September CET!

[summer blizzard]

By dry/wet, I'm talking essentially about about flow patterns. September 2006 was wet in some places, mainly because of torrential downpoars and thunderstorms. So, OK, it was a wettish month, yet synoptically it couldn't be more differant than say September 1995.

In that case, how northerly/southerly has nothing to do with how cyclonic/anticyclonic a month is, making your earlier arguament void.

[Stu_London]

Well, I don't agree with North_Easterly_Blast that a warm and dry September automatically means that you throw in the towel for the coming winter, but I have to say, whats going on in September through the last few years is most extrodinary and without any precedent that I can find. People say that all months are random, and you can't pick out trends for one month based on a previous month, but as far as September goes, I'll say that it is beginning to become obvious that some freakish, cyclical pattern is at work. Look through the Wetter-Zentrale charts, and you see that relatively few Septembers actually start off with big, hot high pressure systems lasting for weeks and weeks on end. I would say that on average, up until 1997, it probably happens about once a decade. Then we come to the mid 90's and suddenly it starts happening year after year after year. And the weird thing is, it doesn't matter what pattern precede's September, through the summer months. We've had poor summers like 2002, 2005 and now 2007 lead to this freakishly anyticyclonic dominated pattern. We've had reasonable summers like 1997 and 1999 lead to freakishly hot/anticyclonic Septembers. And we've had really hot summers, like 2003 and 2006 lead to freakishly hot/anticyclonic Septembers. Its almost like we're caught in a never-ending cycle here, where variability has completely and totally gone out of the window, for just this one month of the year. Its like clock-work.... Its strange, freakish and weird, warming trend and all....

Granted we have an anticyclonic start to the month, however it's not freakishly hot (its actually 0.8C below average up to the 4th). Yes it will warm up between now and the end of the weekend but the forecast after this date is far from certain. September by nature is statistically always likely to be much cooler 2nd half than first too.

It's a while since we've had three cold months on the bounce and I still expect the majority of this month to stay settled.

Things MIGHT start to get more interesting synoptically next month if this general pattern continues.

I've actually punted slightly above average and am happy with that at this stage. I would be very suprised if we saw 15 or 16 point something this year though.

[West is Best]

Morning Stu. I think you are under-punting a little at the moment. This warm spell may not be freakish, but it's not insubstantial either. We're already now into milder nights as the temps begin to lift. I'm confident of 25C being seen in the UK tomorrow, and possibly as early as today (though maybe just short). We may well see some maxima nudge well past that 25C point before Sunday.

As ever, and as you rightly hint, it will all depend on the subsequent synoptics.

I do find it fascinating that with the right synoptics one can still get high temperatures late in the year. 25C has been recorded in the UK everyday from 1st September to 13th October.

[Optimus Prime]

Does look, to me like the pattern maybe merging into something similar to the warm Septembers since 2003.

I'd be surprised if we get a sub 15.0c month.

[Stu_London]

Does look, to me like the pattern maybe merging into something similar to the warm Septembers since 2003.

I'd be surprised if we get a sub 15.0c month.

The 00z run is indeed very settled and very warm, at least in the beginning. There would be cooler nights and days from mid-month which could supress the CET.

Morning Stu. I think you are under-punting a little at the moment.

Well we are 14.4C at the moment and I said we might get to 16C by the 10th. That means an average of 17.1C for the next 6 days. Not sure it can get much higher than that, although heat of this nature tends to cover most of England and Wales rather that the usual south east corner being way ahead of everything else.

[West is Best]

A return to cooler and wetter Septembers will more than likely lead to a switch back to a climatological pattern that leads to colder winters. As long as Septembers continue warm and dry like recent years it is difficult to see a cold winter in the UK, or the winter patterns to deverge away from the patterns of recent winters.

I can see no connection metereologically between a mild September and a mild winter. 1985/6 is a good example of how the connection doesn't work. 01st October 1985 was virtually 30C, only to find Feb 1986 recording a staggering -1.1C for the month: since 1947 only one other month has ever been colder. Not even the winter of 1978/9 produced a month as cold as Feb 1986. And I remember it vividly, just as I remember (quite honestly) sunbathing on 02nd October 1985 in London.

I was chatting to Mrs WIB the other day about the way one can play with sets of statistical data. If you take 1000 surveys over the past 5 years on anything from breastfeeding (on my mind right now) to hair colour you will inevitably find some random quirks: for instance that dog walkers are more likely to eat cheese than cat lovers; or that people who holiday in France are 10x more likely to get prostate cancer. It's because the sample sizes across the surveys are not large enough to stop the quirks that get thrown up, but they are in fact meaningless and unconnected. In this instance you'd need to do a much bigger evaluation than is possible from simply looking at the CET, and ultimately publish something in a peer-reviewed journal where it would get properly scrutinised. At the moment I'm confident that any recent trend of warm Septembers and mild winters is nothing more than a quirk that has no scientific basis. 1985/6 very much suggests that to be the case.

Last night was milder than expected owing to cloud cover.

[stormforce 10]

I can see no connection metereologically between a mild September and a mild winter. 1985/6 is a good example of how the connection doesn't work. 01st October 1985 was virtually 30C, only to find Feb 1986 recording a staggering -1.1C for the month: since 1947 only one other month has ever been colder. Not even the winter of 1978/9 produced a month as cold as Feb 1986. And I remember it vividly, just as I remember (quite honestly) sunbathing on 02nd October 1985 in London.

I was chatting to Mrs WIB the other day about the way one can play with sets of statistical data. If you take 1000 surveys over the past 5 years on anything from breastfeeding (on my mind right now) to hair colour you will inevitably find some random quirks: for instance that dog walkers are more likely to eat cheese than cat lovers; or that people who holiday in France are 10x more likely to get prostate cancer. It's because the sample sizes across the surveys are not large enough to stop the quirks that get thrown up, but they are in fact meaningless and unconnected. In this instance you'd need to do a much bigger evaluation than is possible from simply looking at the CET, and ultimately publish something in a peer-reviewed journal where it would get properly scrutinised. At the moment I'm confident that any recent trend of warm Septembers and mild winters is nothing more than a quirk that has no scientific basis. 1985/6 very much suggests that to be the case.

Last night was milder than expected owing to cloud cover.

wib,

your absolutely right, there is sbsolutely no connection whatsoever,1986 was a good example.

i think your on for your bet with teits, ireland to reach 23c by friday so maybe 26 or 27c a possiblity in england.

[Gavin P]

Yet, 1995/1996 was the last proper, bona-fida cold winter, and was preceded by a cool and wet September and a warm and dry October. Everybody says that autumn weather patterns play no part in what will happen in the coming winter, but we can't escape that one truth that the autumn of 1995 was a unique pattern that hasn't repeated since and neither has a winter like 95/96.

Edited September 5, 2007 by Gavin P

[stormforce 10]

Yet, 1995/1996 was the last proper, bona-fida cold winter, and was preceded by a cool and wet September and a warm and dry October. Everybody says that autumn weather patterns play no part in what will happen in the coming winter, but we can't escape that one truth that the autumn of 1995 was a unique pattern that hasn't repeated since and neither has a winter like 95/96.

yes 1995 was exceptional for extremes, one of the hottest summers, i vividly remember october being a lovely warm settled month and then out of the blue came a severe cold plunge xmas eve, jeepers creepers i thought my ears were gonna fall of xmas morning daytime maxima -5c.

[North-Easterly Blast]

There is certainly a correlation between warm Septembers and mild winters, especially in the modern day. It was certainly true for the 1900s and up to present that a warm September certainly increased the chance of the forthcoming winter being milder than average. Here we go:

Septembers of 14.5 or greater and the following winter CET since 1900:

2006: 16.8 Winter 06-07: 6.4

2005: 15.2 Winter 05-06: 4.2

2004: 14.9 Winter 04-05: 5.2

2000: 14.7 Winter 00-01: 4.5

1999: 15.6 Winter 99-00: 5.4

1998: 14.9 Winter 98-99: 5.4

1991: 14.7 Winter 91-92: 4.6

1989: 14.7 Winter 89-90: 6.2

1985: 14.6 Winter 85-86: 2.9

1981: 14.5 Winter 81-82: 2.6

1980: 14.7 Winter 80-81: 4.5

1961: 15.2 Winter 61-62: 3.5

1959: 14.9 Winter 59-60: 4.6

1958: 15.1 Winter 58-59: 3.5

1949: 16.3 Winter 49-50: 5.1

1947: 14.9 Winter 47-48: 5.1

1941: 14.5 Winter 41-42: 2.2

1934: 14.6 Winter 34-35: 6.1

1933: 14.9 Winter 33-34: 3.2

1929: 15.3 Winter 29-30: 4.6

There were 20 14.5*C+ Septembers in the last 100 years, and four of them have led to cold winters, and two others slightly below average. The other 14 warm Septembers all led to mild winters, or average at best. Also before anyone mentions the warm September = mild winter theory not working in 1981 and 1985 I will point out that those two Septembers were not entirely in the same mould as recent warm Septembers. Sep 1981 was also a very wet September, and certainly not the warm anticylonic type of recent years, and Sep 1985 had a much warmer second half than the first half, and indeed the last time that September had a warmer second half compared to the first half, whereas most recent Septembers have often seen warm or even hot dry weather in the first half.

There is certainly a correlation in the modern day between a warm September and a cold winter. Whatever the summer weather pattern has been in the last ten years, whether it be very hot (2003, 2006) or reasonably warm and dry (1999, 2005), or warm and wet (2004), or rather cool and wet (1998, 2007), we just get stuck in a rut of a warm and anticyclonic September.

[Gavin P]

I still have a feeling this September isn't going to be massively above average. Despite the coming warm spell, I think we'll, more often than not, have relatively cool 850's, and chilly nights will be common. I don't forsee a CET much above 14.5c, which is a risky prediction, given that its looking very anticyclonic right now.

Edited September 5, 2007 by Gavin P

[West is Best]

There is certainly a correlation between warm Septembers and mild winters, especially in the modern day.

Fred, you have done nothing more than produce a series of stats. That's not a demonstration of a correlation surely?. No? Can you start explaining exactly how or why metereologically there is this alleged correlation? Otherwise I shall continue to be confident it is nothing more than a statistical quirk - the kind of quirk which gets broken.

At the moment the correlation argument you have presented is little more than saying that because ball No 14 has come up more often than any other in the lotto it is likely to go on doing so.

Let's have some sound metereology on this: jetstream? SST's? Can you also explain how winter 1985/6 differed metereologically from what the correlation is supposed to be?

Note: friendly questions

Edited September 5, 2007 by West is Best

[summer blizzard]

Sorry, but i'd say your arguament is falling apart...

Only 8 ov those 19 were followed by mild winters.. (Under 50% - no link)

Only 6 were followed by average winters (Under 50% - no link)

When you combine the two, you do get a 60%+ link, however the only conclusions that you can draw from that is that a warm September does not favor a cold winter (CET of 3.9C or less) however nor does it favour a warm one.

As for how anticyclonic/cyclonic they were, that to me is irelevant, September 2006 was cyclonic but the hottest on record, September 2004 and 2005 were anticyclonic but no more than 1.5C above average.

[SP1986]

The very idea that the weather in one month might influence a whole season directly is actually quite amusing in my opinion. Any cold/mild winter this year will be to do with external indirect effects from things like sea surface temperatures, atmsopheric cycles and jetstream movements and other oceanographical/toposherical phenomena. However statistics do sometimes offer us a vague idea of what may be coming - and considering there is no other accurate way of forecasting seasons that we fully comprehend, we'll have to make do with it for now

[Weather-history]

I posted a table, a couple of years ago of Septembers with a CET of 15 and greater and what the following winter CETs were. It was generally mix the results, overall.

[stormforce 10]

whatever way september pans i dont think it will have a direct impact on our oncoming winter,

if synoptics and ssta, cooling of seas, -ao/-nao, these are gonna influence our winter.

[Stratos Ferric]

...

There is certainly a correlation in the modern day between a warm September and a cold winter. Whatever the summer weather pattern has been in the last ten years, whether it be very hot (2003, 2006) or reasonably warm and dry (1999, 2005), or warm and wet (2004), or rather cool and wet (1998, 2007), we just get stuck in a rut of a warm and anticyclonic September.

NEB, for the avoidance of doubt, I'm assuming that you're suggestion of "[certain]correlation" is that there isn't one, there not having been a significantly cold winter in the past ten years.

Sorry, but i'd say your arguament is falling apart...

Only 8 ov those 19 were followed by mild winters.. (Under 50% - no link)

Only 6 were followed by average winters (Under 50% - no link)

When you combine the two, you do get a 60%+ link, however the only conclusions that you can draw from that is that a warm September does not favor a cold winter (CET of 3.9C or less) however nor does it favour a warm one.

As for how anticyclonic/cyclonic they were, that to me is irelevant, September 2006 was cyclonic but the hottest on record, September 2004 and 2005 were anticyclonic but no more than 1.5C above average.

Anyone pondering the statistics here might be well served by seeing my response to GP last night concerning his rather more robustly analysed relationship between AO values through the year. As various posts above suggest there is no correlation, and even if there were it might safely be assumed to be spurious. This type of discussion comes around every 2-3 months, then with abandon as people start to reawaken their wanter hopes with the waning of the summer sun. There is, nowhere in the CET data set, any robust and consistent correlation between the signal given by one month and the outturn a period of time later.

There is rather more surety to be had by looking at where we are in the climatic cycle overall. Month to month, and year to year, we oscillate - generally within fairly narrow bounds - either side of the rolling annual mean. It follows that when climatically we are in a cold period that cold months are more likely; the reverse, of course, is equally true. For that reason alone the odds are not favourably stacked for the forthcoming winter.

[Stratos Ferric]

Yet, 1995/1996 was the last proper, bona-fida cold winter, and was preceded by a cool and wet September and a warm and dry October. Everybody says that autumn weather patterns play no part in what will happen in the coming winter, but we can't escape that one truth that the autumn of 1995 was a unique pattern that hasn't repeated since and neither has a winter like 95/96.

Gavin, I haven't looked, but I know that if I do the same will be absolutely true of every year since 1995 - a unique autumn leading to a unique winter.

Back to the numbers: there will have to be a dramatic change in the progged set-up, and / or a decidely - and unlikely - cold end to the month for September to follow July and August into sub-par territory. 15 plus looks good even at this early stage.

[Gavin P]

I think its a bit early to say that Stratos Ferric. Plenty of time for changes yet. Infact, I reakon the models are hinting at a cool northerly interlude potentially appearing early next week.

[Stratos Ferric]

I think its a bit early to say that Stratos Ferric. Plenty of time for changes yet. Infact, I reakon the models are hinting at a cool northerly interlude potentially appearing early next week.

Indeed they are, but given that yesterday they had the UK in a couple of days of full blown northerly it would appear that the winter game of choice - the GFS teaser that comes to nought - may be back with a vengence. As I observed yesterday, we have a week or so of high probability warmth (including some mild nights) followed by the potential in FI for some [low probability at this distance] cool stuff. Given the projected upcoming warmth it would need quite a correction to get back to par.

[Stu_London]

More likely another in the 15s.

Is it possible SF that you have forgotton the temperature graph falls away quite steeply in the second half of September (almost 2 whole degrees) - to get into the 15Cs we will need to either get to 17C by mid month (unlikely) and then maintain average, or reach somewhere in the mid to late 15s (more likely) by mid month and then for the weather become progressively more above average just to maintain that level.

It's to far out to say categorically but my reading of the models over the last day of so would suggest that 15s (1.5C to 2C above average) is a bit high. Lets not forget we were 0.8C below average to the 4th.

[West is Best]

Is it possible SF that you have forgotton the temperature graph falls away quite steeply in the second half of September (almost 2 whole degrees) -

As this year has demonstrated though Stu, weather doesn't follow the average script. A cooldown is certainly extremely likely but by no means certain at least in terms of minima. A lot depends on the type of pattern if and when the breakdown occurs. A northerly would see a plunge. A zonal flow for instance could keep the minima temps up.

An interesting month certainly. We have a huge temp climb to go through before any cooldown. Hadley was already on 15.1C by the 4th, and yesterday will have seen a lurch upwards on climate uk probably to the same level.

Edited September 6, 2007 by West is Best

[BLAST FROM THE PAST]

Fred, you have done nothing more than produce a series of stats. That's not a demonstration of a correlation surely?. No? Can you start explaining exactly how or why metereologically there is this alleged correlation? Otherwise I shall continue to be confident it is nothing more than a statistical quirk - the kind of quirk which gets broken.

At the moment the correlation argument you have presented is little more than saying that because ball No 14 has come up more often than any other in the lotto it is likely to go on doing so.

Let's have some sound metereology on this: jetstream? SST's? Can you also explain how winter 1985/6 differed metereologically from what the correlation is supposed to be?

Note: friendly questions

Richard

Unless NEB is a Fred too those questions I cannot answer...because I don't agree with the theory.

kind regards

Fred aka BFTP

[West is Best]

Richard

Unless NEB is a Fred too those questions I cannot answer...because I don't agree with the theory.

kind regards

Fred aka BFTP

LOL Put it down to the sleepless nights Fred!

So, North-easterly blast: over to you!

[Stu_London]

As this year has demonstrated though Stu, weather doesn't follow the average script.

It doesn't but statisically September is still more likely to become progressively cooler as the month goes on regardless of what has happened previously.

Therefore to maintain a level that is above average for the month as a whole, means that as time goes by, you need to be further and further above the mean just to maintain the level. The model output seems to be suggesting the reverse of that, although as SF says (quite correctly) the warmer stuff is in the more relaiable timeframe, which means the probability of a warm month as a whole is that much greater.