Given the reduction in solar heating through the NH due to increased cloudiness this summer, it would not be a surprise to see a cold NH winter. It's our turn isn't it after the freakish cold the SH has been experiencing this year and last.

On the subject of cold winters Iset out belows figures showing the probability of a cold winter following the weather types of previous seasons.

This is based on the work of Lionel Smiths seasonal weather.His classification was baesd on Kew and I hav extended using first Kew and then adjusted Heahrow figures.Seasons are classified as verywet,wet,dry,very dry and very warm,warm,cold ,very cold.These are based on seasonal variation to average.There are 16 different combinations of warmth and wetness,spanning 158 years.

The following conclusions emerge

1Overall the weather type of Autumn  has a stronger link to the following winter than Spring  or Summer.

2.The strongest correlation in Autumn is that between  dry,warm Autumns and warm winters (80% probabilty of warm v only 20% cold).Wet warm Autumns give the highest probablity of a cold winter(c 60%)

3.Within Summer c60%of very wet summers were followed by a cold winter( but most recently not 2007/8!)

4.The best indicator of a cold winter within Spring is a very cold spring(80% followed by a cold winter)

 

Looking at 2008 so far spring was wet and warm suggesting  a warm winter,summer is very wet ,so a slight bias to a cold winter.However it is to Autumn we must really look.Last year was a dry warm autumn and I posted then that this indicated a mild winter to follow.If we get another dry  warm  autumn that is bad news for winter cold.A very wet Autumn however would be better news! 

The old 23 year cycle would indicate that this winter will be cold but we'll see.

The problem with that is, you're comparing extremes and lumping months together. What's the link between a completely average winter, or a completely average autumn? And what conclusions can you draw from a wet September, dry October and wet November, for example? Not to mention that months are man made things. A cold spell from January 20th - February 10th could just result in two average months overall!

westv:

The old 23 year cycle would indicate that this winter will be cold but we'll see.

FIrst I've heard of it, what prey tell is the old 23 year cycle? what happened in 1986? (I was only a year old and therefore cant remember)

Alvin Meister:

The problem with that is, you're comparing extremes and lumping months together. What's the link between a completely average winter, or a completely average autumn? And what conclusions can you draw from a wet September, dry October and wet November, for example? Not to mention that months are man made things. A cold spell from January 20th - February 10th could just result in two average months overall!

Alexis, accept that months are lumped together but the seasons they give are the well defined ones.The classificaton is not neccessarily comapring extremes.The Lionel Smith definition of a cold winter is the temperature of the coldest month being 3.2 or less,warm is above that.The very cold winter classification is <2.0,very warm is >4.7.

I have selected innthe commentary the groupings that produce significant variations from average and  arelargein size.for example the dry warm autumn includes dry/warm/, dry/very warm,verydry/warm and very dry/verywarm autumns and  accounts for 50 out of the 158 winters classified.You are right of course that many seasonal classifications produce no useful signals-for example wet but not very wet Autumns produce an average frequency of cold winters.

I dont know what the relationship between individual months and cold winters - there was a claim last autumn by some that a warm dry October favoured  a cold winter but that didnt work for last October! I am sure someone has done the work.My suspicion would be that one month is a less reliable indicator than a whole season.And finally any classifcation is man made - I woudn't like to hav ethe job of  reclassifying data into finding two week cold spells! 

I would not suggest that theLionel smith seasonal classifcation is anything but a rough guide and there are many  seasons which will give no clear signal at all.I have,mentioned that dry warm Autumns deliver around an 80% chance of a warm winters and repsent rougjhly 30% of Autumns.A higher probability of 83% for cold winters  is achieved by very cold springs but they are only a  1 in 9   occurrence.

Finally underlying warming means that the last 20years have seen a lower chance of cold winters,There wahve only been  3 since 1987-1991,1995,1996 and none have been in the very cold catgory.forecasting  very cold winters is a needle in a haystack job!

roger63

 

 

 

Edicius81:

westv:

The old 23 year cycle would indicate that this winter will be cold but we'll see.

FIrst I've heard of it, what prey tell is the old 23 year cycle? what happened in 1986? (I was only a year old and therefore cant remember)

1916/17

1939/40

1962/63

1985/86 

2008/09 ???

Does that answer your question?

Apparently it's called the Hale Cycle.

Edicius81:

westv:

The old 23 year cycle would indicate that this winter will be cold but we'll see.

FIrst I've heard of it, what prey tell is the old 23 year cycle? what happened in 1986? (I was only a year old and therefore cant remember)

 

 

Hi Edicus.

 

I see Westv has answered your question  about the 23 year cycle he was referring to but he didn't tell you what happened in 1986.

February 1986 was the second coldest february of the 20th century after 1947

Alvin Meister:

What conclusions can you draw from a wet September, dry October and wet November, for example? 

Statistically, the conclusion to draw from a wet September followed by a dry October is that the following winter is more likely to be colder than average than mild.

I think that the conclusion is that there is no real link and the weather will do what it wants despite what has gone before. There are enough stats to support any argument really. I don't see this winter being cold, just average, much like last winter 

 Last winter wasn't average....

We went through this last year! Does autumn have an effect on the following winter?  This link is on UKWW posted last October http://www.ukweatherworld.co.uk/forum/forums/thread-view.asp?tid=17542&start=41

Jamie R:

We went through this last year! Does autumn have an effect on the following winter?  This link is on UKWW posted last October http://www.ukweatherworld.co.uk/forum/forums/thread-view.asp?tid=17542&start=41

 

Statistics such as those would be great if taken in isolation but what about if you thrown in warming seas, less sea ice, higher global temps and, ergo. you could make the case the other way

We cant predict tomorrow with enough accuracy let alone 6 months from now

Jamie R:

We went through this last year! Does autumn have an effect on the following winter?  This link is on UKWW posted last October http://www.ukweatherworld.co.uk/forum/forums/thread-view.asp?tid=17542&start=41

Last years thread was specifically about October in isolation-not Autumn as a whole It also I think important to say that in the Lionel Smith classifcation a cold winter means one below average.When people  refer to a cold winter they very often mean a very cold one (Smith defintion coldest winter month< 2.0C.

Looking at very cold winters in the Smith data there are 29 out of 158 years,but none have fallen in the last 20 years.For what it is worth 66% were preceded by a wet autumn-not the same as  60% of wet Autumns generate a very cold winter! As I have mentioned earlier the strongest correlation is actually that between dryer aand warmer than average Autumns   and warmer than  average  winters, 49 ocurrences , 80% of which generated an above temperature for the following winter.The size of sample and the high probability suggest that this one is statistically significant.Of  course you can have a wet October in  dry Autumn and vice versa( October 62 was dry and warmer than average within a wet warm  Autumn).Last years thread did not conclude that the  alleged warm dry October to a cold winter was staitstically significant.    

Moving from pattern matching to cycles.Some go for a   34 year cycle(3 sunspot cycles) of very cold winters  viz 1860/61,1894/95,1928/9,1962/3.We eagerly awaited  1996/7 and it failed to deliver a very cold winter but it did deliver the last cold one.It was aslo a very promising start with cold easterlies setting in before Christmas intensifrying cold and some snow after Christmas but early in the New year the warm Atalantic truimphedThe next yaer in this busted 34 year cycle would be 2020/21 -a long wait!

 

roger63

 

 

roger63:

Moving from pattern matching to cycles.Some go for a   34 year cycle(3 sunspot cycles) of very cold winters  viz 1860/61,1894/95,1928/9,1962/3.We eagerly awaited  1996/7 and it failed to deliver a very cold winter but it did deliver the last cold one.It was aslo a very promising start with cold easterlies setting in before Christmas intensifrying cold and some snow after Christmas but early in the New year the warm Atalantic truimphedThe next yaer in this busted 34 year cycle would be 2020/21 -a long wait!

I think you're close but need to modify your thinking a little. I notice your 34 years (3 cycles) is based on an average of 11.3 years per cycle. As you may be aware the Cycles vary generally (on the 100 year timescale) from 7 to 15 Years as a rule of thumb. This offsets your 34 years actually depending on the configuration of each of the 3 Cycles. For instance the last two cycle of 9.8 and 10.2 and 12.5 (and counting as SC23 isn't deemed over) is actually 32.5 years. But we don't know yet the exact length of SC23. So I would say the '34' year cycle is not 'busted', as it could be 34 +- 4 Years, the range is probably 30-38 years. Although TBH I think this 34 years cycle is an Ocean one probably driven by the Solar Cycle. A more direct Solar Cycle matching would be the 'Hale' Cycle which is 22 Years +- 4 Years (on average) again depending on the configuration of each Solar Cycle within it.