Arguing about models and observations, with respect to global warming

For the past thirty years able people have been trying to develop models of the Earth’s climate, and I’ve written about the models before. If the climate of our planet is technically ‘chaotic’, meaning that elements of it are unpredictable, then modelling it is bound to have have some inaccurate results, to say the least. Roy Spencer has written a good essay on climate and ‘chaos’, which you can read here. There are plenty of others.

Dr Spencer has also recently produced this graph, whose message is clear.


CMIP5 models come from an international collaboration to promote a standard set of model simulations to see how realistic they are in simulating the recent past, provide projections of future climate change on two time-scales — the next twenty years, and beyond 2100 — and provide greater understanding of ‘key feedbacks’, including the carbon cycle and the role of clouds. The CMIP5 set was developed for the IPCC’s AR5.

What is abundantly clear from the graph is that the models aren’t very good at the next twenty years, if we look at their success so far. At the moment only 3 of the 90 are tracking anywhere near two of the standard datasets, UAH (satellite) and HadCRUT4 (which has been ‘adjusted’ so that it shows higher temperatures than does HadCRUT3, whose values were used in the diagram in a recent post). The black-dotted line represents the mean of the set. It’s a very long way away from the observations.

When you point out the disparity between the models and observations to the orthodox you get a variety of responses. ‘Ah, yes,’ you will be told, ‘but remember that climate is the average of weather. The models will be much more accurate about what will happen in thirty years’ time than about what will happen in a month!’

You are likely to suspect a con-job in the making here, for if the models are not accurate about the near term, why should they be accurate about the long term? The orthodox answer is that in the long run the little weather fluctuations will cancel out… Well, maybe, but the evidence doesn’t look persuasive in that graph, which does cover thirty years.

A recent diversionary tactic on the part of the convinced is to suggest that the warming that should be there must have gone somewhere else, like into the deep ocean. There is some observational evidence that the deep ocean is very slightly warmer than it was some little time ago, though in fact we know almost nothing about how warm the deep ocean was, say, fifty years ago. The problem with this hypothesis is first, no one can produce a persuasive explanation about how the heat got into the deep ocean without also warming up the not-quite-so-deep ocean, and second, how or when it will get back out again.

It occurs to one or two puzzled people, like me, that maybe the models provide too high a level of climate sensitivity, a matter I discussed some time ago. That is, the models assume that carbon dioxide has more power in its capacity to affect the climate than it really has. Alas, to say so is to commit a sort of climate blasphemy, for all the orthodox know that carbon dioxide is the control knob of the planet’s temperature. The IPCC has lowered the lower boundary of the likely value of ‘climate sensitivity’ in AR5, which suggests the authors weren’t quite as certain as they used to be. But they won’t give up the ‘control knob’ analogy.

How do they know? Well, they think they know all about all the ‘natural’ causes of climate variation (meaning those not caused by humans), and they’ve accounted for all of them. Therefore, what’s left must have been caused by humans. This is a poor response, because we should never be sure that we know everything about anything, let alone something so vast and complex as the world’s climate system.

So we wait. The models seem to be running too ‘hot’, temperature has levelled out (inasmuch as the precision of the global temperature anomaly means anything), carbon dioxide is increasing the way it has done since we started to measure it in the mid 20th century, the biosphere is improving and greening, food production is growing, and the IPCC is not getting the attention that it has become accustomed to.

Sooner or later, I think, someone in government is going to point out that the carbon dioxide emperor is not wearing any clothes. What then?

Join the discussion 9 Comments

  • Mike O'Ceirin says:

    We understand very well how a jet motor works and have well understood physical laws to compute the stresses involved in a real engine. From this a computer model can be specified, coded and tested to see if it does the calculations properly. I have worked in software for many decades and despite what my ilk pretend this part of the process is not easy. After all this a jet engine designed with it will be tested to destruction. After all there is probably only 95% confidence in it. My opinion is that GCMs are fudged to look right and that they produce the results wanted. In the future they will be seen to be a waste of time,

  • PeterE says:

    Thanks. Very amusing. There is the science and well-meaning scientists but beyond them there are those who would/do benefit from the theory of AGW, like manufacturers of wind turbines, low-lying Pacific islands, Green lobby groups and totalitarians of one kind or another who would like to guide us in how to live. In this context, I recommend the editorial (‘Chronicle’) by Keith Windschuttle in the current October Quadrant. This rightly places AGW in a broader, more sinister context. Scepticism is vital to combat gullibility, with which human history is replete, with tragic consequences.

  • David says:

    So Don, are you still claiming to be agnostic on AGW ? 🙂

    • Don Aitkin says:

      Yes, as you will have seen. Aren’t you?

      • David says:

        No! Currently, I believe the scientific evidence which supports a causal relationship between a human induced increased in CO2 and a problematic rise in global temperatures.

        In my view, while it is ok for a scientist to be agnostic on an scientific issue in the short term, in the longer term he/she should be prepared to make a decision one way or the other. Ultimately it is their professional responsibility to take a position. If they are shown to be incorrect, as new evidence comes to hand, they should change their mind.

        • Don Aitkin says:


          I too think that the physics suggests that increasing carbon dioxide ought, all other things being equal, lead to an increase of a degree C for each doubling of the CO2 proportion in the atmosphere. The ‘hiatus’ (if that is what the current levelling out is) suggests that not all other things are equal. The gloomy stories come from computer models that aren’t very good. On the face of it, warming is good for the biosphere, at least at the rate we’ve seen over the past sixty years.

          So I remain agnostic about AGW’s being a problem for humanity. It has been a blessing so far.

  • Mikeo says:

    Another reference to why Climate Models are a waste of time

    We know fairly well how children behave and many experts claim to be able to predict given a set of stimuli how a child will respond. So let us produce a General Child Model and then everyone will know what Suzy or Johnny will do for the rest of their lives. It really does make no less sense than those who claim using really big computers one can predict behaviour of chaotic systems. What they are outputting anything their authors want.

  • Alister McFarquhar says:

    I too am puzzled

    I presume the models predict future trend having regard to the length of historical data

    and whatever other assumed drivers like CO2 are included

    I think there is no pretence to predict cycles such as the 1942-74 cooling trend within the longer term trend

    So no reason that the prediction 30 years hence will be closer to actual than recently?

  • […] wrote about this subject a few months ago, and I’m going to illustrate what follows with the newest version of a graph that I used […]

Leave a Reply