Is it time to replace the CET?

[Nick H]

I love the way this old chestnut comes around every 9 months or so.

Indeed Shuggee I was thinking exactly the same...except it seems to me to be at least every 2-3 months rather than 9! We might as well wheel out the posts from the NW archive and stick them here right now...these arguments are so well rehearsed.

Edited April 21, 2007 by Nick H

[Stratos Ferric]

I'm surprised nobody has mentioned the snow in the N of the US perhaps heading our way for the start of May: that's about two weeks' time isn't it?

[Nick H]

• 
  

I'm surprised nobody has mentioned the snow in the N of the US perhaps heading our way for the start of May: that's about two weeks' time isn't it?

Maybe they have but in the appropriate thread!

[PersianPaladin]

I love the way this old chestnut comes around every 9 months or so.

As others have said, the CET measures the temp ranges across Central England, simple. Nothing needs to be changed to measure the temps in that area - it does what it says on the tin.

If the argument is that it doesn't represent Cornwall or Ayrshire or Shetland - then use a different measurement - the E&W series or the Scottish averages.

The beauty of the CET is its longevity - 300+ years of measurements from the same longitude and latitude on the globe. Useful for comparison.

I'm not really sure what the argument is...!

You may have a point, and yes Stratos Ferric - I do know that it is the Central England Temperature. All I want to know is why we give it so much credence these days as a true indication for the whole of England when it is just for a certain region which is of course, un-representative.

And as for correllations....well yes, there will be a stronger correllation co-efficient between a temperature in Birmingham and Durham, than say a temperature between Durham and Berlin. However...we need true stats, and therefore I think we should give the Whole Of England temp series just as much attention as the CET (if not more so).

[johnholmes]

I have to agree pretty much in their entirety with the posts by SF.

How either Potty or Pit can say the CET has no correlation to south Yorks I fail to understand.

Comparisons with such a long running data set are relatively easy for most sites in the UK, especially inland rural ones.

Edited April 21, 2007 by johnholmes

[The PIT]

I have to agree pretty much in their entirety with the posts by SF.

How either Potty or Pit can say the CET has no correlation to south Yorks I fail to understand.

Comparisons with such a long running data set are relatively easy for most sites in the UK, especially inland rural ones.

I haven't said anything about CET.

However this thread was already having repeating posts saying basically the same thing over and over again. Hence the clunk click comment.

[Stratos Ferric]

You may have a point, and yes Stratos Ferric - I do know that it is the Central England Temperature. All I want to know is why we give it so much credence these days as a true indication for the whole of England when it is just for a certain region which is of course, un-representative.

And as for correllations....well yes, there will be a stronger correllation co-efficient between a temperature in Birmingham and Durham, than say a temperature between Durham and Berlin. However...we need true stats, and therefore I think we should give the Whole Of England temp series just as much attention as the CET (if not more so).

But PP, a "whole of England" series would be even more unrepresentative for most purposes. In your NE location such an overall average would introduce a bigger westerly bias than is present now; it would be more unrepresenttive in your case than the CET. In any case, it's unnecessary. If you want to see what's happened in a location then go find the station with the longest record in that area.

Also, and just to check, you aren't proposing that the changes we're seeing in the CET are not happening more generally are you?

I don't think any of the people on here whose wiews are reliable are ever guily of attributing and extrapolating from the CET anything that cannot reasonably be atributed or extrapolated. But given that my answer to the question I raise above is that the CET is representative of wider change, then it is a perfectly reasonable surrogate to use for the changing climate in the UK. Yes, on any one day there can be wide variations around the CET composite, but that would be true whatever index were used, even for sites within bounds of the index. Once you construct and index you start averaging, and once you average then you produce a value that actually exists in no one place absolutely.

[johnholmes]

I haven't said anything about CET.

However this thread was already having repeating posts saying basically the same thing over and over again. Hence the clunk click comment.

mon apolgies

[PersianPaladin]

I posted this on Uk-Sci and got this response: -

http://www.ukweatherworld.co.uk/forum/foru...s=3&start=1

Basically it is not truly representative, although due to its long establishment I suppose it has its comparative uses.

[Stratos Ferric]

I posted this on Uk-Sci and got this response: -

http://www.ukweatherworld.co.uk/forum/foru...s=3&start=1

Basically it is not truly representative, although due to its long establishment I suppose it has its comparative uses.

First of all Weatherworld is NOT uk.sci. I doubt you'd have got that sort of response on uk.sci, not without far more along the lines that have already been posted on here rebutting your point. All that that ukww response shows is that you do not have monopoly in being misguided on your views in this matter. How on earth can CET be unrepresentative if (and it does) it draws on a sample of stations within the area that it purports to cover.

We all know it does not represent the NE of England, or the Isles of Scilly, or germany, or Venus, but then it doesn't ever claim that it does. Al of that said, to reiterate the point, CET is a very strong indicator of the climate in Durham and the SW of England once you know the general variability of climate in the UK.

Darn it, I've just trodden in another pile of dog poo.

I posted this on Uk-Sci and got this response: -

http://www.ukweatherworld.co.uk/forum/foru...s=3&start=1

Basically it is not truly representative, although due to its long establishment I suppose it has its comparative uses.

And having now read the post that was posted I see that you've been rather michievious and selective in your cut and paste of the response. Anyone would think you're more interested in winning your point than arriving at the right answer to the question.

[PersianPaladin]

Lets agree to differ on this issue. Clearly the CET is not the ideal accurate average temp mean for the whole of England and it is there wholly for comparative purposes. Hence of course...the clue being in the name.

Edited April 21, 2007 by PersianPaladin

[Guest Mike W]

We should use the CET for means, but not the 30 year one, we should IMO use a current 10 year mean, in this case the current one being 1997 - 2006. But it depend what you wan to reperesent a mean showing where we are or where we are in relation to where we were., if the latter then 71-00 should be used.

[Thundery wintry showers]

I think the main problem is the way the CET is often misinterpreted as a national temperature reading rather than for Central England.

However, that's a problem with people misinterpreting it, not with the idea of the CET itself.

The reason why the CET is so widely used is because the other temperature records generally go back only about 100 years; the CET goes back almost 350 years and therefore gives a longer timespan from which we can monitor changes in climate.

As for 10-year and 30-year means, I'm sure Stratos Ferric has discussed this at length; I see the 10-year mean as a good way of assessing against very recent 'average' conditions, but it is also a method that masks changes in the climate, and doesn't take into account the fact that those who were around in the 1970s and 1980s might think "March 2007 was pretty warm, when you think of the wintry Marches we had in the 70s and 80s".

[PersianPaladin]

I think the main problem is the way the CET is often misinterpreted as a national temperature reading rather than for Central England.

However, that's a problem with people misinterpreting it, not with the idea of the CET itself.

The reason why the CET is so widely used is because the other temperature records generally go back only about 100 years; the CET goes back almost 350 years and therefore gives a longer timespan from which we can monitor changes in climate.

Could well be Ian. Gave me some thought this thread anyway.

[Somerset Squall]

We should use the CET for means, but not the 30 year one, we should IMO use a current 10 year mean, in this case the current one being 1997 - 2006. But it depend what you wan to reperesent a mean showing where we are or where we are in relation to where we were., if the latter then 71-00 should be used.

I think a 10 year mean would be too short, you can obviously get a lot of variation in a 10 year span but a 30 year mean gives a more accurate average temperature IMO.

[Stratos Ferric]

I think a 10 year mean would be too short, you can obviously get a lot of variation in a 10 year span but a 30 year mean gives a more accurate average temperature IMO.

Firstly, CET is just a data series. There is no formal reporting or standard average for CET. The UKMO, in providing a climatic baseline, do use a thirty year reference period. Why they don't continually roll this I don't know, but that aside a sample of thirty data points does provide robustness for inferential stats. Smaller data sets can be used but the confidence intervals start to widen markedly.

A thirty year mean gives a picture over 30 years; a ten year mean over ten years: simple as that. The problem with the use of the thirty year mean at present is that we are not in a flat climate, therefore the start of the series includes data which, for all intents and purposes, has become a climatic anachronism. Yes, there is more potential for volatility to skew data in a shorter data series (hence the degrees of freedom problem alluded to in the first para re inferential stats) but, for now at least, that is certainly no worse than including data in a longer series that is no longer truly representative of our climate.

[Terminal Moraine]

Whichever period of mean temp' is used, it is only a tool for indicating anomalies and is therefore just as valid as any other period provided this is clearly understood.

If I want to know how the current temperature compares with that of, for instance, the 1700s I can take the mean of the last 25 years or so and instantly see the anomalies of each month/season/ year in that century. If I want to know how the temperature of the 1700s compares with other years in that century it would make more sense to take a 25 or 30 year average from that time.

Equally, if I need to know how the current temp' compares with the recent past a sample from the 1700s isn't going to tell me much but the average of the last 25 years certainly will.

The right tool for the right job.

T.M