I was upbeat about the potential for this summer to be at least reasonable ever since I did my composite analysis way back in April. Looking at years that had similar Nov-Dec, Jan-Feb and March periods, the signal was for decent positive height anomalies across the mid-latitudes, with the potential for the Azores High to extend across the UK and out to the NE.
During the first half of June, the high pressure did just that, but at latitudes just a bit too high for temperatures to climb very far away from some western regions. The rest of June was more mixed and to be honest didn't really display much connection with the composite outlook.
July then rallied with a tremendous UK high, which was a variant of the theme depicted by those similar years. I must admit that I never envisioned anything quite so remarkable, but then it's not usually advisable to expect such extremes.
August seems to have seen the pattern flattened into a very benign zonal pattern. This differs to the composite outlook, which suggested that high pressure would shift north and form blocking well to our NE and N - leading to an unsettled month with the wettest weather in the south.
Overall, I'm pleased with how well the composite method has performed, given that it was entirely experimental and not entirely scientific in it's execution. What I did was to identify all the similar years to 2012 for Nov-Dec and then look at the Jan-Feb patterns from those chosen years to pick out which ones were similar to Jan-Feb 2013. In other words, the resulting selection had been similar across both of those two periods, even though in some cases the Nov-Feb anomalies differed. Then I narrowed things down further using the exceptional March of 2013, which proved to be a big deal, as it revealed a much clearer projection for the summer, even though the number of years was reduced to just a few.
Yes - it seems that the extraordinary March, either by itself or in combination with the two previous periods, in some way precluded the better summer pattern - particularly the combination of +ve mid-lat height anomalies and -ve polar height anomalies. Due to the small sample size, I can't say with much confidence whether this discovery is actually true of reality or just a statistical quirk, but it's fascinating nonetheless
Originally Posted by: Stormchaser