Today was the final (half) day of the 2009 ICCC, followed by flying home; I promised to toss up a few thoughts, which I’ll keep terse and to the point. Once I get caught up on other things (including some sleep), I’ll try to provide some more over-arching thoughts on it all. But for this evening, I just want to note highlights along the lines of…. things that were new today.
The opening breakfast keynote was given by former New Hampshire Governor (and White House Chief of Staff 1989 – 1991) John Sununu. Governor Sununu also holds a Ph.D. in mechanical engineering, so he’s certainly professionally qualified to comment upon the whole AGW thing.
He exercised some old memories when he noted that the grand-daddy of all “disaster-movie” computer models is the infamous “Club of Rome” economic/world forecast that was emitted back in 1972. To make a longer story short, the Club of Rome forecasts used complicated equations and input parameters, and predicted a disastrous future for mankind based on shortages of material resources; naturally, the timing of this was very important. But it was quickly noted that the “Club of Rome” models were very sensitive (in outcome) to very small changes in the input parameter values, and that small changes in input parameters could change the outcome qualitatively – predicting very rosy results.
As Steve McIntyre had noted in his Monday presentation, there is a very similar situation with the (admittedly moving target) of the present climate models; small perturbations in input values produce very large output changes – the models lack robustness.
This lack of robustness (and this is me saying this now) is a characteristic of mathematical descriptions that are poor and unpredictive.
Governor Sununu noted that the main “improvement” of the present climate models (vs. the Club of Rome models) is that the new climate models are so complex that they are obscure and too-abstract to debate.
Two very good quotes about the AGW alarmists:
“…. resonating self-adulation….”
“…. a well-orchestrated symphony of effort….”
During the technical sessions, Chris Essex, a mathematician at the University of Western Ontario, gave an intricate (but very interesting) talk on many problems (many of which have been personally battled by this writer) that arise when one attempts to construct mathematical models of complex systems.
Since this is supposedly a more general audience that will be reading this, I’ll just note one salient point. He noted that the climate models are empirical – that is, they are constructed based on data and the development of equations connected directly to that specific data…. rather than from previously-known (and well-understood) equations that could be used in this problem. Models of this sort rely heavily on “parametrization” (that is, not just the choice of parameters used, but also on the specific quantitative values chosen for them for use in the model)…. when these parameters do not in fact correspond to anything solid and previously-known. This is often not a show-stopper…. since engineering models can be based on controlled experiments; however, climate models cannot be based on controlled experiments, and can only be back-fit to the available recorded data from the past.
Now we’ll switch to my thoughts from experience on this sort of thing. “Models” of this sort can indeed be useful – provided you are trying to “dehydrate” a simple description that will only apply to the extant data that you have on the table. Since the model equations are a purely mathematical construct that truly cannot be used beyond the boundaries of where one possesses data…. one cannot really use them outside of those bounds. For example, the ability to adjust a model to fit previously-measured data endows it with essentially no predictive abilities beyond just that data. Back-fitting does not endow a model with predictive ability – in fact, it rarely does so, and pretending that it does leads to comical results.
This is clearly the situation with the “prediction” curve that was presented yesterday by Professor Akasofu and that I included via a photograph that I took of this slide. If one splices out a small set of data and fits a curve to it, the “model” that was created applies only to the region of the data – any extrapolations (or predictions) outside of that region of data are completely devoid of content.
There’s another subtle aspect of mathematical modeling that I wanted to note this evening, since I’ve encountered it so many times in the past and I realized in New York that (when it comes to the “official” climate models) I’m seeing it again.
It can be relatively easy to take a very small part of a problem, construct a mathematical model for that aspect, and produce a good result. This can be done for several different small problems, producing equally good results.
However, all too often, the obvious “next step” is to combine all those smaller piece-models together to try to model something much larger. This seems logical, but in reality it’s a nonsense-based disaster. Since the pieces are at best semi-empirical and have no relationship to each other, the throwing-together of the little models into one big model always produces results which are comical. They just don’t predict anything because they produce silly results.
I can think of several notable disasters in electrical engineering that I’ve seen first-hand that resulted from this kind of non-understanding of this sort of thing.
The “climate models” all seem to have been constructed in a similar fashion – tiny little things were modeled with good fitting to what data was available. But then all the pieces have been crammed together into a much bigger model. I would expect this to produce insane and laughable results – and this is indeed the case. It’s just that the IPCC calls these sorts of results “scientific predictions” (sic)….
I’ll close with one item from today’s closing lunch (which I had to leave early in order to catch my flight home). Jay Lehr from Heartland opined that he had drawn five important going-forward conclusions from the event – what he called the “RTLCC” quintet:
1) Those not jumping onto the AGW bandwagon should be called not “climate skeptics,” but “climate Realists;
2) We need to mobilize quickly to start talking about “cap and trade” for what it truly is – “Tax and trade”;
3) The collateral damage that will be caused by “mitigation methods” in the developing world is not a matter of aesthetics or even a living-standard cut; for the developing world, this is truly a matter of Life and death;
4) To effectively counter the AGW propaganda machine, it is necessary to talk to the Consumer in short sound bites;
5) If right now we were to find a magic technology that would let us Capture and make-disappear all of the CO2 being emitted, the total effect on the world’s temperature would be…. nothing; CO2 has a very weak effect on temperature (compared to other factors), and variations in the climate and temperature are driven by nature, not by man.
I’ll stop there for today (ZZZZZZZZ) and hope that these dispatches and videos we’ve been contributing from New York have been useful and enlightening. As I noted at the top, when I can get my sleep supply replenished and catch up on other things, I’ll try to go back through my notes and pen some broader but more over-arching thoughts.
One thing I will note to conclude for tonight. Perhaps the most enjoyable aspect of this event was that the “science” parts of the event were truly in the form of a “scientific conference.” Different people presented different views of various causes-and-effects; there were disagreements, but they were all reasonable scientific differences that then point (because of the discussions) to ways to further parse, study, and analyze the problem. This is how “science” is truly done.
For some reason, there was an expectation in certain quarters (such as the New York Times) that this event would be an event like that farce going on in Copenhagen – where today the latest “scare-’em-gasless” model result is a forecast (sic) of a one-meter rise in sea level by the end of the century (one suspects that the tweaking of a couple of parameters would lead to a forecast of half the seas disappearing, but I digress). The purpose in New York was to compare studies and results, and discuss them.
It’s not to get everyone together to agree on the lock-step talking points to take home.
We can leave that streetwalker version of science to Al Gore, James Hansen, and that whole roster of poseurs….