Mean max temps in winter - inherent mild bias?

[Summer8906]

With the current setup in mind, in which it's possible places in the southwest might have three consecutive calendar days with very mild temps even though the mild only lasts a little over 24 hours, it struck me that mean max temps in winter might have significant mild bias.

What I mean by that is that even if a very mild spell lasts only, say, 30 hours, it could potentially impact three consecutive 24-hour periods (whatever 24-hour period is chosen) and have quite a profound impact on mean max.

For example, imagine a 15-day cold spell interrupted by two very mild episodes, each lasting 30 hours (this could perhaps be not too dissimilar to the current setup, given that New Year's Day also started mild but became cold). By chance, each mild episode has an onset at 2100 and clears at 0300, 30 hours later. This means that 6 out of the 15 days feature very mild temps.

Let's say the mild episodes see temps constantly at around 13C, due to being warm sectors, but the remainder of the time daytime temps are no higher than 5C (and night-time temps much colder).

We end up with 6 days of max 13C, and 9 days with max 5C. This averages out at 8.2C, which if we imagine a typical southern England location is slightly above average (assuming 7C as average for this time of year), even though 13 out of the 15 daytimes featured temps no higher than 5C!

If we used a different measure and only counted temps towards the mean max if they occur in the "lighter" half of the day, 0600 to 1800, then only two days are affected by the mild and we end up with a significantly lower mean, 6.0C.

Arguably the second measure is a more accurate reflection of the typical daytime temps of this 15-day period. I can see for example the first half of January this year "appearing" milder than it actually was due to the short mild spells over-counting towards the mean max due to being spread over multiple days, despite being short.

Are there any techniques used to mitigate what seems to me to be inherent mild bias (due to a mild spell impacting the max for whole calendar days even if it only includes a small portion of that day)? In years to come the records might give the false impression that today, tomorrow and Monday were all very mild days in the southwest whereas actually, Sunday will be the only really mild day, with today and Monday mostly cold.

Generally transient mild periods are more common than transient cold periods IMX, so you're more likely to get the mean max impacted by transient mild than transient cold.

Restricting the reporting period to 0600-1800 would have a significant mitigating effect, for example.

In summer of course the max almost invariably occurs during the lighter half of the day, as in summer, high temps are governed more by the presence or otherwise of sunlight compared to winter when airmass is all important.

Edited January 4 by Summer8906

[B87]

8.2c would be average for most of the south in January. Lets have a look at January average highs in some locations, from West to East.

Scilly: 9.9c

Plymouth: 9.0c

Exeter Airport: 9.1c

Portland: 8.7c

Swanage: 9.0c

Southampton: 8.8c

Martyr Worthy: 7.8c

Solent: 8.6c

Farnborough: 8.0c

North Heath: 8.6c

Kew: 8.6c

Bodiam: 8.3c

Faversham: 7.9c

[Summer8906]

 B87 I'd still consider anything above 8C mild for January (as discussed in another thread) but my main point is that a mild spell that covers a short part of the day then makes that day "look" mild. For example yesterday will go down in the record books as extremely mild when in fact the daylight hours were largely not.

Interesting Martyr Worthy is the coldest of all these though, given it's not too far from the coast and a good way west. South Hampshire always seems to be significantly milder in winter than places further north or east. Colder maxes than Bodiam, for example? And colder than Farnborough or Kew? I'd have expected somewhere in Surrey or inland Sussex or Kent to be colder. Also, North Heath seems incredibly mild for its location.

 

Edited January 7 by Summer8906

[reef]

There's no mild bias as the data used for the averages that are being compared to is collected in the same way.

Your 30 hour mild spell example starting at 2100 and ending at 0300 30 hours later would only have two days with high maxima as recordings are 0900-0900. This would only lead to one day having a mild minima, as just the second "day" would be affected.

Edited January 7 by reef