UPDATE 8/20: For whatever reason, Frank Abel has completely taken down his post. Most likely it did not amuse some particular persons.
==================================We all know that seasonal forecasts looking 90 days or so into the future are not easy to make and have considerable uncertainty. But that hasn’t kept some bold meteorologists from going out and issuing 3-month seasonal forecasts that even include the week-by-week weather development for the period, and claiming up to 80% probability they will be right. I’m not aware of any technology that would allow that.
And who can forget the forecasts of “barbecue summers”?
Increasingly a number of serious meteorologists are now getting annoyed by these crackpot crystal ball forecasts, which are eagerly spread by the media and never turn out to be correct.
One vocal critic is German meteorologist Frank Abel here, who is a weatherman at MeteoGroup Deutschland GmbH in Berlin, part of the biggest private European weather company.
Abel took two prominent summer-2011 forecasts made by 2 popular German meteorologists back in March and May respectively, and tells us today how they panned out. Abel:
In summary what they did get right was (like with a horoscope) purely by chance.”
He then quotes the Institute for Weather and Climate Communication (IWK) in Hamburg concerning the forecasts made by the 2 meteorologists.
Frank Böttcher of the IWK told me what he thought: ‘What does that tell us? Their projections are worse than if they had been picked randomly.”
Indeed Abel and dozens of other meteorologists had gotten so fed up with the dubious seasonal forecasts that they banded together back in April and drew up the so-called Hamburg Declaration on Longterm Prognoses, which has since been signed by well over 100 meteorologists. Here it is in English (my translation):
Hamburg Declaration On Longterm Weather Prognoses
by the Meteorologists of the 6th Extreme Weather Congress
We, the signers of the Hamburg Declaration, do hereby declare:
Long-term prognoses for approaching seasons are among the major targets of meteorology. Generating these prognoses are still at the very early stages of development. For producing such prognoses, research institutes are using super computers, without which, from today’s scientific perspective, it would be impossible to calculate the long-term developments. The results provide information on the monthly-basis anticipated deviations from the mean value and the probability of their occurrence. Such forecasts are currently possible only for very large extended regions.
Using today’s state of the art science, only the following type of forecast for June would be possible: ‘With a probability of 65%, it will be 0.3° to 0.5°C cooler than normal in southern Germany.’ Considerably more accurate forecasts for weeks and days for specific local areas, in the view of the undersigned, are not meteorologically and scientifically tenable and thus serve to damage the reputation of serious meteorologists. Thus the undersigned hereby request that making and publicizing of such prognoses, which give the public the incorrect appearance that it is possible to make accurate forecasts with our current level of knowledge, be avoided.
Well how about that – 65% probability a forecast for a period only 2 or 3 months into the future will be correct? Now what should that tell us about forecasts made for 5, 10, 20, 50 or even 100 YEARS down the road? Precisely – also just “horoscopes”, and them actually happening would be “purely by chance.”
Declaration On Climatic Forecasts
I now propose a Declaration on Climatic Forecasts. Serious meteorologists admit that seasonal forecasts are tricky and fraught with huge uncertainties. So isn’t it only appropriate that climatologists stop fooling the public and admit that their decadal and 100-year forecasts are also highly uncertain?
And please spare us the bullcrap that weather forecasts and climate prognoses are completely different. If anything, for obvious reasons, climatic prognoses are even far more difficult and thus uncertain.
One only has to consider that there is absolutely no agreement today among scientists on the strengths of forcings from the vast array of factors that act on our climate.
Science begins with honesty.
For German readers also see: Peter Braun at klimazwiebel.blogspot.com.
11 responses to “Meteorologist Blasts “Horoscope” Longterm Forecasts – Over 100 Sign “Hamburg Declaration On Long-Term Prognoses””
I see the link at Scienceblogs provided in Abel’s story no longer functions. I can only speculate that the 2 meteorologists he criticised may have contacted Abel and wasted no time threatening legal action (I’m only guessing -could be just a technical error with the link).
Again, I think Abel is dead on with the uncertainty surrounding these seasonal forecasts. They do indeed give the false impression that the meteorologists are able to do something that we all know is not possible. There’s no was anyone is going to convince me that they can tell me what the weather will be like at the end of September with 80% certainty.
you really need to read up on the topic. Compare voting: you can’t predict how person x will vote, but you can make pretty good estimates on how 80 million people will vote, within a few percent. climate projections are similar: they will never try to predict in what year a drought/flood etc. will happen, because they can’t; and nobody claims they can. but they can give you a picture of what will change in the general picture. even these predictions are not very good, and maybe never will be; the important point is that current science is very sure that weather extremes will increase: not only from models, but from comparing the geological record with ice cores, for example.
you don’t need to know which bone you’ll break when you’re rushing towards a wall with 120 mph; it’s sufficient to know hitting it will be harmful to make the smart decision to hit the brakes.
And with that comment, Hartmann, you have managed to show that you are far more ignorant on the subject of climate than I could have ever imagined.
Peter, you’re talking about a statistical approach. It turns out that the climate models cannot model convective fronts well BECAUSE of their statistical approach. One of the little caveats climatologists love to be silent about.
Fair to laude Abel and his colleagues. Indeed forecasting has not improved for decades. They think they can do it with statistics. Instead, what they need are ocean water temperatures from several depths levels. When I had to draw the day weather map while sailing the North Atlantic as third Mate (long time ago), they had been so accurate that you could set the time according the changing weather conditions. That changed the closer you came to the continents. As forecasters have no information about the acute sea temperature structure, e.g. the sea areas around Britain and to depth down to 20, 30,50m), they use SST and statistical records. The more you have, the better and longer term forecast could be made. Long term forecast requires ocean data (to get them is one problem, to acknowledge it in the first place is another and currently the more serious).
Back in 1942 the oceanographer Sverdrup told the meteorologists:
“It might appear, therefore, as if the oceanic circulation and the distribution of temperature and salinity in the oceans are caused by the atmospheric processes, but such a conclusion would be erroneous, because the energy that maintains the atmospheric circulation is to a great extent supplied by the ocean.”
From “Oceanography for Meteorologists” published in 1942 (Sverdrup, 1942).
The deal with people like Hartmann is that it really doesn’t matter what the climate does. It could put the brakes on by itself, go into reverse, and smash into the other wall, and of course it would be because of human activity. Unfalsifiable science is not science – it’s pure quackery from charlatans.
Message to readers: I’m only going to put up with quacks and charlatans for so long here before editing them out.
It is often forgotten to mention in the forcast issue the band width of the predicted values. The broader that width the higher the probability of a correct prediction. I can predict with near certainty that after two years in a certain village in Germany the temperature will be between plus and minus one hundred degrees Celsius. This is a forcast but some nobody is interested in. It would be better to maintain one probability level of ninety percent for example, and to mention the band width of the values. After a few weeks that width will approach the long term range, which is a forcast not better than the general picture. By the way, every amateur can earn a good score without super computers. Predict that the weather tomorrow will be the same as today, and you have a probability of eighty percent that it is correct.
If you watch the BBC web based 5 day forecast you will see that they can’t even get that right.
It will be changed on a Day to Day and sometimes half day basis. They modify it as they go along.
In fact the 24 hour forecast is hardly ever right.
That’s of course because it’s the Met Office who supply that forecast. The same Met Office which told us we’d be getting warmer winters with less snow, even as they brushed the snow off their front step, warmly clad. The predictors of less rainfall and a “Barbecue Summer”, not mentioning we’d have to cover our barbecues to keep the rain off.
Looks like a new La Nina is starting right now.
See WUWT’s ENSO page for more.
There is an interesting chart in this paper from the Texas State Climatologist on the Prediction capability of weather and climate for the past and future. He shows a peak at a hundred years out. WUWT!