrong”, you should at least be clear about what “law” you think is being proved wrong.

    Following this through, if a weather forecast is wrong, what “law” does this disprove? The answer is: none. The forecaster has simply been defeated by chaos. If you predict that tossing a coin 10 times will give you 5 heads and 5 tails, and instead you get 3 heads and 7 tails, you’ve not disproved any law, you’ve simply found that random things are hard to predict.

    Natural internal variability of the climate system can be thought of as long-term weather (some use the phrase “macro-weather”). i.e.: it’s a chaotic system. The timescales are longer because much of it relates to ocean processes, which are generally slower than in the atmosphere – so another way of looking at it might be “ocean weather”. The potential for predictability comes from these longer timescales.

    Of course it’s not just about internal variability. On decadal timescales we’d also expect external forcings to have an influence – eg. changes in solar irradiance, greenhouse gas concentrations or aerosol concentrations (either anthropogenic or natural). There’s two uncertainties in this part of the problem: (i) the strength of the net forcing (GHGs and other forcings) and (ii) the response to this net forcing.

    So, while it is true that global mean temperatures have not warmed as predicted by Smith et al (2007) in recent years, it’s hard to say why until we understand more about the various processes. Was long-term warming offset by internal variability? Was the response to external forcing smaller than expected? Was the forcing itself smaller than expected? Or some combination of all 3? Simply saying “it’s wrong” is not how science works – science works by understanding why predictions don’t match reality. That requires taking bold steps to make predictions in the first place.

    It’s a shame that you’ve not cited Smith et al (2012), “What is the current state of scientific knowledge with regard to seasonal and decadal forecasting?”. This gives an open and frank discussion about this emerging area of science and its acknowledged limitations at the present time, along with the progress so far and the reasons for thinking that further progress can be made.

    Decadal climate forecasting is an enormous challenge – but if it can be done successfully then clearly there will be huge benefits. What if we could forecast major droughts, heat waves or cold spells a long time in advance? Imagine the benefits to everyone, from vulnerable people to big businesses. Surely that’s a great thing for science to aim for?

  10. Paul, excellent post. I note you said on twitter you will be responding to Richard Betts later so I’ll hang back and mop up any points you don’t get round to raising concerning his response. There’s certainly plenty to go at. It’s interesting to see for example, that solar is the first external forcing he mentions, yet his organisation still uses spuriously low figures for solar forcing with spuriously narrow error bars and spuriously small uncertainty ranges. There is also a willful misunderstanding and dismissal of the climatic effects of parts of the irradiance spectrum which vary far more than TSI does as a whole.

    The co2 fetishists are deliberately ignoring important climate variables, and evidence of cyclic climatic phenomena which can inform phenomenological models which have a better track record and hindcasting ability than co2 driven GCM’s. Which is why their models don’t work and their reputation is going down the pan.

  11. Richard forecasts from the Met Office that imply global warming is going to accelerate are awaited with baited breath by the myriads of environmentalists and metro-elites who want to impose their world view on the rest of us. If they haven’t broken a law of physics they’ve broken a sacred trust with the people of the UK and elsewhere by being reckless enough to forecast a 0.3C rise in temperature over a ten year period when no such confidence is justified in forecasting the behaviour of a chaotic system. That confidence is then carried into policies such as the CCA written one year later on the back of this and other wildly inaccurate forecasts. An act which ensures the energy prices for all the people of the UK will rise no matter what the wholesale costs are to wean us off the GHGs that are supposedly going to cause catastrophic weather.

    They used a hypothesis with its physics and a methodology to produce these forecasts and both appear to be wrong. What Feynmann said, if your experiment is wrong your hypothesis is wrong. By definition a law of physics is a hypothesis whose predictions have been consistently shown to be true in the empirical data, but can still be killed by one experiment proving it wrong.

    So, yes you’re right, they didn’t break a law of physics, they showed their hypothesis is wrong. And what’s surprising for me is that the hypothesis will continue to be supported by the climate science community as all hands go to the pump to produce papers saying it was right if other factors had been taken into consideration. Like, for instance, a distinguished and respected physicist coming onto this blog to tell us that they made a testable prediction and then diverting our attention by saying they hadn’t broken any laws of physics when Feynmann tells us the tests are to prove the hypothesis, not the law of physics-:)

  12. So, while it is true that global mean temperatures have not warmed as predicted by Smith et al (2007) in recent years, it’s hard to say why until we understand more about the various processes.

    But it is not hard to tell Smith et al (2007) thought they knew how the system works … and they were wrong.

  13. Hi Richard, thanks for engaging.

    I absolutely agree that predict and test is the way to go. Predict and test, predict and test all out in the open for everyone to see. It sure concentrates the mind. The trouble is you open yourself up to onlookers realising you don’t know as much as you think you do, which is what has happened here. Sure, bravo for trying, I do respect people putting their neck on the line, but you then have to accept when things don’t go as planned.

    And of course I understand your coin toss analogy. I can’t predict the winner of the next horse race but I can weigh up the odds and see where they’re wrong and hence make a long term profit.

    So how do we tell the difference between being “defeated by chaos” and just plain old being wrong. I’d suggest two ways:

    1. Most important of all is an historical track record of being right over a sufficient timescale. Sure you can get individual forecasts wrong but overall they should show “skill”. This is absent from Climate predictions as far as I can tell. (SteveMc could audit my results and he’d find a year on year profit over 10 years, i.e. skill).
    2. Secondary is an honest unbiased appraisal of the forecasts. All I see is obfuscation: trying to palm off hindcasts as forecasts, changing the baselines after the event; changing the metrics being measured after the event (ocean heat!). The handling of previous forecasts in AR5 just screams dishonesty to me. You’re hiding the fact that you were wrong. Bad tipsters do this all the time.

    Sorry to be harsh but there’s a saying, “if you don’t know who the mugpunter is then it’s probably you”.

    “That requires taking bold steps to make predictions in the first place.”

    Had to laugh at this one, you’re not serious are you? So it’s bold to practice the scientific method and make a prediction, then even if you’re totally wrong you just make a new one and carry on picking up your salary with no accountability. Yes, very bold. Again, I make predictions with real cash every day and if I’m wrong my kids don’t eat so I’m not impressed by that “boldness”. Sorry.

    “you should at least be clear about what “law” you think is being proved wrong.”

    OK, I’ll go for the one that says CO2 is a major driver of global temperatures. That’s what all this messing with the world economy is all about isn’t it?

    “What if we could forecast major droughts, heatwaves …Surely that’s a great thing for science to aim for?”

    So would searching for unicorns at the bottom of the garden…and we are a similar distance from finding them.

    It’s where opportunity cost comes in. How many billions have been bunged at Climate research and to what end. Sure we’ve got lots of fancy, expensive computer models and you’ve all got well paid interesting jobs but we hardly seemed to have moved on from Arrhenius in practical understanding. What a monumental waste of resources. Scrap the lot except the data recording arm (and come back in 100 years when we have more reliable data to play with) and put the money into Thorium research or similar.

  14. I think one of the most curious aspects of the whole climate change controversy, that will be remarked upon by future commentators, will be the abandonment of nuclear power, because of failures involving bad design and poor siting of plants built in the sixties.

    The surest path to low cost reliable energy, that also reduced CO2 (whether it needs to be reduced or not) was rejected by the people most loudly concerned with CO2 buildup.

    I expect that the country that is currently least concerned with CO2 (China) is where the system of modular designs produced in giant factories will be perfected and they will then be able to dominate the market for nuclear plant sales and installation.

  15. Climate Change 07
    Video of Met Office’s Vicky Pope PREDICTING 0.3C by 2014… (1st 30 seconds)

    and making strong statements (her words) later on

  16. This is a great post, and my only complain is how you let Mr Betts off way too easy. As the Met Office’s “Head of the Climate Impacts strategic area”, he’s one of the main people taking the AGW theory (which had some reasonable basis to start with, but which has been disproved at this point) and extrapolating upon it with the far more speculative and apocalyptic theories. History tells us that modestly warmer temperatures are a good thing; it’s this guy’s job to make warming a boogey-man.

    He arrogantly tells us of his bemusement, and goes on to try to change the subject, rather than dispute anything you say.Like the politician that he is, he claims to be on board with Feinman, but then proceeds to violate everything that Feinman says. As you point out, they “bravely” make “testable predictions”, but then they grant themselves error bounds that include almost all possible outcome – but fail anyway! Then they fail to recognize their failure, and continue on with their “settled science” claims, regardless.

    He absurdly demands that you identify exactly which “law” is being violated, when the context makes it obvious. When it suits his purposes, he will criticize you as disputing “settled science”, but when reality comes to call, he suddenly doesn’t recognize the very science he has staked his career on. “Law? What law?” he says.

    Then he comically declares, implicitly, ALL scientific hypotheses as impervious to being disproved. If the reality doesn’t match the law, just call it “natural variation”. This is intentionally obtuse, and ignores the basic principles of hypothesis testing that any undergraduate in a technical curriculum takes.

    He goes on to contradict himself, suggesting that “the potential for predictability comes from these longer timescales” after having just called it “chaotic”. At the same time he’s using the excuse that it’s unpredictable, he’s saying there’s hope for predicting. At the same time he’s predicting the End of the World – according to his models – he’s saying there’s “natural variation” that is apparently dominating the model, and which he says is random. If the outcome is random, and dominated by the random part, what good is his model? It’s very odd stuff for a guy in his profession to say.

    Pathetically, he evades the step where you say “We were wrong”, even implying – in contradiction to Feinman – that you can’t just say you were wrong until you figure out WHY you were wrong. Implicitly, he’s actually admitting that AGW was wrong, while claiming that it can’t be thrown out unless a better theory is put forward. While that line of reasoning works with the uneducated (including some WITHIN technical fields – I’ve seen it myself) it’s EXACTLY what Feinstein is saying is a huge NO-NO in real science. Mr. Feinstein is a poor excuse for a scientist, and not being straight with us, or with himself.

  17. It’s a shame that you didn’t explain decadal forecasting very clearly. The aim is to forecast changes over the next 10 years arising from all causes, including natural internal variability as well as any long-term forcing.

    The natural variability that we were told back then was minor. But that now is suddenly post hoc, ergo propter hoc very important.

    Scientists don’t make predictions. They make predictions with an error range. I don’t know what the Met Office one was for this study because, inexplicably, they didn’t publicise it, even in their own releases. So having made a very unscientific prediction to the public — in that it ignored an error range — you are now complaining that we aren’t taking into account inevitable errors!

    I suggest the Met Office can avoid this future embarrassment in it’s predictions by firmly stating 95% error margins in all media releases and correspondence. And insist that journalists and politicians include them in any conclusions reached from the studies.

    (You have a bind then, of course. Put in real errors and it will be obvious that you actually have almost no skill in prediction at all. Put in ones that are too small and you will virtually always be wrong. But that’s science for you.)

  18. Richard,
    Thank you for your substantial comment. I will look at the Smith 2012 paper. Of course the Feynman clip is not really directly applicable, since it is concerned with formulating a physical law, which is not the same thing as what climate science is doing (unless the ‘law’ is that man-made emissions dominate the climate!). Nevertheless, the principles of the science are the same: you (1) make some simplified model of the real world, (2) use that model to make predictions, (3) compare the model predictions with reality, (4) use that information to revise and improve the model. My gripe with the 2013 paper is that it provides no evidence that stage 3 has taken place at all. If it has happened, it has been behind closed doors, swept under the carpet. As Maurizio said on twitter, launching another set of predictions of sharp warming in the next few years without mentioning the failure of the previous one is a pointless exercise.
    As you know very well, it’s been said many times, climate science needs to be scrupulously open and honest in view of the scrutiny it is under and the mistakes that have been made in the past.

    There is another Feynman quote that is very relevant here, from his famous Cargo Cult speech. Climate science has failed and continues to fail to live up to these standards.

    “It’s a kind of scientific integrity, a principle of scientific thought that corresponds to a kind of utter honesty–a kind of leaning over backwards. For example, if you’re doing an experiment, you should report everything that you think might make it invalid–not only what you think is right about it: other causes that could possibly explain your results; and things you thought of that you’ve eliminated by some other experiment, and how they worked–to make sure the other fellow can tell they have been eliminated.

    Details that could throw doubt on your interpretation must be given, if you know them. You must do the best you can–if you know anything at all wrong, or possibly wrong–to explain it. If you make a theory, for example, and advertise it, or put it out, then you must also put down all the facts that disagree with it, as well as those that agree with it.”

  19. Everyone else, thank you for your comments, and apologies for not responding to them all. Please be polite and respectful to Richard Betts, and please stay on topic.

  20. Richard Betts,

    I too admire you for putting your head above the parapet in the lions den and I hope you relay the slings and arrows that you receive back to Mission Control.

    That said, I am utterly bemused by your closing statement:

    Decadal climate forecasting is an enormous challenge – but if it can be done successfully then clearly there will be huge benefits. What if we could forecast major droughts, heat waves or cold spells a long time in advance? Imagine the benefits to everyone, from vulnerable people to big businesses. Surely that’s a great thing for science to aim for?

    Surely you realise that heat waves and cold spells, unless they last for a very long time, are weather and therefore are meaningless in the context of a decadal forecast. Even droughts, unless they last for more than a year, I would class as weather. In reality, a decadal forecast is not much use to most people. They are obviously of interest to planners involved in building infrastructure but I suspect most of those planners take account of everything that can possibly happen rather than betting the farm on a forecast of which way things are likely to go. I am afraid my plans for a resort on the South Coast based on a forecast that the South of England would resemble the South of France in the future are not looking too good after 15 years of static temperatures. Some of the yucca plants around here did not fare too well during the winter of early 2010.

    I think huge benefits would accrue from long-range weather forecasting; decadal forecasts – not so much.

  21. Strange comment. Smith 2007 was the first paper ever that tried to put together the techniques for decadal forecasts. There are numerous problems to solve for such non-climatological simulations. Of course this must be published and discussed. But you shouldnt mistake that as a forecast but rather the first step to establish a technique for doing real forecasts. Still it is not clear if this can be brought to an operational level with a skill sufficient so that e.g. Australian farmers should change planting certain crops for the next two years (just an example).
    You were asking (or ironically commenting on) for the skills of the initialized model runs. You can find that in the paper. The skill of an initialized run exists for about 4-5 years. This is demonstrated in numerous hindcast runs. 4-5 years with some skill might be just enough to make a prediction for warmer/colder next ten years since more that half of the simulation would be without any relation to the real ocean state.
    “I think huge benefits would accrue from long-range weather forecasting; decadal forecasts – not so much.”

    Depends also on the region. Persistency is much larger in the tropics. In any case, this is all brand new. This decadal prediction idea exists for only 7 years or so. If you suppressed any technique which is not yet running perfectly after 7 years we would still live in caves.

  22. Georg, thank you for your comment. Maybe you came here from Kalte Sonne who have just translated this post into German –

    There is not really any difference between the words ‘forecast’ and ‘prediction’. Also I think the sceptics will not be impressed by hindcasts. To convince them that the models are skillful, accurate forecasts will be needed. The Smith 2013 paper shows temperature going up in the next few years. It will be interesting to see if this happens. My forecast is that it will not.

  23. @Paul
    “There is not really any difference between the words ‘forecast’ and ‘prediction’. ”
    There is, but it is not my point here

    “Also I think the sceptics will not be impressed by hindcasts.”
    They should. The important thing in science is the method. If it works in the past, it might work in the future. That’s how weather services improved their forecast over decades. Slowly but steadily. A Bayesian approach of making things better.

    “The Smith 2013 paper shows temperature going up in the next few years. It will be interesting to see if this happens.”
    If it will not happen, science goes on. There might be a real limit of predictability or the today’s calculated limits are underestimated or there are systematic errors or in the model or in the assimilation methods of data before the forecast starts.

    If it will happen no sceptics will change his/hers opinion. So that’s the not so interesting case.

    “My forecast is that it will not.”
    Which would be interesting if you could argue why that is.

  24. You are getting muddled. You brought up the word ‘forecast’ in the climate context, then you give a link to an article that only uses ‘forecasts’ for weather. You said we should not mistake it as a forecast, but it was a forecast – they used that word themselves in the abstract of both papers.

    Hindcasts can be useful in showing where the models is wrong. But the only way to demonstrate that a model is right is by forecasts. Otherwise there is a danger of fooling yourself. Again there is a lot of wisdom about this in Feynman’s lectures. Everyone should go back and study these, especially climate scientists.

  25. I never worried about the distinction forecast vs predictions. I just showed you that they are used in a different context in meteorology/climatology.

    When I said: This is not a forecast, I meant this is not a forecast like your daily forecast from the Metoffice, ie a mature end product of a scientific development. Nobody in the business took it like this. This is not a forecast to decide if you take your umbrella for the afternoon walk and it’s not a forecasts that would really allow you to judge the models (the most important problem is the assimilation technique and only then it might teach us something on model physics/problems)

    Smith 2007 is the first (as far as I know) forecast/prediction simulation on decadal scales ever. Absolutely nobody sold this as as “now comes the forecast for the next ten years”. You are looking on a developing subject of science.

    I also want to make sure that you understand that for the success/failure of these simulations the CO2 part of the models is completely irrelevant. You could keep the CO2 constant in all simulations.The forcing is way too small. Even the often discussed atmospheric feedback processes are not very important. What matters is how much of the heat that was mixed into the ocean the last 100 years of rising CO2 concentrations before is brought back to the surface. So the test of the models is on the dynamical part much less on the radiative part.

    “Again there is a lot of wisdom about this in Feynman’s lectures. Everyone should go back and study these, especially climate scientists.”

    For sure. However it is much too early to come up with a real evaluation of decadal predictions. Also the parts of the models (clouds, feedbacks, etc) that sceptics are so much obsessed about are not really tested.
    So here Feynman must wait for the moment.

  26. Absolutely nobody sold this as as “now comes the forecast for the next ten years”.

    That is not true. Vicky Pope sold it this way, see the video linked.
    I wonder what you hope to achieve by coming here and making untrue statements. All you have achieved is that another climate scientist, who I had never heard of before, belongs in the ‘cannot be trusted’ category.

  27. Hi Georg,

    Thanks for your replies, but I read them and RichardB’s to some extent and just think you need to get out of your academic bubble a lot more. (I appreciate that’s what you’re trying to do here, bravo).

    For those who have had to make a living in the outside world, the idea that you can just play around for years/decades with unvalidated and frankly useless models with no accountability is just surreal.

    Now I would say you’re on a fools errand trying to predict the behaviour of a chaotic system over a decade or more but if you disagree fine. Feel free to back your hunches and invest in your research …but do it with your own resources. See if you can get good enough with a Bayesian approach for those Australian farmers to pay you for your skill in deciding what crops to plant. Good luck with that. You could make a fortune. However, I wonder if they, as your paying customers, will be impressed as you by your hindcasts.

    I presume you’re aware of Von Neumann’s elephant:
    (different John Cook!)

    Piers Corbyn has to pay his own way at Weather Action and he seems, anecdotally at least, to be a lot more successful than the mainstream guys (I would really like to see a true comparison).

    Unfortunately, once you are conducting this wild goose chase on the public purse I feel justified in calling you to account rather than indulging you. Your project should also have to compete against all the other demands for public investment. And from where I’m standing, researching decadal forecasts would be near the bottom of a very big pile. Sorry.

    Regarding acting on non-forecasts, didn’t the Aussies build hugely expensive desalination plants because of the forecast droughts…just before the floods arrived? And are you seriously expecting me to believe that if your non-forecast was randomly correct you wouldn’t be shouting it from the rooftops as evidence of your skill? Come on.

    Watch that Vicky Pope video again to see how cocksure she is even when she’s wrong. Just imagine if she got something right.

    We’ve also seen it all before. Look at Arctic Ice as an example. It’s about the only prediction that has remotely come true and it is everywhere in the media as evidence of impending doom ignoring the mass of other failed predictions. Again you must be aware of the Texas sharpshooter fallacy:

    Make enough predictions, some of them will be right (although very few it seems for Climate “Science”).

    Oh and maybe there’s just a little bit of hubris comparing your forecasting efforts to those advances that got us out of the caves. Maybe?

    Sorry to be blunt, thankyou for engaging, no offence intended.

    P.S. I absolutely would look to change my opinion if you started to get a decent track record of correct predictions (rather than the odd random hit). Why wouldn’t I? That’s what it’s all about. Not looking likely though is it and there’s already a backlog of wrong ones.

    P.S2. As you’re in the field, can you tell me the purpose of the Hurricane forecast above that is “70% chance of between the min and max of the last 10 years”? Thanks.

  28. Steven Goddard is very prolific in finding failed predictions of climate scientists in newspaper archives. Here is an amazing one from James Hansen in 1986.
    “Average global temperatures would rise by one-half a degree to one degree Fahrenheit from 1990 to 2000 if current trends are unchanged, according to Dr. Hansen’s findings. Dr. Hansen said the global temperature would rise by another 2 to 4 degrees in the following decade.”

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s