Climate Change Home

Predictions of winter 2006/2007 in the UK & Europe

By Andrew Masterman 9 November 2006

There have been large oscillations in the winter climate of Europe over the last 100 years which are highly correlated with the strength of the westerlies across the Atlantic. The most common measure of the strength of the westerlies is the North Atlantic Oscillation (NAO)  but others exist such as the P index derived from the Lamb daily weather types which are classifications of the surface pressure pattern across the UK (PSC indices defined here). Figure 1 shows changes in the five year running mean of the P index since 1885 and a highly cyclical pattern is obvious and westerliness in winter  is evidently on the decline since the turn of the millennium after a  period of strong westerlies in the 1990s.

 

Figure 1 Five year running means of  winter P index (high value indicates strong westerlies across the UK) and winter C index (high value indicates cyclonic or  low pressure conditions across the UK) 1884 to 2006.

Figure 1 also shows changes in the five year running mean of the C index which is a measure of anticyclonicity/cycloncity across the UK with high values indicating more cyclonic/low pressure conditions. This index is also on the decline and explains why recent winters have been generally dry whereas there were some very wet winter months around the turn of the millennium. So recent winters have had a combination of weak westerlies and anticyclonic conditions.

Figure 2 shows that the decline in the westerlies (lower NAO) since the turn of the millennium has been associated with a decline in winter temperature across Europe with winter 2005/2006 being the 26th coldest winter since 1900 (see European Temperatures winter 2005/2006).

Figure 2  Relationship between the winter NAO and mean winter temperature anomaly (from 1961-90 average) for Europe (r = 0.651, P<0.001). See here for list of ten sites used to calculate Europe winter temperature anomaly.

 

In the  winter forecast for 2004/2005, six main surface pressure patterns in the North Atlantic/Europe region were identified: positive NAO (lower pressure than average near Iceland & higher than average near Azores); negative NAO (higher pressure than average near Iceland & lower than average near Azores); positive PNA (higher pressure than average in east Atlantic/western Europe & lower than average in eastern Europe); negative PNA (lower pressure than average in east Atlantic/western Europe & higher than average in eastern Europe); anticyclonic (A) (high pressure centred over or near UK); and cyclonic (C) (low pressure centred over or near UK). In practice, the infinite variations in monthly surface pressure patterns  means it can be difficult to allocate all monthly surface pressure anomaly charts decisively to these six categories but nonetheless, it is a valid exercise to do so and provides more information than the declining westerliness and cyclonity trends alone show.

Table 1 shows the PSC quintiles (PSC indices defined here) for each winter and each month from December to March  for the last four years and in the case of the last three years, monthly surface pressure anomaly charts from Philip Eden's website, www.climate-uk.com have been used to allocate each winter and winter month to one of  (in some cases a combination of) the six pressure patterns described above.

Table 1  PSC quintiles for each winter and each month December to March 2002/2003 to 2005/2006 and classification of monthly surface pressure anomaly charts (pattern) from Philip Eden's website www.climate-uk.com (pressure anomaly charts are in top-right of pages linked to from the Pattern fields below) from January 2004.

Winter*  Pattern December  Pattern January   Pattern February    Pattern March   Pattern
____________________________________________________________________________________________________
2002/2003 P1S3C2   P2S1C3   P1S3C2   P2S4C1   P2S1C1  
2003/2004 P3S1C2 P2S2C1   P5S1C5      C P2S4C1 +PNA P3S3C3     A
2004/2005 P4S1C2  +NAO P5S2C1 +NAO P4S1C3    +NAO P2S1C3 +PNA P3S4C4 +PNA/-NAO
2005/2006 P1S1C2       A P3S2C3 +PNA P2S3C1      A P1S1C4 +PNA/-NAO P3S2C5 -NAO
_____________________________________________________________________________________________________

Quintiles 1-2 denote below average, quintile 3 average and quintiles 4-5 above average.

Pressure Patterns (see discussion above):  NAO: positive or negative North Atlantic Osillation  index

                                                                PNA:  positive or negative Pacific North American index

                                                                 A:  anticyclonic                                   C:  cyclonic

* Winter defined as December to February.

Table 1 shows that three of the four most recent winters (Dec-Feb) had average or below average westerliness (low P), all four had average or below average southerliness (low S) and all four had above average anticlonicity (low C) confirming the trends shown in Figure 1. These same biases are evident in the monthly data for December to March 2003-2006 with 75% of months having average or below average westerliness and the comparative figures for the S & C indices are 81%. Winter 2004/2005 is the odd winter out with above average westerliness in December and January resulting in the whole winter being classified as positive NAO  but pressure patterns in the late winter period were very different with February and March having  pressure anomalies classified as +PNA/-NAO. Winters 2003/2004 and 2005/2006 also showed the same tendency for +PNA, -NAO or A in  February & March.  This is evident in Figure 3 which presents mean temperatures in Central England in the months of  December to March 1900 to 2006 and shows a cooling in February to average temperature values in three of the last four years. And March 2006  in contrast to most recent Marches which have been very mild was in fact 0.6 C below the 1961-90 average.

 

Figure 3  Monthly values of Central England Temperature (CET) December to March 1900-2006.

 

Figure 2 showed the marked decline in winter temperatures across Europe since the 1990s with winter 2005/2006 being the 26th coldest since 1900. The decline in temperature has been mostly in February with three of the last four Februaries being colder than average in central and eastern Europe (2003, 2005, 2006) but the UK has mostly missed the cold air. The reason for this is the pressure patterns across Europe discussed above and winter 2005/2006 provided some excellent examples of how pressure patterns dictate temperatures across Europe.

Figure 4 shows the global temperature anomalies in January 2006 when record low temperatures were recorded in the Baltic States and Siberian cold of ~-30 C occurred in Moscow. Most of Europe was colder than average in January 2006 while northern Britain and Scandinavia had above average temperatures. The surface pressure anomaly chart  for January 2006 (top right) from Philip Eden's website shows pressure higher than normal across the UK and eastwards across Scandinavia and pressure lower than average across Greenland. The effect of the high pressure block over the UK and Scandinavia was to divert Atlantic low pressure systems up the east side of Greenland into the Arctic instead of the normal situation of them continuing eastwards across the UK and Europe. This kept mild Atlantic air out of Europe and western Russia and allowed more continental cooling than normal across Europe and Russia so temperatures fell very low with monthly anomalies up to 6 C below the 1961-90 average across western Siberia. This temperature distribution was associated with the anticyclone pressure pattern across the eastern Atlantic/ UK (Table 1).

Figure 4 January 2006 global temperature anomalies according to the NCDC

 

Figure 5 shows the global temperature anomalies in March 2006 when the whole of Europe including northern Britain and Scandinavia was colder than average. The surface pressure anomaly charts for March 2006 shows pressure higher than average across Iceland & Greenland and lower than average in the central North Atlantic/Azores region which is the classic negative NAO set-up which brings cold weather to Europe in winter.

 

Figure 5 March 2006 global temperature anomalies according to the NCDC

 

These two examples (January & March 2006) show the large effect of the position of a blocking high in winter on the temperatures experienced in a small part of the globe called the UK and illustrates why it is very difficult to predict winter temperature in the UK: you need to be sufficiently omnipotent to know precisely where the highs and lows will sit in each winter month to obtain an accurate temperature prediction. To attempt to predict this is a tall order. The more rational and realistic approach is to try and predict the pressure pattern across Europe in general terms and from this make some tentative predictions about temperature, rainfall, gales & snowfall across the UK. This is the approach of the Useful Info winter forecasts.

Returning to recent winters and the classification of the surface pressure patterns of individual months (Table 1), only March 2006 had the classic negative NAO set-up which delivers cold across the whole of Europe. In the remaining ten months for which charts are available, three were + PNA, two were +PNA/-NAO hybrids, two were anticyclonic (A), two were + NAO and one was cyclonic (C) . Thus in total, 73% of  these 11 months with charts had either + PNA, A, -NAO or +PNA/-NAO hybrid and this provides a more detailed insight into the nature of recent winter months than the simple decline of the P & C indices in Figure 1 shows. And this also provides an explanation for why the UK has not experienced the cold Februaries of recent years (and January 2006) which most of Europe had: the blocking high which has been a characteristic of recent winters was too close to the UK keeping the cold air to the east and south. Figures 1 & 2 show that although westerliness is on the decline, current westerliness is not as low as in the 1980s and 1960s when there were a number of winter months with the classic negative NAO set-up similar to March 2006.  The 20 winter months since 1950 with the most negative NAO will be looked at later to see whether there is an obvious reason why recent winters have not seen a strongly negative winter NAO month when the whole of Europe is very cold.

 

NORTH ATLANTIC OSCILLATION (NAO) FORECAST

The NAO was discussed in detail in the winter 2004/2005 forecast and a number of factors which dictate NAO variations were identified. The current status of these factors is discussed below:

ATLANTIC SEA SURFACE TEMPERATURE (SST) ANOMALIES: these are shown in Figure 6 which compares global SSTs observed on 4 November 2006 with data from a similar time in the last three years. In all four years, the North Atlantic is much warmer than average and shows the characteristic tri-pole of anomalies: the current combination of warm anomalies in the tropical and northern parts and negative/average anomalies in between favours a negative NAO winter. Rodwell, Rowell & Folland ( 1999) identified positive SST anomalies off the eastern seaboard of the USA and in the Greenland Sea north of Iceland, and  negative SST anomalies south of Greenland and east to Iceland as favouring positive NAO winters while the opposite anomalies favour negative NAO winters. In 2006 according to the rules of  Rodwell et al (1999),  the neutral anomalies off the eastern seaboard of the US are ambiguous while the strong warm anomalies south of Greenland east to Iceland favour negative NAO and the small warm anomalies in the Greenland Sea north of Iceland favour positive NAO.  Another paper emphasises SST anomalies south and east of Newfoundland as the critical region of the North Atlantic: negative SST anomalies there favouring negative NAO and positive anomalies positive NAO (Ratcliffe & Murray, 1970).  With warm anomalies east of Newfoundland, this paper argues for a positive NAO winter.

This technique of studying Atlantic SST anomalies in various regions does tend to give conflicting signals and Table 2 below documents the current status of the various North Atlantic regions in 2006 and in the three previous years.  Given that in the previous three years, the negative NAO signals verified twice with the other year being neutral (similar to - NAO in sense of weak  westerlies), the most misleading region has been the eastern seaboard of the US where modest warm anomalies in two of the three years have favoured positive NAO but to be fair, these anomalies were fairly small. South & east of Newfoundland has also been misleading or unhelpful in the previous three years. The positive NAO signal from the north of Iceland region in 2004/2005 although not verifying for the winter as a whole did appear to foretell a very positive NAO period from December to mid-January 2005 with deep lows in the Greenland and Nordic seas so it is perhaps unwise to dismiss this region in 2006 although current anomalies are not as large or as extensive as in November 2004.

 

Table 2 Status of North Atlantic Sea Surface Temperature Anomalies in different regions in  November 2003-2006 and the observed winter NAO from 2003/2004 to 2005/2006.

 
      Status of  Atlantic SST anomaly with respect to NAO forecast
_____________________________________________________________
%  favouring Observed
Tripole E of US   S of Greenland N of Iceland S & E of Newfoundland negative NAO  Jan-Mar NAO
___________________________________________________________________________________________________
2003/2004  -NAO  +NAO  -NAO  -NAO +NAO 60 +0.04
2004/2005  -NAO  -NAO  -NAO +NAO neutral 60 -3.14
2005/2006  -NAO  +NAO  -NAO  -NAO neutral 60 -2.46
2006/2007  -NAO  neutral  -NAO  +NAO  +NAO 40  ?
___________________________________________________________________________________________________

 

So North Atlantic SST anomalies do show the characteristic tripole of anomalies and warmth south of Greenland which favour a negative NAO winter suggesting that the recent trend to below average westerlies is likely to apply in winter 2006/2007 and these two particular characteristics have with hindsight proved reliable over the last three winters as  predictors for winter NAO (Table 2). However, the modest positive anomalies in the Greenland Sea and stronger positive anomalies east of Newfoundland suggest positive NAO but SST anomalies in these two regions have been unreliable as predictors of the winter NAO. But with only 40% of the SST anomaly regions indicating a negative NAO winter, this is the lowest percentage of the last four years.

STRATOSPHERIC VOLCANIC DUST: there have been no major volcanic eruptions in 2006 and the latest Weekly Volcanic Activity Report  shows no major eruptions so stratospheric volcanic dust does not favour a positive NAO winter.

SOLAR ACTIVITY/SUNSPOTS: the mean monthly sunspot count for Jan-Sep 2006 is very low at 15 and sunspots are currently low.

NAO PROFILE THROUGH 2006.

The purple line in Figure 6 represents the long term average (1900-2004) monthly  NAO through the year and shows a peak in winter and a secondary peak in August. The dark blue line is the average for the 1989-2004 period which has been characterised by persistent above average temperatures all year round and winters with strongly positive NAO (ie persistent westerly winds) and a notable feature of this period is that Septembers & Octobers had anomalously negative NAO. In contrast, the light blue and yellow lines represent two cold periods with negative NAO winters (ie. persistent northerly & easterly winds), 1963-69 & 1979-87 respectively, and both these periods had below average NAO in August and  above average NAO in September & October but in November & December  in the case of 1963-69,  the NAO dropped below average.

In February & March 2006, NAO values were well below the long term averages indicating that winter 2005/2006 was clearly of the colder negative NAO type. Later in the year, the very negative August 2006 value is characteristic of the 1963-69 cold period but the very negative September value is characteristic of the mild 1989-2004 period so there are some conflicting signals here.

 

Figure 6    Monthly values of NAO in 200 6 relative to longer periods: 1900-2004;  1963-1969 (a cold period with negative NAO winters), 1979-1987 (a cold period with negative NAO winters)  and 1989-2004 (a warm period with positive NAO winters).   Data from the Climatic Research Unit.


OTHER NAO FORECASTS

The Met Office Winter NAO forecast is slightly positive and this is based on Atlantic SST anomalies in May. The University of London NAO forecast is also slightly positive showing that there are weak indications for a westerly-type winter across Europe but neither forecast is arguing for a highly positive NAO winter like many in the 1990s.

 

OTHER TELECONNECTIONS

El Nino/Southern Oscillation Index (SOI)

Table 3 in last winter's forecast showed that weak westerlies (low P index) across the Atlantic/Europe tended to occur when the winter SOI was between -2.0 and +2.0 and the following explanation was put forward.  El Ninos (SOI < -2.0) are associated with a strong sub-tropical jetstream across the North Pacific which leads to strong westerlies across North America which would engender a strong westerly flow across the North Atlantic while La Ninas (SOI >2.0) are associated with a strong polar jet stream across the Pacific which again leads to strong westerlies across North America and subsequently across the Atlantic. However, SOI values in between lead to  weaker westerlies across the North Pacific and North America as neither the sub-tropical nor polar jetstream across the North Pacific are strong  and this results in weaker westerlies across the North Atlantic & Europe. And weak westerlies across the North Atlantic/Europe provide a window of opportunity for cold arctic airstreams to come south and westwards across Europe should the state of atmosphere favour arctic outbreaks to mid -latitudes which is the case when negative NAO is favoured by the various factors listed above.

This theory can be tested by consideration of the last four winters shown in Table 1.The Dec-Feb SOI values for winters 2002/2003 to 2005/2006 were -3.0, 0.5, -4.9 & 1.4 respectively. While both 2002/2003 & 2004/2005 qualified as El Ninos on the basis of their negative SOI values, neither were classified as strong El Ninos but weak to moderate ones. In weak to moderate El Ninos, Pacific warmth associated with stronger sub-tropical jetstream across the Pacific often fails to reach the eastern US leading to late winters which are often cold and snowy. Therefore weak to moderate El Ninos also mean weak westerlies across the North Atlantic/Europe which was the case in all months during  winter 2002/2003 but only in February in winter 2004/2005 (Table 1). Winter 2004/2005 was a winter of two halves with the first half having strong westerlies across northern Europe but the latter half was very different with weak westerlies which is characteristic of weak to moderate El Nino winters. In winter 2003/2004, only January had above average westerlies (P quintile 5) which is consistent with the observed winter SOI value of 0.5 being neutral. And winter 2005/2006 was exceptionally blocked/weak westerlies (winter P quintile 1) which is consistent with the observed winter SOI value of 1.4. So the last four winters, two of which were weak to moderate El Ninos do support the hypothesis that the state of El Nino/SOI in the Pacific does have a downstream effect in the North Atlantic & Europe via influencing the strength of the westerlies.

The latest El Nino/SOI forecast is for El Nino conditions to persist through winter 2006/2007 although it doesn't suggest a strong El Nino is expected. The SST anomalies in the equatorial Pacific (Figure 6d) are characteristic of a weak to moderate El Nino with greatest warming in the central region as opposed to greatest warmth being concentrated in the eastern Pacific with a strong El Nino. The consensus appears to be that the current El Nino may have already peaked so a strong El Nino is not expected. The implication of this is that another winter of below average westerlies is likely in the North Atlantic & Europe .


NORTH AMERICAN WINTER FORECASTS

The winter weather pattern across North America is important to the winter weather across Europe for two reasons:

1) via  El Nino/SOI and the effect on the strength of the westerlies as just discussed.

2) the timing and longitude of buckling in the jetstream

Cold arctic airmasses are able to come south to mid-latitudes when the polar jetstream buckles (meanders across a wide range of latitudes) and the jetstream may buckle when it slows down which is why weak westerlies associated with slower jetstreams are necessary for prolonged cold to develop in Europe. However, a buckled jetstream as well as causing cold arctic air to move south also causes warm air from lower latitudes to move north so the longitude of the buckle in the jetstream determines whether Europe is in a warm or cold pattern. These buckles in the jetstream are known as Rossby waves in meteorological jargon and come in different forms depending on how many buckles there are around the northern hemisphere. The three wave pattern is one of the most stable and when the three buckles are located in the longitudes of the Far East, eastern North America and  Europe, cold weather in Europe is likely. Therefore cold weather in Europe is often correlated with cold weather in the eastern United States when the jetstream dives south-eastwards across North America around a ridge in western North America and a trough in eastern North America. This is known as a positive PNA (Pacific North American) pattern which has been discussed already and highlighted in the last two winter forecasts. This positive PNA pattern via the influence of the buckling jetstream tends to get mirrored in Europe with a ridge in the central/eastern Atlantic and a trough in Europe and this is a recurrent pattern we have seen in the last four winters with an anticyclone over or near to the UK and a trough in SE Europe meaning eastern Europe has seen much more cold weather than western Europe.  It is Greece and the Balkans in recent winters which have seen the coldest and snowiest weather.

North American winter forecasts are of course strongly influenced by the current El Nino. The NOOA winter 2006/2007 forecast predicts warmer than average temperatures over most of the US away from the SE quarter where near average temperatures are expected.  The Accuweather.com winter 2006/2007 forecast  predicts a similar overall temperature distribution but predicts colder than normal pattern for eastern and south-eastern US as opposed to near normal as in the case of NOAA. The rationale for this is that Accuweather expect the current El Nino to weaken as we go into winter and weak El Ninos usually lead to cold winters in the east and south-east of the US owing to a recurrent positive PNA pattern. The Accuweather forecast expects the buckling of the jet to be strongest (highly positive PNA) in January & February which is likely to correlate with cold weather in Europe for the reasons outlined above.

 

THE TOP TWENTY NEGATIVE NAO MONTHS SINCE 1950

In Figure 1, the tendency for recent winters to have both weak westerlies and to be anticyclonic was identified and in Table 1, more detail of the weather patterns across Europe in recent winters was revealed with both anticyclonic and positive PNA patterns being recurrent. And it was explained using the examples of January & March 2006 that when the blocking high is located near to the UK under an anticyclonic pattern, the cold air is confined to eastern & southern Europe, but when the blocking high is up near Greenland (negative NAO), cold air is advected into western Europe too meaning the whole of Europe is cold.

Table 1 showed that March 2006 is the only month in the last three winters to have the classic negative NAO set-up which brings cold into western Europe but looking at calendar months only can be misleading. For example, the period mid-February to mid-March 2005 was also the classic negative NAO set-up with a strong Greenland high but February 2005 was classified as +PNA and March 2005 as a +PNA/-NAO hybrid. So recent winters have shown a tendency to negative NAO but the timing of these episodes has been from mid-February into March when northern hemisphere temperatures are rising rapidly with the strengthening sun, so temperatures experienced in the UK have not been below freezing day and night. However, the effects of a strongly negative NAO pattern earlier in the winter from late December to early February are much colder with multi-day periods when temperatures even on low ground in the south of Britain are below freezing day and night. Therefore it has been the late timing of the negative NAO spells in recent years which have meant western Europe has had mild winters while eastern Europe has had cold winters.

It is relevant here to list the twenty months (Dec-Mar) since 1950 with the most negative NAO and to examine the status of various atmospheric parameters in each of those months and winters (Table 3). This list has been compiled using means of  daily NAO in calendar months and not  any consecutive 30 day means of daily NAO as the latter are unavailable on the Internet. Therefore these twenty months will not necessarily be the most negative 30 day periods from Dec-Mar since 1950 but the exercise is nevertheless a valid one. These twenty months are 9% of the total of 228 months Dec-Mar 1950-2006.


 

Table 3  The status of various meteorological measurements in the twenty months
from December to March 1950-2006 with the most negative NAO

    Observeda     Observed      Observed       Annualb
     CET    Anomaly C Observed NAO   Dec-Feb NAO Observed SOI   Dec-Feb SOI No of Hurricanes Observed Sunspots   Sunspots in    preceding year
____________________________________________________________________________________________________
Jan 63 -5.9 -4.09 -6.65 1.1 1.5 3 20 38
Feb 86 -4.8 -4.02 -2.99 -1.6 -0.6 4 23 18
Mar 62 -2.9 -3.78 1.49 -0.4 3.1 3 46 54
Dec 76 -2.7 -3.63 -5.71 -0.6 -0.2 6 15 16
Dec 95 -2.4 -3.33 -6.72 -0.8 0.1 11 10 18
Jan 96 0.5 -3.27 -6.72 1.0 0.1 11 12 18
Jan 79 -4.2 -3.22 -5.92 -0.7 -0.2 5 167 93
Feb 69 -2.8 -3.16 -6.53 -1.1 -3.1 4 112 106
Dec 63 -2.1 -3.09 -2.29 -1.6 -2.4 7 15 38
Feb 65 -0.7 -3.03 -4.26 0.1 -1.0 6 14 11
Feb 56 -4.0 -2.96 -3.21 1.5 3.9 9 124 38
Jan 85 -3.0 -2.87 -1.59 -0.5 0.1 5 17 46
Mar 96 -1.2 -2.57 -6.72 0.7 0.1 11 9 18
Jan 77 -1.0 -2.36 -5.71 -0.7 -0.2 6 16 13
Feb 05 0.6 -2.25 0.84 -4.1 -4.9 9 29 40
Dec 01 -1.1 -2.25 3.07 -1.2 0.1 9 132 120
Dec 89 0.2 -2.23 6.38 -0.7 -3.3 7 166 100
Feb 55 -2.6 -2.21 -1.94 1.8 2.6 8 21 5
Jan 87 -3.0 -2.12 1.06 -0.9 -4.6 4 10 13
Mar 84 -1.0 -2.12 5.09 -0.9 0.6 3 84 67
___________________________________________________________________________________________________

a Central England Temperature anomaly from 1961-90 average
b mean monthly sunspot count for current calendar year for December but preceding calendar year for January to March

The main features of Table 3 are summarised below:

So there is some evidence that strongly negative NAO winter months have neutral SOI and occur in winters with an overall negative NAO pattern and neutral SOI pattern after above average hurricane seasons and in years close to the sunspot minimum.

The status of these predictors in 2006 are:

So according to Table 3, the various parameters do favour the potential of  a strongly negative NAO month occurring at some point during winter 2006/2007 but an important question is the timing of such a spell of negative NAO. If it occurs during late winter as in the last two winters, the UK may well have another mild winter overall with a cold late February and March. If it occurs earlier in the winter say in mid-January, then the whole winter may be colder than average with some multi-day periods below freezing day and night. The Accuweather North American 2006/2007 winter forecast does highlight January & February as a period when buckling in the jetstream across North America may be pronounced which would teleconnect to buckling of the jetstream downstream in Europe at the same time. So perhaps winter 2006/2007 may feature a classic negative NAO set-up earlier  in the winter than in recent years meaning a much colder winter in western Europe than in recent years.

AUTUMN WEATHER

In last winter's forecast, the weather patterns in October and November across the UK were examined to test whether there is a relationship between autumn weather patterns and weather patterns in the following winter. This exercise was confined to the S index (measure of southerliness/northerliness across the UK) and the conclusion was that there was no relationship. The same exercise has been repeated below for the months of October & November using the P index which is a measure of the westerliness across the UK and is highly correlated with the NAO. The period of study was 1950 to 2005

Table 4 shows the results for Octobers 1950 to 2005 and there is a strong tendency for Octobers with above average westerliness (high P/high NAO) to be followed by winter months with below average westerliness (low P) and to some extent above average cyclonicity (high C index). This is consistent with Figure 6 which showed that the warm years of the 1990s had mild westerly winters and Octobers with negative NAO/weak westerlies.

Table 4  Number of PSC quintiles in winters months following Octobers with below average (P1/P2) and
above average (P4/P5) P indices : winters 1950/51 to 2005/2006.

                                                            Number of winter months with PSC quintiles
         _______________________________________________________________________________      
             
                             P                            S                             C
  ___________________________    __________________________   __________________________
             
  Below Avg  Above Avg Below Avg  Above Avg Below Avg Above Avg
_________________________________________________________________________________________________
1-2 14 40 17 28 14 22
3 7 10 12 16 13 17
4-5 18 19 10 25 12 30
_________________________________________________________________________________________________

No. of winters following: P1/P2 Octobers: 13      P4/P5 Octobers: 23

 

Table 5 shows the results for Novembers 1950 to 2005 and shows a tendency for Novembers with above average westerliness (high P) to be followed by winter months with below average southerliness (low P) and there is also a tendency for Novembers with below average westerliness (low P) to be followed by winter months with below average westerliness (low P).

Table 5  Number of PSC quintiles in winters months following Novembers with below average ( P1/P2) and
above average ( P4/P5) P indices : winters 1950/51 to 2005/2006.

                                                         Number of winter months with PSC quintiles
    _________________________________________________________________________________
             
November                         P                          S                             C
Quintiles _______________________     ______________________      _______________________
       
      Below Avg         Above Avg       Below Avg          Above Avg       Below Avg            Above Avg
________________________________________________________________________________________________
1-2 30 39 21 39 18 31
3 10 12 14 22 15 29
4-5 17 36 22 26 24 27
__________________________________________________________________________________________________

                                   No. of winters following: P1/P2 Novembers : 19      P 4/P5 Novembers : 29

 

RECENT TRENDS IN PSC INDICES

Figure 7 shows PSC quintiles from January 2004 to October 2006 (PSC indices defined here) and the logic of this part of the forecast is that PSC quintiles may reveal what biases the atmosphere has in the region of the UK in recent months and one can speculate that such biases may continue in the winter 2006/2007. Of course, such biases may not continue but nevertheless it is a valid exercise. In this case, Figure 7 shows that eight of the ten months so far in 2006 have had below average P indices (low westerlies), five have had above average C indices (high cyclonicity) and three of the last four months have had above average S indices (high southerliness). Obviously, a continued bias to southerliness implies a mild winter, but a continued bias to weak westerlies and above average cyclonicity would suggest cold or average temperatures.

Figure 7  Quintiles of PSC indices for each month January 2004 to October 2006 .

 

TO BE WESTERLY OR NOT TO BE WESTERLY

This is the key question to be asked by anybody who bravely attempts to produce a winter forecast for the UK and Europe. In the case of winter 2006/2007, the various predictors presented above do show some conflicting signals:

This conflict is also evident in some other winter forecasts for 2006/2007. The latest Met Office forecast is for average or above average temperatures and rainfall  in line with their slightly positive NAO forecast although they have acknowledged considerable uncertainty and have included the caveat that their next update could revert to the colder and drier outlook for northern Europe which was their forecast in September. The unofficial US National Weather Service Seasonal Outlook goes for a cold winter for much of Europe although it appears to forecast that the UK will be on the edge of the cold air like last winter. And the International Research Institute for Climate & Society forecast is for a mild winter for much of  Europe.

Sitting on the fence is not satisfactory option here so which way to jump? The Useful Info winter forecast for 2006/2007 goes for a below average westerly winter on the logical grounds that there are more predictors favouring a negative  NAO winter than a positive NAO winter. And Figures 1 & 2 clearly indicate that westerliness is falling since the 1990s and the current SST tripole in the Atlantic is certainly of the form which favours negative NAO (Figure 6d; Table 2) and this has proved reliable in the last three winters in predicting  westerliness/NAO. With hindsight, the warm anomalies in the Greenland Sea in winter 2004/2005 (Figure 6b) did appear to foretell the positive NAO months of December 2004 & January 2005 (Table 1), so it is perhaps unwise to dismiss the more modest current warm anomalies in the Greenland Sea. So the following caveat is added: these modest Greenland Sea positive anomalies may possibly lead to December 2006 having above average westerliness but as the Greenland Sea cools with time, their effect will diminish.  So there is more confidence of  negative NAO patterns later in the winter as has been the case in recent winters.

So the overall pattern across Europe and indeed North America this coming winter looks similar to the last four winters in general terms although each year has had some specific differences. In Europe, three of the last four winters have been cold in eastern Europe and mild in the west while in the US, three of the last four winters have been mild over most of the nation but colder than average in the north-east with much above average snowfall. This consistent pattern either side of the Atlantic in recent years owes to Rossby waves/buckles in the jetstream aligning cold troughs in the longitudes of eastern North America and eastern Europe and this is likely to be the case again this winter. Owing to the current weak to moderate El Nino, the closest analogues of the last four winters are 2002/2003 which averaged over 1 C colder than normal in eastern Europe and 2004/2005 which averaged 0.7 C warmer than normal overall but featured a colder than average February everywhere apart from the UK. Winter 2004/2005 as mentioned earlier (Table 1) has been the odd winter out of recent winters with positive NAO in both December & January and a possible reason was the anomalous warmth in the Greenland Sea (Figure 6b; Table 3). If these two anomalous months are put to one side, the remainder of months (Dec-Mar) have below average westerliness and were classified as either anticlonic (A), +PNA, -NAO or  +PNA/-NAO hybrids (Table 1). In these months, a buckled jetstream across the Atlantic/Europe associated with a recurrent +PNA pattern across North America have been characteristic and similar weather patterns across Europe seem likely during winter 2006/2007. However, it was pointed out earlier that the timing of the strongly negative NAO spells in recent winters has been in late February and March which has meant that western Europe has been milder than average while eastern Europe has been cold owing to the influence of the persistent high centred near the UK as in January 2006 (Figure 4). Is 2006/2007 going to follow the same pattern?

In the author's opinion, it is probable that a high pressure located over or near to the UK is likely to be a recurrent pattern again this winter. This pattern has been a feature of the first nine days of November with two cold blasts associated with the highly buckled jetstream pushing arctic airmasses down the eastern flank of a high near the UK into eastern Europe and a long way south into Greece & the Balkans. Both cold and snow in this region have been characteristic of recent winters. The question of negative NAO spells is more challenging. The negative NAO Top Twenty (Table 3) suggests that there are a number of factors (weak to moderate El Nino, low sunspots, above average hurricane season) which favour the possibility of strongly negative NAO spell occurring during winter 2006/2007 but this forecast has not presented any original evidence that such spells will occur earlier in the winter than in recent years. The most likely evolution is for such a pattern to kick-in ~ 20 February as in the last two winters but August, September and October 2006 have shown an unusual bias to negative NAO (Figure 6; Chart) which indicates that there is currently a propensity for high latitude blocking and should this propensity continue through winter 2006/2007, then strongly negative NAO spells in Dec-mid Feb period could mean a cold winter for the UK & the rest of Europe. The Accuweather US Forecast  identifies January & February as a period of high latitude blocking and highly buckled jetstreams on the basis of the timing of the Madden Julian Oscillation this winter which would imply a risk of negative NAO spells at this time. So perhaps this winter will be different than recent winters when negative NAO spells occurred in late winter.

Another weather pattern that may repeat itself is the stationary mid-Atlantic low occurring during a period of negative NAO which leads to very wet and mild spells with southerly winds across the UK. These have occurred in both recent El Nino winters (2002/2003 & 2004/2005) and we have just recently seen an example in late October 2006. The recent trend of months having a strong bias to southerly winds (Figure 7) does perhaps support this prediction.

So a concise summary of all the above is that the overall northern hemisphere pattern during the last four winters has supported weak westerlies across the North Atlantic and as a consequence the occurrence of cold weather in Europe with  negative NAO spells but owing to particular conditions in these winters, these negative NAO spells have occurred late in the winter and a persistent high located near the UK has meant most cold weather has been confined to eastern Europe. Now the overall northern hemisphere pattern which supports weak westerlies across the North Atlantic also applies for this winter 2006/2007 and in the first nine days of November 2006, there have already been two arctic blasts directed towards  the Balkans & Greece which has been characteristic of recent winters. So another winter which is cold in the east and average or mild in western Europe is a good bet. However, Table 3 suggests conditions are ripe for a period of strongly negative NAO with a weak El Nino and low sunpsots in particular, so there is a significant risk that the timing of the NAO spells may be earlier than in recent winters meaning some really cold airmasses get into the UK and western Europe too this winter. Both Figure 1 & 2 show that previous periods of low westerlies in the 1960s and 1980s did have some winters in which the average NAO was more negative than recent ones and in these cases very cold air associated with strongly negative NAO spells occurred in the UK and western Europe. There is no reason to suspect that this current cycle of low westerlies won't have winters which are similar. The overall northern hemisphere pattern suggests that winter 2006/2007 could be such a winter.

Now follows the Useful Info Winter forecast for 2006/2006.


THE GENERAL WEATHER PATTERN

THE WEATHER

PROSPECTS FOR THE SCOTTISH SKI CENTRES

Before embarking on this part of the forecast, it is germaine to consider what the ideal weather pattern for prolonged snowcover in the Scottish mountains is and what sort of snow conditions weather from the main compass points bring.

The ideal weather pattern would be one which would ensure below zero temperatures at an altitude of 500 to 1000m from 1 December through to 31 March resulting in no thawing of snow cover and ensuring that all precipitation fell as snow. But the inconvenient reality is that it is very rare for temperatures to stay below zero for a single month never mind a whole ski season at this relatively low altitude. The Scottish mountains are simply not high enough to ensure sub-zero temperatures in winter, the UK is surrounded by seas with temperatures at their lowest in winter ~ 4 C in the North Sea and ~8 C in the Atlantic and the UK is in the path of Atlantic lows which drag up warm air from the tropical Atlantic which is present in their warm sectors.  So the reality is that the Scottish ski centres are located at altitudes and geographic locations which are highly unfavourable for persistent snow cover. Periods with temperatures above zero are inevitable and combined with high orographic rainfall which is characteristic in Scotland, major thaws are also inevitable. These are the main reasons why snow cover on the Scottish mountains is not permanent in winter and not global warming/climate change.

Snowfall from different compass points:

This forecast and future Scottish skiing forecasts will be presented in terms of the probability of these wind directions occurring.

Scottish snowfall forecast 2006/2007:

The above suggests good potential for some heavy snowfalls and prolonged snow cover but with a risk of a few devastating thaws associated with wet southerly spells. However, the previous Scottish skiing forecasts for the last two winters by the author predicted above average incidence of northerlies but both of the December to mid-February periods were largely snowless and good skiing conditions only prevailed from mid-Feb to mid-March in 2004/2005 and from late Nov to early Dec and late Feb to late April in 2005/2006. In 2004/2005, December & January had above average incidence of westerlies leading to mild, wet & very windy conditions with poor snowcover: with hindsight, this probably owed to the warm anomalies in the Greenland Seas (Figure 6b). In 2005/2006, the season started early with a northerly in late November which dumped snow in spades at the eastern centres giving a week of top to bottom riding and two weeks of cover on upper and middle runs. However, once this had melted by ~ 7 December, there was weeks and weeks of mainly dry weather until late February which meant no or poor snowcover for most of this period with only limited skiing available at the eastern centres through January and much of February before conditions dramatically improved in late February. This dry period was associated with a persistent anticyclone situated close to the UK, a feature which is likely again this winter but the El Nino is likely to mean it is less persistent this winter and that there will be more Atlantic lows passing through the UK which should mean snowfall at intervals between the dry anticyclonic periods. This of course doesn't guarantee good snow conditions in the peak season of late December to mid-February but there is no reason to suppose this winter will again have little or no snow in the peak season. All it takes is a 24-48 hour snowstorm in late December to provide weeks of snow cover on the upper and middle runs even without further snowfall, particularly in a winter with below average westerliness which is the pattern we are in at the moment (Figures 1 & 2).

To be honest, predicting snowcover on the mountains of Scotland is a tall order but this doesn't mean it cannot or should not be attempted. This section is a pioneering attempt to do so and like the whole winter forecast should not be taken too seriously.

References

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