BBC reports UK Met Office revises 2017 temperature forecast cooler

BBC here – posted by the UKMO on Christmas Eve – a huge departure from their forecast of a year ago – thanks to Wayback machine.
I have stiched together the two forecasts so readers can best see the huge difference.

One point puzzles me and that is the difference in the white curves after 2007, marked on my graphic as 07. UKMO says – “Previous predictions starting from June 1960, 1965, …, 2005 are shown as white curves,..”. OK but is anybody aware that the post 2007 downturn was ever published as a forecast ?

The above little comparison shows the difference in the white curves better.
UK MP Mr Graham Stringer gives his views on the UKMO Christmas revision.

2 comments to BBC reports UK Met Office revises 2017 temperature forecast cooler

  • David Brewer

    You have to really look carefully to see what the Met Office has been up to here.

    Their forecast of a year ago included three predictions – white lines with red borders around them to give an uncertainty range. These were stated to have been predictions from June 1985, June 1995 and June 2005.

    If you look carefully you can see what happened to these. The first prediction from 1985 was too high. By the time it runs out in 1993 or 1994 it is about a tenth of a degree higher than reality. The second prediction was just a little bit too high by the time it ran out in 2003 or 2004. The third prediction, from 2005, is way too high already – about 2 tenths of a degree above the figures already in for 2011. So they overestimated 3 times running.

    Now have a look at the version they just published this Christmas.

    Wow, now they have predictions from as far back as 1960 and they are all just about spot-on! What happened? Simple. They just cut every forecast off after 4 years. The forecasts are for 1961-65, 1965-69, 1970-74 etc. Since they started each forecast from whatever the temperature was at the time, and only put in 4 years of forecasts, they could hardly go wrong, since interannual variability is greater than their predicted trend.

    So, last year they let the forecasts go on for 10 years or so, by which point you could see that they were not reliable, and biased on the high side. Now they just cut every forecast off after 4 years and start again with another one. The result is to give a highly misleading impression of reliability to the forecasts.

  • David Brewer

    Two further points.

    1. It seemed odd that the Met Office was making 5 or 10 year forecasts back in the 80s, let alone the 60s. So I looked into this. Of course, they were not. Instead:

    Hindcasts:…10-year predictions, 1965, 1970, … 2000
    First forecast: 2005-2015

    In other words, all the “forecasts” before 2005 were hindcasts made after the outcome was known!

    2. The 2005 prediction, the first real one, did indeed predict a slight drop up to 2009. That was only to be expected given that 2005 was above trend. But, as late 14 September 2009 the Met Office insisted:

    A return to warming is subsequently predicted, with at least half of the years after 2009 predicted to be warmer than the warmest year currently on record.

    That prediction was included in last year’s version of the chart, but has now been flushed down the memory hole, with a new and radically different prediction being substituted. According to it, none of the first 8 years after 2009 is expected to break the 1998 record.

Leave a Reply