William M. Briggs

Statistician to the Stars!

Pajamas Media: Actually, Weather Is Climate

Today’s post is at Pajamas Media: Actually, Weather Is Climate


Once again, thanks to the editors David Steinberg and Aaron Hanscom for coming up with a title and a tagline. Which is “It is statistically appropriate to point to this year’s frigidity as evidence that the theory of man-made global warming is suspect.”

They even provided a picture of a cute dog. How can you go wrong?

The post answers the question: how come all those environmental groups get to use hot weather as evidence of global warming, but we’re not allowed to use cold weather as evidence that catastrophic man-made global warming is false?

For those too busy, the answer is: we are allowed.


  1. If you toss a coin I can’t predict the outcome very well but if you toss it 1000 times I can predict the average number of heads reasonably well. Why doesn’t this argument work for weather and climate? Since climate is “weather aggregated”.

    And if it does work then why isn’t reasonable that the models don’t predict weather very well but are OK for climate? (Can we set aside for the moment that they don’t appear to be predicting climate very well either?)

  2. Rich, foolishly I’m going to take a stab at that. [Matt has answered it before on this site, if my ancient brain remembers correctly, so we’ll see if I’ve learned anything]. I think it has to do with the “predictable” element you are seeking to set aside. Unlike coin flips, which are true guesses with only two outcomes available, a climate model has many possible outcomes and therefore could be useful if it could be relied upon. There are climate models that are more accurate than AGW, and thus more reliable. But Those That Care only wish to push – and believe in – their own [failed up till now] AGW model. That makes it analogous to a religion, imo. But you knew that.

    Several great comments on the PJM site, Matt. Good post!

  3. If I understand you correctly, climate is a statistical yield from an aggregation of weather event-moments – the standard container of these moments is usually considered to be a 30 year period. But you could choose other aggregations. I once argued that 10 years is also valid for certain purposes. Or how about 500 years, which would give you a different degree of variation. Right now, the CAGW crowd is using 1850 as a starting point for a 160 year climate grouping for the grand scheme of warming acceleration trends. But these are all artificial (humanly selected) constructs and thus carry their own inherent warming or cooling trend bias depending on what you are looking for. Is this right?

  4. Briggs

    January 22, 2010 at 5:17 pm


    Exactly so.


    My heart soars like a hawk.

  5. Out in the nether regions of long tailed distributions, anything, and I mean anything, can happen. We could have declining temps for 30 years straight, and that would not disprove an AGW model with a long tail.

    I’m sure that concept has been discussed here before, but I don’t remember when exactly.

  6. Lot’s of comments on this over at WUWT
    What are the “climate models that are more accurate than AGW” that 49erDweet mentions?
    By the way coins occasionally land on edge 😉

  7. Contrary to popular belief, the more flips there are in the coin flipping game, the less accurate the prediction of the average number of heads will be. While this should be obvious, most people don’t get it.

  8. According to GISTEMP, global temperatures for December 2009 were 0.1 degrees warmer than the 2000–2008 average.

    What data are you using?

  9. Doug M: put me down with most people then.

    Briggs: I’m still on the ground peering at the sky.

    49erDweet: That didn’t really help. You don’t have a link to that earlier post by any chance?

  10. “Man-made global warming is just one of many possible theories of climate. Another is the Business-as-Usual Theory (BUT), which states that whatever happened last year will more or less happen this year, and so on into the future.”

    Very true indeed. The cyclical nature of the climate is underestimated by the narcissistic nature of man.

  11. Sorry for the time lag. Family stuff.
    Phil’s Dad: I googled – well wiki’d – “climate models” and came up with links to seven true models at the bottom of the page. If you add BUT [see Jim S, above] that’s eight. And I don’t trust wiki editorial policies to believe their entry include others that scoff at or challenge the “accepted” narrative that’s now proved to have relied on seriously fudged data.
    Rich: [and Phil’s Dad:] Too much week-end going on for me right now but I will look into Matt’s archives the first of the week and promise to post back here, on this thread, with a few links from the past that should satisfy many of your questions.

  12. Rich,

    Suppose you flip 2 coins. If I predict you will get one head, I will be right on half the time and be within one filp with 100% confidence.

    Suppose you filp 1000 coins. If I predict you will get 500 heads. I will be right on 2% of the time, and be within 30 filps 95% confidence.

    A professional can make more accurate forecasts the weather tommorrow than the weather next week. And, it is predict where the dow jones will be next week than where it will be next year.

  13. Mr. Briggs: Thank you for your excellent essay, Actually, Weather is Climate. The slogan that weather is not climate has always bothered me. Your idea that climate is just a summation of weather really speaks volumes. Also I have thought for a long time that the statements made in defense or rebuttal about AGW or any other climate theory cannot be established by statistical correlations in an open thermodynamic system such as climate. Probabilistic inferences of events that are likely to occur do not prove that an assumed physical model is the cause of what has occurred. In a closed physical system, the mathematical model predicts what is happening and what will happen, ie. planetary motion. Worse than that are the many extrapolations into the future that are made by scientists and engineers that are based on statistical correlations for physical model that is possible to the exclusion of all other possible models. For climate predictions, one can apply confidence limits to a temperature prediction, but it may be meaningless because the processes may not be for a stochastic phenomena to start with and the principles of chaos theory can come into play. I spent several years in the Nuclear Reactor Field where the safety of the reactor is vital to human. We conducted numerous fault tree analyses with wild guesses about the likelihood of a failure of a nuclear component. The value the studies were not that one proved nuclear reactor was safe, but that the discipline of identifying failure modes and assessing the sensitivity of the design to the faults eliminated a lot of design flaws.
    I am so happy that you have cleared the air on the issue proving AGW by statistical analysis. Personally, I support BUT and am willing live as man has for centuries with the consequences. Unfortunately, there are a number of scientists willing to risk their reputations to statistically prove that by 2035 all the glaciers will have melted away as a result of a few more ppm of CO2 released into the atmosphere and that there is an army of politicians and bureaucrats who want to take my money to reduce the release of CO2. All of this activity is based on a statistical correlation between the CO2 concentration in the atmosphere and the average global temperature and a series of unscientific anecdotal observations. What would you call that?

  14. Briggs

    January 24, 2010 at 8:00 pm


    Thanks for the kind words. I’m very interested in the way you went about guessing probability of failures for reactors. You’ve probably heard the joke, but there’s not many experiments you can run!

  15. Doug M

    Your example is more akin to predicting a particular temperature. I still think my prediction of an average gets more accurate and that, after all, is the claim made by modellers: they predict climate which is average weather.

Leave a Reply

Your email address will not be published.


© 2016 William M. Briggs

Theme by Anders NorenUp ↑