Climate Change Home

Predictions of winter 2004/2005 in the UK & Europe

It is a month now since Metcheck caused a real hullabaloo with their seasonal forecast of an average to slightly colder than average winter in the UK.   Other weather forecasting companies including the UK Met Office responded by arguing that forecasting beyond a month ahead is 'very difficult' (BBC News) implying seasonal forecasts for the UK and Europe are little better than guesswork.  It is certainly true that seasonal forecasting for Europe is far more difficult than other regions of the world, especially the Tropics, but this does not mean winter 2004/2005 is a completely blank canvas and that current meteorological science can't provide some information on the likely weather of the coming winter  Some UK weather forecasting companies routinely issue seasonal forecasts and some amateur meteorologists also believe such forecasts are possible and regularly produce seasonal forecasts.  And without a doubt there is tremendous interest from both businesses and individuals for winter forecasts as winter weather such as severe gales, heavy rain, snowfall, fog and frost affect us all.   This article includes the author's first attempt at a winter forecast for the UK and the analyses suggest that at the very least, it is possible to identify recurrent weather patterns that are likely to be a feature and from these, it is possible to suggest some detail

The North Atlantic Oscillation

It can be argued that the starting point of any winter forecast for the UK and Europe should be the North Atlantic Oscillation (NAO) which is one of the most prominent and recurrent patterns of atmospheric circulation variability and it dictates climate variability from the eastern seaboard of the United States to Siberia, especially in winter (Hurrell, Kushnir, Ottersen, & Visbeck, 2003). The NAO is a see-saw of atmospheric pressure between the Arctic regions and mid-latitudes of the northern hemisphere: when it is positive, pressure is relatively low in the Arctic and relatively high in the mid-latitudes of the Atlantic/Europe which means strong westerly winds across the UK & Europe whereas when it is negative, pressure is relatively high in the Arctic and low in mid-latitudes of the Atlantic/Europe and cold airmasses are able to move south from the arctic to affect the UK and Europe.   There is some debate whether or not the NAO is a phenomenon confined to the Atlantic basin or whether it applies to northern hemisphere as a whole but there is considerable evidence to support the hemispheric viewpoint and that the NAO is really the Northern Hemisphere Annular mode (Thompson, Lee,  & Baldwin,  2003).  This evidence to support the hemispheric viewpoint includes a number of different theories proposed as explanations for variations in the NAO.  In particular, the observation that NAO variations are correlated with changes in the polar vortex in the stratosphere and that these changes in the stratosphere precede changes in the NAO by one to two weeks: a stronger polar vortex leads to stronger westerlies in the troposhere and a positive NAO whereas a weaker polar vortex leads to weaker westerlies in the trophosphere and a negative NAO.  

The importance of the stratosphere to the NAO is highlighted by the effect of  volcanic dust and increased solar activity which both favour positive NAO in winter (Gillett, Graf & Osborn, 2003).   Figure 1 shows changes in the December to March NAO since 1825 and it is obvious that since the 1960s when the winter NAO was remarkably low, there has been a dramatic increase. Most scientists involved in NAO research believe that increasing greenhouse gases are at least partly responsible for this upward trend because the effect of greenhouse gases is to warm the troposphere and cool the stratosphere which increases the meridional temperature gradient in the stratosphere and strengthens the polar vortex (Gillett et al, 2003). However, Figure 1 shows the possible influence of both volcanic aerosols and sunspot maxima on the winter NAO in recent decades.   Both El Chichon in 1982 and Mount Pinatubo in 1991 were volcanic eruptions of sufficient strength to eject volcanic dust into the stratosphere and because these volcanoes are both located in the tropics, the ash was able to spread right around the world resulting in a global climate effect (see Volcanoes & their effect on climate). The proposed mechanism for volcanic aerosols causing positive winter NAO is 'the absorption of near-infrared solar and longwave terrestrial radiation by the volcanic aerosol leads to strong heating in the stratosphere at low latitudes. Since over the winter pole, solar radiation is absent and longwave irradiance from the surface is smaller than at lower latitudes, the meridional temperature gradient in the stratosphere is strengthened which makes the polar vortex stronger' (Gillett et al, 2003). This effect only lasts for a couple of winters after the eruption as most volcanic dust ejected into the stratosphere has fallen through to the troposphere and subsequently to the surface within two to three years. Therefore the positive NAO winters of 1983 and 1984 may be linked to El Chichon and those of 1992-93 linked to Mount Pinatubo (Figure 1). Also, both these periods coincided with peaks in the 11-year sunspot cycle which may also have contributed to the high NAO values in these winters. The association of peaks in winter NAO with peaks in sunpots is very obvious since 1950 but less convincing prior to this. The proposed mechanism for increased sunspot counts affecting the NAO again involves the stratosphere. Peaks in the 11 year sunspot cycle are associated with a small increase of around 0.1% in the total solar irradiance but these increases are greater in the ultra-violet band which are mostly absorbed by ozone in the stratosphere (Gillett et al, 2003). Thus the increased ultra-violet radiation at sunspot peaks strengthens the polar vortex and results in positive NAO winters.

                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                EC     MP

Figure 1  Winter NAO (December to March) from 1825 to 2004 and 5 year running mean of annual sunspots counts and two recent major volcanic eruptions which were of sufficient magnitude to eject volcanic dust into the stratosphere: EC, El Chichon 1982 & MP, Mount Pinatubo in 1991


So although most scientists involved in NAO research believe that increased greenhouse gases have played a role in the increasing trend in winter NAO since the 1960s, two natural forcings, volcanic ash in the stratosphere and sunspot peaks, have certainly contributed to some of these winters with positive NAO, particularly in the early to mid 1990s when the eruption of Mount Pinatubo coincided with a peak in the 11 year sunspot cycle.  The early to mid 1990s peak in the winter NAO looks unusual in the context of the record since 1825 but the natural forcings provide a satisfactory explanation without blaming this peak on increasing greenhouse gases. Another feature of Figure 1 which suggests that increasing greenhouse gases are not the overriding factor dictating winter NAO is the obvious trend to more negative winter NAO since the early to mid 1990s peak.

Scientists working in the NAO field do not agree on what causes the large fluctuations in the winter NAO with some arguing that it is driven entirely by processes within the atmosphere (Thompson, Lee & Baldwin, M, 2003) while others believe that North Atlantic Sea Surface Temperature (SSTs) anomalies play a major role in NAO variations (Czaja, Robertson & Huck, 2003). Climate models ambiguously support both these schools of thought and as a result there is considerable uncertainty what governs natural variability and how increasing greenhouse gases may or may not affect it.   While changes in atmospheric circulation are generally considered to determine SST anomalies in the extra-tropics, there is considerable support for the theory that changes in the Thermo-Haline Circulation (THC) result in characteristic SST anomalies in the North Atlantic and these have a significant feedback on the atmosphere, thereby influencing the NAO.   These characteristic Atlantic SST anomalies are the so-called Atlantic tripole which has two opposing modes: warm anomalies in the northern and tropical regions and cooler than average in between and cold anomalies in the tropical and northern regions and warm anomalies in between.  The former mode is considered to favour negative NAO and the latter positive NAO.

Figure 2 shows current global SST anomalies and compares them with the chart 13 months ago (the chart for the same week was not saved last year).   In both cases, the Atlantic SST anomalies show the Atlantic tripole which favours negative NAO patterns ie warm anomalies in the tropical and northern North Atlantic and colder anomalies in between.   However there are some differences.

Overall, it can be said that currently there definitely is not a strong signal from the Atlantic for a positive NAO winter but that there is a clear Atlantic tripole signal for negative NAO patterns.   Once very cold air starts flooding off Canada and the northern US states in the coming weeks, it seems very likely that the Atlantic off the eastern seaboard of North America will cool further with marked negative anomalies developing as the winter progresses which would be a very strong signal for negative NAO.

Multi-decadal phases are a feature of the winter NAO and these are evident in Figure 1 which shows positive NAO winters being predominant in the first few decades of the twentieth century and the last and negative NAO winters being more common from the 1940s to 1970s.   Both changes in Atlantic SST anomalies (Visbeck et al. 2003) & solar activity (Gillett et al, 2003) are thought to play major roles in determining these multi-decadal phases.   Judging from Figure 1, we may be at the beginning of a negative phase implying colder UK and European winters in the next few decades.

Some organisations have already issued NAO forecasts for the coming winter.   The UK Met Office issues a forecast based on North Atlantic SSTs in May and their forecast is for a positive NAO winter.   However, this technique has performed very badly in recent years in which it mostly predicted the wrong sign of the NAO and therefore it is best ignored.  University College, London issues a NAO forecast based on June/July northern hemisphere subpolar surface air temperature and June/July northern hemisphere snow cover.   Their forecast is 'the North Atlantic Oscillation is four times more likely to be below-average than above-average this coming winter'.

With respect to the NAO in the coming winter, it is also relevant to look at what the NAO has been doing in recent weeks.   This website shows that the NAO has been negative most of the time since 1 October 2004 and is forecast to turn negative again during this week (15 November 2004) which indicates that whatever governs NAO variations has favoured negative NAO in recent weeks.  


The Useful Info Winter Forecast for the UK

This is the author's first attempt at issuing a proper seasonal forecast so it remains to be seen whether it has any practical predictive potential.   It is based on northern hemisphere teleconnections and attempts to determine whether or not there are likely to be recurrent weather patterns in the UK region based on what these teleconnections are telling us about the overall northern hemisphere circulation.

First, it is helpful to recognise four different weather patterns in the North Atlantic and this article has already discussed two of them.   In addition to positive and negative NAO patterns, Hurrell et al. (2003) identified two other patterns by classifying the sea-level pressure charts for the North Atlantic region for winter months 1900 to 2001 (102 years x 3 months = 306 cases).  Positive NAO patterns accounted for 17% of cases, negative NAO 29%, anticyclone centred around the UK 32% and low pressure centred around the UK 22%.

 Massacand & Davies (2001) examined the pressure patterns in the upper trophosphere (potential vorticity) across Europe in Januaries 1980 to 1999 and identified three pairs of alternative pressure patterns: the NAO; PNA; & MAT.   The PNA is a well known teleconnection which applies to North America with the positive mode resulting from a strong ridge/anticyclone across western North America and a trough/low pressure in eastern North America and the negative mode being the opposite pattern. Massacand & Davies (2001) found that the PNA pattern of North America tended to be mirrored over Europe with a positive mode featuring a ridge off western Europe and trough in eastern Europe and the negative mode the opposite.  The MAT pair of patterns was essentially the anticyclone or low pressure patterns (near to the UK) of  Hurrell et al. (2003).

Relating these six patterns to the track of the jetstream is helpful as the jetsream tells us the direction of travel of Atlantic depressions and where cold and warm airmasses are located in relation to the UK: cold airmasses are found to the north of the jetstream and warm airmasses to the south.  

This forecast attempts to predict which of these six patterns are going to be predominant this winter by consideration of various teleconnections around the northern hemisphere and Atlantic SSTs.

EL NINO-SOUTHERN OSCILLATION (ENSO) - this teleconnection is by far the most important and determines the pattern across the Pacific and North America.   The latest ENSO forecast is for warm episode conditions to extend into early 2005 although a strong El Nino is not being forecast.   The winter forecasts for the United States reflect this eg NOAA by showing the positive PNA pattern: strong ridge across western US with warm anomalies and a trough over south-eastern US with cold anomalies.  

POSITIVE PNA - There is strong agreement amongst the various weather forecasters that this positive PNA pattern will be predominant across North America which if correct means 2004/2005 will be the third successive winter with this pattern meaning a cold winter in eastern parts of North America.   In addition to ENSO, Joe Bastardi, a forecaster at Accuweather.com, argues there are perhaps two other factors which support this positive PNA pattern.   The first is the overall warm Atlantic (Figure 2b) which in the absence of a strong ENSO signal means an above average upward motion of air across the Atlantic which tends to slow the westerlies and allow blocking to develop in the Atlantic and drains cold air south-eastwards across North America.   Atlantic blocking implies a meandering jetstream which permits cold arctic air to move south.   The second factor is the active hurricane season in the Carribean during 2004 which meant low pressure and upward motion of air across the the southern United States and high pressure across high latitudes of North America. The assumption is that this pattern will persist during the coming winter and again this means cold air drains south-eastwards across North America.   So there is strong confidence for the positive PNA pattern to dominate in North America.

This confidence for a positive PNA pattern across North America argues for the same pattern to predominate across Europe with an anticyclone off western Europe and a trough in eastern Europe.   And this pattern certainly was a feature of the last two winters

The following analyses use PSC indices which are derived from the Lamb Daily weather types.   A summary of the Lamb Daily Weather types & PSC indices can be found here:

P is the index of progressiveness and measures the difference in frequency of days of progressive and blocked synoptic types over the UK. It is positive when there is a bias towards progressive types. Progressive types are westerly, northwesterly and southwesterly types while blocked types are the other directional types.

S is the index of southerliness and measures the difference in frequency of southerly and of northerly type days over the UK. It is positive when the bias is southerly

C is the index of cyclonicity and measures the difference in frequency of cyclonic and anticyclonic type days over the UK.  It is positive when cyclonic days predominate

The PSC quintiles for the last two winters are discussed below.   Quintile 3 indicates average, quintiles 1 & 2 below average and quintiles 3 & 4 above average.

PSC quintiles for winter 2002/2003: P1 S3 C2;   Dec: P2 S1 C3   Jan: P1 S3 C2   Feb: P2 S4 C1;

PSC quintiles for winter 2003/2004: P3 S1 C2;   Dec: P2 S2 C1   Jan: P5 S1 C5   Feb: P2 S1 C1;

Of the six winter months, five had below average P indices i.e the pattern was blocked with meandering jetstreams across the Atlantic, four had below average S indices indicating above average incidence of northerlies and three had below average C indices indicating anticyclonic conditions.

These PSC quintiles of the last two winters are consistent with a postive PNA pattern across Europe with the anticyclone off western Europe giving the UK blocked anticyclonic conditions with an above average incidence of northerlies.

The Tables below examine the relationship between the PNA index in North America and the weather patterns around the UK using PSC quintiles.     Table 1 shows the PSC quintiles of winter months from 1950/1951 to 2003/2004 in which the PNA pattern across North America was strongly negative: 11 winters in total referred to as Group A.   Table 2 shows the comparative data for winters with moderately positive PNA across North America: 8 winters referred to as Group B.   And Table 3 shows the comparative data for winters with strongly positive PNA pattern across North America: 8 winters referred to as Group C.   Each Table shows the PSC quintiles for each winter month in the group plus a summary showing the number of years with PSC quintiles in the following three categories: below average (quintile 1 & 2); average (quintile 3); and above average (quintiles 4 & 5).   This enables any trends in weather patterns to show themselves.   For example, Decembers in strongly negative PNA winters (Table 1) showed a bias to above average incidence of southerlies and anticyclonicity across the UK. The CET anomalies for each month are also displayed in each Table.


Table 1 PSC quintiles and CET anomalies (from 1961-90 averages) for the eleven years with strongly negative PNA values (below -3) -referred to as Group A

                         
                                     PSC quintiles -  T is CET Anomaly  
  __________________________________________________________________  
  December   January   February  
                         
  P S C T P S C T P S C T
_________________________________________________________________________________________
1952 5 3 3 +0.8 4 1 4 -1.1 5 1 2 -0.4
1956 2 4 3 +0.7 3 2 4 -0. 2 1 2 1 -4.0
1957 4 5 2 +1.0 4 3 2 +1.7 3 4 5 +1.5
1966 4 2 5 0.0 1 5 3 -0.9 2 5 4 +1.9
1969 1 5 2 -1.7 1 5 5 +1.7 1 2 4 - 2 .8
1971 1 2 2 -0.4 1 5 4 +0.7 2 4 2 +0.7
1972 4 3 2 +1.9 1 5 3 +0.1 2 5 3 +0.5
1979 1 5 4 -0.8 2 2 3 -4. 2 2 3 3 - 2 .6
1982 1 3 5 +4.4 1 4 3 -1. 2 4 5 3 +1.0
1989 5 2 2 + 2 .8 5 3 3 + 2 .3 5 3 4 + 2 .1
1990 1 4 1 +0. 2 5 5 4 + 2 .7 5 4 5 +3.5
Totals for each quintile group                  
1- 2 (low) 6 3 6   6 3 1   6 3 3  
3 (average) 0 3 2   1 2 5   1 2 3  
4-5 (high) 5 5 3   4 6 5   4 6 5  
No of months with below average CET 3       5       4
              10/33 (30%) months with below average CET
_________________________________________________________________________________________


Table 2 PSC quintiles and CET anomalies (from 1961-90 averages) for the eight years with moderately positive PNA values (between 1.00 & 2.99) -referred to as Group B

                         
                                     PSC quintiles -  T is CET Anomaly  
  __________________________________________________________________  
  December   January   February  
                         
  P S C T P S C T P S C T
_________________________________________________________________________________________
1958 5 2 4 -0.2 2 2 5 -0.4 4 1 3 +0.9
1963 3 1 2 -2.9 1 1 1 -5.9 1 3 3 -4.5
1964 1 4 1 -2.1 1 4 1 -0.4 1 5 2 +0.7
1987 5 2 3 +1.5 1 4 1 -3.0 2 4 2 -0.2
1992 2 3 1 0.0 1 3 1 -0.1 5 3 2 +1.6
1995 5 4 3 +1.7 4 2 4 +1.0 5 2 4 +2.7
2001 1 4 5 +1.1 1 4 4 -0.6 2 2 3 +0.6
2003 1 5 4 +0.9 2 1 3 +0.7 2 4 2 +0.1
Totals for each quintile group                  
1- 2 (low) 4 3 3   7 4 4   5 3 4  
3 (average) 1 1 2   0 1 1   0 2 3  
4-5 (high) 3 4 3   1 3 3   3 3 1  
No of months with below average CET 3       6       2
              11/24 (46%) months with below average CET
_________________________________________________________________________________________


Table 3 PSC quintiles and CET anomalies (from 1961-90 averages) for the eight years with strongly positive PNA values (above 2.99 ) -referred to as Group C

                         
                                     PSC quintiles -  T is CET Anomaly  
  __________________________________________________________________  
  December   January   February  
                         
  P S C T P S C T P S C T
_________________________________________________________________________________________
1961 4 2 4 -0.8 1 5 4 +0.1 5 4 3 +3.1
1970 2 3 4 -1.4 1 5 3 -0.1 4 1 5 -0.9
1977 1 2 3 -2.7 1 1 5 -1.0 3 4 5 +1.4
1978 2 3 3 +1.4 3 2 5 -0.4 1 3 4 -1.0
1981 5 1 3 +0.9 4 1 1 +1.1 2 4 3 -0.8
1983 4 2 3 -0.3 5 2 3 +2.9 1 2 1 -2.1
1986 3 3 4 +1.6 3 1 5 -0.3 1 3 2 -4.9
1998 3 2 3 +1.1 3 4 3 +1.4 5 3 2 +3.5
Totals for each quintile group                  
1- 2 (low) 3 5 0   3 5 1   4 2 3  
3 (average) 2 3 5   3 0 3   1 3 2  
4-5 (high) 3 0 3   2 3 4   3 3 3  
No of months with below average CET 4       4       5
              13/24 (54%) months with below average CET
_________________________________________________________________________________________


Tables 1 to 3 contain a lot of information which is summarised in Table 4 which shows the total winter counts of the number of months in the three categories for each group and for groups B & C combined.   Group A shows a tendency to blocked winter months with above average incidence of southerly winds.   Group B shows a tendency for blocked and anticyclonic winter months.  And Group C shows a tendency for winter months with above average incidence of northerlies and cyclonic weather patterns.   Pooling the data from Groups B & C gives a definite bias for both blocking (low P index) and northerlies (low S index) but there was no apparent trend in the C index.  The tendency for low P & S indices in winters with a positive PNA pattern across North America is consistent with the PNA pattern being mirrored across Europe but the lack of an anticyclonic bias (low C index) is surprising.   The most likely explanation which is consistent with the observed tendency for low P & S indices is that the whole positive PNA pattern across Europe may be further west in some winters which would mean the anticyclone off western Europe is in the central Atlantic and the low pressure in eastern Europe is centred in western/central Europe.   Thus winters in the UK with a positive PNA pattern across North America may be either 'anticyclonic/east PNA' or 'cyclonic/west PNA' depending on how close the Atlantic anticyclone is to the UK.   Although Chi-squared tests gave non-significant results when the ratios of PSC indices in Group A & Groups B + C were compared, these tendencies to certain weather patterns make synoptic sense and are real but the non-significant results indicate that there is considerable variability in the patterns within each group.


Table 4 Number of PSC quintiles for each Group in three categories: low (quintile 1 & 2); average (quintile 3); & high (quintiles 4 & 5).

                                      Number of months    
  __________________________________________________________________________
               
  Group A   Group B   Group C   Groups B + C
P S C P S C P S C P S C
_______________________________________________________________________________________
1- 2 (Low) 18 9 10   16 10 11   10 12 4   26 22 15
3 (Average) 2 7 10   1 4 6   6 6 10   7 10 16
4-5 (High) 13 17 13   7 10 7   8 6 10   15 16 17
________________________________________________________________________________________

Tables 1 to 3 also list the number of months with below average Central England Temperature (CET) and it is apparent that there is a greater incidence of below average CET months in Groups B & C than in Group A which is consistent with the higher incidence of northerlies in Groups B & C.

Given the recent tendency to negative NAO, the Atlantic SST anomalies which favour negative NAO and the University College London forecast for a negative NAO winter, it is appropriate to create a subset of winters with a positive PNA pattern across North America which also had a negative NAO pattern in the Atlantic.   Table 5 shows this data in the same format as Tables 1 to 3: winters in Groups B & C which had negative winter NAO (Dec-Feb).   And Table 6 shows the pooled data from Table 5 for the whole winter.   The following is a summary of the results in Table 5:


Table 5 PSC quintiles and CET anomalies (from 1961-90 averages) for the eleven years in Groups B & C with negative winter NAO

                         
                                     PSC quintiles -  T is CET Anomaly  
  __________________________________________________________________  
  December   January   February  
                         
  P S C T P S C T P S C T
_________________________________________________________________________________________
1958 5 2 4 -0.2 2 2 5 -0.4 4 1 3 +0.9
1963 3 1 2 -2.9 1 1 1 -5.9 1 3 3 -4.5
1964 1 4 1 -2.1 1 4 1 -0.4 1 5 2 +0.7
1970 2 3 4 -1.4 1 5 3 -0.1 4 1 5 -0.9
1977 1 2 3 -2.7 1 1 5 -1.0 3 4 5 +1.4
1978 2 3 3 +1.4 3 2 5 -0.4 1 3 4 -1.0
1987 5 2 3 +1.5 1 4 1 -3.0 2 4 2 -0.2
1986 3 3 4 +1.6 3 1 5 -0.3 1 3 2 -4.9
1998 3 2 3 +1.1 3 4 3 +1.4 5 3 2 +3.5
2001 1 4 5 +1.1 1 4 4 -0.6 2 2 3 +0.6
2003 1 5 4 +0.9 2 1 3 +0.7 2 4 2 +0.1
Totals for each quintile group                  
1- 2 (low) 6 5 2   8 6 3   7 3 5  
3 (average) 3 3 3   3 0 3   1 4 3  
4-5 (high) 2 3 5   0 5 5   3 4 3  
No of months with below average CET 5       9       5
              19/33 (58%) months with below average CET
_________________________________________________________________________________________

 

 

Table 6 Data from Table 5 pooled to give winter totals of PSC quintiles in the three categories

     Number of months with PSC quintiles
  ____________________________________
       
  P S C
________________________________________________________
1-2 (low) 21 14 10
3 (average) 7 7 9
4-5 (high) 5 12 13
________________________________________________________


The fact that Table 6 does not show much of a tendency for above average incidence of northerlies may seem odd as Table 4 clearly showed this bias for northerlies but the reason is that negative NAO patterns imply an above average incidence of easterlies.   Indeed, this sample of months in Table 5 includes quite a few with notable cold easterly spells: December 1962-February 1963 (the coldest winter since 1740); December 1963; December 1976; January 1977; February 1978; February 1986; & January 1987

The question of whether or not easterly spells of weather will occur this winter is an important one. Easterlies bring the UK and Europe's coldest winter weather and generally give more prolonged cold (day and nights with temperatures continuously below freezing) than northerlies. However, easterly spells have been very rare in winters post 1988 with the last easterly winter being 1995/96.

Figure 3 displays CET anomalies for all winter months December 1659 to February 2004 and shows this post-1988 lack of cold months/easterlies and a much longer period when cold easterlies were rare: the first three decades of the twentieth century.   In contrast, the second half of the twentieth century had a much higher incidence of cold easterly type months and this high frequency was at its greatest in the 1978 to 1987 period.   This recent cold period includes a number of notably cold months which rival the coldest winter months in the last 340 years: January 1979; February 1979; January 1980; December 1981; January 1982; January 1985; February 1986. In fact, the coldest day of the 20th century was recorded in many parts of  England & Wales on 12 January 1987 during an extraordinary easterly spell.

 

Figure 3 Mean Central England Temperature (CET) anomaly (from 1961-90 averages) in each winter month:
December 1659 to February 2004


So will cold easterlies return to the UK during winter 2004/2005?   The factors which permit this pattern across Europe have clearly been absent recently. Table 6 strongly suggests that the incidence of cold easterly months is linked to negative NAO patterns and a combination of a strong positive PNA pattern across North America and a strong negative NAO pattern across the Atlantic gives a significant likelihood of a cold easterly month developing across Europe.   At least 29% of months in Table 6 (charts of all these months have not been studied but the following months certainly do include significant easterly spells: December 1962-February 1963 (the coldest winter since 1740); December 1963; December 1976; January 1977; February 1978; February 1986; & January 1987) had significant cold easterly spells which is a remarkably high percentage.

The final clue to the likely weather patterns in the coming winter is recent trends in PSC indices.  Figure 4 shows the PSC quintiles in each month from January 2002 to September 2004.  The most striking feature of Figure 4 is the fall in the S index over the three years showing a decrease in the incidence of southerlies and increase in the incidence of northerlies: eight months in both 2002 & 2003 had above average incidence of southerlies (quintiles 4 or 5) compared with just two months in 2004 so far.  The other two indices do not show such clear trends: blocking (low P indices) has been predominant throughout while the C index shows that 2003 & 2004 have been much more anticyclonic than 2002 but there is a trend to more cyclonic months just recently with June, August and October (not shown) having above average cyclonicity.   These trends to increasing northerliness and cyclonicity may indicate that the PNA pattern across Europe may be more of a cyclonic/west PNA pattern than the anticyclonic/east PNA pattern of the last two winters which implies a colder and wetter winter.

 

Figure 4 Quintiles of PSC indices for each month January 2002 to September 2004.

 

So now follows the first Useful Info winter forecast for the UK which is based on the various factors and trends discussed above:


The General Weather Pattern:

  • the confidently predicted positive PNA pattern across North America means a ridge off western Europe and trough in eastern Europe is likely to be a recurrent and dominant synoptic pattern in winter 2004/2005 (High confidence)

  • the increasing trend to northerliness over the last year and slight increase in cyclonicity (Figure 4) suggests that this positive PNA pattern across Europe may be shifted further west than in the last two winters meaning the UK spends longer periods in cold airstreams and more cyclonic conditions will occur but the main thrust of the cold northerlies and more unsettled conditions will probably again occur in central/eastern Europe - the bias for cyclonicity shown in Table 5 also suggests a more cyclonic pattern if strongly negative NAO develops (Moderate confidence)

  • the NAO since 1 October 2004 has been mostly negative, weather forecasting models are predicting negative NAO patterns to develop during the week commencing 15 November 2004, University College London NAO forecast is for a negative NAO winter, current SSTs off the eastern seaboard are currently near average and the expected positive PNA pattern implies cold air persistently draining south east across North America in the coming winter which is likely to cause large negative SST anomalies to develop off the eastern seaboard of the US (the main area for Atlantic depression development) which would be a very strong signal for increasingly negative NAO during the winter - a negative NAO implies cold arctic air moving south in the Atlantic/Europe region favouring some long cold spells with snowfall associated with northerly/north-easterly/easterly winds and also the possibility of a very wet and cyclonic pattern (trough over the UK) with a risk of heavy snowfall becoming established across southern Britain (Moderate confidence)

  • a negative NAO pattern means long periods of positive NAO with persistently mild conditions for weeks on end with high rainfall in north western regions (which characterised the 1990s winters) is unlikely and although two volcanic eruptions have been in the news lately (Mount St Helens in the US & Grímsvötn in Iceland), neither volcano is in the tropics and neither were of sufficient strength to eject volcanic ash into the stratosphere, so volcanic dust is not a natural factor favouring positive NAO this winter. With respect to solar activity, although there was a recent large sunspot, solar activity is not currently high, nor forecast to be (see Spaceweather ) so this other natural factor does not favour positive NAO either (High Confidence)

  • periods with an anticyclone located over the UK are also likely to be common giving some long dry spells with frost and fog (High Confidence)

The Weather:

  • TEMPERATURE: likely to be a colder winter than recent ones with longer cold spells from the north but the proximity of the anticyclone off western Europe means mild Atlantic air coming over the top of the high pressure will occur quite frequently. The decrease in the S index during 2004 (Figure 4) suggests this winter is unlikely to have the very mild southerly spells which occurred in the last two winters.   There is a possibility that strongly negative NAO develops in January & February which may bring some very cold easterly spells.
    75% Confidence Range for winter CET Anomaly (from 1961-90 average): -1.0 to +0.5 C

  • RAINFALL: the proximity of the anticyclone off western Europe means a drier than average winter is likely. However, if the 'cyclonic/west PNA' pattern develops, wetter cyclonic conditions may prevail at times which may mean eastern regions have average rainfall. If a strongly negative NAO pattern develops, wet cyclonic conditions may develop in southern regions. Northern & western Scotland and Northern Ireland are likely to have well below average rainfall

  • SNOWFALL: as was the case in the last two winters, northerly spells may bring some snowfall to many with heavy snowfalls likely in northern Scotland and in western and eastern coastal regions exposed to the north. A shifting west of the anticyclone at times could bring more general heavy snowfall to eastern regions.   There is a low risk that a strongly negative NAO pattern may bring a risk of heavy snowfall to many regions if it develops at the same time as a mobile cyclonic pattern across southern Britain - average to above average snowfall

  • GALES: the blocked anticyclonic pattern means gales and severe gales are likely to have a below average incidence. This is consistent with the University College London North Atlantic Storminess Forecast.   However, if a strong negative NAO develops, a mobile cyclonic pattern may evolve across southern Britain giving a risk of gales here.

Prospects for the Scottish Ski Resorts

This will be a regular feature of the Useful Info winter forecast for the UK. And the prospects for winter 2004/2005 look very good with frequent northerly spells bringing regular heavy snowfalls to the Scottish ski resorts and the blocked weather pattern with low frequency of gales suggests ski days lost to severe gales should not be a big problem. Last winter, the northerly spells with heavy snowfall were soon followed a few days later by very mild southerly winds and heavy rain which melted the heavy snowfalls before they had been allowed to consolidate. The decrease in the S index during 2004 (Figure 4) suggests that very mild southerly spells will not follow the northerlies this winter allowing consolidation of the snowpack and hopefully much greater and more persistent snowcover in highland Scotland. And the high probability of a drier than average winter in northern Scotland should minimise melting from rainfall and mean many dry sunny days making for very pleasant conditions for skiing. (Confidence Moderate)

Is an adjustment for global warming required?

Higher global temperatures imply that weather at all times of the year and from all compass points will become warmer.  If the dramatic warming trend of the surface temperature record and sensational claims that northern hemisphere temperatures are higher now than at any time in the last one thousand years (according to the tree ring reconstruction of Mann, Bradley & Hughes (1999)) are correct, clear evidence of such warming would be apparent when you adjust a temperature record like the CET for wind direction/weather pattern.  The author has thoroughly examined this issue in two studies of the CET (UK Climate Change since 1881 and the 1989 to 2002 warm period in detail and Are the warm years in the UK since 1988 caused by a high incidence of warm weather patterns or global warming?) and the conclusion was that there was little evidence that weather associated with winds from all points of the compass was getting warmer in the last thirty years and what evidence there was was mainly in the spring and autumn.   The marked post-1988 warming of winters and summers is very satisfactorily explained by an increase of westerly winds in winter and an increase in southerly winds and anticyclonicity in summer.  The hypothesis that significantly higher global temperatures, in fact the highest for at least one thousand years, can selectively affect spring and autumn but not winter and summer is absurd and physically impossible.  If such dramatic global warming has occurred, it should be apparent in all seasons.   Table 7 confirms the lack of a significant long term warming trend in winter for a number of sites which have temperature records going back to 1860 but it shows long term warming trends in all other seasons.  

Table 7 Seasonal and annual temperature change explained by a linear trend ( C ) in a number of regional temperature records for the UK area and for the northern hemisphere 1861-2000 and 1901-2000 - From Jones & Lister (2004).

SMT - Scottish Mainland Temperature
SIT - Scottish Island Temperature
NIT - Northern Ireland Temperature
CET - Central England Temperature
Berg - Bergen, Norway
Tor - Torshaven Faroe Islands
Vale - Valentia, Ireland
NH - Northern Hemisphere Surface Temperature Record

* linear trend significant at 5% level

1861-2000 Temperature Change explained by linear trend ( C )  
  ____________________________________________________
                 
  SMT SIT NIT CET Berg Tor Vale NH
____________________________________________________________________
DJF 0.33 0.31 0.30 0.42 0.80 -0.07 0.03 0.75*
MAM 0.78* 0.73* 0.85* 0.69* 1.46* 0.63* 0.37 0.64*
JJA 0.62* 0.53* 0.81* 0.62* 0.50 0.82* 0.07 0.37*
SON 1.06* 1.02* 1.14* 1.29* 1.16* 0.58* 0.55* 0.63*
Annual 0.69* 0.64* 0.77* 0.75* 0.97* 0.50* 0.26* 0.73*
_____________________________________________________________________

1901-2000 Temperature Change explained by linear trend ( C )  
  ________________________________________________
                 
  SMT SIT NIT CET Berg Tor Vale NH
____________________________________________________________________
DJF 0.00 0.08 -0.17 0.12 0.27 0.15 0.16 0.75*
MAM 0.81* 0.60* 0.75* 0.59* 0.75* 0.47 0.63* 0.78*
JJA 0.81* 0.60* 0.81* 0.79* 0.71* 0.44* 0.33 0.56*
SON 0.82* 0.61* 0.69* 1.12* 1.04* 0.48 0.48 0.51*
Annual 0.62* 0.48* 0.53* 0.67* 0.70* 0.39* 0.41* 0.65*
____________________________________________________________________

The fact that the northern hemisphere surface temperature record shows a significant warming trend in winter suggests it may be inaccurate.  Winter in the northern hemisphere is the season when the westerlies in the mid-latitudes are strongest and the sites listed in Table 7 are all in the mid-latitudes and couldn't possibly be unaffected by a significant rise in northern hemisphere temperatures.  It is an impossibility.   What makes winter unique compared to the other three seasons is that north of around 50o N, surface heating from the sun during the day is minimal so the warming trends in the other season are likely to be connected to an increase in sunshine reaching the surface.   Increased anticyclonicity is certainly one possible factor (and this factor does account for the warm summers across Europe post-1988) but another major cause of reduced sunshine in towns and cities of developed countries in the last 300 years has been the widespread use of coal as a fuel source but this practice has dramatically reduced since the 1960s.   Surely the dramatic improvement in air quality in developed countries since the 1960s might possibly explain a major part of the observed warming in spring, summer and autumn but it would have no effect in winter because the warming effect of sunshine at this time of year is minimal anyway.

This is a bit off an off-topic rant for an article about winter forecasts for the UK but the lack of any significant warming trend in winter (Table 7) means that no adjustment for global warming is scientifically justified for UK winter forecasts: northerlies, easterlies southerlies and westerlies are not significantly warmer now than they were 150 years ago in spite all the hype about global warming.   The dramatic warming since the cold 1980s winters (Figure 3) owes to a much higher incidence of warm weather patterns.   If cold weather patterns are dominant in this coming winter, then the UK and Europe will be cold.


Other Winter Forecasts for the UK & Europe

The Useful Info Winter Forecast for the UK is the author's first attempt and shouldn't be taken too seriously.  It should be considered alongside other winter forecasts and areas of common agreement identified.   Here are some winter forecasts, issued by both professional weather forecasting companies and amateur meteorologists.   On the last day of autumn (30 November 2004), of the 11 forecasts listed, all but three were for average to slightly below average temperature with the remainder being mild forecasts which were all issued by the UK Met Office.   Weather forecasting is not a democracy in which the majority prediction will prove to be correct. The UK Met Office with their mild forecast may of course prove to be correct but the consensus view is that winter 2004/2005 in the UK/Europe is likely to be an anticyclonic dryish winter with near or a little below average temperature with a number of northerly cold spells involving snowfall and several forecasts go for long easterly cold spells which have been absent from most winters post 1987 and if these occur, it will be a real shock to many of us:

  • Metcheck - average to slightly colder than average 'winter' (November to January) and three likely cold spells were ambitiously forecasted: November 22 - November 27; December 4 - December 11; January 22 - January 25. Northerly and north-easterly spells are predicted together with a possible stormy period in January.

  • UK Met Office NAO forecast is for a positive NAO which implies a mild winter

  • UK Met Office European Headline - warmer than average temperatures across northern and eastern UK & average across other parts

  • BBC Monthly Outlook After a dry start, the weather will gradually become rather more unsettled during December, with milder weather for most towards Christmas

  • US National Weather Service - unofficial seasonal outlook - is updated daily and on the 15 November update, showed temperature forecasts for average December & January and a cold February in Europe but interestingly the jetstream forecast has consistently been suggesting a negative NAO pattern with a southerly tracking jetstream

  • University College, London NAO forecast is for a negative NAO which implies a cold winter in Europe

  • UKWeatherworld average to slightly below average temperature with northerly/north-westerly spells

  • Net-Weather average temperature but with fairly regular snowfall

  • Tom Presutti average temperature but above average snowfall in most areas

  • International Research Institute for Climate Prediction average temperatures for the UK & most of Europe

  • The Weather Outlook average/slightly below average temperature with above average snowfall

  • Ian Currie mild December but with a white Xmas, average temperature January & February

 REFERENCES

Czaja, A., Robertson, A. W. & Huck, T.  (2003).  The Role of Atlantic Ocean-Atmosphere Coupling in Affecting North Atlantic Oscillation Variability.  In The North Atlantic Oscillation: Climatic Significance and Environmental Impact. Geophysical Monograph, 134,  147-172.

Gillett, N. P., Graf, H. F. & Osborn, T. J. (2003).  Climate Change and the North Atlantic Oscillation.  In The North Atlantic Oscillation: Climatic Significance and Environmental Impact. Geophysical Monograph, 134,  193-209.

Hurrell, J. W., Kushnir, Y., Ottersen, G. & Visbeck, M. (2003) An overview of the North Atlantic Oscillation.  In The North Atlantic Oscillation: Climatic Significance and Environmental Impact. Geophysical Monograph, 134, 173-192.

Mann, M. E., Bradley, R. S. & Hughes, K. S. (1999). Northern Hemisphere Temperatures During the Past Millennium: Inferences, Uncertainties, and Limitations. Geophysical Research Letters, Vol 26, No 6, pp 759

Massacand, A. C. & Davies, H. C. (2001). Interannual variability of European winter weather: the potential vorticity insight. Atmospheric Science Letters, 2, Issue 1-4, p52-60

Jones, P. D. & Lister, D. H. (2004). The development of monthly temperature series for Scotland and Northern Ireland. Report for Scotland & Northern Ireland Forum for Environmental Research (enter sniffer code CC01 to access report).

Ratcliffe, R.A.S & Murray, R. New lag associations between North Atlantic sea temperature and European pressure applied to long range weather forecasting. Q. J. R. Met. Soc., London, 96, 1970 pp 226-246.

Rodwell, M. J. (2003) On the Predictability of North Atlantic Climate.  In The North Atlantic Oscillation: Climatic Significance and Environmental Impact. Geophysical Monograph, 134, 173-192.

Rodwell, M.J., Rowell, D.P. and Folland, C. K. 1999: Oceanic Forcing of the wintertime North Atlantic Oscillation and European climate. Nature, 398, 320-323

Thompson, D. J, Lee, J. S.  & Baldwin, M. P. ( 2003)  Atmospheric processes governing the Northern Hemisphere Annular Mode/North Atlantic Oscillation. In The North Atlantic Oscillation: Climatic Significance and Environmental Impact. Geophysical Monograph, 134, 81-112.

Visbeck, M,  Chassignet, E. P.,  Curry, R. G., Delworth, T. L.,  Dickson, R. R & Krahmann, G.  2003.  The Ocean's Response to North Atlantic Oscillation Variability. In The North Atlantic Oscillation: Climatic Significance and Environmental Impact. Geophysical Monograph, 134,  113-145.