On Sunday, 29 May 2016 01:48:19 UTC+1, RW wrote:
On Wednesday, 25 May 2016 20:07:51 UTC+12, Dawlish wrote:

Dunno. There is no definition. When I/we/the neighbours and dog walkers I meet feel it's summer, probably the best way! 😀 (real smiley face!) No consensus yet!

As an academic exercise - if you have a site with a fairly long record - and one is just considering temperatures (mean daily for each day, or perhaps the mean daily maximum), then assign the top 25% of the ranked calculated results to summer, the bottom 25% to winter, and the rest to spring and autumn. An interesting alternative is to actually divide the range of these daily values rather than the rankings - top 25% to summer, bottom 25% to winter, etc. With this method one needs either a fairly long record or a smoothing method (say 5-day means) to avoid letting an outlier day or two at the ends of the range distort the boundary values too much. I tried this latter for some of our NZ sites and got some sensible results

Without lpping off the medieval warming bit I presume?

- "spring" was considerably longer than "autumn", as expected. The variances in summer and winter lengths were trickier to explain in some cases.

Having said all that, to really assess my town's seasons, I consider sunshine and rainfall to also be very important, so would need to use some sort of weighting formula - haven't bothered to do this - perhaps a retirement project.

We should all have suffocated with nearly perfect agricultural grade cabon dioxide levels by then.