Autumn and Winter discussion...

[Stratos Ferric]

Suffice to say IMO there is nothing to learn from someone who bases their view on percentages and probabilities other than knowing their personal opinion of what is likely to happen. Statistics tell you what has happened but they can be skewed in any direction you wish, to prove any point you wish, and they do not take account of the wider picture which influences future events and they are only as good as their latest breath.

Tamara

Tamara, don't disagree with any of your opening paragraphs; any comments I make about crowing are based on past winters, and are an early call this year for more focus on robust assessment of why, and why not, forecasts are or are not proving to be accurate.

Yor swipe at staistics though is a naive one. Without statistics how can you prove that a forecast method works? es statistics can be skewed, but so can airliners be flown into tower blocks, and cars used to run people over. In the right, and responsible hands, statistics describe occurrences, and help to determine trends which are otherwise just point events. Witout staistics many on here would welcome each winter with equal prospects for cold as if we were still int he 1700s. We are not, the baseline has moved on somewhat; like it or not cold seasons, and cold interludes, are less likely at present than they once were, even in the fairly recent past.

And do you not suppose that the statistics that you seem to deride are at the very heart of the short term models we all spend our time slavishly obsessing over for the merest hint of something "blue" on the distant charts? You cannot model forwards without understanding cause and effect, and that requires at the very least basic statistics and understanding of probabilities to link a precedent to an antecedent.

Don't get me wrong, as I've illustrated elsewhere today, it's impossible to extrapolate forwards from current data sets to useful detail about winter outcomes, but I'm not sure that your inference that stats cannot be extrapolated forwards would meet with anything other than derision from the actuarial communities who spend their lives doing precisely that in order to make hidden decisions that effect all of our lives.

At the very least, at present, like it or not, our climate is warming. That much can be proven only by statistical assessment (of whatever data you choose; temperature, growing season, species dispersion, glaciation etc.). Given the climatic flywheel this baseline makes any prognosis for a colder winter less likely than was the case when I was younger. I don't like this any more than you obviously do not, but I'm not going to shout and holler as if doing so makes this fact less true. For sure, we could still get a cold winter, and will still get some cold interludes, and statistical extrapolation doesn't allow this detail to be forecast, but - and please point me at the solution if I'm wrong in the following assertion - LRFs, whatever their provenance, are ot curretnly achieving this either.

So I agree, continue to make efforts to improve LRFs, but be clear that doing so requires an approach that allows for post-event dissection, otherwise we go on doing the same old things, expecting a different outcome, which is one definition of madness.

[acbrixton]

Yes! I wholeheartedly agree. As learning tool, some of them are excellent! As a means of saying, accurately, what the winter, or parts of it will be like, at this distance.........lorryload of salt time, for almost every one!

Precisely Paul: a well argued LRF has, to my mind, two useful functions...first it assists the education/understanding of the less knowledgeable (such as me) and secondly, provided that lessons are learnt from each LRF's success or failure and are applied to future LRFs, slow and incremental improvements to technique may, over time, gradually improve their reliability.

I believe that the MO describes its seasonal forecasts as "experimental" (and the GFS applies that term to forecasts beyond T144): those (gifted) amateurs who produce LRFs would do well to bear that in mind.

Regards

ACB

[The PIT]

I'm just thinking isn't this Autumn like last autumn in the setup. A block over Europe thats tending stick around keeping us in a flow of mainly SW. This week we're going to have a long feed pushing warm air northwards like we id last year. Okay theres a strong hint that the high preasure will build and feed us an easterly later this week but the blocking is so similar.

[Dawlish]

Statistics tell you what has happened but they can be skewed in any direction you wish, to prove any point you wish, and they do not take account of the wider picture which influences future events and they are only as good as their latest breath.

Tamara

No, sorry, that's not true. With respect, that just shows your personal lack of understanding about the use of statistics. Just because you are trotting out the layman's view of the use of statistics, it does not, in any way shape, or form, make it true. To tar a whole branch of mathematics with that kind of completely dismissive attitude beggars belief.

Let me put the use of statisticss in a little bit of context.

Most areas of scientific study use statistics to either explain their results, or to predict outcomes. It is not an exaggeration to say that scientific research would be lost without the advances in statistics that have taken place this century. NASA

Statistical analysis is important in many areas of everyday life ranging from actuarial work in insurance and finance to quality control, image processing and the analysis of scientific and technical data. Medical and social statistics are increasingly important in identifying causes of illness and long-term trends in society. - City University, London, prospectus.

I could go on for a long time, but I think I'd be better off asking you to do this....

........It isn't often that I quote Wikipedia, but as an introduction to statistics, this will do fine. Snowp, before you trot out that nonesense about being able to skew statistics... "in any direction you wish, to prove any point you wish, and they do not take account of the wider picture which influences future events and they are only as good as their latest breath"., I heartily suggest you read it and seeing just what statistics can do and how lacking scientific method would be without them.

I honestly hope that has put this erroneous view about the use of statistics to bed.

Paul

Edited October 8, 2006 by Dawlish

[Stratos Ferric]

Thank you for your reply

My point about statistics was not to suggest that they cannot be extrapolated forwards to base future predictions/outcomes if used properly - because indeed they can. It was to suggest though, that a line should be drawn when they are not used properly or as a tool alone without taking into account newer factors or shorter term drivers which may flesh out and put a different perspective of future predictions and possible outcomes. In that way, a statistic will tell you what has happened but it is not neccessarily a guide to what the future may hold.

I have been/still am trying very hard to avoid personaling this, but there is a very small element of the forum who rather crow about stats/probabilities and their supremacy over personal research, presentation and conclusions (LRFs in this context) and discredit any other offering beyond statistical evidence and probablities as inferior. An impression of 'I am always right' and 'the most accurate'.

I would repeat that there is more than just accuracy that is at stake here - as I tried to demonstrate earlier on.

Your other point which once again appears to hint about me somehow refusing to accept an overall warming trend has been done to death on here - and I have put forward my answer to this one up to three times, in detail, just this week which would actaully be for your 'benefit' so to speak, as well as one or perhaps two other individuals. It is disappointing that it seems not to have been read properly, taken on board or understood - because your comments do not reflect accurately what I really think.

Tamara,

Apologies if I have missed your explanation regarding your stance on warming; I'm pretty sure I manage more than most on here to read threads diligently and completely, but nobody can be expected to read every post I think.

I assume you're refering to our bookie friend when mentioning individuals using stats to set predictions against which they then take bets. Not quite sure what your complaint is here, but w.r.t. Dawlish he seems to be reasonable clear about the specifics he offers odds on, and I'm not sure that he often relies on stats to make the punts asmany of them are in the near distance where stats really aren't adding any value. Fair enough on his odds for winter though; we may not like the apparent crudeness of the method, but if it yields results then it's hard to argue with. You can suggest there is "more than accuracy" at stake, and I guess you mean method that can be transferred to other situations, and it's hard to argue with that, apart from that it presupposes that we will ever be able to forecast that far into the future. We may not, just as there are people in the lab trying to produce superconductors at absolute zero; we know what 0K equates to, but making it happen in practice is a different matter entirely. Just because we can imagine something it doesn't have to be the case that we can definitely make it reality.

[TEITS]

Thank you for your reply

My point about statistics was not to suggest that they cannot be extrapolated forwards to base future predictions/outcomes if used properly - because indeed they can. It was to suggest though, that a line should be drawn when they are not used properly or as a tool alone without taking into account newer factors or shorter term drivers which may flesh out and put a different perspective of future predictions and possible outcomes. In that way, a statistic will tell you what has happened but it is not neccessarily a guide to what the future may hold.

You are absolutely right SP.

Assuming our future weather based on previous year's is a rather basic way of looking at our climate and isn't taking all the factor's in that make our climate. As an example our winter's since the 80's have been largely dominated by a positive NAO resulting in a rampant Atlantic with little high lattitude blocking. Now suppose the next decade bring's a return of this high lattitude blocking with more frequent occurences of continental/artic airmasses our winter's would be far different to those in the 80's.90's and would render the stats on the 80's.90's as being pretty useless.

So you are quiet right SP Statistics are ok if used properly but you need to look at all the influences that make our climate. I believe you should judge each year on it's own merits and when it's comes to what to expect this winter I prefer to look at the sign's in Autumn e.g SST's/Teleconnections rather than look at stats for the previous years to base a forecast for the oncoming winter.

Let me say this how many people on this forum expected August to be as poor as it was after the blazing hot June/July??

[Shiny_Bottom_1]

I cant believe we are arguing about statistical analysis of weather :huh:

If i'm honest, and going on what I know of statistics (which may not really be alot), both of you (SP and Dawlish) are actually correct to points in your own views.

Statistics are, yes, a good way of looking into the future. They give probabilities of outcomes.

However, statistics are also measurable, statistically (if you get what I mean). Surely a statistic is reliable upon its base and fixed factors? If there are less fixed base values or factors (i.e. factors which are not part of a trend, but values that are just historic), then the statistics are less reliable.

The reliability is also dependant on the confidence of the future. Where the values are again, not part of a trend, then the statistical confidence decreases.

Because the weather is not a fixed variable, then to use statistics as a 'complete' arguement is not a fair method. But in turn, to ignore the statistic (even though they are not using fixed variables), is also not taking into account all methods of analysis.

Statistics are good indicators in life and/or science, but they are like many things, variable in confidence (just like various LRF forecast methods).

Well...maybe?

[kold weather]

I have to agree Chris, it seems people are getting more heated about statistics then about discussing the actual possiblties of the weather this winter!!!

LRF are never going to be 100% accurate but they can give a good guide as to what is at least possible. Despite the [problems they have come a long way since even 10 years ago. Heck in 1998 the computers forecasted a La Nina...next year we got the second strongest El nino ever recorded!!!

Stats are intresting...but right now they seem to be the be all and end of all of some peoples arguements!

Edited October 8, 2006 by kold weather

[WhiteFox]

So you are quiet right SP Statistics are ok if used properly but you need to look at all the influences that make our climate. I believe you should judge each year on it's own merits and when it's comes to what to expect this winter I prefer to look at the sign's in Autumn e.g SST's/Teleconnections rather than look at stats for the previous years to base a forecast for the oncoming winter.

The problem is that you are assuming that statistical analysis only involves looking at basic variables for the previous years. It's worth pointing out that the complex teleconnections used by Steve and GP are based on statistical analysis. It just emphasises the point that statistics is key in every branch of science.

[TEITS]

The problem is that you are assuming that statistical analysis only involves looking at basic variables for the previous years. It's worth pointing out that the complex teleconnections used by Steve and GP are based on statistical analysis. It just emphasises the point that statistics is key in every branch of science.

Which is why I say using stats are ok to a certain degree. SST's have a major impact on our weather and yet we know how quickly these can change and im sure GP/SM are closely montioring these over the next 6 weeks.

[AtlanticFlamethrower]

This winter? 70% chance warmer than average, 10% chance average, 20% chance colder than average.

Statistics made to make the best of an environment in which there is incomplete information cannot obscure the fact cold winters don't just happen by accident and that somethings happen some way rather than another for reasons. After so many failed seasons, and the synopic winter of 2005/6 (Ou est Mons. Bartlett?) signs are this year may finally deliver synopics and the cold. Statistics are a tool that keeps our expectations in perspective but it is no more than a tool of the bigger perspective. Knowing what will happen this winter requires specific forecasting. LRFs are not perfect but are aimed at this specific winter. I seem to remember last year the LRFs predicted the pattern change - more blocked easterly set-up with southerly running jet - so I'm not going to dismiss them as hopecasts.

[WhiteFox]

Which is why I say using stats are ok to a certain degree. SST's have a major impact on our weather and yet we know how quickly these can change and im sure GP/SM are closely montioring these over the next 6 weeks.

Of course, and in the absence of perfect information we have to make assumptions all of the time. It is the level of assumptions we have to make which will influence predicted outcomes. One of these areas is of course SSTs.

As we can never have access to perfect information there will always be errors in all forecasts, which probably brings us back full circle to some of the original points! I don't think there will ever exist a time when we can give any forecast for a few months ahead with any certainty other than probabilities. We've seen how difficult it is to forecast days ahead past day 5, so the best that any other methods can hope to achieve, given that they are governed by the same statistical laws as conventional models, is to provide implied probabilities of patterns rather than specifics. That's not to say that different methods do not have any value. One thing I enjoy about Steve Murr's and GP's forecasts is the way they re-analyse using short- to medium-term forecasts and explain what would need to happen for the forecasts to return true.

Measurements will improve both in accuracy and multitude, but it is impossible to measure everything everywhere at the same time. Sampling will remain the boundaries of accuracy.

[phantom forecaster]

A really fascinating read in this thread.

LRF? Stats/odds? I'm not educated enough (meteorologically wise) to comment either way. However as a prefessional lurker and non-contributor, I find the LRF's that are produced by the regulars on here compelling reading as are the counter arguments to them.

As for the stats, well, Ladbrooks are pretty good at providing them so I'll wait for their annual odds on a white Christmas! I guess you never see a poor bookmaker so stats must have their place

My preference will always be for cold and my hope will always be for snow and watching my kids grow up with year after year of snowless winters down south only hightens that preference.

Yido,

Season ticket holder myself, do you go regularly?

[Stratos Ferric]

Chris,

Much sense in that post, and ahelpful summary of where this argument seems to have progressed to.

Statistics are, yes, a good way of looking into the future. They give probabilities of outcomes.

However, statistics are also measurable, statistically (if you get what I mean). Surely a statistic is reliable upon its base and fixed factors? If there are less fixed base values or factors (i.e. factors which are not part of a trend, but values that are just historic), then the statistics are less reliable.

Reliable extrapolation of trend requires consistency of context, and I think when SP, SM (both) and others object what they are pointing to is the short-term drivers of weather. I don't think those of us who use stats as part of our projections (whatever form they take) are blind to this. What's more I can no more project detailed LR synoptic outcomes using stats tan I can paint detail on a watercolour with a 4" brush. That's not the point though, what I can do is propose the general "wash" of the season, and as Dawlish pointed out earlier, using the principles of, for example, statistical process control, the likely limits of "performance" can be inferred.

With better understanding of the other drivers, e.g SSTs, and robust longitudinal data sets, then the use of extrapolation will improve, both in accuracy an in use towards proposing pressure fields as well as actual outcomes. There is o reason why the different approaches cannot co-exist, and every likelihood that over time improvements in synoptic LRFs will be proven and reinforced only by use of stats to identify and prove hypotheses.

What IS clear is that if we get unusual forcing, or the "all things equal" stop being equal, then any extrapolation becomes less reliable, just as you suggest.

Hi TEITS

Me too on this - and that is where the info GP etc provide is so invaluable - irrespective of whether it eventually proves right and wrong. This is where the short term drivers - teleconnections, SST's etc are at least as valuable as the past statistics, as you say.

Tamara

Tamara,

They are only VALUABLE in a forecasting sense if they are proven to be correct in identifying outcomes. They are valuable learnign tools as and when they are dissected, irrespective of whether the outcome was as projected or not. If neither of these conditions hold true then from my perspective they are not valuable; they might be interesting to read, or colourful to look at, or atract a whole host of other eye-candy type plaudits, but reference forecasting I can't see how they're valuable.

Perhaps what we need is a dedicated partition for forecasts to be posted, and for outcomes to be analysed, so that lessons can be learned anbd captured deliberately.

None of this is about hounding individuals, and it should not degenrate into a slanging match, but equally people should not be too "precious" about their forecasts. Whatever anyone on here might fancy for their projections, as one of our number sensibly pointed out earlier, all LRFs should be treated as experimental.

Statistics made to make the best of an environment in which there is incomplete information cannot obscure the fact cold winters don't just happen by accident and that somethings happen some way rather than another for reasons. After so many failed seasons, and the synopic winter of 2005/6 (Ou est Mons. Bartlett?) signs are this year may finally deliver synopics and the cold. Statistics are a tool that keeps our expectations in perspective but it is no more than a tool of the bigger perspective. Knowing what will happen this winter requires specific forecasting. LRFs are not perfect but are aimed at this specific winter. I seem to remember last year the LRFs predicted the pattern change - more blocked easterly set-up with southerly running jet - so I'm not going to dismiss them as hopecasts.

AFF, is it just me or am I right in suspecting that precisely the same sort of claims are made every year? I hoenstly cannot recall an autumn in the six or so years that the web has been a very active resource when chat rooms have not been full of people seeing some sign or other in the synoptics that are suggestive of cold weather. What I am certain of is that if you put expectation on one side of the scales, and delivery on the other, the expectation side would be firmly on the deck.

Yes, last year's forecasts were pretty good, but there were several failures prior to that. I certainly shan't be reading any runes until well into next month, but ahead of that, the likely limits to the forthcoming winter seem to me to be set by the stats suggested in my posting in the "what does October tell us?" thread. Like many on here, I do hope the fundamentals change, but I wouldn't be betting on it happeneing.

[ukmoose]

Suffice to say IMO there is nothing to learn from someone who bases their view on percentages and probabilities other than knowing their personal opinion of what is likely to happen. Statistics tell you what has happened but they can be skewed in any direction you wish, to prove any point you wish, and they do not take account of the wider picture which influences future events and they are only as good as their latest breath.

Tamara

Tamara,

I didn't mean to imply that you, or anyone else, can learn anything about weather forecasting from people (perhaps like me) who are interested in past weather patterns. I don't lay claim to being able to teach anyone anything about predicting the weather.

You and Teits seem to be in agreement (apologies if I'm wrong) that the past doesn't mount to a hill of beans when looking at the future. I just can't agree with you there - without past weather patterns we would have no climatological studies at all. Climate is based on previous statistics gathered over the years. In essence, climate is statistics. Perhaps I am more interested in weather patterns and climate than actual forecasting, but I do enjoy reading people's forecasts and can certainly appreciate the blood sweat and tears that must go in to producing them. I also think the forum needs these forecasters because we are all interested in what might happen and cold lovers like me hope they come up trumps.

However, I don't think we can get away from what has been happening to our weather since about 1990. It is not long in statistical terms but it does throw up a lot of data that is hard to ignore and the facts are there for all to see (and they are measurable facts). We are getting warmer and periods of cold are becoming less frequent and less severe whether we like it or not. (I don't). Therefore, I feel reasonably confident in predicting that the winter ahead (December, January and February) will be mild with temperatures above the long term average. I base this purely on what has happened in recent years - not sexy I know, but let's see what happens.

Having said that, I will continue to read people's forecasts with interest and will acknowledge that they all know a lot more about the weather than I do; theywork damned hard at putting their forecasts together and I enjoy reading them. I just think history is against us.

ukmoose

.

[AtlanticFlamethrower]

AFF, is it just me or am I right in suspecting that precisely the same sort of claims are made every year? I hoenstly cannot recall an autumn in the six or so years that the web has been a very active resource when chat rooms have not been full of people seeing some sign or other in the synoptics that are suggestive of cold weather. What I am certain of is that if you put expectation on one side of the scales, and delivery on the other, the expectation side would be firmly on the deck.

Yes, last year's forecasts were pretty good, but there were several failures prior to that. I certainly shan't be reading any runes until well into next month, but ahead of that, the likely limits to the forthcoming winter seem to me to be set by the stats suggested in my posting in the "what does October tell us?" thread. Like many on here, I do hope the fundamentals change, but I wouldn't be betting on it happeneing.

Over recent years forecasts have got technically better and more accurate. For example, when I first came to this site forecasters seemed to interpret SST anomalies and there effect on synoptics in opposite ways. It is now agreed warm anomalies encourage ridging, cold troughing, but two years ago there were people forecasting on the basis it was the opposite way round! Remember when cold anomalies off Western Europe were the purported reason for a persistent Bartlett high (label Bartlett is I think incorrect for this image - I use the word to express high pressure of western Europe pumping warm air our way from the Azores).

That is one example where LRF were once compromised by lack of basic understanding of weather processes.

An example showing improvement of techniques is the use of teleconnections to forecast - when I came to this site teleconnection theory was just beginning to take off. After last year I think the LRFs have matured and their use of teleconnections makes them worth their place at the table.

Those who think our warm winters owe mostly to climate cycles and the solar maximum are going to watch out for a cold turn. With this belief it makes sense that every year they would expect cold winters to return to end the warm winter. It makes sense they would look for signs every year. It is only natural their forecasts would be skewed to cold by their looking for a return of cold. This explains why every year some forecast a return to cold. The fact these forecasts have been wrong does not make the idea that a return to colder winters is coming wrong. It certainly suggests their forecasting techniques are not quite right and these techniques are being refined every year.

I personally find it hard to believe that even someone who believes in runaway man-made AGW can scoff at forecasts of a very cold winter in UK. We are so near the North Pole the right synoptics combined with a cold pool would plunge us into a big freeze. Were this winter to be as blocked as last year it would be clear that the major flip happened at the end of winter 2004/5, continued through 2005/6 and intensified further in 2006/7.

Edited October 8, 2006 by AtlanticFlamethrower

[SMU]

Hers something funny to think about-

Lets say for FUN this greenland high And the Scandi trough were to last from next weekend until End of Feb with a record breaking Autumn & Winter- I wonder how long it would be for the traditional cold rampers on here to start saying they wanted some warmer weather....

Maybe before the start of the official Winter!!!!!

lol

S

[BLAST FROM THE PAST]

Those who think our warm winters owe mostly to climate cycles and the solar maximum are going to watch out for a cold turn. With this belief it makes sense that every year they would expect cold winters to return to end the warm winter. It makes sense they would look for signs every year. It is only natural their forecasts would be skewed to cold by their looking for a return of cold. This explains why every year some forecast a return to cold. The fact these forecasts have been wrong does not make the idea that a return to colder winters is coming wrong. It certainly suggests their forecasting techniques are not quite right and these techniques are being refined every year.

I personally find it hard to believe that even someone who believes in runaway man-made AGW can scoff at forecasts of a very cold winter in UK. We are so near the North Pole the right synoptics combined with a cold pool would plunge us into a big freeze. Were this winter to be as blocked as last year it would be clear that the major flip happened at the end of winter 2004/5, continued through 2005/6 and intensified further in 2006/7.

AFT

Re the skewed to cold I don't think it is that basic, certainly not with some. I ceratinly for one lean towards natural cycles and certainly believe solar cycles are more important than some will accept. But IMO there was a change in longterm synoptics that happened in feb 05. We are 18 months down the line and so far the change is holding out. What high latitude blocking does IMO is increase the likelihood of colder winters but not guarantee them as the one could be on the wrong side of the block The LRF folk who use and display their technique are not 'skewed by cold' but are displaying what is likely to occur with that pattern in place. No one is forecasting 47/63 but an 80s style winter is in the offing this time round.

BFTP

[Gray-Wolf]

Heck in 1998 the computers forecasted a La Nina...next year we got the second strongest El nino ever recorded!!!

And this May NASA /NOAA released a joint statement that the 'La-Nina' wasn't going to impact on thge Atlantic Basin Hurricane season!!! Didn't they forcast the September 'switch'??

Maybe not a lot changes ,even with our advances, when the goal posts are constantly shifting (in a warming world?).

I obviously haven't been watching these things as long as you guys but I've never seen the Tropical Atlantic run backwards before!!, not to the point of running a 'retro' flow across Africa anyhows. To just witness what is occuring on the ground makes me wonder about 'short track' TM airmasses being a common feature this winter?

[Dawlish]

I personally find it hard to believe that even someone who believes in runaway man-made AGW can scoff at forecasts of a very cold winter in UK. We are so near the North Pole the right synoptics combined with a cold pool would plunge us into a big freeze. Were this winter to be as blocked as last year it would be clear that the major flip happened at the end of winter 2004/5, continued through 2005/6 and intensified further in 2006/7.

I would agree entirely, AF. An acceptance of GWUK should in no way preclude the possibility of a colder than average winter, it only reduces it, as the trend extends. I put the chances of a colder than average UK winter, in any one year at 20%, ie 5/1 against, or a 1 in 5 chance. Translated further, it means that I would expect, on average, a colder than average winter to come around one in every 5 years. The trend is continuing, however and the weight of evidence points to further global warmth (no certainty, just the weight of evidence). Those percentages held good through the 1990s. Since then we have had 4 warmer and one average(ish) winter. I don't really expect to change my odds on a colder than average winter until the end of the decade, but the trend is advancing and it wouldn'y take a statistical marvey to figure in which direction those odds are likely to be heading. given the last 10 UK winters (9 warmer, one average(ish)).

Paul

[West is Best]

I personally find it hard to believe that even someone who believes in runaway man-made AGW can scoff at forecasts of a very cold winter in UK. We are so near the North Pole the right synoptics combined with a cold pool would plunge us into a big freeze. Were this winter to be as blocked as last year it would be clear that the major flip happened at the end of winter 2004/5, continued through 2005/6 and intensified further in 2006/7.

I don't think being so near the north Pole really comes into it. The prevaling flow in the UK is south-westerly, and we are first in line from the Atlantic. Add to this the gulf stream and you can have all the cold you like to our north and east, if the west / north-west isn't right it won't come our way ... at least not for any sustained periods. The Atlantic will always return. Even with a blocked set up last year we didn't get a properly cold winter. The final nail in the cold coffin.

Whilst local variations to the general AGW trend are always possible, and it would be foolish to deny this possibility entirely, the chances become rarer with each passing year. As more and more warm records tumble around the globe it begins to look more and more desperate for those clinging to cold hopes for the UK. Personally I'd rate the chances of a below average winter now in the UK as 1/25 or 4%. This is reinforced by the full-blown return of the Atlantic this autumn.

I'm looking to record mild possibilities rather than the other way.

Edited October 9, 2006 by West is Best

[Weather-history]

This is reinforced by the full-blown return of the Atlantic this autumn.

What makes you say that, Richard? Looks blocked.

Believe it or not we are half way through autumn at the end of the coming weekend, if the Atlantic is coming this autumn it better get a move on!

[West is Best]

Let me say this how many people on this forum expected August to be as poor as it was after the blazing hot June/July??

As you know Dave the blocking was there for August - just 300 miles to our west. However, August may have seemed poor to you and some others but this is where statistics provide a useful corrective to mistaken personal perceptions. At 16.1C August was within 0.1C of the 1971-200 average, and comfortably warmer than both the 1961-1990 average (15.8C) and the 100 year rolling average (15.9C). It was of course a relative dull month, and whilst wet in much of eastern England it remained bone dry in much of the west (30% of rainfall totals in our parts!). All in all it was therefore an average month. The mood was lightened by a thumping warm September breaking the all-time record.

Statistics do have their uses, providing the source is rigorously tested.

Edited October 9, 2006 by West is Best

[Glacier Point]

Welcome back sceptical West.

Your claim that the Atlantic is full blown is slightly erroneous. The story so far......

Low pressure manifested to our SW all the way to the Azores, and much above heights recorded over Greenland. Not a classic autumnal zonal pattern with low pressures crashing in on a strong jet.

The Autumn so far has been one of strong meridional flows and a very weak southward displaced west-east pattern. This could all reverse spectacularly over the coming weeks however, but right now it is no way full blown.

GP

[West is Best]

What makes you say that, Richard? Looks blocked.

Believe it or not we are half way through autumn at the end of the coming weekend, if the Atlantic is coming this autumn it better get a move on!

Front after front has been crashing across our parts though Kevin, and moving through the east e.g. last night. Seems pretty Atlantic dominated to me!