At left is NOAA’s 30-day temperature forecast for December, 2009, and, at right, are the actual temperatures for December (click to enlarge). You’ll note the location of the warmest forecast temperatures (burnt orange colors) was where the coldest temperatures actually occurred (purples). This forecast was made in mid-November.
The British Met Office, which uses a different technique than NOAA, also missed badly. Their miss comes as part of a string of terrible forecasts that has caused some to question whether the Met Office should get out of the long range forecasting business.
My experience with the NOAA and Met Office forecasts seems to indicate little, if any, skill and I can see the point of some critics to take these back into experimental mode until consistent skill can be demonstrated, which might be years in the future.
That said, if we can’t get the 15 to 45 day forecast correct what makes us think we can forecast the climate 50 months or 50 years into the future?!
I’ll have more on the topic of confidence in long range climate projections later today or tomorrow.