Sunday, July 21, 2013

The Inability of Climate Science to Make Accurate Forecasts

I'm very much a cheerleader for atmospheric science and, especially, meteorology's ability to make accurate forecasts of major storms. However, climate science is a different matter.

Last month, at the American Meteorological Society's meeting in Nashville, I presented this graphic which shows the IPCC's 2001 forecast is far off. Current temperatures are below the 95% probability range, and nowhere near the centerline of the forecast. A similar graphic can be made of their 2004 forecasts. The actual temperatures are the small dot with the arrow pointing to it.

Since that meeting, others are reporting on the same topic. Lucia Lijegren created the graphic below. I have added the arrow that shows actual temperatures (gray lines) are outside of the all of the models' forecasts on cold side.
This week, Steve McIntyre plotted the most recent three climate forecasts from the British Met Office. Here is the 2010 forecast in red with the subsequent actual temperatures in black. Wildly incorrect.

Here is their 2011 forecast. Again, way off.

Finally, here is their 2012 forecast. So far, so good. But, note that they now forecast cooling starting in about four years. They also keep constricting the number of years into the future for which they make forecasts.
I refer you back to yesterday's posting pertaining to Dr. Roy Spencer's testimony before the U.S. Senate Thursday (scroll down) as to why missing these forecasts at this time is especially significant. Clearly, the hypothesis that climate science can make accurate forecasts of earth's future temperature is falsified. 

Even The Economist, until the last six months a reliable cheerleader for the catastrophic global warming hypothesis, has caught on. This from the July 20 edition:

Still, over the past year, several other papers have suggested that views on climate sensitivity are changing. Both the 2007 IPCC report and a previous draft of the new assessment reflected earlier views on the matter by saying that the standard measure of climate sensitivity (the likely rise in equilibrium temperature in response to a doubling of CO2 concentration) was between 2°C and 4.5°C, with 3°C the most probable figure. In the new draft, the lower end of the range has been reduced to 1.5°C and the “most likely” figure has been scrapped. That seems to reflect a growing sense that climate sensitivity may have been overestimated in the past and that the science is too uncertain to justify a single estimate of future rises.
If this does turn out to be the case, it would have significant implications for policy.
I view this as an opportunity for climate science: Take a deep breath, concede what has become obvious, and start over with a blank slate. That would give climate science the opportunity to to become a true asset to society rather than propagandists for the same old -- incorrect -- "the world is ending" theory.

NOTE: This posting is the third in a three part series resulting from the U.S. Senate's hearing Thursday. Here are links to the other two parts:

Dr. Roger Pielke, Jr.'s testimony on disasters.

Dr. Roy Spencer's testimony on why temperatures have not risen as forecast.


  1. Superb commentary. But why do I feel that the liberal political community will not listen to even the slightest challenge to these inaccurate forecasts put out by so-called experts?

    And of course, "there is no debate" (allowed)...

  2. And this in my opinion is why those who believe in AGW ( anthropological global warming) as a serious threat get into trouble with the general public.

    When you insist that your original ideas (FORECASTS) from 10 years ago... from 15 years ago... are still valid and that "everything is going exactly according to plan/ forecasts" you end up looking like a fool.

    Such intransigence and dogma in the face of actual reality ends of drawing a stronger reaction than the actual science itself because it looks like to the general public and the unknowing skeptic ... that the issue is NOT the science but that you cannot admit you are wrong .... or that some of your ideas have to be changed.

    Once you start down this road those who believe in AGW are no longer practicing actual sound science. You CANNOT win argument against some idiot on right wing radio talk show hist is proclaiming a cold day in July means there is no such thing as global warming.

    You cannot win an argument against those who think its all "a LIBERAL CONSPIRACY" (see the whacko poster above).

    Those who believe in AGW as a serious threat ... need to change their tactics. They need to start preaching science and uncertainty in science and accept the fact that the early ideas are wrong (or need some adjustment) and remind everyone that is the way science is SUPPOSE to work.

  3. As much as I think the first poster has lost it - I think you don't see the climate change side with as clear of a vision.

    The entire reason many people have given up on *trusting* the AGW side is precisely because they stopped using science except as a means to whip up political money.

    The first poster was correct 99% of the time no debate at all is allowed - and the (yes) radical claims that every natural disaster is caused by global warming only makes it look more and more like the entire AGW 'side' is trying to sell nothing more than homeopathic water (diluted to 1 part per million of course).

    Seriously the Al Gore vid was shocking - the evidence *seemed* to make sense - but the climate models (which all the red alert emergency comes from) turned out to be wrong. When that *started* I figured we'd adjust and see why. Instead the entire thing has doubled down on stupid and when the smoke clears it will take a *calamity* for the public to trust climate science again.

    Or perhaps a model that can predict at least 5 years into the future with a 90% certainty. That'd be a good place to start rebuilding faith.

  4. You didn't test the data, you took a PNG and slapped a dot on it. What you claim to be outside the 95% CI is, as labeled on the chart, outside the model envelope of solutions which is not a CI. So I'm assuming you did something basic like a t-test to establish your own CI that the little dot was in actuality outside of a 95% CI (not the envelope) and this was not due to random chance? Because I don't see any words in your post that addresses that. I'm sure your work wasn't as sloppy as just haphazardly using Paint to put a dot on there, right?

  5. Mr. Martin:

    Regarding the IPCC graph:

    The graphic comes straight from an IPCC report. For the AMS versions, I left all of the identifying info on the graphic. For the blog presentation I but stripped off the identifying info so it would be legible.

    The IPCC (in the document's text) describes these as 95% confidence intervals.

    I used the HADCRUT3 database of world temperatures which is the one used by the IPCC (apples to apples).

    There is no need to establish a CI, they state one. Current temperatures are either in it or not. They are not.

    With regard to the BMO forecasts, I took the work of Steve McIntyre (click on link) but separated out the three forecasts into three graphs for ease of interpretation.

    The thing that pro-GW people seem to want to forget is these are not my forecasts nor are they my confidence intervals. The IPCC and BMO are strongly "pro" global warning organizations. The point of this post is simply to test the quality of their forecasts, which is very poor.

    Thanks for the question.



Note: Only a member of this blog may post a comment.