Annual CET 2007

[Stratos Ferric]

I'd hate for things to go wrong like they did for you in September.....

Though it's looking okay at present...things can always change mind.

[acbrixton]

uk_cet_monthly_comparisons_081007.doc

Whilst looking at the table/graph showing projected UK average temperatures I wondered whether one could establish a relationship between the UK monthly averages and the monthly CET. The table attached attempts to do this: UK figures are from the UKMO Areal Series for 1971-2000 and the CET is UKMO Hadley 1971-2000. The monthly variance ranges from 0.7 in late winter to 1.6 in mid summer.

Obviously there would be a danger in trying to convert UK forecast averages for individual days to a precise CET equivalent as synoptic details might render such an exercise pointless but as a general rough and ready guide it perhaps is of use...

Regards

ACB

[Stratos Ferric]

My take on where we are coming up towards the year end.

The likeliest corridor looks like 10.65-10.75 now. October is running comfortably above par, and if it comes in high then there's a straight line trend developing out of summer. It's not unprecedented in recent times to oscillate quickly out of cold to warm then back to cold again, but it's only happened once in the last decade, interestingly enough in late autumn 2005, so there is a wee bit of form. December is also a tough call, still one of the most variable months year to year, and one with relatively little warming in the modern climate.

There's still a small chance of exceeding last year's high water mark, but I'd put it below 5%, not least because it would probably require a near recod December. More realistically the upper margin looks around 10.8, the lower margin presently around 10.55.

[Stu_London]

My take on where we are coming up towards the year end.

The likeliest corridor looks like 10.65-10.75 now. October is running comfortably above par, and if it comes in high then there's a straight line trend developing out of summer. It's not unprecedented in recent times to oscillate quickly out of cold to warm then back to cold again, but it's only happened once in the last decade, interestingly enough in late autumn 2005, so there is a wee bit of form. December is also a tough call, still one of the most variable months year to year, and one with relatively little warming in the modern climate.

There's still a small chance of exceeding last year's high water mark, but I'd put it below 5%, not least because it would probably require a near recod December. More realistically the upper margin looks around 10.8, the lower margin presently around 10.55.

Now it's well known that you have a mild bias SF, but sneaking an extra 0.1C on 2006 isn't really on is it.

[eddie]

I still don't think we should write of a final CET below 10.5C just yet. It certainly wouldn't require any kind of record breaking cold.

[Snowly does it]

I still don't think we should write of a final CET below 10.5C just yet. It certainly wouldn't require any kind of record breaking cold.

Its possible, to achieve it we would need to lose 4.3 degrees against last years output - if we can get down to 11.5 for October then 2.8 to lose from November and December is certainly do-able - although an outside bet I'd say. Really depends how much we can shave off last October's 13 (we are currently 0.3 degrees ahead of 2006 to the end of September)

Edit - before any bright sparks point it out, I am fully aware it actually depends on how much we can shave off all 3 months, but I see this months as the crucial one for 10.5 hunters.

Edited October 13, 2007 by snowmaiden

[Stratos Ferric]

Now it's well known that you have a mild bias SF, but sneaking an extra 0.1C on 2006 isn't really on is it.

Mmm, used some old numbers there: have to say I thought 10.92 looked high but forgot to recheck. Anyway, makes not a jot of difference on the projection for this year, other than bringing the record within touching distance.

Now it's well known that you have a mild bias SF, but sneaking an extra 0.1C on 2006 isn't really on is it.

Stu, the weather has a mild bias; I just report what's happening - witness the various projections I produced for last month which, I do recall, came in pretty much bang on the money. If I am perceived to have a bias it is because numptys like you persist in labelling me as having such.

I still don't think we should write of a final CET below 10.5C just yet. It certainly wouldn't require any kind of record breaking cold.

Not record breaking, but it would require a run which would be cold, and in the context of the period of cold we've already had over summer almost unprecedented in the past twenty years. Everything needs to be viewed in the context of the climate we now have, not that which occurred before Queen Victoria came to the throne I'm afraid.

[Snowly does it]

Not record breaking, but it would require a run which would be cold, and in the context of the period of cold we've already had over summer almost unprecedented in the past twenty years. Everything needs to be viewed in the context of the climate we now have, not that which occurred before Queen Victoria came to the throne I'm afraid.

Actualy 11, 7, 5.3 would bring us in at 10.5 SF - so 3 slightly above (71-00) average months would do it - not really cold by any context, average to above average would be nearer the mark than 'cold'

In context, October and December almost spot on their 10 year means and November 0.6 below its ten year mean - so OK, one month coming in slightly below the recent mean

Double edit - actually you can add 0.1 to any of those months of your choice and still achieve 10.50.

Edited October 13, 2007 by snowmaiden

[Stratos Ferric]

Actualy 11, 7, 5.3 would bring us in at 10.5 SF - so 3 slightly above (71-00) average months would do it - not really cold by any context, average to above average would be nearer the mark than 'cold'

In context, October and December almost spot on their 10 year means and November 0.6 below its ten year mean - so OK, one month coming in slightly below the recent mean

Double edit - actually you can add 0.1 to any of those months of your choice and still achieve 10.50.

On the assumptions you've made, yes, but you can see I'm projecting October to come in higher than 11. On that basis we need a slightly cool Nov + Dec. As I said above, on its own its possible, but in the context of what's happened recently it looks slightly less likely. Far from impossible, just, in my opinion, unlikely.

[Snowly does it]

On the assumptions you've made, yes, but you can see I'm projecting October to come in higher than 11. On that basis we need a slightly cool Nov + Dec. As I said above, on its own its possible, but in the context of what's happened recently it looks slightly less likely. Far from impossible, just, in my opinion, unlikely.

OK yes on the assumption of an 11.5 October then the 71-00 average required to get about 10.50 - maybe 0.1 below on one of those two months.

I'd err on the side of no at the moment as well but I think sub 10.6 is very much more likely. November and December in line with the only recent 'cold years end' we have to look at - 05, sees a final outturn (if Oct = your 11.5) of 10.4 which should probably be regarded as the bargain basement from here?

Edited October 13, 2007 by snowmaiden

[Stu_London]

On the assumptions you've made, yes, but you can see I'm projecting October to come in higher than 11.

Are you using the same spade you used to dig September's hole or have you bought a new one for October

Remember Stratos when it comes to the weather, to ASSUME generally makes an ASS of U and ME (but fair to say mainly U when it comes to temperature projections)

[SP1986]

The human condition is fatal, we WILL wipe ourselves out.

Anyhow - I think before we start thinking what the years CET will be, we need to get rid of October and November, or at least a time when everyone can agree instead of fuelling a fire of sorts.

That said its ok to assume, for now, perhaps more stictness can come in around mid December.

[summer blizzard]

How many cumulative degrees do we need to lose on 2002 (10.63) to come in lower?????

[Snowly does it]

How many cumulative degrees do we need to lose on 2002 (10.63) to come in lower?????

We are currently 0.4 lower than 2002 however it featured a coolish October - 10.1 8.5 5.7 was the run in, if October were say 11.5 as SF had postualted above we would be a degree ahead but could equal it by losing a degree off the mild November of 2002 - 7.5, 5.7 not an ureasonable ask to equal 2002s 10.60, I favour slightly below this.

Edited October 13, 2007 by snowmaiden

[eddie]

Not record breaking, but it would require a run which would be cold, and in the context of the period of cold we've already had over summer almost unprecedented in the past twenty years. Everything needs to be viewed in the context of the climate we now have, not that which occurred before Queen Victoria came to the throne I'm afraid.

In order to get below 10.5C the 6 month anomaly from August to December, with respect to the 1971-00 baseline, would need to be below 0.33C

Here are the 6 month periods in the last 20 years which satisfy this requirement:

I can't see anything unprecedented being required.

Edited October 13, 2007 by eddie

[Stratos Ferric]

In order to get below 10.5C the 6 month anomaly from August to December, with respect to the 1971-00 baseline, would need to be below 0.33C

Here are the 6 month periods in the last 20 years which satisfy this requirement:

I can't see anything unprecedented being required.

I don't think anybody is suggesting anything unprecedented as being required.

The data is interesting Eddie, but is slightly flawed. What's at issue here is NOT the last six months of the year, but the remaining two. We can do nothing about July - Oct (I'm assuming, here, October comes in around 11.5 by the way). It's the same as, say, calculating the odds on flipping eight straight heads; the odds start to look a bit different once you're six flips into the series. IN any case, I'm not sure I'd be looking for series at ANY point in the year to make the case; better to compare like with like: there are far more ways of turning six consecutive months against a particular parameter than there are of turning six specific months (in fact, there are 12 times as many permutations available).

So, comparing like with like, and using my assumption, we require no more than 11.1 cumulative degrees in N+D to come in below 10.5. This, too, is far from unprecedented, but shares one interesting bias - like your own numbers - when we look at the detail. Very few of the instances have occurred in the last ten years, but the last ten years is very much the time we inhabit.

It's occurred twice; 11.1 in 2001, and 10.6 in 2005. 2005 is an alluring analogue because it followed a coolish summer, and brief warming in early autumn. 2001 was not that dissimilar. However, 1998 and 2002 both look like even better pattern matches for the temperature cycle we've seen this year, and in both cases the final two months came in above the required mark.

We'll have a better idea in a couple of weeks' time, but right now I'm happy to place the 95% line around 10.50C. By recent standards, and given that there's an indication that we're trending up at present, a back end cool enough to dip under that mark looks less likely that a finish further up the scale.

[eddie]

Very few of the instances have occurred in the last ten years, but the last ten years is very much the time we inhabit.

Yes the last ten years is very much the time we inhabitand yes we are on an upward trend but you of all people shouldn't need reminding that the 10 year mean is not always going to go upwards, even with AGW there are still going to be peaks and troughs due the underlying variability of our climate.

For example, look at the 10 rolling mean for July. It just dropped 0.2C. The August one just dropped 0.3C! Does this mean that next year your prediction for August will be lower than this year?

I will agree that my data was rather cherry picked to suit my point but I think yours could be too.

[Stratos Ferric]

Yes the last ten years is very much the time we inhabitand yes we are on an upward trend but you of all people shouldn't need reminding that the 10 year mean is not always going to go upwards, even with AGW there are still going to be peaks and troughs due the underlying variability of our climate.

For example, look at the 10 rolling mean for July. It just dropped 0.2C. The August one just dropped 0.3C! Does this mean that next year your prediction for August will be lower than this year?

I will agree that my data was rather cherry picked to suit my point but I think yours could be too.

Not at all, and I haven't once said the ten year mean must go remorelessly upwards. My point is exactly as I stated it: the required outturn has happened twice in the last ten years, all other things being equal, and extrapolating from that, there's about a 20% chance of the same happening this year. I have stated clearly why I think the chances this year are lower. I have not picked the data to make my point Eddie: I have made my point based on what I consider to be an appropriate choice of data. There's a very important difference. You rather betray yourself I'm afraid with your chosen use of 'cherry picked'.

A few years ago DTP turned even the most colour blind people into would be publishers. The slow march of Excel, whilst it makes the processing of data easier, does not assist in the determination of sensible analysis.

Edited October 14, 2007 by Stratos Ferric

[eddie]

A few years ago DTP turned even the most colour blind people into would be publishers. The slow march of Excel, whilst it makes the processing of data easier, does not assist in the determination of sensible analysis.

Could you please clarify what you mean by this please?

[Stratos Ferric]

Could you please clarify what you mean by this please?

What I mean is that Excel allows people to manipulate numbers and draw graphs far more easily than otherwise, however it does not explain to people the principles of sound analysis of data. Just because Excel allows people to calculate things, doesn't mean that the calculation performed is in any way a sensible test of a particular hypothesis.

Not having a go at your general assessment by the way, but beware, as you put it, 'cherry picking' to make a point. If one cherry picks then one isn't really making a point at all, other than the hackneyed old one about it being possible to prove anything with statistics. This isn't strictly speaking true, any more than it's true that you can do any job with a hammer. There's lots of things that you can do with one, even if plenty of them do have tools more fit for purpose.

Edited October 17, 2007 by Stratos Ferric

[The PIT]

What I mean is that Excel allows people to manipulate numbers and draw graphs far more easily than otherwise, however it does not explain to people the principles of sound analysis of data. Just because Excel allows people to calculate things, doesn't mean that the calculation performed is in any way a sensible test of a particular hypothesis.

Not having a go at your general assessment by the way, but beware, as you put it, 'cherry picking' to make a point. If one cherry picks then one isn't really making a point at all, other than the hackneyed old one about it being possible to prove anything with statistics. This isn't strictly speaking true, any more than it's true that you can do any job with a hammer. There's lots of things that you can do with one, even if plenty of them do have tools more fit for purpose.

It also depends on Excel to give correct answers which it doesn't in certain circumstances and no I'm talking about the recently found bug in 2007.

Anyway if this month comes in around 11.5C i.e above normal and the remaining months average it's still going to be an impressively warm year.

[Stratos Ferric]

It also depends on Excel to give correct answers which it doesn't in certain circumstances and no I'm talking about the recently found bug in 2007.

Anyway if this month comes in around 11.5C i.e above normal and the remaining months average it's still going to be an impressively warm year.

I wasn't aware of any issues with the formulae in Excel. Pray tell...

This year could still go either way. I'm fairly sure we won't now be coming in under 10.5, and the real interest now is probably whether we land no.2 spot in the UK all-time list, and so herald in, if not a long run of cooling, at least a lot of misguided contributions on here re how we are now cooling because this year is cooler than last year...

[West is Best]

This year could still go either way. I'm fairly sure we won't now be coming in under 10.5,

Gosh that's astonishing bold. Rash even.

[BLAST FROM THE PAST]

This year could still go either way.

About as much chance of being highest as SA beating England on Saturday. And I am a Welsh lad .....C'mon England

BFTP

[Stratos Ferric]

Gosh that's astonishing bold. Rash even.

I don't think so. I seem to recall giving my reasoning higher up. How can saying I'm fairly sure in any way, shape or form be considered 'rash'. You are a laugh sometimes WiB, I think you've had a melodrama transfusion from the same source as Mr Sussex, late of this parish.

This year could still go either way.

About as much chance of being highest as SA beating England on Saturday. And I am a Welsh lad .....C'mon England

BFTP

Agree, highly unlikely that we'll land above 10.82.