On Sunday, 4 October 2015 18:07:30 UTC+1, xmetman wrote:
I've been analysing the occurrence of frost using the data in the daily Central England Temperature [CET] series from 1878 to 2015. It may not come as a surprise to some to learn that the first frost is getting later and the last frost getting earlier. The first frost of the autumn/winter at the start of the series in 1878 occurred around the 29th of October, but 137 years later this has now slipped by over two weeks to the 13th of November. Likewise the last frost of any winter/spring has slipped back 11 days from the 16th of April in 1878 to the 5th of April in 2014.

Charts etc

How is a frost defined in the daily CET? Is it that the average minimum at the 3 or 4 (correct?) stations used is below 0°C or does it need just one station to be below 0°C? My guess is that it is the former and this would mean that the first CET frost of autumn would be later than the first frost at any individual CET station. It would seem to be the explanation of why my own figures over 32 years show an earlier first frost and a later last one than CET. The dates are 21 April and 4 November, both with a wide range. The standard deviations are 14 and 18 days respectively.

Tudor Hughes, Warlingham, NE Surrey, 557 ft, 169 m.