It’s mid-May and the weather has been okay, if a bit cold. But the big question on people’s minds at this time of year must be: What’s the summer going to be like in the UK? Are we going to get as much rain as the previous three years?
In the past one source of information was the Met Office, you could go to their web site and then click your way to a seasonal forecast. For a brief period you could even get a forecast for several years forward if you knew where to look. But not anymore, the longest forecast you’ll easily find is an average for 15-30 days ahead.
The Met Office announced earlier this year that it was not going to continue publishing its seasonal forecasts. They said the reason was that these forecasts were difficult to get right. In my opinion, that cannot be the real reason, because it has been known for a long time in the forecasting community that a seasonal forecast for the UK is not easy.
Why did they bother publishing their seasonal forecasts initially then? Well apart from the fact that it was supposed to be good publicity for the Met Office(!), some information is better than no information. They also served it up as a probabilistic forecast so you knew the odds if you were taking a chance on the forecast being right.
The beginning of the end for the publication of the forecasts probably came with the ill-fated decision to announce the summer of 2009 to be a “BBQ summer”. If you look at the appraisal of the forecast it was not as bad as the media made out, it just rained more than they said. But there was a probability attached to that rainfall forecast, so they never ruled out the chance of above average rainfall amounts.
If you ask me, the whole problem lies with the public expecting a deterministic forecast. Unlike many other European meteorological agencies, the Met Office does not publish daily probabilistic forecasts along the lines of “today there is a 70% chance of rain”. All this amounts to the fact that, if the British public were educated about probabilistic forecasts on a daily basis, then they would be a lot better at understanding the probabilistic nature of seasonal forecasts. And I would have had a summer forecast to look at and muse over.