May CET

[Thundery wintry showers]

Well, most of these are educated guesses to be fair- Beng in particular even got the month's general pattern right.

Looks like I was 0.8C out; so not completely wrong, but not close enough to be termed 'accurate' either.

Excellent work with the graphs by the way.

Edited June 1, 2007 by Thundery wintry showers
Rather stupid original post about December 2006- I've realised that a dash or blank in the error column means no error, not no entry!

[HighPressure]

Wow, I came second overall depsite a fairly poor April if the truth be known as is reflected in my rise from 18th place, but the other months were close, though it has to be said well done to Phil, 0.1C away for two months is pretty good going!

Now just got to keep that up for the rest of the year, easier said...

This was your prediction done on 18th April:

I'll go for 12.1C, a very easterly month in all, a fairly changable month because of possible Lp's skirting away to our south and far far NW which will give fronts possibly to N. part of the UK.

I got to say that was a really good call in the middle of a hot spell, I wish I was that good

[West is Best]

The Manley CET figure for May has been confirmed at 12.0c.

Sorry, this has probably already been mentioned as I'm not around much but the proper official CET for May has already been published by the Met Office Hadley centre. It is 11.9C.

http://www.metoffice.gov.uk/research/hadle.../HadCET_act.txt

Well done to those who got close!

(p.s. call me a stickler but I shall continue to use the official figure in discussions beyond this fun competition. No offence to Philip who also does a brilliant job with the Manley emulation.)

Edited June 2, 2007 by West is Best

[J10]

Sorry, this has probably already been mentioned as I'm not around much but the proper official CET for May has already been published by the Met Office Hadley centre. It is 11.9C.

http://www.metoffice.gov.uk/research/hadle.../HadCET_act.txt

Well done to those who got close!

(p.s. call me a stickler but I shall continue to use the official figure in discussions beyond this fun competition. No offence to Philip who also does a brilliant job with the Manley emulation.)

But we are using the Manley CET for this competition, and this is the official basis for all scores.

However as you point out, there have been slight differences in the CET for March and May, and IF we had run the comp by Hadley rules, these are the scores that would have occured.

No Change in the top 2 but Stu London would have been in 3rd, with Duncan McAlister in 4th.

[Snowly does it]

Indeed, Hadley have declared and the Spring record remains intact, by a 'full' tenth of a degree, lol!

The Hadley figure does surprise me as they had it down as 12 to the 30th so some adjustment must have taken place as there is no way the 31st saw a drop in the CET, nevertheless the figure has been given.

At the risk of infuriating Richard, the following are facts and figurettes from May.

This was the second warmest Spring in the CET series

May came in 91st in the series since 1659 so is top end second quartile

The month is the first since October 2006 to be lower than its preceeding equivalent

Its the first month since August 2006 to be lower than its own rolling ten year mean

The lowest anomoly to the 71-00 series since August

The month ranks as 'above average' (by the narrowest margain possible).

It was a very wet month as shown by Mr Data's figures above.

Aside from the deluge, the remarkable thing about May is how stubbornly consistent it is being with a run of returns in a neat bundle.

[Weather-history]

Sorry, this has probably already been mentioned as I'm not around much but the proper official CET for May has already been published by the Met Office Hadley centre. It is 11.9C.

That figure looks suspicious, its at odds with its own daily data output. One of them is wrong.

[Snowly does it]

That figure looks suspicious, its at odds with its own daily data output. One of them is wrong.

I imagine they have factored in their adjustments - February for example fell by 0.2 degrees, and from the daily to the 30th of 12.0 I would have expected a final figure of 12.1, so perhaps they have factored in the adjustments they make to the Hadley figure?

[cheeky_monkey]

hey looks like im 1st in the table of those who have had a punt for every month...not bad considering i jus go on gut feeling and nothin else...now watch me plummet down the table through the summer!!..lol

[North-Easterly Blast]

The May CET of 11.9 that has just been published is actually 11.87 rounded up.

So May 2007 was actually 0.3*C below the average for the modern day post 1997 average, about average by the warmed post 1988 average, but still 0.6*C above average by the long term 30 year 1971-2000 standards.

[cheeky_monkey]

what is this modern post 1997 rubbish comming from..it bad enuff with the post modern 1988 guff...jus getting silly now!

[Optimus Prime]

Not quite sure the full extent of the meaning a even larger teapot. The 1960s was modern yet it wasn't a modern climate? Surely 1997 isn't modern because it's in the past before the 21st century? And what's classed as modern?

[West is Best]

below the average for the modern day post 1997 average,

Yes let's kick this nonsense into touch once and for all. The post-modern post-1997 average is utter guff: an entirely spurious mean that has no basis whatsoever in metereology. Either use the 1961-1990 mean or the 1971-2000 mean; or reference to the 'longer term average'; but no more of this silliness please. A baseline of just 9 years is statistically too open to margins of error to have any meaning, quite apart from seeming to invite manipulation to prove this or that theory about any single month's performance.

Edited June 3, 2007 by West is Best

[eddie]

Yes let's kick this nonsense into touch once and for all. The post-modern post-1997 average is utter guff: an entirely spurious mean that has no basis whatsoever in metereology. Either use the 1961-1990 mean or the 1971-2000 mean; or reference to the 'longer term average'; but no more of this silliness please. A baseline of just 9 years is statistically too open to margins of error to have any meaning, quite apart from seeming to invite manipulation to prove this or that theory about any single month's performance.

A quick question for you oldies on here; before 1990 what was the basline against which temperatures were measured?

Also, who chooses when these baseline periods get changed? When we get to 2011 will we start using the 1981 to 2010 mean or does this have to be decided by the 'mean' police?

[phil nw.]

A quick question for you oldies on here; before 1990 what was the basline against which temperatures were measured?

Also, who chooses when these baseline periods get changed? When we get to 2011 will we start using the 1981 to 2010 mean or does this have to be decided by the 'mean' police?

Yes at the start of a new decade the previous 30years stats. are used,so the example you give is correct.

[Stu_London]

Yes at the start of a new decade the previous 30years stats. are used,so the example you give is correct.

Not so actually, there is a lag of 5 years - The met office only started using the 1971-2000 averages as a benchmarck from January 2006. For the rest of the decade previously, the 1961-1990 averages were used.

Therefore in January 2016, the 1981-2010 average will be used as a benchmark.

Unless of course the whole benchmarking procedure is overhauled which is possible

[Snowly does it]

Yes let's kick this nonsense into touch once and for all. The post-modern post-1997 average is utter guff: an entirely spurious mean that has no basis whatsoever in metereology. Either use the 1961-1990 mean or the 1971-2000 mean; or reference to the 'longer term average'; but no more of this silliness please. A baseline of just 9 years is statistically too open to margins of error to have any meaning, quite apart from seeming to invite manipulation to prove this or that theory about any single month's performance.

Can't agree with that at all. Why on earth would anyone use the 61-90 average? Might as well use the 1931-1960 average or the 1691-1720 baseline.

It is perfectly valid to take other timescales and review months, seasons or years against them, especially in a dynamic period, such as the one we appear to be in. as none of us are here (as far as I am aware) due to punitive net-imprisonment, employment by the Met Office or as vengeful statistics demons there is no reason to stick rigidly to a 30 year assessment. Yes, that should be the first point of review (and ONLY the 71-00 averages), but then there is room for interpretation of the month outside that restrictive box.

In fact no twisting of figures or averages in the case of May 2007 will make it (in terms of temperature) anything other than an utterly unremarkable month sitting in the top half, but outside the top quarter.

[Stu_London]

Yes let's kick this nonsense into touch once and for all. The post-modern post-1997 average is utter guff: an entirely spurious mean that has no basis whatsoever in metereology. Either use the 1961-1990 mean or the 1971-2000 mean; or reference to the 'longer term average'; but no more of this silliness please. A baseline of just 9 years is statistically too open to margins of error to have any meaning, quite apart from seeming to invite manipulation to prove this or that theory about any single month's performance.

Have to agree for the most part - I become very suspicious of statistics being manipulated in this way, besides niether post 1988 nor post 1997 hardly really contain enough cricital mass to make them worthwhile.

Also agree with SM that 1971-2000 is the one to use - that is what everyone else uses.

[West is Best]

Can't agree with that at all. Why on earth would anyone use the 61-90 average? Might as well use the 1931-1960 average or the 1691-1720 baseline.

Well you began this new-fangled 9 yr mean nonsense Charlotte so of course you would agree with it!

You are entirely at liberty to do whatever you want with statistics. But if you publish it publicly you must expect it to be challenged. It is completely spurious, without any backing in metereology, and from a staticician's point of view risks being wayward and inaccurate. We've only just recently all moved to the 1971-2000 mean which is a perfectly acceptable, statistically meaningful, time period against which to test any given month's performance during the rest of this decade. Suggest you get used to it! And I'm not just saying that because of the way you manipulate alleged 'low' months against this false mean: by using an insufficiently long baseline you risk being challenged the other way when (like April) we get a statistical freak month. It is necessary to have a sufficient base for any mean to have meaning - pun intended. 30 years is the bare minimum (as you may know I also like to compare against the long-term i.e. 100 year average).

Edited June 3, 2007 by West is Best

[Snowly does it]

Well you began this new-fangled 9 yr mean nonsense Charlotte so of course you would agree with it!

10 years Richard, 10 years!

I acknowledge the limited statistical usefullness of my preferred statistic du jour.

To give you an example of another way I look at it which is not to give cold-credence to a month that does not perhaps deserve it is that the run of months that had failed to dip below the 10 year mean was August to May - individually meaningless against the short time-span but when looked at as a group it starts to look like the 'ceiling' has been smashed in the ways house prices in a street have a ceiling until suddenly everything pops over the top and a new baseline emerges. Perhaps thats why I found May a relief even though it was patently and undeniably an above-average month (note, only just and certainly not under any stretch of the imagination a well-above average month).

Its a question of what should we have expected versus what should we expect and the longer term mean is no help in determining the new reality. Its very useful at confirming it though.

No matter, challenge away, I enjoy your challenges as you are very passionate about what you say and believe.

[West is Best]

No matter, challenge away, I enjoy your challenges as you are very passionate about what you say and believe.

lol - you handle me well! I'm a bit OTT sometimes. I went to bed last night thinking I'd overdone it, so apologies.

Let's agree to disagree on the new 10yr mean you want to introduce! I happen to think it so limited asa baseline as to be open to huge distortion. If I had more time I'd play a game with a given month showing it to be warmest and coldest according to how you manipulate the stats to suit ... but I haven't! Agree to disagree?!

[Snowly does it]

lol - you handle me well! I'm a bit OTT sometimes. I went to bed last night thinking I'd overdone it, so apologies.

Let's agree to disagree on the new 10yr mean you want to introduce! I happen to think it so limited asa baseline as to be open to huge distortion. If I had more time I'd play a game with a given month showing it to be warmest and coldest according to how you manipulate the stats to suit ... but I haven't! Agree to disagree?!

lol, don't worry, I am a country girl, we can handle it.

I agree to disagree yes, I might also incorporate the thirty year rolling mean in to how I am looking at stats to give it a longer baseline.

[summer blizzard]

Personally, i look at the 1970-2000 average for comparison, though given the leap in warmth over the past year, i can understand why also looking at a 10 year mean is understandable, though i do agree that comparing to an outdated mean is useless..

[Optimus Prime]

I think the 20th century 100 year mean should be used. But then you could argue the warming of the 90's (except October and December and perhaps June) offset the generally persistantly unvaried temperatures pre 1990.

Trevor Harley has released his review of May;

May. One of the wettest on records in many places in the south. The late May Bank Holiday was terrible - cold and wet. Temperatures overall were slightly above average, but in some places (e.g. here, East Scotland) May was cooler than April, and in many places maxima in May were lower than maxima in April (for the first time since 1987). The month was saved by many mild nights. The highest temperature of the month was 25.7C at Heathrow on the 24th; tje ;pwest -5.7C at Kinbrace (Sutherland) on the 27th. It was the wettest May since 1967, and the fourth wettest on record, with an England and Wale average total 121 mm, nearly twice as much as average. Sunny in the north and west but dull in the east, where it was the dullest since 1991. The dry weather of March and April continued into the start of the month: some parts of east Kent had 38 consecutive days without rain up to the 5th. Parts of Shropshire had 79 mm of rain on the 13th. The 28th was the coldest day so late in May since 1948: the maximum was only 6.4C at Whipsnade.

http://www.personal.dundee.ac.uk/~taharley...ather_years.htm

[damianslaw]

I think the 20th century 100 year mean should be used. But then you could argue the warming of the 90's (except October and December and perhaps June) offset the generally persistantly unvaried temperatures pre 1990.

Trevor Harley has released his review of May;

May. One of the wettest on records in many places in the south. The late May Bank Holiday was terrible - cold and wet. Temperatures overall were slightly above average, but in some places (e.g. here, East Scotland) May was cooler than April, and in many places maxima in May were lower than maxima in April (for the first time since 1987). The month was saved by many mild nights. The highest temperature of the month was 25.7C at Heathrow on the 24th; tje ;pwest -5.7C at Kinbrace (Sutherland) on the 27th. It was the wettest May since 1967, and the fourth wettest on record, with an England and Wale average total 121 mm, nearly twice as much as average. Sunny in the north and west but dull in the east, where it was the dullest since 1991. The dry weather of March and April continued into the start of the month: some parts of east Kent had 38 consecutive days without rain up to the 5th. Parts of Shropshire had 79 mm of rain on the 13th. The 28th was the coldest day so late in May since 1948: the maximum was only 6.4C at Whipsnade.

http://www.personal.dundee.ac.uk/~taharley...ather_years.htm

I enjoyed reading the above summary as the words unseasonable warmth or one of the warmest were not used something we have become too accustomed too in recent analogies of recent months.

[acbrixton]

Yes let's kick this nonsense into touch once and for all. The post-modern post-1997 average is utter guff: an entirely spurious mean that has no basis whatsoever in metereology. Either use the 1961-1990 mean or the 1971-2000 mean; or reference to the 'longer term average'; but no more of this silliness please. A baseline of just 9 years is statistically too open to margins of error to have any meaning, quite apart from seeming to invite manipulation to prove this or that theory about any single month's performance.

Richard, not sure that I wholly agree with you here:

1. You are of course right to state that the system of 30 year averaged are useful as they are the internationally acepted method of recording acerages and then comparing the performeance of individaul months/years agsinst those averages.

2. I also agree that a 30 year average will be more "reliable" in that it is relatively free from distortionds produced by individual months/years.

3. However the use of an alternative mean, such as a 10 year rolling average, in conjunction with ythe 30 year average has a limited use in recording recent trends especially where, as at present, there appears to be a recent and noteable trend to a "rising norm" in the climate of the UK and where the current 30 year average inadequately shows that trend as its early data are of a time well before that apparent warming phenomenon.

4. Obviously those who cite a 10year roling average should do so fully recognising its statistical limitations still less implying that it is of equal, let alone, greater integrity/application than the conventional 30 year averages.

Kind Regards

Andrew

[P.S. and woefully ot: did ou see recipes for 2 salads as requested in the CWA thread?]