An honest question, but what deviation from the mean who constitute a warm or cold tag?

One method is to split a sample into 5 groups of equal size. For winters the warmest fifth could be referred to as mild, second fifth 'milder than average', third section 'average' etc, or alternatively very mild/mild/average/cold/very cold.

Using this type of method 'average' means close to average, not the precise value of the mean or median. A similar classification can be used for precipitation. 

An honest question, but what deviation from the mean who constitute a warm or cold tag?

One method is to split a sample into 5 groups of equal size. For winters the warmest fifth could be referred to as mild, second fifth 'milder than average', third section 'average' etc, or alternatively very mild/mild/average/cold/very cold.

Using this type of method 'average' means close to average, not the precise value of the mean or median. A similar classification can be used for precipitation. 

John have looked at this approach.Couple of observations.With 5 groups of equal size there  are 25 different combinations of temerature and precipitaion types compared to 16 on the simple very warm,warm,cold,very cold and similar for precipitation.This means smaller numbers of years in each classifcation type.Second there is clear understanding of anomalies versus mean ie warmer/colder.I,am not sure that  the inclusion of "average" band makes things clearer-although I know that the Met Offoce use this quintile approach in their contingency planning seasonal forecast.

CFS throws up the cold January ......................again

http://modeles.meteociel.fr/modeles/cfs/run/monthly/runs/2013110606/cfs-8-1-2014.png?06

Fixed

Hi all,

Here's my latest seasonal update looking at NASA's experimental model;

http://www.gavsweathervids.com

Seems to be going for quite a cold late winter period, though the Dec-Feb anomaly comes out slightly colder than average as well.

I didn't know NASA had a seasonal forecasting model TBH.

Good vid Gav

Cheers

---

http://modeles.meteociel.fr/modeles/cfs/run3/cfsnh-0-894.png?12

JFF JFF JFF

A cold flow mid December

 

---

http://modeles.meteociel.fr/modeles/cfs/run3/cfsnh-0-894.png?12

JFF JFF JFF

A cold flow mid December

 

About as far as you can get from snow for Santa though!

http://modeles.meteociel.fr/modeles/cfs/run3/cfsnh-0-1164.png?12

---

http://modeles.meteociel.fr/modeles/cfs/run3/cfsnh-0-894.png?12

JFF JFF JFF

A cold flow mid December

 

About as far as you can get from snow for Santa though!

http://modeles.meteociel.fr/modeles/cfs/run3/cfsnh-0-1164.png?12

http://modeles.meteociel.fr/modeles/cfs/run4/cfsnh-0-1266.png?18

JFF JFF JFF

Some wintry stuff soon follows

---

Italian researchers are predicting a mild winter for the UK. Arctic Oscillation expected to be positive based on correlation with the October Pattern Index [OPI]. Long term correlation said to be to be at least +0.83. OPI value for October 2013 is +1.6
Plenty reading below for those interested:

http://www.centrometeotoscana.it/forum/index.php?topic=7371.0 
http://www.centrometeotoscana.it/forum/index.php?topic=7473.0 
http://www.americanwx.com/bb/index.php/topic/41379-october-pattern-index-predicting-winter-ao-from-october-with-90-accuracy/ 
http://www.americanwx.com/bb/index.php/topic/41639-seasonal-winter-20132014-forecast-based-on-opi-index/ 

http://www.americanwx.com/bb/index.php/topic/41639-seasonal-winter-20132014-forecast-based-on-opi-index/ 

It's fairly new research is this though they did call last winter right if my memory serves me right, again with any LRF there are just far too many variables to come up with an accurate forecast time after time.

http://www.americanwx.com/bb/index.php/topic/41639-seasonal-winter-20132014-forecast-based-on-opi-index/ 

Interesting pattern matching October to the AO.Although the time series is short 1979-2013 the correlation is high.

The next question is how well does the AO link to European winter temperatures?I dont know the answer for the AO but the Met office work  for the NAO suggests that the correlation between the NAO and winter temperature is r =0.6 meaning that the NAO alone accounts for c40% % of winter temperature.Also was involved in a paper published in Weather magazine (March 2013) which demonstrated that across much of Europe there is  a closer link between NAO and winter temperature in NAO + years than  NAO -ve years.

---

The MetO very confident prediction of a warm November issued on its website just 8 days ago is going horribly wrong with temps during the first 8 days below average especially in the north where its been particularly chilly with night frosts and hill snow.

Now the MetO have dropped the warm November forecast all together with an updated MRF indicating a cool month, all this with such high Atlantic SST's

Confidence significantly decreases regarding which weather type is most likely to dominate through this forecast period. There are currently no strong signals indicating significantly above or below average temperatures or rainfall. However, the projections suggest that precipitation amounts are more likely to be just above the seasonal average than below. This would suggest that the weather remains changeable with spells of showers or longer spells of rain, particularly in the north and west, interspersed with drier and clearer periods, giving a greater threat of overnight frost and fog. Temperatures are also more likely to be just below average than above, and this would give a greater threat of wintry precipitation, particularly over hills.

Updated: 1201 on Fri  8 Nov 2013

Model output continues to show cool zonality which again is surprising given the high SST's, where is the cold air coming from?

Its back to the drawing board with any winter forecast IMO.

Andy

The key phrase in the MetO MRF sums up any winter forecast 'Confidence significantly decreases'.

That says it all!

Andy

---

The MetO very confident prediction of a warm November issued on its website just 8 days ago is going horribly wrong with temps during the first 8 days below average especially in the north where its been particularly chilly with night frosts and hill snow.

Now the MetO have dropped the warm November forecast all together with an updated MRF indicating a cool month, all this with such high Atlantic SST's

Confidence significantly decreases regarding which weather type is most likely to dominate through this forecast period. There are currently no strong signals indicating significantly above or below average temperatures or rainfall. However, the projections suggest that precipitation amounts are more likely to be just above the seasonal average than below. This would suggest that the weather remains changeable with spells of showers or longer spells of rain, particularly in the north and west, interspersed with drier and clearer periods, giving a greater threat of overnight frost and fog. Temperatures are also more likely to be just below average than above, and this would give a greater threat of wintry precipitation, particularly over hills.

Updated: 1201 on Fri  8 Nov 2013

Model output continues to show cool zonality which again is surprising given the high SST's, where is the cold air coming from?

Its back to the drawing board with any winter forecast IMO.

Andy

The key phrase in the MetO MRF sums up any winter forecast 'Confidence significantly decreases'.

That says it all!

Andy

Lol, I think they should stick to the 7-10 day timeline as I feel there longer range stuff is just as woeful as it's always been. Still at anything up to 10 days they are the best in the business.

The MetO very confident prediction of a warm November issued on its website just 8 days ago is going horribly wrong with temps during the first 8 days below average especially in the north where its been particularly chilly with night frosts and hill snow.

Now the MetO have dropped the warm November forecast all together with an updated MRF indicating a cool month, all this with such high Atlantic SST's

Confidence significantly decreases regarding which weather type is most likely to dominate through this forecast period. There are currently no strong signals indicating significantly above or below average temperatures or rainfall. However, the projections suggest that precipitation amounts are more likely to be just above the seasonal average than below. This would suggest that the weather remains changeable with spells of showers or longer spells of rain, particularly in the north and west, interspersed with drier and clearer periods, giving a greater threat of overnight frost and fog. Temperatures are also more likely to be just below average than above, and this would give a greater threat of wintry precipitation, particularly over hills.

Updated: 1201 on Fri  8 Nov 2013

Model output continues to show cool zonality which again is surprising given the high SST's, where is the cold air coming from?

Its back to the drawing board with any winter forecast IMO.

Andy

The key phrase in the MetO MRF sums up any winter forecast 'Confidence significantly decreases'.

That says it all!

Andy

It does and for time scales beyond 2 weeks it is very rarely more than 75%. In my view all the indications are for a mild winter in NW Europe but that obviously doesn't mean it will be mild just more likely than recent winters. Seasonal forecasts are useful for those who understand them and particularly for those who have weather dependent businesss where seasonal forecasts can be of value over a decade.

As you said in the past Tom, the atmosphere is in constant chaos.

---

As you said in the past Tom, the atmosphere is in constant chaos.

Exactly but some scenarios can still be more likely than others but as you say constant chaos.

The MetO very confident prediction of a warm November issued on its website just 8 days ago is going horribly wrong with temps during the first 8 days below average especially in the north where its been particularly chilly with night frosts and hill snow.

Now the MetO have dropped the warm November forecast all together with an updated MRF indicating a cool month, all this with such high Atlantic SST's

Confidence significantly decreases regarding which weather type is most likely to dominate through this forecast period. There are currently no strong signals indicating significantly above or below average temperatures or rainfall. However, the projections suggest that precipitation amounts are more likely to be just above the seasonal average than below. This would suggest that the weather remains changeable with spells of showers or longer spells of rain, particularly in the north and west, interspersed with drier and clearer periods, giving a greater threat of overnight frost and fog. Temperatures are also more likely to be just below average than above, and this would give a greater threat of wintry precipitation, particularly over hills.

Updated: 1201 on Fri  8 Nov 2013

Model output continues to show cool zonality which again is surprising given the high SST's, where is the cold air coming from?

Its back to the drawing board with any winter forecast IMO.

Andy

The key phrase in the MetO MRF sums up any winter forecast 'Confidence significantly decreases'.

That says it all!

Andy

It does and for time scales beyond 2 weeks it is very rarely more than 75%. In my view all the indications are for a mild winter in NW Europe but that obviously doesn't mean it will be mild just more likely than recent winters. Seasonal forecasts are useful for those who understand them and particularly for those who have weather dependent businesss where seasonal forecasts can be of value over a decade.

I still say all seasonal models have an inbuilt bias built into them this is why height and temp anomalies differ so much when viewing any of the models ouput. A prime example would be one model showing  +high anomalies over Greenland and trough to our South, yet the temp anomalie shown for us ends up above average.

 

I still say all seasonal models have an inbuilt bias built into them this is why height and temp anomalies differ so much when viewing any of the models ouput. A prime example would be one model showing  +high anomalies over Greenland and trough to our South, yet the temp anomalie shown for us ends up above average.

The models are intiated using the current conditions globally and predict forward a few months to try to predict which regions will be colder than average by a degree or more and which warmer. Global mean anomalies only change by a few tenths of a degree at the most on that timescale due to ENSO changes mainly. I don't see how there can be a bias unless the model misses a big ENSO event.

What are all these mild or very mild winter indicators then? Now The mild November outlook by Met office has been canned will more forecasters jump ship and sit on the fence?? It's a crucial turning point for those producing a winter forecast. Regarding roger's posts yesterday - For pattern matchers a cool November could give us an average overall autumn reducing the chances of a mild dec/jan/ feb period. Especially when taking into account the cold spring-cold subsequent winter theory.

http://www.weatherweb.net/cfs/cfses4avmaxukweek8.gif

http://www.weatherweb.net/cfs/cfses4avmaxukweek9.gif

http://www.weatherweb.net/cfs/cfses4avmaxukweek10.gif

http://www.weatherweb.net/cfs/cfses4avmaxukweek11.gif

http://www.weatherweb.net/cfs/cfses4avmaxukweek12.gif

JFF JFF JFF JFF JFF

A cold January from CFS

---