greatbatch2015 .pdf



Nom original: greatbatch2015.pdfTitre: Tropical origin of the severe European winter of 1962/1963

Ce document au format PDF 1.4 a été généré par dvips 5.83 Copyright 1998 Radical Eye Software / Acrobat Distiller 7.0 (Windows), et a été envoyé sur fichier-pdf.fr le 16/07/2016 à 12:06, depuis l'adresse IP 89.82.x.x. La présente page de téléchargement du fichier a été vue 513 fois.
Taille du document: 3.3 Mo (13 pages).
Confidentialité: fichier public


Aperçu du document


Quarterly Journal of the Royal Meteorological Society

Q. J. R. Meteorol. Soc. 141: 153–165, January 2015 A DOI:10.1002/qj.2346

Tropical origin of the severe European winter of 1962/1963
R. J. Greatbatch,a * G. Gollan,a T. Jungb and T. Kunza
a

b

GEOMAR Helmholtz Centre for Ocean Research, Kiel, Germany
Alfred Wegener Institute, Helmholtz Centre for Polar and Marine Research, Bremerhaven, Germany

*Correspondence to: R. J. Greatbatch, GEOMAR | Helmholtz Zentrum f¨ur Ozeanforschung Kiel, D¨usternbrooker Weg 20, 24105
Kiel, Germany.
E-mail: rgreatbatch@geomar.de

A set of relaxation experiments using the European Centre for Medium-Range Weather
Forecasts (ECMWF) atmospheric model is used to analyze the severe European winter of
1962/1963. We argue that the severe winter weather was associated with a wave train that
originated in the tropical Pacific sector (where weak La Ni˜na conditions were present)
and was redirected towards Europe, a process we suggest was influenced by the combined
effect of the strong easterly phase of the Quasi-Biennial Oscillation (QBO) and unusually
strong easterly winds in the upper equatorial troposphere that winter. A weak tendency
towards negative North Atlantic Oscillation (NAO) conditions in December, associated
with extratropical sea-surface temperature and sea-ice anomalies, might have acted as a
favourable preconditioning. The redirection of the wave train towards Europe culminated
in the stratospheric sudden warming at the end of January 1963. We argue that in February
the sudden warming event helped maintain the negative NAO regime, allowing the severe
weather to persist for a further month. A possible influence from the Madden–Julian
Oscillation, as well as a role for internal atmospheric variability, is noted.
Key Words:

1962/1963; tropical impact; stratospheric impact; NAO

Received 15 May 2013; Revised 23 January 2014; Accepted 5 February 2014; Published online in Wiley Online Library 26
March 2014

1. Introduction
The winter of 1962/1963 was particularly severe in northern
Europe. Taking winter to be the average over December,
January and February (DJF), with a winter mean temperature
of −0.3 ◦ C, 1962/1963 is the third coldest in the Central
England Temperature (CET) record going back to the winter
of 1659/1660 (Manley, 1974; Parker et al., 1992), exceeded
only by the winters of 1683/1684 and 1739/1740 (see
http://www.metoffice.gov.uk/hadobs/hadcet/). January 1963 is
the fifth coldest January in the CET record going back to 1659,
with a mean temperature of −2.1 ◦ C, and February, with a
mean temperature of −0.7 ◦ C, is the seventh coldest February
in the CET record, also going back to 1659. The severe weather
was notable for its persistence (see Burt, 2013), beginning in
earnest in the UK at Christmas 1962 and lasting with little respite
until the beginning of March 1963 (Brown, 2006; Dent, 2013).
Indeed, the winter of 1962/1963 lives long in the memory of
those who experienced it, including the first author of this article.
The severe weather was not confined to the British Isles but
extended across much of Europe (see Figure 1), with the peak
anomaly for the DJF average temperature (referred to the period
1960/1961–2001/2002) occurring over Germany and Poland at
more that 6 ◦ C below normal.
Severe winters over the UK and northern Europe are typically
associated with the negative phase of the North Atlantic
Oscillation (NAO) (e.g., Greatbatch, 2000; Hurrell et al., 2003)
c 2014 Royal Meteorological Society


and the winter (DJF) NAO index is, indeed, negative for
1962/1963, although not as low as in the recent winter of
2009/2010.∗ Another factor was the East Atlantic (EA) pattern,
which, in combination with the negative NAO index, was also
in a (negative) phase in 1962/1963, favourable for cold winter
weather over Europe (Moore and Renfrew, 2011). The question
arises as to whether the negative phase of both the NAO and the
EA pattern arose simply by chance, or was the likelihood of severe
winter weather in Europe enhanced by other factors such as the El
Ni˜no/Southern Oscillation (ENSO) phenomenon in the tropical
Pacific (e.g., Fraedrich and M¨uller, 1992; Greatbatch et al., 2004;
Ineson and Scaife, 2009), North Atlantic sea-surface temperature
(SST) anomalies (e.g., Ratcliffe and Murray, 1970; Rodwell and
Folland, 2002) or events in the stratosphere (e.g., Baldwin and
Dunkerton, 2001)?
The winter of 1962/1963 is topical, partly because the time
of writing is 2013, 50 years after the event, but also because
the recent decade has seen a recurrence of cold winter weather
in Europe (see, for example Petoukhov and Semenov, 2010;
Cohen et al., 2012) following the period of predominantly mild
winters in Europe during the 1980s and early 1990s. Folland et
al. (2012) claim that two factors were important for determining

The winter (DJF) of 2009/2010 currently holds the record for the lowest value
of the winter NAO index in the record going back to the winter of 1863/1864 and
provided by Hurrell at https://climatedataguide.ucar.edu/climate-data/hurrellnorth-atlantic-oscillation-nao-index-pc-based.

154

R. J. Greatbatch et al.
DJF 1962/63

72°N

60°N

48°N

36°N
24°N
12°N
90°W

60°W

30°W



30°E

60°E

(a)

5

m s−1

Figure 1. DJF mean 2 m temperature anomaly with respect to the reference
period 1960/1961–2001/2002 (all data from the ERA-40 reanalysis). Red (blue)
contours indicate warm (cold) anomalies. The contour interval is 1 ◦ C and the
zero contour is omitted.

0

−5

1965

1970

1975

1980

1985

1990

1995

2000

62/63
climatology

(b) 10

m s−1

5
0
−5
−10
01/11/62

01/12/62

01/01/63

01/02/63

01/03/63

01/04/63

Figure 2. (a) DJF mean of the zonal-mean zonal wind at 150 hPa averaged over
the equatorial zone between 5◦ N/S (all data from the ERA-40 reanalysis). (b)
Blue: Daily mean of the zonal-mean zonal wind at 150 hPa averaged over the
equatorial zone between 5◦ N/S, from 1 November 1962–31 March 1963. Red:
Daily climatology of the same index. The black dashed lines indicate ±1 standard
deviation about the climatological mean, both computed using the whole ERA-40
data set.

the character of the winter in 1962/1963. One is the tripole SST
pattern in the North Atlantic from the previous May, identified
by Rodwell and Folland (2002), and which in May 1962 favoured
a negative NAO the following winter. The other is the strong
easterly phase during the winter of 1962/1963 (see figure 1 of
Folland et al., 2012) of the Quasi-Biennial Oscillation (QBO),† a
stratospheric phenomenon (see Andrews et al., 1987) which was
also identified by Jung et al. (2010c) as contributing to the cold
European winter of 2005/2006. It is well known that the negative
(that is easterly) phase of the QBO in winter favours a weakened
polar vortex and, by implication, the negative NAO (Holton and
Tan, 1980; Boer and Hamilton, 2008). The mechanism is thought
to involve the northward displacement of the zero wind line in
easterly QBO winters (Holton and Tan, 1982), a factor we think
played a role in 1962/1963.
Another feature of the winter of 1962/1963, to our knowledge
not previously noted, is the occurrence of unsually strong easterly
winds in the zonal mean along the Equator in the upper
troposphere. Figure 2(a) shows the time series of winter (DJF)
mean zonal-mean zonal wind at 150 hPa averaged around the

When referring to the QBO in the following, we refer to the index used by
Folland et al. (2012) defined in terms of the zonal wind at 30 hPa.

c 2014 Royal Meteorological Society


Equator between 5◦ N and 5◦ S. The winter of 1962/1963 clearly
stands out as by far the most easterly over the plotted time period.
Furthermore, from Figure 2(b), which shows the corresponding
daily index, it is the clear that there was a shift to easterly winds in
the zonal mean along the Equator at 150 hPa in mid-December,
immediately preceeding the onset of the severe weather in Europe
and remaining not only easterly but anomalously so, by one
standard deviation or more, until early March 1963, when the
cold weather finally came to an end. Similarly to what happens
in the stratosphere in association with the QBO, anomalously
strongly easterly winds in the equatorial zone at 150 hPa (where
the mean zonal wind is weakly westerly in ERA-40) can influence
the propagation of planetary waves towards Europe, a factor we
think played an important role in 1962/1963.
Interestingly, the tropical Pacific during the 1962/1963 winter
exhibited weak La Ni˜na conditions (see the time series of the
Ni˜no3 and Ni˜no3.4 SST indices at http://www.cgd.ucar.edu/
cas/catalog/climind), as in fact had been the case for the previous
winter as well. According to the canonical response to ENSO over
the Euro-Atlantic sector (Fraedrich and M¨uller, 1992), a La Ni˜na
in the tropical Pacific favours the positive NAO, contrary to what
was observed. Nevertheless, we cannot rule out the possibility
that the La Ni˜na conditions that winter played a role in the
dynamics of the 1962/1963 winter. More likely, however, is the
possibility that the Madden–Julian Oscillation (MJO) influenced
the dynamics of the 1962/1963 winter over Europe (Cassou, 2008;
Lin et al., 2009), a topic beyond the scope of the present article but
discussed briefly at the end of section 4. It is known, for example,
that the MJO is connected to anomalies in the zonal-mean zonal
wind along the Equator (Madden and Julian, 1994; Slingo et al.,
1996; Lee, 1999; Hoskins et al., 1999), which, as noted above, was
unsually strong and easterly during the winter of 1962/1963.
In order to analyze the 1962/1963 winter, we follow the
approach of Jung et al. (2010c) and use ensembles of experiments
using a recent version of the European Centre for MediumRange Weather Forecasts (ECMWF) atmospheric model. These
experiments have been previously used by Greatbatch et al. (2012)
to examine influences on Northern Hemisphere winter mean
circulation anomalies due to the Tropics, extratropical SST and
sea-ice and the stratosphere. The approach is to use a relaxation
technique, described in detail in Jung et al. (2010b,c), to constrain
the atmospheric circulation in parts of the model domain to be
close to the ERA-40 reanalysis. Greatbatch et al. (2012) examined
the 42 winters 1960/1961–2001/2002, whereas here we focus on
a single winter (1962/1963) and also use monthly mean output
to study the time development of features influencing the severe
weather. Section 2 discusses the model set-up and the different
experiments and section 3 focuses on the results. Section 4
provides some discussion and, finally, section 5 gives a summary.
2. Methods
2.1.

Experimental set-up

The numerical model used is a version of the ECMWF atmosphere
model, very similar to the model described in Jung et al. (2010a,b).
The horizontal and vertical resolution are the same as used for the
ERA-40 reanalysis (thereby avoiding the need to perform interpolation when carrying out the relaxation –see below); in particular,
the model uses spectral truncation T159 with 60 levels in the vertical extending up to 0.1 hPa, with about half the levels located
above the tropopause (Untch et al., 1998). Each model experiment consists of 12 ensemble members, each ensemble member
being run forward from initial conditions taken from the ERA-40
reanalysis at six-hourly intervals from around 1 November 1962
and the analysis carried out on the following winter, i.e. December, January and February (DJF). All model experiments, along
with their abbreviations, are discussed in detail below. For those
experiments that include the time series of observed SST and
sea-ice, the data were taken from the ERA-40 reanalysis.
Q. J. R. Meteorol. Soc. 141: 153–165 (2015)

Severe European Winter of 1962/1963
Throughout this article, ‘anomalies’ for a particular experiment
refer to departures of the ensemble mean or individual ensemble
members from the mean winter state of the same experiment
carried out for all winters from 1960/1961–2001/2002, as in
Greatbatch et al. (2012), the winter mean being the average over
all ensemble members and all years comprising the experiment.
By defining anomalies in this way, we ensure that the anomalies
reflect the anomalous conditions during the individual winters,
with no contribution from the different model climates associated
with each experiment. Likewise, for the ERA-40 reanalysis,
anomalies refer to the departure from the mean over all winters
1960/1961–2001/2002.
2.2.

Relaxation formulation

In the relaxation experiments, the model is drawn toward the
ERA-40 reanalysis over a specific region; this is achieved by
adding an extra term

−λ(x − xref )

(1)

to the ECMWF model. The model state vector is represented by
x and the reference field toward which the model is drawn by
xref . The strength of the relaxation is determined by λ = aλ0 ,
where a defines the geographic region and model levels where
the relaxation is applied. Here λ0 = 0.1 h−1 and corresponds to
a time-scale of 10 h. The parameters that are relaxed are zonal
velocity u, meridional velocity v, temperature T and the logarithm
of the surface pressure, ln ps (ln ps is not relaxed in stratospheric
relaxation experiments). xref is taken from the six-hourly ERA-40
reanalysis data, linearly interpolated in time to each model time
step. When applying masks (in grid-point space) to localize the
relaxation, care was taken to reduce adverse effects close to the
relaxation boundaries. In particular, the transition from relaxed
to unrelaxed regions in the tropical relaxation cases is smoothed
using a hyperbolic tangent function. The smoothing is such that
the relaxation coefficient λ goes from λ0 to 0 within a 20◦ belt
in latitude. Boundaries of 20◦ N and 20◦ S stated in the text below
refer to the centre of the respective 20◦ belt (see figure 1 of Jung
et al., 2010b). Changes of λ are also smoothed in the vertical in
the stratospheric relaxation case. Here, the relaxation coefficient
goes from λ0 to 0 in a vertical layer encompassing about 13
model levels, as in Jung et al. (2010b,c). The values of λ at 500,
200, 50 and 20 hPa are given by 1.1 × 10−7 λ0 , 2.3 × 10−6 λ0 ,
0.018λ0 and 0.5λ0 , respectively (see figure 2 of Jung et al.,
2010b). In the case of stratospheric relaxation, the design of the
relaxation zone was chosen to test the influence of large-scale
stratospheric circulation anomalies, which have been observed to
appear first in the upper stratosphere and subsequently spread
downward into the lower stratosphere, where they are believed
to affect tropospheric weather regimes (Baldwin and Dunkerton,
2001) and hence the winter mean circulation in the troposphere
(Jung and Leutbecher (2007) have shown that the ECMWF
model reproduces such behaviour). The use of a smooth, rather
than abrupt, transition zone is designed to reduce the spurious
reflection of upward-propagating planetary waves, although such
effects are difficult to eliminate completely and should be borne
in mind when interpreting the results.
For discussion of the relaxation technique as applied to the
Tropics, readers are referred to Hoskins et al. (2012). These
authors note that the 10 h time-scale used here for the relaxation
is sufficient to be able to evaluate the impact of the Tropics
on the extratropics, our aim here. In addition, Jung (2011) has
compared the relaxation technique with 4D Var data assimilation,
both being applied separately in the same selected regions of
the model domain and using a version of the ECMWF model
very similar to the one employed here. He found that both
techniques give very similar results. The relaxation technique,
however, has the advantage of being computationally much less
expensive.
c 2014 Royal Meteorological Society


2.3.

155

Model experiments

The model experiments are as follows.
(1) OBS-NO. In this case, the model sees observed SST and
sea-ice at the lower boundary and no relaxation is used.
(2) CLIM-TROPICS. In this case, climatological SST and seaice are specified at the lower boundary and relaxation
is used between 20◦ N and 20◦ S throughout the whole
depth of the model atmosphere, including the tropical
stratosphere.
(3) OBS-TROPICS. In this case, observed SST and sea-ice
are specified at the lower boundary and relaxation is used
between 20◦ N and 20◦ S as in CLIM-TROPICS. Since
the relaxation completely overwhelms the specification of
observed SST in the Tropics, this experiment effectively
looks at the additional information gained compared with
CLIM-TROPICS, by specifying the observed SST and seaice in the extratropics.
(4) CLIM-STRAT. In this case, climatological SST and sea-ice
are specified at the lower boundary and relaxation is used
in the stratosphere north of 30◦ N, excluding the Tropics.
A 20◦ wide transition zone is applied, centred on 30◦ N.
A concern regarding CLIM-STRAT is the quality of the ERA-40
reanalysis, given that winter 1962/1963 took place prior to the
satellite era. However, this is not to say that the reanalysis in
1962/1963 is completely unconstrained in the stratosphere, since
there is certainly an influence from the underlying troposphere,
as can be shown using model experiments that employ relaxation
only in the troposphere (not shown here) as well as data from
radiosondes. Also, the stratosphere as represented in the reanalysis
is a dynamically consistent realization of the stratospheric state
within the model used for the reanalysis. We believe, therefore,
that the results we report in section 3 at least indicate the influence
of the stratosphere on the underlying troposphere within the
context of the model used for the reanalysis. Furthermore, since
the model we use is a version of the ECMWF model with the
same resolution as used for the reanalysis, the dynamics of the
model we use is very similar to that used for the reanalysis. Similar
considerations apply to other changes in the data stream used for
the reanalysis (e.g. data from the Tropics, which are much more
abundant today than was the case in 1962/1963).
2.4.

Testing the statistical significance of the results

In the following analysis, we show pattern correlations between
the ensemble mean model anomaly for each of the different
model experiments and the corresponding anomaly fields from
the ERA-40 reanalysis. To test the statistical significance of the
results, we use the full set of ensemble mean anomaly fields from
the model experiments covering the 42 winters from 1960/1961
to 2001/2002 (see Greatbatch et al., 2012). These ensemble mean
anomaly fields have been correlated with the reanalysis anomaly
fields in all possible combinations, mixing up the years. The
resulting distribution of pattern correlations is very similar for
each experiment when considering winter means or monthly
means separately, enabling us to combine experiments and
produce the threshold correlations for 90, 95 and 99% significance
shown in Table 1. It should be noted that the pattern correlations
are computed only north of 30◦ N to avoid regions in which
relaxation is applied and also that, for the pattern correlations at
50 hPa, the experiment CLIM-STRAT, which includes relaxation
at that level, is excluded. The 95% significance level is defined,
for example, as the pattern correlation above which only 5% of
the randomly produced pattern correlations are found. If the
pattern correlation between the model ensemble mean and the
corresponding field in the same winter (here 1962/1963) from
the reanalysis exceeds this threshold, we say that the pattern
correlation is significantly different from zero at the 95% level.
Q. J. R. Meteorol. Soc. 141: 153–165 (2015)

156

R. J. Greatbatch et al.

Table 1. Significance thresholds for the pattern correlations shown in the figures.
NH, NA and NP refer to the Northern Hemisphere north of 30◦ , the North
Atlantic sector and the North Pacific sector, respectively.
NH

NA

NP

Z500 DJF mean
90% level
95% level
99% level

0.41
0.51
0.66

0.54
0.64
0.79

0.68
0.77
0.90

Z500 DJF monthly means
90% level
95% level
99% level

0.39
0.48
0.63

0.53
0.64
0.80

0.64
0.75
0.88

Z50 DJF monthly means
90% level
95% level
99% level

0.65
0.75
0.87

0.79
0.87
0.94

0.72
0.82
0.92

the strong negative NAO pattern stands out, with positive
height anomalies near Iceland and negative height anomalies
over Europe to the south. Viewed hemispherically, the anomaly
pattern has something of the appearance of the negative (be it
westward shifted) of the so-called ‘cold ocean warm land’ pattern
(COWL: Wallace et al., 1996), with positive height anomalies over
the northern and western parts of the North Pacific as well as over
the northern North Atlantic. The positive height anomalies over
the North Pacific sector appear to be northward and, particularly,
westward shifted compared with expectations, given the La Ni˜na
conditions in the equatorial Pacific noted earlier (compare Figure
3(b) with figure 3(b) of Hoerling et al., 1997) and there are
notable negative anomalies further south over the North Pacific
in the reanalysis. Looking at the model results, the features over
the North Pacific are generally captured in the two cases that use
tropical relaxation (CLIM-TROPICS and OBS-TROPICS). This
is not the case in the experiment (OBS-NO), which uses only the
time series of observed SST and sea-ice at the lower boundary,
suggesting that ocean/atmosphere interaction processes are not
being correctly represented (at least somewhere) in the Tropics
in OBS-NO (see the discussion of this issue in Greatbatch et al.,
2012).
Viewed over the hemisphere as a whole, the best model
performance is clearly shown by CLIM-TROPICS, the experiment
that has relaxation in the Tropics but uses climatological SST and
sea-ice outside the Tropics. Indeed, the pattern correlation for
CLIM-TROPICS reaches or exceeds the 99% significance level
over the Northern Hemisphere as a whole and over both the North
Atlantic and North Pacific sectors separately. The amplitude is
reduced compared with the observations, but this is no surprise
given that what is shown from the model is the ensemble mean,

3. Results
Seasonal mean

Figure 3 shows seasonal (DJF) mean anomalies in 500 hPa
geopotential height (Z500) from ERA-40 and the ensemble mean
of the model experiments. The Z500 field from ERA-40 is also
shown without removing the mean (Figure 3(a)) from which the
strongly blocked pattern over and to the west of the British Isles is
clearly evident. Looking at the anomalies from the reanalysis,

o

180 W

o
45 N

o
45 N

E

E

o

o

o

W

60

o

60

W

o

o

o

o
60 N
o
75 N

75 N

o

W

60

60 N
o

60

o

0

0

0

(d) CLIM−TROPICS
0.66* / 0.85* / 0.95*

(e) OBS−TROPICS
0.62* / 0.69* / 0.85*

(f) CLIM−STRAT
0.21 / 0.30 / −0.10

o

o

o

o
30 N

o
45 N

o
45 N

o
45 N

o

o

60 N

o

60 N

o

o
75 N

o

o

60

W

W
o

0

o

0

oW
60

o

60

o

60

E

75 N

E

75 N

60 o
E

60 N

12 o
0W

30 N

o

180 W

o
0E
12

30 N

o

o
0E
12

180 W

o

o
0E
12

180 W

60

60

E

75 N

5280
5520

o

5760

60 N

o

60

o

12 o
0W

o
45 N

12 o
0W

o
30 N

o
0E
12

o
30 N

o
0E
12

o
30 N

o
0E
12

12 o
0W

180oW

o

180 W

12 o
0W

(c) OBS−NO
0.39 / 0.60 / 0.40

(b) ERA−40 anomaly

(a) ERA−40 total

12 o
0W

3.1.

o

0

Figure 3. 1962/1963 DJF mean 500 hPa geopotential height plotted over the region north of 20◦ N. (a) Full field from ERA-40 (contours are labelled in m in the plot).
(b) ERA-40 anomaly field with respect to the period 1961–2002. The contour interval is 15 m and the zero contour is omitted. (c) As (b), but for the ensemble mean
of the experiment OBS-NO. The pattern correlations between the ERA-40 anomaly and the anomaly of the ensemble mean, shown as NH/NA/NP, are given for the
Northern Hemisphere north of 30◦ N (NH), the North Atlantic sector (NA: 90◦ W–40◦ E, 30–80◦ N) and the North Pacific sector (NP: 160◦ E–140◦ W, 30–65◦ N),
respectively. Pattern correlations that are significantly different from zero at the 95% level or higher are highlighted with an asterisk (see Table 1). (d) As (c), but for
CLIM-TROPICS. (e) As (c), but for OBS-TROPICS. (f) As (c), but for CLIM-STRAT.

c 2014 Royal Meteorological Society


Q. J. R. Meteorol. Soc. 141: 153–165 (2015)

Severe European Winter of 1962/1963
whereas the real world corresponds to a single realization, an issue
we return to at the end of this subsection. Adding extratropical
SST and sea-ice in OBS-TROPICS acts to increase the amplitude
the model response over the North Atlantic, but at the expense
of reducing the pattern correlation. The increase in amplitude
is consistent with Folland et al. (2012), who claim that the
North Atlantic SST was an important influence on the overlying
atmospheric circulation in winter 1962/1963 (see the discussion
in section 4). The decrease in pattern correlation, which now
fails to meet the 99% significance level over not only the North
Atlantic but also the Northern Hemisphere as a whole and the
North Pacific sector, suggests the possibility of a deficiency in
the way the model captures the influence of extratropical SST
and sea-ice in OBS-TROPICS. The impact of the extratropical
SST and sea-ice is also evident in OBS-NO, which also exhibits a
westward-shifted, negative NAO pattern over the North Atlantic,
with the pattern correlation reaching the 90% significance level
over the North Atlantic sector.
Viewing the winter as a whole, there appears to be no role for
the mid-latitude stratosphere, as revealed by CLIM-STRAT. It
should be noted that the severe weather started around Christmas
1962 and it was not until the end of January 1963 that a
sudden warming event took place in the stratosphere (Finger
and Teweles, 1964). This means that the scenario pointed out by
Baldwin and Dunkerton (2001), in which a weakened polar vortex
spreads downwards from the stratosphere, ultimately affecting the
troposphere, was not operative in the early part of the cold spell,
although it may have played a role in causing the severe weather
to persist throughout February, as we discuss below.
It was noted that the amplitude of the ensemble mean
model response in CLIM-TROPICS is lower than that of the

corresponding winter mean field from the reanalysis. It is always
important to remember that what the ensemble mean shows is
the model response to the imposed forcing (in the case of CLIMTROPICS, the relaxation towards the reanalysis in the Tropics)
and that the variability internal to the model has been filtered out
by the ensemble averaging. The real world, on the other hand, as
represented by the reanalysis, is a single realization and so includes
a contribution from atmospheric internal variability. Indeed, the
discrepancy in amplitude between the ensemble mean field in
CLIM-TROPICS and the corresponding field from the reanalysis
suggests that internal atmospheric variability played a role in
making the winter of 1962/1963 such an extreme event. It seems
likely that extreme winters such as 1962/1963 probably occur
when both a forced response, e.g. from the Tropics, and internal
variability happen to interfere constructively. The degree to which
such constructive interference can occur other than by chance is
an interesting question. For example, the tropical circulation, to
which the model is constrained in CLIM-TROPICS, may well itself
have been influenced by the state of the extratropical circulation in
1962/1963. There is then the interesting possibility that a positive
feedback might have occurred in 1962/1963 between the tropical
and extratropical circulations. That the extratropical circulation
can be an important influence on the Tropics has been shown,
for example, by Vitart and Jung (2010). Furthermore, Lin et al.
(2009) note an intriguing two-way interaction between the MJO
and NAO during the winter season. The possible role of the MJO
in 1962/1963, in connection with the anomalously strong easterly
winds in the upper equatorial troposphere that winter, was noted
in the Introduction.
A closer look at the individual ensemble members is given
by Figure 4, which shows a comparison between the winter

(a) ERA−40 62/63 DJF

(b) 0.61* / 0.77* / 0.80*

o

o

o
30 N

o
45 N

o
45 N

o
75 N

o

o

60

60

W

W

60 o
E

o
75 N

o

0

0o

(c) 0.41 / 0.75* / 0.16

(d) 0.44 / 0.65 / 0.93*

o

o

180 W

180 W

o
45 N

o
45 N

o
75 N

o

o

W

W

60 o
E

o
75 N

60

60

0o

o
60 N

60 o
E

60 N

12

o

0o
W

o
30 N

o
0E
12

o
30 N

o
0E
12

0o
W

o
60 N

60 o
E

60 N

12

o

0o
W

o
30 N

o
0E
12

12

180 W

o
0E
12

0 oW

180 W

12

157

0o

Figure 4. As Figure 3, but for three individual ensemble members from the experiment CLIM-TROPICS. (a) Winter mean anomaly from ERA-40; (b) winter mean
anomaly from the ensemble member with the highest pattern correlation over the Northern Hemisphere north of 30◦ N. This case also has the highest pattern
correlation over the North Atlantic sector. (c) Winter mean anomaly for the ensemble member with the second highest pattern correlation over the North Atlantic
sector; (d) winter mean anomaly for the ensemble member with the highest pattern correlation over the North Pacific sector. The contour interval is 15 m and the
zero contour is omitted.

c 2014 Royal Meteorological Society


Q. J. R. Meteorol. Soc. 141: 153–165 (2015)

R. J. Greatbatch et al.
(a) 12/1962,
ERA−40 total
o

180 W

180 W

12 o
0W

12 o
0W

E

E

o

o

W

W

60

o

60

o

75 N

o

W

60

o
60 N
o
75 N

o

60

o

(e) OBS−TROPICS
0.11 / −0.06 / −0.21

(f) CLIM−STRAT
0.23 / −0.16 / 0.48

180oW

180oW

180oW

o

30 N

o

30 N

o
30 N

o
45 N

o
45 N

o
45 N

o

60 N
o

o
75 N

E

o

E

o

o

W

60

o

W

W

0o

60

o

60

o

60

0o

E

75 N

75 N

o
60 N

60

60 N

o

60

o

12 o
0W

(d) CLIM−TROPICS
0.27 / −0.10 / 0.66

o
0E
12

0o

o
0E
12

0o

o
0E
12

0o

12 o
0W

60

E

5520

o
60 N

o
45 N

o
0E
12

45 N

5760

75 N

5280

o

o
0E
12

45 N

o

o
0E
12

o

o
60 N

o
30 N

o
30 N

o
30 N

12 o
0W

o

o

180 W

12 o
0W

(c) OBS−NO
0.15 / 0.28 / 0.20

(b) ERA−40 anomaly

60

158

0o

Figure 5. As Figure 3, but for the December 1962 monthly mean 500 hPa geopotential height.

(DJF) mean anomalies for 1962/1963 from the reanalysis (top left
panel) and three of the 12 ensemble members comprising CLIMTROPICS (the experiment with the best performance compared
with the reanalyis in the ensemble mean). It should be noted that
the same contour interval is used for the model anomaly fields
as for the reanalysis, from which it is immediately clear that the
amplitude of the model anomalies is comparable to that in the
reanalysis when looking at individual ensemble members. The
pattern correlations with the reanalysis are also shown, exactly
as in Figure 3. The significance levels for the pattern correlations
are almost the same as in Table 1 for the ensemble means and
were computed as described in section 2.4, but with ‘ensemble
mean’ replaced everywhere by ‘ensemble member’. Considering
the 12 ensemble members as a whole, seven members have a
pattern correlation over the Northern Hemisphere north of 30◦ N
that exceeds the 90% significance level, clearly much in excess
of what might be expected by chance (12/10, i.e. roughly one
ensemble member). Over the North Atlantic sector, seven out of
the 12 ensemble members have a pattern correlation that exceeds
the 90% significance level, with another two ensemble members
almost reaching that level. Over the North Pacific sector, six
members exceed the 90% significance level and, of these, five
exceed the 95% significance level, again far more than can be
expected by chance (a further two ensemble members almost
reach the 90% threshold).
Overall, these results imply an important role for the Tropics
in the dynamics of the 1962/1963 winter, an issue we now explore
further by looking at each of the months of December, January
and February separately below.
3.2.

Monthly mean anomalies

We now look at monthly mean anomalies in both the ERA40 reanalysis and the model experiments. Starting first with
c 2014 Royal Meteorological Society


December 1962 (Figure 5), we first note that none of the ensemble
mean fields from the model reaches the 90% significance level,
either hemispherically or in either of the North Pacific or North
Atlantic sectors, with the exception of CLIM-TROPICS, which
reaches the 90% significance level in the North Pacific sector.
Indeed, the best model performance is clearly CLIM-TROPICS in
the North Pacific sector, where the model shows similar features
to those seen in the reanalysis. Interestingly, the ensemble mean
fields from the tropically relaxed case (CLIM-TROPICS) and the
stratospherically relaxed case (CLIM-STRAT) both show some
similar features over the North Pacific and North American
sectors, even though CLIM-STRAT contains no information
about the Tropics (apart from what might be left from the
initial conditions). The signal is, nevertheless, much more clearly
pronounced in CLIM-TROPICS than it is in CLIM-STRAT. These
features, the low anomaly in Z500 over the North Pacific and
the dipole of positive anomalies over western North America
and the far northeastern part of Russia, are also present in the
reanalysis. Even over the western part of the North Atlantic
sector, both CLIM-TROPICS and CLIM-STRAT exhibit similar
anomalies, even though here they do not agree with those seen in
the reanalysis. Rather, to get closer to the anomalies seen in the
reanalysis over the North Atlantic in December, it is necessary
to include the influence from extratropical SST and sea-ice, as in
OBS-NO and OBS-TROPICS, both of which indicate weakened
westerly winds over the North Atlantic, although the pattern
correlations are clearly not significant.
Returning now to CLIM-TROPICS, the model results for
December 1962 (Figure 5) suggest that a signal is emerging from
the tropical Pacific sector. Furthermore, the fact that CLIMSTRAT exhibits similar features to CLIM-TROPICS over the
North Pacific/North America sector where, as we have seen,
CLIM-TROPICS exhibits some skill when compared with the
reanalysis, suggests that the stratosphere might be helping to
maintain the signal from the Tropics. This could, perhaps, be
Q. J. R. Meteorol. Soc. 141: 153–165 (2015)

Severe European Winter of 1962/1963
(a) 1/1963,
ERA−40 total
o

o

o

180 W

180 W

o
45 N

o

75 N

60 N

5760

60 N

o

12 o
0W

o
45 N

o

12 o
0W

o
45 N

52
5 5 2 80
0

o
30 N

o
60 N

o
0E
12

o
30 N

o
0E
12

o
30 N

o
0E
12

0 oW

(c) OBS−NO
0.47 / 0.62 / 0.36

(b) ERA−40 anomaly

180 W

12

159

o
75 N

o

75 N

E

o

60

E

o

60

o

60

W

W

W

o

o

o

(e) OBS−TROPICS
0.70* / 0.69* / 0.78*

(f) CLIM−STRAT
0.02 / 0.31 / −0.41

180oW

180oW

180oW

o

30 N

o

30 N

o
30 N

o
45 N

o
45 N

o
45 N

o

o
75 N

E

o

o

60

o

60

W

o

0

o

W

W
o

0

60

o

60

o

60

E

75 N

E

75 N

o
60 N

60

60 N

o
0E

o

o

o
0E

60 N

o
0E

o

12 o
0W

(d) CLIM−TROPICS
0.73* / 0.87* / 0.79*

12 o
0W

0

12

0

12

0

12

12 o
0W

o

60

o

60

o

60

E

5040

o

0

Figure 6. As Figure 5, but for January 1963.

associated with the easterly phase of the QBO in the stratosphere
that, as noted earlier, was a feature of this winter (Folland et al.,
2012), or the easterly zonal-mean zonal wind that was already
becoming established in the upper equatorial troposphere in
December (Figure 2(b)).
Turning to January 1963 (Figure 6), we see that the best model
performance compared with the reanalysis is given by CLIMTROPICS (the tropically relaxed case). The pattern correlation for
this experiment exceeds the 99% significance level over both the
Northern Hemisphere as a whole and the North Atlantic sector
and exceeds the 95% significance level over the North Pacific
sector. Indeed, CLIM-TROPICS is clearly capable of capturing
the main features seen in the reanalyis, although with less skill over
the Asian continent than in other sectors. Looking at the signal
emanating from the tropical Pacific in CLIM-TROPICS, we see the
strong suggestion that this signal is redirected (reflected) towards
Europe from the region over the southwestern US. Furthermore,
the part of the wave train that appears to be redirected towards
Europe (but not the part emanating from the tropical Pacific) is
also captured in CLIM-STRAT, suggesting that the stratosphere
plays a role in the redirection of the wave train emanating from
the Tropics and that CLIM-STRAT is capturing the residual effect
of this redirection (reflection) process (see also Figure 10 and the
discussion thereon).
These results point to the importance of the zero wind line and
the influence of both the strong easterly phase of the QBO (Holton
and Tan, 1982) and the unusually strong easterly winds in the
upper equatorial troposphere during January 1963 (Figure 2(b)).
It is well known that zero wind lines can act as reflectors of quasistationary Rossby waves (see Killworth and McIntyre, 1985).
We therefore suggest that the events of winter 1962/1963 were
influenced by a wave train that emanated from the tropical Pacific
region and was redirected (reflected) towards Europe by the
critical layer associated with the zero wind line that, in turn,
was influenced by both the strong easterly phase of the QBO
c 2014 Royal Meteorological Society


and the unusually strong easterly zonal-mean zonal winds in the
upper equatorial troposphere that winter. We note that wave
reflection from the zero wind line in the Tropics is also a feature
of stratospheric sudden warming events (Dunkerton et al., 1981).
In fact, it seems likely the wave reflection process noted here
culminated in the stratospheric sudden warming that took place
at the end of January 1963 (Finger and Teweles, 1964).
Before leaving January, we note that the influence from
extratropical SST and sea-ice (as seen by comparing OBSTROPICS with CLIM-TROPICS) acts to modify the model
response from that seen in CLIM-TROPICS but, as for the
winter mean, acts to reduce the pattern correlation in the North
Atlantic sector. The evidence from OBS-NO, the case forced only
by observed SST and sea-ice without relaxation, is that, as in
December, the extratropical SST and sea-ice also favoured the
negative NAO. The pattern correlation reaches the 90% level over
the North Atlantic sector in this case.
Turning now to February (Figure 7), we see that the Tropics
(CLIM-TROPICS) continued to exercise an important influence,
especially over the North Atlantic sector where the pattern
correlation with the reanalysis exceeds the 95% significance
level in CLIM-TROPICS. Over the hemisphere as a whole, the
pattern correlation is much reduced compared with January but,
nevertheless, has its main features in common with the reanalysis
over the North Pacific and North American sectors. The negative
NAO signal in CLIM-TROPICS appears to be enhanced by
including the effect of the extratropical SST and sea-ice, as can
be seen by comparing OBS-TROPICS with CLIM-TROPICS.
Indeed, the pattern correlation over the North Atlantic exceeds
the 95% significance level in OBS-TROPICS and also in OBSNO. Furthermore, February is the only one of the three months
in which CLIM-STRAT exhibits a strong performance, with the
pattern correlation exceeding the 95% significance level over
the North Atlantic sector. This suggests that the stratospheric
sudden warming at the end of January (Finger and Teweles, 1964)
Q. J. R. Meteorol. Soc. 141: 153–165 (2015)

160

R. J. Greatbatch et al.
(a) 2/1963,
ERA−40 total
o

180 W

180 W

o
45 N

o
45 N

o
45 N

60 N

o
60 N
o
75 N

o

o

E

o

o

60

W

W

W

60

o

60

o

60

o

o

o

o

(e) OBS−TROPICS
0.48* / 0.75* / 0.68

(f) CLIM−STRAT
0.34 / 0.70* / −0.30

180oW

180oW

180oW

o

30 N

o

30 N

o
30 N

o
45 N

o
45 N

o
45 N

o

60 N

o
60 N

o

o
75 N

o

60

E

o

o

W

W

W

60

o

60

o

60

o

60

o

0

E

75 N

E

75 N

o
0E

o
0E

o

12

60 N

o
0E

o

12 o
0W

(d) CLIM−TROPICS
0.45 / 0.76* / 0.66

12 o
0W

0

12

0

12

0

60

60

E

75 N

E

75 N

60

o

o

5180
5420
5660

60 N

o
0E
12

o

12 o
0W

o
30 N

12 o
0W

o
30 N

o
0E
12

o
30 N

o
0E
12

12 o
0W

o

o

180 W

12 o
0W

(c) OBS−NO
0.32 / 0.70* / −0.29

(b) ERA−40 anomaly

o

o

0

0

Figure 7. As Figure 5, but for February 1963.



What is shown is the projection of daily zonal-mean geopotential height
anomalies from the ERA-40 reanalysis on to the pattern associated with the
first EOF of December–March monthly mean zonal-mean geopotential height
anomalies, as in Kunz and Greatbatch (2013).
c 2014 Royal Meteorological Society


1962/63
z[km]

did indeed play a role in causing the severe weather to persist
from January into February, following the ideas of Baldwin and
Dunkerton (2001) in which a weakened polar vortex in the
stratosphere can spread downwards and affect the troposphere
(see also Baldwin and Dunkerton, 1999). However, the events
of the late winter of 1963 were very different from those of
the late winter and early spring of 2013, despite the fact that,
as in 1963, a stratospheric sudden warming took place at the
end of January 2013. Whereas in 1963 there was a change to
a positive NAO regime that took place in early March and
brought an end to the severe winter, in 2013 a negative NAO
regime persisted in the troposphere well into April, leading to
the coldest March in the UK since at least 1962 and perhaps
since 1892 (see the Central England Temperature record at
http://www.metoffice.gov.uk/hadobs/hadcet/). It is interesting, in
this respect, to look at Figure 8 (see also the panel for 1962/1963
in Plate 2 of Baldwin and Dunkerton (1999), reproduced as figure
12 of Greatbatch (2000)). The figure shows the daily index for the
Northern Annular Mode (NAM: Thompson and Wallace, 2000)
over a range of pressure levels in the atmosphere.‡ It is clearly seen
that whereas the annular mode remained in a negative (red) state
in the stratosphere in March 1963, implying a weakened polar
vortex, the positive annular mode (blue) became established in
the troposphere in early March, bringing an end to the severe
winter weather. The examples of 1963 and 2013 indicate the
caution that should be applied when drawing implications about
the troposphere from the occurrence of warming events a month
or more previously in the stratosphere: very different outcomes
are possible.

40
20

Nov

Dec

Jan

Feb

Mar

Apr

Figure 8. The daily NAM index during the winter 1962/1963, computed as
in Kunz and Greatbatch (2013). The vertical coordinate is log pressure height
z = −H ln(p/ps ), where p is pressure, ps is reference pressure, taken to be
1000 hPa, and H = 7 km is a scale height for the atmosphere. The calculation is
carried out at each height z separately and the time series have been normalized
to have zero mean and variance 1 (see Kunz and Greatbatch, 2013). The contour
interval is 0.5 standard deviations, yellow/red colours indicating negative values
(weakened vortex) and blue indicating positive values (strengthened vortex).

To understand the success of CLIM-TROPICS in February
1963 over the North Atlantic sector, it is helpful to look at the
monthly mean plots of the ensemble mean geopotential height at
50 hPa (Z50) in the stratosphere shown in Figures 9–11. Starting
with February (Figure 11), the reanalysis shows a much weakened
polar vortex, as indicated by the positive height anomalies
centred over the northwestern part of Greenland, as one would
expect after a sudden warming event. Since CLIM-STRAT uses
relaxation in the stratosphere, it is not a surprise that this case
reproduces the observed anomalies at 50 hPa. Of the remaining
experiments, OBS-NO (the case that does not include relaxation)
shows negative Z50 anomalies, in contrast to the reanalysis.
On the other hand, both CLIM-TROPICS and OBS-TROPICS,
which include tropical relaxation, both show a weakened polar
vortex, the pattern correlation with the reanalysis exceeding the
99% significance level over the North Atlantic sector in CLIMTROPICS and the 95% level over the hemisphere as a whole,
with somewhat reduced pattern correlations in OBS-TROPICS.
The reduced amplitude compared with the reanalysis indicates,
not surprisingly, that there is variability in the stratospheric polar
Q. J. R. Meteorol. Soc. 141: 153–165 (2015)

Severe European Winter of 1962/1963
(a) 12/1962,
ERA−40 total

(c) OBS−NO
0.11 / 0.75 / −0.87

(b) ERA−40 anomaly

o

o

o

180 W

180 W

180 W

o
45 N

20300
20060

E

o

E

o

o
o

60

o

60

o

60

E

75 N

W

W

W

60

o
60 N
o
75 N

o

19580
19820

75 N

60 N

60

0

o

60

54

60 N

o

12 o
0W

o
45 N

o

12 o
0W

o
45 N

o
0E
12

o
30 N

o
0E
12

o
30 N

o
0E
12

o
30 N

20

12 o
0W

161

20540
0

0

(d) CLIM−TROPICS
−0.04 / 0.50 / −0.73

(e) OBS−TROPICS
0.57 / 0.71 / 0.38

(f) CLIM−STRAT
0.94* / 0.99* / 0.92*

o

o

o

o
30 N

o
45 N

o
45 N

o
45 N

o
75 N

o

E

E

o

o

60

W

W

W

60

o

60

o

60

o

60

E

75 N

75 N

o

o

o

0

0

o
60 N

60

60 N
o

o

o
0E

60 N

o

o
0E

o
0E

o

180 W

12 o
0W

30 N

12 o
0W

30 N

o

12

180 W

o

12

180 W

12

12 o
0W

o

o

o

0

0

Figure 9. As Figure 5, but for the December 1962 monthly mean 50 hPa geopotential height.

(a) 1/1963,
ERA−40 total
o

180 W

180 W

o
45 N

o
45 N

60 N

E

o

E

o

o

60

W

W

W

60

o

o

60

E

20460

75 N

60

o

o

o

o

(f) CLIM−STRAT
0.90* / 0.87* / 0.78

180oW

180oW

180oW

o

30 N

o

30 N

o
30 N

o
45 N

o
45 N

o
45 N

o

o
75 N

E

o

o

o

60

W

o

0

o

W

W
0o

60

o

60

o

60

E

75 N

E

75 N

o
60 N

60

60 N

o

0

Figure 10. As Figure 9, but for January 1963.

c 2014 Royal Meteorological Society


o
0E

o

o

o
0E

60 N

o
0E

o

12 o
0W

(e) OBS−TROPICS
0.81* / 0.64 / 0.76

12 o
0W

(d) CLIM−TROPICS
0.70 / 0.80 / 0.46

12

0

12

0

12

0

60

60

201

00

75 N

o
60 N
o
75 N

o

60

o

o
0E

60 N

o

o
0E

o
0E

o

12 o
0W

o
45 N

12 o
0W

o
30 N

12

o
30 N

12

o
30 N

12

12 o
0W

o

o

180 W

12 o
0W

(c) OBS−NO
0.61 / 0.28 / 0.79

(b) ERA−40 anomaly

Q. J. R. Meteorol. Soc. 141: 153–165 (2015)

162

R. J. Greatbatch et al.

o

180oW

o

30 N

o

30 N

o
30 N

o
45 N

o
45 N

o
45 N

60 N

12 o
0W

o

o
75 N

o

o

o

o

o

0

0

0

(d) CLIM−TROPICS
0.85* / 0.97* / 0.80

(e) OBS−TROPICS
0.74 / 0.68 / 0.82*

o

60

E

E

o

o

60

W

W

W

60

o

60

o

60

o

(f) CLIM−STRAT
0.99* / 0.99* / 0.97*

o

o

o
45 N

o
45 N

o
45 N

E

o

60

E

o

o

0o

W

o

0

o

W

W

60

o
75 N

60

o

60

o

60

E

o
75 N

o
60 N

o
0E

o
75 N

60 N

o
0E

o
0E

60 N

o

12 o
0W

o
30 N

12 o
0W

30 N

o

180 W

12

30 N

o

12

180 W

o

12

180 W

60

60

E

75 N

20360

20120

75 N

o
60 N

o
0E
12

60 N

180 W

o
0E
12

o

12 o
0W

180oW

o

12 o
0W

(c) OBS−NO
−0.35 / −0.73 / −0.26

(b) ERA−40 anomaly

o
0E
12

12 o
0W

(a) 2/1963,
ERA−40 total

0o

Figure 11. As Figure 9, but for February 1963.

vortex between the different ensemble members. We suggest
that the success of CLIM-TROPICS and to some extent OBSTROPICS at reproducing the anomalies in Z500 in February is
related to the weakened vortex in the stratosphere during this
month in these experiments and that the downward-spreading
scenario noted by Baldwin and Dunkerton (2001) is likely to be
operative. Some confirmation of this idea is found by looking
at the Z500 plots from these experiments (especially CLIMTROPICS) shown in Figure 7. It is notable that the negative NAO
dipole in this plot is orientated in the north/south direction,
different from in January (Figure 6), when the features over the
North Atlantic appear as part of a wave train, with the anomalies
orientated southwest–northeast. This difference and the strong
performance of CLIM-STRAT in the North Atlantic sector noted
when discussing Figure 7 argue strongly that, whereas in January
the anomalies over the North Atlantic were part of a wave train
originating in the Tropics, in February they appear to be locally
generated and separated from the wave train (now much weaker),
which nevertheless still seems to be present leaving the tropical
Pacific sector. Note that the upper tropospheric equatorial zonalmean zonal wind in February 1963 remained unusually strong
and easterly, as had been the case in January 1963 ( Figure 2(b)),
favouring the presence of this wave train; the reader is referred to
a work of Gollan and Greatbatch (2014).
Looking at the Z50 anomalies for December (Figure 9), both
CLIM-TROPICS and OBS-TROPICS indicate a weakened polar
vortex, whereas the reanalysis suggests a vortex that is displaced
towards the northeastern part of Russia. In particular, both
CLIM-TROPICS and OBS-TROPICS miss the negative height
anomalies over the northeastern part of Russia in the reanalysis,
although negative height anomalies are present further west over
the Eurasian continent in OBS-TROPICS, the case that includes
extratropical SST and sea-ice). Looking at January (Figure 10),
we see that the pattern correlation with the reanalysis in OBSTROPICS exceeds the 95% significance level over the hemisphere
c 2014 Royal Meteorological Society


as a whole. We can also see the wave train, which we have
argued is reflected towards Europe from near the southwestern
USA, in both CLIM-TROPICS and OBS-TROPICS. Furthermore,
the orientation of these anomalies is better represented in CLIMTROPICS and OBS-TROPICS than it is in the case with relaxation
in the stratosphere (CLIM-STRAT), arguing that these features
do not have their origin in the extratropical stratosphere (where
the relaxation is applied in CLIM-STRAT) but rather in the
Tropics. These same features are present in the reanalysis, but
with enhanced amplitude. It is rather remarkable that this wave
train is present in the stratosphere at 50 hPa, even though the
wave train is almost certainly of tropospheric origin, indicating a
role for the stratosphere in its dynamics.
4. Discussion
4.1.

The role of North Atlantic SST anomalies

We noted that in December 1962 extratropical SST and sea-ice
favoured negative NAO conditions over the North Atlantic, albeit
a weak effect (see the discussion regarding Figure 5, in particular
concerning OBS-TROPICS and OBS-NO). Figure 12 shows the
SST anomalies in each month (i.e. anomalies from the mean over
the ERA-40 period in each month) over the North Atlantic. It is
clear that, as the winter develops, cold SST anomalies near the
British Isles become increasingly intense, as one would expect in
response to the anomalously cold atmospheric conditions. There
is also a band of weak cold anomalies that extends southwestwards
from Europe across the Atlantic and which also intensifies into
January and February, almost certainly due to direct forcing from
the atmosphere (Cayan, 1992, see also figure 2 in Rodwell et al.
(1999)). The warm anomalies in the band from Atlantic Canada to
Iceland strengthen in the observations from December–January,
also consistent with forcing by the negative NAO state of the
Q. J. R. Meteorol. Soc. 141: 153–165 (2015)

Severe European Winter of 1962/1963
(a) Dec 1962

72°N

60°N

48°N

48°N

36°N

36°N

24°N

24°N

60°W

30°W



30°E

60°E

60°N

48°N

36°N
24°N
12°N
90°W

60°W

30°W



30°E

60°E

(c) Feb 1963

72°N

60°N

48°N

24°N

60°W

30°W

60°W

30°W



30°E

60°E

the summer thermocline, at which time it can produce the
required atmospheric response directly. However, it is clear that in
1962/1963 the situation was not as simple as this. Interestingly, the
cold anomalies in May near Atlantic Canada persisted throughout
the summer of 1962 (not shown), only to be replaced by warm
anomalies in this region during the autumn and into December
(see Figure 12). Indeed, by December, only a small remnant of the
cold SST anomalies remains in the mid-Atlantic Bight region. It
is, therefore, unlikely that the mechanism put forward by Rodwell
and Folland (2002) was responsible for the tendency towards
negative NAO conditions in December 1962, even though we
have noted such a tendency, albeit weak, in this article and have
associated that tendency with the influence of extratropical SST
and sea-ice. More importantly, our experiments imply that the
dominant influence on the winter of 1962/1963 over Europe, aside
from internal atmospheric variability, came from the Tropics,
with North Atlantic SST anomalies playing a minor role at best.
4.2.

36°N

12°N
90°W

12°N
90°W

Figure 13. As Figure 12, but showing SST anomalies for May 1962.

(b) Jan 1963

72°N

May 1962

72°N

60°N

12°N
90°W

163



30°E

60°E

Figure 12. Monthly mean SST anomalies for December 1962, January 1963 and
February 1963 using data from the ERA-40 reanalysis and referred to the mean over
the ERA-40 period (1958–2002). Yellow/red contours indicate warm anomalies,
while green/blue contours indicate cold anomalies. The contour interval is 0.5 ◦ C
and the zero contour is omitted.

overlying atmosphere (Cayan, 1992). As shown by Rodwell et al.
(1999), the SST pattern that is forced by the NAO in turn tends to
feed back positively on to the NAO (although the model response
shown in Rodwell et al. (1999), e.g. Figure 3(a), is clearly shifted
northeastward compared with the NAO), a conclusion that is
generally consistent with our model results.
Looking at the SST anomaly pattern in May 1962 (Figure 13),
we see a pattern of anomalies quite similar to the negative
of that shown in figure 5 of Rodwell and Folland (2002),
especially the cold SST anomalies off Atlantic Canada. This
pattern of SST anomalies in May is associated by these authors
with the negative NAO in the following winter, consistent with
observations and an important factor in the statistical prediction
for that winter given by Folland et al. (2012) (see their figure
2. Note, however, that these authors can only account for about
50% of the north European temperature anomaly that winter,
leaving room for other factors such as the Tropics or internal
variability to play a role). The argument given by Rodwell and
Folland (2002), for connecting SST anomalies in May with the
atmospheric circulation the following winter, is that the SST
anomaly pattern re-emerges in the early winter from beneath
c 2014 Royal Meteorological Society


Implications for predictability

It is a good question to ask whether, on the basis of our
analysis, the winter of 1962/1963 might have been predictable
in advance? Interestingly, Namias (1964) shows a successful
forecast for this particular winter based on a very simple formula;
namely that, at any location, the change in the atmospheric
circulation from one winter to the next is taken to be the
same as the change in the atmospheric circulation between
the preceding autumns. The formula is based on the idea that
the climate system sometimes exhibits, quoting from Namias
(1964), ‘long-period climate fluctuations lasting more than one
year’, a possible example being slow variations in SST due to
ocean dynamics (Bjerknes, 1964). However, in the case of the
1962/1963 winter, the success rests on the fact the NAO index
in autumn 1962 was much more negative than was the case
in 1961 (see https://climatedataguide.ucar.edu/climate-data/
hurrell-north-atlantic-oscillation-nao-index-station-based), on
the basis of which the formula predicts a much more negative
NAO index in winter 1962/1963 than in 1961/1962. Such a
difference, however, was not associated with low-frequency,
decadal variability of the climate system. Indeed, the key
to being able to predict such dramatic differences as that
between winters 1961/1962§ and 1962/1963 probably rests
with other factors, the QBO and the zonal-mean zonal wind
in the upper equatorial troposphere being good candidates.
A related point is clearly the importance of representing
stratospheric processes in models used for seasonal prediction, as noted by Folland et al. (2012) and by Scaife
§
Interestingly, although the NAO index for January and February 1962 is positive, the index for December 1961 is strongly
negative: see https://climatedataguide.ucar.edu/climate-data/hurrell-northatlantic-oscillation-nao-index-pc-based.

Q. J. R. Meteorol. Soc. 141: 153–165 (2015)

164

R. J. Greatbatch et al.

(http://www.rmets.org/events/long-range-forecasting-problemmythology-science-and-progress/european-seasonal-forecast).
Not only do we suggest that the QBO (a stratospheric phenomenon) played a role in the dynamics of the 1962/1963 winter,
but we have also suggested that the scenario described by Baldwin
and Dunkerton (2001) played a role in causing the severe weather
to persist throughout February. In this scenario, anomalies
present in the stratosphere typically spread downwards and can
impact the underlying troposphere for up to several months after
an event such as a stratospheric sudden warming.
We have noted that consistently, during the 1962/1963 winter,
the Tropics played a key role in our model experiments. Winter
1962/1963 was characterized by weak La Ni˜na conditions in the
tropical Pacific, as noted earlier. However, La Ni˜na conditions
are not usually associated with the negative NAO (Fraedrich and
M¨uller, 1992; Ineson and Scaife, 2009). Nevertheless, the wave
train that we have identified in December 1962 and January 1963
as being implicated in the severe winter weather consistently
appears to have its origin in the North Pacific sector. One
possibility is that the wave train is related to the MJO (see Lin et
al., 2009, 2010). A detailed discussion of the possible role of the
MJO is beyond the scope of the present manuscript, but clearly
remains a topic for future research. We have argued that an
important role in shaping the winter of 1962/1963 was played by
the zonal-mean zonal wind in the equatorial upper troposphere.
This wind was in an extreme easterly state during the winter of
1962/1963 (Figure 2) and anomalies can be driven by the MJO
(Madden and Julian, 1994; Slingo et al., 1996; Lee, 1999). Other
factors can also play a role in driving anomalies in the zonal-mean
zonal wind in the upper equatorial troposphere, e.g. nonlinear
wave breaking associated with Rossby waves of extratropical
origin (Lee et al., 2007) or variations in the Hadley Cell (Hoskins
et al., 1999; Lee, 1999). For further discussion concerning the
influence of anomalies in the equatorial zonal-mean zonal wind
on the extratropics, readers are referred to the work of Gollan and
Greatbatch (2014).
Finally, we note that Garfinkel et al. (2012) report a link
between the MJO and stratospheric sudden warmings in which
the phase of the MJO can be related to the occurrence of sudden
warming up to one month ahead. However, it seems unlikely that
such a mechanism operated in 1962/1963, despite the sudden
warming that occurred at the end of January. The reason is
that the anomalies shown in figure 3 of their article have a
rather different, more polar-orientated, character compared with
those we have reported here and also have positive, rather than
negative, anomalies in 500 hPa height over the North Pacific
(compare figure 3 in their article with Figure 6 here).
The results presented here argue that an important ingredient
for successful prediction of seasonal means over Europe is
successful prediction of conditions in the Tropics during the
same season. They also argue that the impact of conditions
in the tropical Pacific over Europe can depend crucially on the
ambient atmospheric state, e.g the phase and strength of the QBO,
the zonal-mean zonal wind in the upper equatorial troposphere
(related to the MJO), atmospheric conditions influenced by
extratropical SST and sea-ice anomalies, which, as we have seen,
favoured (albeit weakly) the negative NAO in December 1962 (and
indeed throughout the winter 1962/1963), and other random,
unpredictable factors. In this respect, it is worth noting that,
while Jung et al. (2010c) noted an important influence from the
Tropics on the cold European winter of 2005/2006 (including a
role for the QBO), when looking at winters as a whole over the
period 1960/1961–2001/2002 Greatbatch et al. (2012) found only
a weak (albeit significant) influence from the Tropics on the NAO
on interannual time-scales.
5. Summary
We have used a set of relaxation experiments, following Jung
et al. (2010c), applied to the ECMWF atmospheric model to
c 2014 Royal Meteorological Society


analyze the severe European winter of 1962/1963. We argue that
the severe winter weather, which set in around Christmas time
in 1962, was associated with a wave train that emanated from
the tropical Pacific sector (where weak La Ni˜na conditions were
present) and was redirected towards Europe in association with
the zero wind line, which was, in turn, influenced by the combined
effect of the strong easterly phase of the QBO and easterly zonalmean zonal winds in the upper equatorial troposphere that
were in an extreme state that particular winter. A tendency
towards negative NAO conditions associated with extratropical
SST and sea-ice anomalies might have acted as a favourable
preconditioning over the North Atlantic/European sector. The
wave redirection (reflection) towards Europe culminated in the
stratospheric sudden warming at the end of January 1963 (Finger
and Teweles, 1964). We have argued that in February the sudden
warming event helped to maintain the negative NAO regime,
following the ideas of Baldwin and Dunkerton (2001), allowing
the severe weather to persist for a further month. A similar
situation occurred in 2013, when there was a stratospheric sudden
warming at the end of January. However, unlike in 2013, when
the subsequent negative NAO episode continued through to
April, in 1963 the severe weather came to an end in March with
the onset of a positive NAO regime in the troposphere, despite
the persistence of the weak vortex in the stratosphere. A role
of internal atmospheric variability in the dynamics of winter
1962/1963 is also noted.
Overall, our results argue that a successful seasonal prediction
system for Europe should include a good representation of the
stratosphere, as noted by Folland et al. (2012), as well as skill
in the Tropics. Indeed, although the (weak) La Ni˜na conditions
present in the tropical Pacific during the 1962/1963 winter would
not normally be associated with the negative NAO and severe
weather over Europe (Fraedrich and M¨uller, 1992; Ineson and
Scaife, 2009), we have seen that the source of the wave train that
ultimately led to the severe weather in 1962/1963 was most likely
the tropical Pacific sector. A possible player in the dynamics of
this wave train is the MJO (cf. Lin et al., 2009), a topic for future
investigation.
Acknowledgements
We are grateful to ECMWF for the provision of the model and
the use of computer facilities to carry out the model runs reported
here and to Soumia Serrar for technical assistance. During the
time when this work was carried out, GG was supported partly
by GEOMAR and partly by the DFG under ISOLAA, a project
within the Priority Programme 1158, and TK was supported by
GEOMAR. RJG is grateful to GEOMAR for continuing support.
We are also grateful to two anonymous reviewers for helpful
comments that led to a significantly improved manuscript. This
article grew out of a poster presented by the first author at the
meeting of the Royal Meteorological Society held in London
in March 2013 on long-range weather forecasting. The winter of
1962/1963 was particularly featured at this meeting. RJG benefited
from a discussion with Chris Folland at this meeting concerning
the dynamics of the 1962/1963 winter. We are also grateful to Lisa
Neef, who motivated our interest in the zonal-mean zonal wind
at the Equator.
References
Andrews DG, Holton JR, Leovy CB. 1987. Middle Atmosphere Dynamics.
Academic Press: San Diego, CA.
Baldwin MP, Dunkerton TJ. 1999. Propagation of the Arctic Oscillation from
the stratosphere to the troposphere. J. Geophys. Res. 104(:): 30937–30946,
doi: 10.1029/1999JD900445.
Baldwin MP, Dunkerton TJ. 2001. Stratospheric harbingers of anomalous
weather regimes. Science 294: 581, doi: 10.1126/science.1063315.
Bjerknes J. 1964. Atlantic air–sea interaction. Adv. Geophys. 20: 1–82.
Boer GJ, Hamilton K. 2008. QBO influence on extratropical predictive skill.
Clim. Dyn. 31: 987–1000, doi: 10.1007/s00382-008-0379-5.
Q. J. R. Meteorol. Soc. 141: 153–165 (2015)

Severe European Winter of 1962/1963
Brown P. 2006. The severe winter of 1962–63 over Great Britain: Synoptics
patterns and resulting weather. Int. J. Meteorol. 31: 63–72.
Burt S. 2013. ‘The bitter winter of 1962/63 in the British Isles’. In Royal
Meteorological Society Meeting, March 2013. London. http://www.rmets.org/
events/long-range-forecasting-problemmythology-science-and-progress/
bitter-winter-1962/63-british.
Cassou C. 2008. Intraseasonal interaction between the Madden–Julian
Oscillation and the North Atlantic Oscillation. Nature 455: 523–527, doi:
10.1038/nature07286.
Cayan DR. 1992. Latent and sensible heat flux anomalies mover the northern
oceans: Driving the sea surface temperature. J. Phys. Oceanogr. 22: 859–881,
doi: 10.1175/1520-0485(1992)0220859:lashfa2.0.CO;2.
Cohen JL, Furtado JC, Barlow MA, Alexeev VA, Cherry JE. 2012. Arctic
warming, increasing snow cover and widespread boreal cooling. Environ.
Res. Lett., doi: 10.1088/1748-9326/7/1/014007.
Dent J. 2013. The 1962–1963 winter as observed at Belstead Hall (Suffolk)
and through investigation of the synoptic charts. Weather 68: 18–23, doi:
10.1002/wea.2040.
Dunkerton T, Hsu CPF, McIntyre ME. 1981. Some Eulerian and Lagrangian
diagnostics from a model stratospheric warming. J. Atmos. Sci. 38: 819–843,
doi: 10.1175/1520-0469(1981)0380819:sealdf2.0.CO;2.
Finger FG, Teweles S. 1964. The mid-winter 1963 stratospheric warming
and circulation change. J. Appl. Meteorol. 3: 1–15, doi: 10.1175/15200450(1964)0030001:tmwswa2.0.CO;2.
Folland CK, Scaife AA, Lindesay J, Stephenson DB. 2012. How potentially
predictable is northern European winter climate a season ahead? Int. J.
Climatol. 32: 801–818, doi: 10.1002/joc.2314.
Fraedrich K, M¨uller K. 1992. Climate anomalies in Europe associated with
ENSO extremes. Int. J. Climatol. 12: 25–31, doi: 10.1002/joc.3370120104.
Garfinkel CI, Feldstein SB, Waugh DW, Yoo C, Lee S. 2012. Observed connection between stratospheric sudden warmings and the Madden–Julian
oscillation. Geophys. Res. Lett. 39: L18807, doi: 10.1029/2012GL053144.
Gollan G, Greatbatch RJ. 2014. On the extratropical influence of the equatorial
zonal-mean zonal wind in the upper troposphere during boreal winter. J.
Clim. (In press).
Greatbatch RJ. 2000. The North Atlantic Oscillation. Stoch. Env. Res. Risk
Assess. 14: 213–242, doi: 10.1007/s004770000047.
Greatbatch RJ, Lu J, Peterson KA. 2004. Nonstationary impact of ENSO
on Euro-Atlantic winter climate. Geophys. Res. Lett. 31: 4–7, doi:
10.1029/2003GL018542.
Greatbatch R, Gollan G, Jung T, Kunz T. 2012. Factors influencing northern
hemisphere winter mean atmospheric circulation anomalies during the
period 1960/61 to 2001/02. Q. J. R. Meteorol. Soc. 138: 1970–1982, doi:
10.1002/qj.1947.
Hoerling M, Kumar A, Zhong M. 1997. El Ni˜no, La Ni˜na, and the nonlinearity
of their teleconnections. J. Clim. 10: 1769–1786, doi: 10.1175/15200442(1997)0101769:enolna2.0.CO;2.
Holton JR, Tan HC. 1980. The influence of the equatorial Quasi-Biennial
Oscillation on the global circulation at 50 mb. J. Atmos. Sci. 37: 2200–2208,
doi: 10.1175/1520-0469(1980)0372200:tioteq2.0.CO;2.
Holton JR, Tan HC. 1982. The Quasi-Biennial Oscillation in the northern
hemisphere lower stratosphere. J. Meteorol. Soc. Jpn. 60: 140–148.
Hoskins BJ, Neale R, Rodwell M, Yang GY. 1999. Aspects of the large-scale
tropical atmospheric circulation. Tellus 51: 33–44, doi: 10.1034/j.16000889.1999.00004.x.
Hoskins B, Fonseca R, Blackburn M, Jung T. 2012. Relaxing the tropics to an
‘observed’ state: Analysis using a simple baroclinic model. Q. J. R. Meteorol.
Soc. 138: 1618–1626, doi: 10.1002/qj.1881.
Hurrell J, Kushnir Y, Ottersen G, Visbeck M. 2003. An overview of the North
Atlantic Oscillation. Geophys. Monogr. –Am. Geophys. Union 134: 1–36.
Ineson S, Scaife AA. 2009. The role of the stratosphere in the European climate
response to El Ni˜no. Nat. Geosci. 2: 32–36, doi: 10.1038/ngeo381.
Jung T. 2011. Diagnosing remote origins of forecast error: Relaxation versus
4D-Var data-assimilation experiments. Q. J. R. Meteorol. Soc. 137: 598–606,
doi: 10.1002/qj.781.
Jung T, Leutbecher M. 2007. Performance of the ECMWF forecasting system
in the Arctic during winter. Q. J. R. Meteorol. Soc. 133: 1327–1340, doi:
10.1002/qj.99.
Jung T, Balsamo G, Bechtold P, Beljaars ACM, Kohler M, Miller MJ, Morcrette
JJ, Orr A, Rodwell MJ, Tompkins AM. 2010a. The ECMWF model climate:

c 2014 Royal Meteorological Society


165

Recent progress through improved physical parametrizations. Q. J. R.
Meteorol. Soc. 136: 1145–1160, doi: 10.1002/qj.634.
Jung T, Miller MJ, Palmer TN. 2010b. Diagnosing the origin of
extended-range forecast errors. Mon. Weather Rev. 138: 2434–2446, doi:
10.1175/2010MWR3255.1.
Jung T, Palmer TN, Rodwell MJ, Serrar S. 2010c. Understanding the
anomalously cold European winter of 2005/06 using relaxation experiments.
Mon. Weather Rev. 138: 3157–3174, doi: 10.1175/2010MWR3258.1.
Killworth PD, McIntyre ME. 1985. Do Rossby wave critical layers
absorb, reflect or over-reflect? J. Fluid Mech. 161: 449–492, doi:
10.1017/S0022112085003019.
Kunz T, Greatbatch RJ. 2013. On the northern annular mode surface signal
associated with stratospheric variability. J. Atmos. Sci. 70: 2103–2118, doi:
10.1175/JAS-D-12-0158.1.
Lee S. 1999. Why are the climatological zonal winds easterly in the equatorial
upper troposphere. J. Atmos. Sci. 56: 1353–1363, doi: 10.1175/15200469(1999)0561353:watczw2.0.CO;2.
Lee S, Son SW, Grise K, Feldstein S. 2007. A mechanism for the poleward
propagation of zonal mean flow anomalies. J. Atmos. Sci. 64: 849–868, doi:
10.1175/JAS3861.1.
Lin H, Brunet G, Derome J. 2009. An observed connection between the
North Atlantic Oscillation and the Madden–Julian Oscillation. J. Clim. 22:
364–380, doi: 10.1175/2008JCLI2515.1.
Lin H, Brunet G, Mo R. 2010. Impact of the Madden–Julian Oscillation on
wintertime precipitation in Canada. Mon. Weather Rev. 138: 3822–3839,
doi: 10.1175/2010MWR3363.1.
Madden RA, Julian PR. 1994. Observations of the 40–50 day tropical
oscillation –a review. Mon. Weather Rev. 122: 814–837, doi: 10.1175/15200493(1994)1220814:ootdto2.0.CO;2.
Manley G. 1974. Central England temperatures: Monthly means 1659 to 1973.
Q. J. R. Meteorol. Soc. 100: 389–405, doi: 10.1002/qj.49710042511.
Moore GWK, Renfrew IA. 2011. Cold European winters: Interplay between
the NAO and the East Atlantic mode. Atmos. Sci. Lett. 13: 1–8, doi:
10.1002/asl.356.
Namias J. 1964. A 5-year experiment in the preparation of seasonal outlooks. Mon. Weather Rev. 92: 449–464, doi: 10.1175/15200493(1964)0920449:aeitpo2.3.CO;2.
Parker DE, Legg TP, Folland CK. 1992. A new daily central England
temperature series, 1772–1991. Int. J. Climatol. 12: 317–342, doi:
10.1002/joc.3370120402.
Petoukhov V, Semenov VA. 2010. A link between reduced Barents–Kara seaice and cold winter extremes over northern continents. J. Geophys. Res. 115:
D21111, doi: 10.1029/2009JD013568.
Ratcliffe RAS, Murray R. 1970. New lag associations between North
Atlantic sea temperatures and European pressure applied to longrange weather forecasting. Q. J. R. Meteorol. Soc. 96: 226–246, doi:
10.1002/qj.49709640806.
Rodwell MJ, Folland CK. 2002. Atlantic air–sea interaction and seasonal predictability. Q. J. R. Meteorol. Soc. 128: 1413–1443, doi:
10.1002/qj.200212858302.
Rodwell MJ, Rowell DP, Folland CK. 1999. Oceanic forcing of the wintertime
North Atlantic Oscillation and European climate. Nature 398: 320–323,
doi: 10.1038/18648.
Slingo JM, Sperber KR, Boyle JS, Ceron J-P, Dix M, Dugas B, Ebisuzaki W,
Fyfe J, Gregory D, Gueremy J-F, Hack J, Harzallah A, Inness P, Kitoh A,
Lau WK-M, McAvaney B, Madden R, Matthews A, Palmer TN, Park C-K,
Randall D, Renno N. 1996. Intraseasonal oscillations in 15 atmospheric
general circulation models: Results from an AMIP diagnostic subproject.
Clim. Dyn. 12: 325–357, doi: 10.1007/BF00231106.
Thompson DWJ, Wallace JM. 2000. Annular modes in the extratropical
circulation. Part I: Month-to-month variability. J. Clim. 13: 1000–1016,
doi: 10.1175/1520-0442(2000)0131000:amitec2.0.CO;2.
Untch A, Simmons A, Hortal M, Jakob C. 1998. Increased stratospheric
resolution in the ECMWF forecasting system. ECMWF Newsl. 82: 2–8.
Vitart F, Jung T. 2010. Impact of the Northern Hemisphere extratropics on
the skill in predicting the Madden Julian Oscillation. Geophys. Res. Lett. 37:
1–6, doi: 10.1029/2010GL045465.
Wallace JM, Zhang Y, Bajuk L. 1996. Interpretation of interdecadal trends in
Northern Hemisphere surface air temperature. J. Clim. 9: 249–259, doi:
10.1175/1520-0442(1996)0090249:ioitin2.0.CO;2.

Q. J. R. Meteorol. Soc. 141: 153–165 (2015)


Aperçu du document greatbatch2015.pdf - page 1/13
 
greatbatch2015.pdf - page 2/13
greatbatch2015.pdf - page 3/13
greatbatch2015.pdf - page 4/13
greatbatch2015.pdf - page 5/13
greatbatch2015.pdf - page 6/13
 




Télécharger le fichier (PDF)


greatbatch2015.pdf (PDF, 3.3 Mo)

Télécharger
Formats alternatifs: ZIP



Documents similaires


greatbatch2015
caesar2018nature 1
ahlstrom 2012 erl
walthers decembre 2018
walthers mai 2018 1
climate change 2013 the physical science basis

Sur le même sujet..