Content uploaded by Martin Wild
Author content
All content in this area was uploaded by Martin Wild
Content may be subject to copyright.
Available via license: CC BY 3.0
Content may be subject to copyright.
Atmos. Chem. Phys., 8, 6483–6498, 2008
www.atmos-chem-phys.net/8/6483/2008/
© Author(s) 2008. This work is distributed under
the Creative Commons Attribution 3.0 License.
Atmospheric
Chemistry
and Physics
Diurnal temperature range over Europe between 1950 and 2005
K. Makowski, M. Wild, and A. Ohmura
Institute for Atmospheric and Climate Science, ETH Zurich, Zurich, Switzerland
Received: 6 March 2008 – Published in Atmos. Chem. Phys. Discuss.: 9 April 2008
Revised: 13 June 2008 – Accepted: 30 September 2008 – Published: 13 November 2008
Abstract. It has been widely accepted that diurnal temper-
ature range (DTR) decreased on a global scale during the
second half of the twentieth century. Here we show however,
that the long-term trend of annual DTR has reversed from
a decrease to an increase during the 1970s in Western Eu-
rope and during the 1980s in Eastern Europe. The analysis
is based on the high-quality dataset of the European Climate
Assessment and Dataset Project, from which we selected ap-
proximately 200 stations covering the area bordered by Ice-
land, Algeria, Turkey and Russia for the period 1950 to 2005.
We investigate national and regional annual means as well
as the pan-European mean with respect to trends and rever-
sal periods. 17 of the 24 investigated regions including the
pan-European mean show a statistical significant increase of
DTR since 1990 at the latest. Of the remaining 7 regions, two
show a non-significant increase, three a significant decrease
and two no significant trend. Changes in DTR are affected by
both surface shortwave and longwave radiation, the former of
which has undergone a change from dimming to brightening
in the period considered. Consequently, we discuss the con-
nections between DTR, shortwave radiation and sulfur emis-
sions which are thought to be amongst the most important
factors influencing the incoming solar radiation through the
primary and secondary aerosol effect. We find reasonable
agreement between trends in SO2emissions, radiation and
DTR in areas affected by high pollution. Consequently, we
conclude that the trends in DTR could be mostly determined
by changes in emissions and the associated changes in in-
coming solar radiation.
Correspondence to: K. Makowski
(makowski@env.ethz.ch)
1 Introduction
Satellite and ground based measurements for Europe show
that the mean surface air temperature has increased overall
during the second half of the last century (Trenberth et al.,
2007). For the 1950s and 1960s, a characteristic phase of
roughly no increase or even decrease is apparent. Since the
late 1970s an accelerated increase in the mean temperature
was observed. The slow increase of the mean temperature
followed by a rapid increase is especially evident during the
summer months (Trenberth et al., 2007) where the incoming
shortwave radiation is one of the most dominant factors for
the daily temperature development. This leads to the hypoth-
esis that changes in the incoming solar flux at the surface had
a discernible influence on the mean temperature development
between 1950 and 2000 (Wild et al., 2007). Measurements of
shortwave radiation at the surface, from stations around the
globe, have shown that the incoming flux has significantly
decreased and subsequently increased in many of the inves-
tigated stations within the last 4 to 5 decades (Ohmura and
Lang 1989, Gilgen et al., 1998, Liepert and Kukla, 1997;
Stanhill and Cohen, 2001; Roderick and Farquhar, 2002;
Pinker et al., 2005; Wild et al., 2005).
The diurnal temperature range (DTR) is considered a suit-
able measure to investigate the counteracting effects of long-
wave and shortwave radiative forcing, because the diurnal
minimum is closely related to the longwave radiative flux,
while the diurnal maximum is predominantly determined by
shortwave radiation (Fig. 1a). It is known that the DTR
has been decreasing since the 1950s on a global scale due
to a strong increase of the diurnal minimum (Karl et al.,
1984, 1993; Kukla and Karl, 1993). Comparison of GCM
simulations with observations have shown that the DTR de-
crease has been underestimated due to a strong increase in
the modeled maximum temperature (Braganza et al., 2004).
The change in DTR has formerly been addressed mainly
as consequence of cloud cover development, precipitation,
Published by Copernicus Publications on behalf of the European Geosciences Union.
6484 K. Makowski et al.: Diurnal temperature range over Europe between 1950 and 2005
29
Figure 1. Sketch of mean diurnal temperature (T) cycle under a) weak anthropogenic
radiative influence, dominant radiative processes are denoted by arrows, b) enhanced
shortwave radiative cooling – “global dimming” (represented by the black arrows) and
long-wave radiative warming (represented by the grey arrows), and c) weakening
shortwave radiative cooling – “global brightening” (thinner black arrows) and continued
long-wave radiative warming.
(DTR: diurnal temperature range; T-MAX/-MIN: daily mean maximum/minimum; T-
MEAN: daily mean temperature, SW: shortwave, LW: longwave, ¨ T1/3: overall amount
of warming from state a) to c) or e.g. 1950 to 1990 respectively)
a) c) b)
Fig. 1. Sketch of mean diurnal temperature (T) cycle under (a) weak anthropogenic radiative influence, dominant radiative processes are
denoted by arrows, (b) enhanced shortwave radiative cooling – “global dimming” (represented by the black arrows) and long-wave radiative
warming (represented by the grey arrows), and (c) weakening shortwave radiative cooling – “global brightening” (thinner black arrows) and
continued long-wave radiative warming. (DTR: diurnal temperature range; T-MAX/-MIN: daily mean maximum/minimum; T-MEAN: daily
mean temperature, SW: shortwave, LW: longwave, 1T1/3: overall amount of warming from state (a) to (c) or e.g. 1950 to 1990 respectively).
change in irrigation and surface albedo or water vapor feed-
back (Stenchikov and Robock, 1995; Easterling et al., 1997;
Dai et al., 1997, 1999; Stone and Weaver, 2002; Vose et al.,
2005; Engelhart and Douglas, 2005). Many of the cited pub-
lications have concluded that neither of these factors alone is
likely to be the unique explanation of the observed changes in
DTR (Easterling et al., 1997). We argue that shortwave radia-
tion directly or via feedbacks is a major factor for the changes
in DTR since only the shortwave radiation – modulated by
the atmospheric aerosol burden – could exert a strong and
sufficiently homogeneous effect to change DTR on a global
scale (Liu et al., 2004; Wild et al., 2007).
The decrease of the solar flux and its relative cooling ef-
fect can been seen as a blocking action against the increase of
temperature caused by the greenhouse effect. Consequently
the diurnal maximum temperature remains constant while the
diurnal minimum is forced to increase (Fig. 1b). The re-
covery of surface solar radiation results in a removal of the
blocking on diurnal temperature development thus leading to
an increase of DTR and daily maximum respectively, thereby
revealing the full extent of global warming (Wild et al., 2007)
(Fig. 1c).
In the present study a detailed investigation of this issue
is conducted focusing on the European area where the best
coverage with observational data can be found.
2 Data and methods
We chose the data products of the European Climate Assess-
ment and Dataset Project (ECA&D-P) for an internally con-
sistent investigation of the DTR evolution during the recent
decades. It contains freely available data for more than 600
stations with minimum and maximum temperature measure-
ments in daily resolution for different periods between about
1800 and today (Klein Tank et al., 2002).
Because the change of incoming radiative flux at the sur-
face is considered very important to DTR development and
is measured since the 1950s, the complete second half of the
last century is investigated in this study. From the ECA&D-
P dataset, all stations with data for the period 1951 up to
2003 (or up to 2005 where available) were selected and na-
tional means were calculated. The time series of a station
was dismissed if it had more than five years with data gaps
or if two or more consecutive years were affected by these
gaps. In addition, each time series was checked for jumps
in the DTR. If jumps of more than half degree were caused
by filled data (from neighboring stations, performed during
ECA&D-P), then the value was replaced by an interpolated
value, on monthly basis, if the measurement from the original
site was available in the previous and following year. In total
less than 0.5% of the monthly values used in this study were
interpolated due to missing data. Systematic errors probably
due to data submission were found in all stations in Iceland,
Denmark and Romania. For Iceland all data after 1998, for
Denmark all data after 2002 and for Romania all data had to
be discarded and these were replaced by data which we ob-
tained directly from the respective national meteorological
service. For the region of the former Republic of Yugoslavia
(FRY) all data prior to 1956 had to be dismissed due to qual-
ity issues. For Poland, data are available only since 1966.
The temporal coverage was still considered sufficient to in-
vestigate decadal changes in DTR.
To obtain a sufficient number of stations for the calculation
of regional annual means (Fig. 2), station measurements were
grouped either nationally or by averaging over several small
nations. Netherlands, Luxembourg and Belgium were drawn
together as BeNeLux, likewise the states of the FRY; Estonia,
Lithuania and Latvia were grouped as Baltic States, Slovakia
and the Czech Republic to former Czechoslovakia (FCZS).
Conversely, Germany was divided into an eastern and a
western part according to the border line from the pre-1989
Atmos. Chem. Phys., 8, 6483–6498, 2008 www.atmos-chem-phys.net/8/6483/2008/
K. Makowski et al.: Diurnal temperature range over Europe between 1950 and 2005 6485
30
a c d b
e g h f
i k l j
q s t r
m o p n
u w x v
Fig. 2. Time series of annual mean DTR for each investigated region. All y-axes are scaled to 3 degrees Celsius for a better comparability.
Graphs are geographically arranged – except: surrounding regions of Europe as well as European mean are in the last row. The order of the
best suitable polynomial trend model according to Table 1 is indicated in brackets next to the name of the region and the investigated period.
Thick, grey, solid line presents 7 year running mean. The thick, black, dashed line shows the fitted trend model, if no black line is plotted
none of the models was significant above the 90% level.FCZS – former Czechoslovakia, FRY – Former Republic of Yugoslavia.
www.atmos-chem-phys.net/8/6483/2008/ Atmos. Chem. Phys., 8, 6483–6498, 2008
6486 K. Makowski et al.: Diurnal temperature range over Europe between 1950 and 2005
Fig. 3. Distribution of statistical significant, fitted DTR trend models. Blue – linear (all trends are negative), red – second order (all trends
show first a decrease, then an increase), green – third order (all trends show first an increase, then a decrease, then an increase) orange –
forth order polynomial. Numbers are the year of reversal from decrease to increase in the 7 year running mean, derived from the annual
mean DTR of region/country where denoted. The uncertainty of the actual year of reversal can be inferred from Fig. 4. The trend model is
not significant (>90%) if the numbers are in brackets, consequently the investigated region is also not color-coded (Spain, Eastern-Germany,
Benelux). Blue crosses represent stations investigated.
period. This is to take account of the different development
of atmospheric aerosol burden in the two countries, which
depends mainly on the industrial emissions within the range
of some tens to hundreds of kilometers upstream. The overall
resulting data coverage is indicated by the small blue crosses
shown in Fig. 3, identifying 189 (168 with coverage 1956–
2003) out of the original 604 ECA&D-P stations satisfying
the criteria described above.
Subsequently these national annual mean time series were
fitted by polynomials up to fourth order, to facilitate the char-
acterization and quantification of the DTR trend (Fig. 2).
The rational for fitting polynomial trend models was investi-
gated by applying multiple regression analysis and control of
lagged autocorrelation within the residuals to assure station-
ary white noise. The regression analysis, followed by calcu-
lation of statistical significance level (1 – P-values; given in
%) based on a standard T-test was performed for every co-
efficient for fits between first order (linear) and fourth order
polynomials as summarized in Table 1 (for further details see
reading example, Appendix A1).
For most regions R2increased together with the statistical
significance of the fitted model, hence making it easy to de-
cide which of the investigated models performs best. If the
comparison of R2and p-values (significance levels) showed
an ambiguous result (see e.g. Table 1, line 5, Denmark), the
residuals were checked in more detail and the model with
the lowest autocorrelations in the residuals was selected (not
shown). Note that only models with no significant autocor-
relations were accepted. The annual mean time series to-
gether with the fitted trend curve and the seven year running
mean trend for all investigated regions as well as the Euro-
pean mean are shown in Fig. 2.
Further information for each time series was obtained by
estimating the year of reversal from decreasing to increasing
DTR (applies not to regions with linear trends). The estima-
tion was performed by calculating the minimum in the seven
year running mean (Fig. 2) for the period 1965 to 1995. For
the example of Finland (Fig. 2c), the running mean given as
gray bold line shows a clear local minimum in 1989 (com-
pare year given in Fig. 3 and diamond at the row “Finland”
in Fig. 4). The particular period 1965 to 1995 was chosen
since it embraces the whole era of reversal from dimming to
brightening (Wild et al., 2005). The results are presented in
Fig. 3 and Fig. 4. The numbers displayed in Fig. 3 give the
year of the minimum of DTR. If they are printed in brackets
no significant trend could be estimated (compare Table 1).
In addition to the minimum DTR value between 1965 and
1995, all values within the lowest 10% of the difference be-
tween maximum and minimum (7 year running mean) value
within that period have been calculated to give additional in-
formation on the distinctness of the reversal. In Fig. 4 the
years below and equal to the 10th-percentile are marked with
dashes, diamonds show the minimum (for further details on
the method see detailed example, Appendix A2).
Atmos. Chem. Phys., 8, 6483–6498, 2008 www.atmos-chem-phys.net/8/6483/2008/
K. Makowski et al.: Diurnal temperature range over Europe between 1950 and 2005 6487
Table 1. Data for each investigated region, including overall Europe and each trend type. Order: Western Europe N–S; Eastern N–S,
surrounding regions, Europe; Columns from left to right: (1) Name of the region, (2) R2and significance codes for each coefficient (–: <90%,
o: 90%–95%, x: 95%–99%, xx: 99%–99.999%, xxx: >99.999%), columns (3) to (5) equal to column (2) but for higher order polynomial
fits, (6) number of stations for the mean calculations, and (7) data period (for more details see example in Appendix A1). R2’s in bold denote
the best suitable model according to R2, significance and residuals (not shown), which was subsequently used in Fig. 2. FCZS – former
Czechoslovakia, FRY – Former Republic of Yugoslavia.
1st 2nd 3rd 4th No. Period
Norway 0.15 xx 0.17 –
–0.26 o
x
x
0.32 x
x
x
o
4 51-05
Sweden 0.15 xx 0.15 –
–0.16 –
–
–
0.20 –
–
–
–
5 51-03
Finland 0.01 – 0.25 xxx
xxx 0.27 –
–
–
0.28 –
–
–
–
3 51-05
Great Britain <0.01 – 0.14 xx
xx 0.14 –
–
–
0.14 –
–
–
–
3 50-05
Denmark 0.06 o0.18 xx
x0.30 –
x
xx
0.33 –
–
–
–
5 50-03
East Germany (DDR) <0.01 –<0.01 –
–<0.01 –
–
–
<0.01 –
–
–
–
12 50-05
West Germany (BRD) <0.01 – 0.12 x
x0.12 –
–
–
0.12 –
–
–
–
13 51-05
BeNeLux <0.01 –0.02 –
–0.04 –
–
–
0.06 –
–
–
–
9 51-05
Alpine 0.01 – 0.24 xxx
xxx 0.25 –
–
–
0.25 –
–
–
–
2 50-05
France 0.02 – 0.11 x
x0.11 –
–
–
0.11 –
–
–
–
25 50-05
Italy 0.03 –0.42 xxx
xxx 0.52 xxx
xxx
xx
0.59 –
–
x
xx
4 51-03
Spain 0.02 –0.05 –
–0.05 –
–
–
0.08 –
–
–
–
9 51-05
Portugal 0.05 o0.05 –
–0.06 –
–
–
0.12 –
o
o
o
3 50-05
www.atmos-chem-phys.net/8/6483/2008/ Atmos. Chem. Phys., 8, 6483–6498, 2008
6488 K. Makowski et al.: Diurnal temperature range over Europe between 1950 and 2005
Table 1. Continued.
1st 2nd 3rd 4th No. Period
Russia & Belarus 0.02 –0.02 –
–0.22 xx
xxx
xxx
0.22 –
–
–
–
36 50-03
Baltic States 0.13 xx 0.15 –
–0.16 –
–
–
0.17 –
–
–
–
9 50-04
Poland 0.07 o 0.17 o
x0.17 –
–
–
0.17 –
–
–
–
2 66-05
FCZS 0.02 –0.03 –
–0.04 –
–
–
0.16 x
xx
xx
x
3 51-04
Ukraine 0.37 xxx 0.37 –
–0.37 –
–
–
0.37 –
–
–
–
9 51-05
FRY 0.08 x0.11 –
–0.13 –
–
–
0.26 x
x
xx
xx
4 56-04
Romania 0.01 –0.05 –
–0.13 x
x
o
0.16 –
–
–
–
19 61-05
Iceland 0.07 o0.07 –
–0.18 x
x
x
0.22 –
–
–
–
4 51-05
Algeria 0.17 xx 0.19 –
–0.37 xx
xxx
xxx
0.37 o
–
–
–
3 50-05
Turkey <0.01 –0.04 –
–0.13 –
o
x
0.13 –
–
–
–
3 50-04
Europe 0.01 – 0.14 x
x0.21 –
–
o
0.21 –
–
–
–
168 56-03
3 Results
In the following section we discuss annual means of the
DTR records, starting with the regional averages as described
above, followed by a description of the European mean. The
data records and polynomial fits determined in this analysis
are compiled in Fig. 2.
By the use of regional averages we aim to underline the
hypothesis that DTR is affected by changes in regional emis-
sions influencing shortwave radiation reaching the ground.
Detailed information for each country or region can be found
in Table 1 and Figs. 2–4. A complete description of all re-
gions (except the European mean, see below) shown in Fig. 2
is provided in Appendix B.
Atmos. Chem. Phys., 8, 6483–6498, 2008 www.atmos-chem-phys.net/8/6483/2008/
K. Makowski et al.: Diurnal temperature range over Europe between 1950 and 2005 6489
For most of Western Europe a distinct reversal from de-
creasing to increasing DTR is visible. The fitted polynomial
trends are significant in the Great Britain, Germany, Poland,
Finland, France, Italy and Switzerland/Austria. For Spain
and the Benelux an alike development of decrease and in-
crease in DTR can be seen from the running mean but the
fitted polynomial trends miss the 90% significance level.
Circumjacent countries, as Portugal, FRY, FCZS and Nor-
way show trends significant at the fourth order polyno-
mial with pronounced periods of increasing DTR in recent
decades. In North-Eastern Europe a region covering Sweden,
the Baltic States and the Ukraine, with a continued decrease
in DTR can be identified. All decreasing linear trends are
significant at the 99% level.
The countries located farther away from central Europe,
namely Russia, Belarus, Turkey, Algeria and Iceland show
trends which are best described by a third order polynomial.
All coefficients for all trends are significant at the 95% level
except for the first (not significant) and second coefficient
(90% level) of Turkey. In addition to the presently increasing
DTR a prominent feature in the annual mean time series of
the above mentioned countries is a second, earlier increase
in DTR between 1950 and 1960 which is addressed in more
detail in the discussion section.
Overall, the farther away the country is located from cen-
tral Europe the more recent is the time of reversal from de-
crease to increase of DTR. The earliest can be found in the
UK (1967) and Germany (1967), the latest in Iceland (1987),
Turkey (1990) and Russia (1992). From Fig. 4 extended pe-
riods of reversal can be seen in Romania, Norway and Den-
mark. For Romania, which is an outlier compared to the sur-
rounding nations, the early appearing of the lowest value in
the 7 year running mean is put more into perspective by the
“error bars” in Fig. 4, equally true for the late reversal in
Norway and the early one in Denmark.
For the European geographical mean between 1956 and
2003, 168 stations were used. To avoid biases, series shorter
than this period have been excluded. The European trend
is best described by a second order polynomial (Fig. 2x).
Both coefficients are significant at the 95% level. The rever-
sal from decrease to increase takes place in the early 1980s.
This overall character of the averaged European DTR is even
strengthened if shorter data series such as those from Roma-
nia and Poland were included (alternative mean not shown).
4 Discussion
The extent of the DTR is determined by many different fac-
tors, such as surface solar radiation or sunshine duration,
cloud cover connected with changes in large scale circula-
tion or aerosols, soil moisture and water vapor content of the
atmosphere.
Change in water vapor for example leads to an asymme-
try in the DTR (Stenchikov and Robock, 1995) by changing
65
70
75
80
85
90
Great Britain
East Germany
Romania
Italy
Poland
FRY
Alpine
FCZS
West Germany
Spain
Benelux
France
Norway
Denmark
Algeria
Iceland
Turkey
Finnland
Portugal
Russia
Europe
Fig. 4. Reversal of 7 year running mean DTR trends. Diamonds
represent the year of reversal of DTR as calculated from 7 year run-
ning mean trend. Dashed lines show additionally the period covered
by values within the lowest 10% of the amplitude of maximum DTR
minus minimum DTR (of seven year running mean values) for the
period 1965–1995. For more details see example in Appendix A2.
FCZS – former Czechoslovakia, FRY – Former Republic of Yu-
goslavia.
longwave and shortwave downwelling fluxes. A continued
increase in water vapor due to anthropogenic influence would
lead to a slightly reduced downwelling shortwave and in-
creased downwelling longwave radiation at the surface. This
would consequently lead to a continued reduction of DTR
which, however, we did not observe in the investigated area,
and therefore do not consider water vapor as a major factor
influencing DTR in Europe.
Soil moisture plays an important role by damping the DTR
as energy is consumed by evaporation during the daytime and
released by condensation during the nighttime. However, ac-
cording to Robock and Li (2006), long-term changes in soil
moisture are coupled to changes in solar radiation and tro-
pospheric air pollution respectively at least on regional scale
in Russia and the Ukraine, where long-term records of soil
moisture data are available.
For the inter-annual variability of DTR the total amount of
cloud cover as well as the cloud optical properties play an im-
portant role again by altering longwave and shortwave down-
welling fluxes (Karl et al., 1993). Clouds alter DTR mostly
by damping the daytime maximum via a strong reduction of
surface solar radiation, while the influence on the nighttime
minimum seems to be rather small (Dai et al., 1999). Apart
from local convection, long-term changes in cloud cover can
be connected to large scale circulation patterns and aerosols.
However, the correlation between DTR and cloud cover in
Europe for the period 1910 to 1990 is only 0.35 according to
Dai et al. (1997).
www.atmos-chem-phys.net/8/6483/2008/ Atmos. Chem. Phys., 8, 6483–6498, 2008
6490 K. Makowski et al.: Diurnal temperature range over Europe between 1950 and 2005
For the long-term influence of changes in large scale cir-
culation Sanchez-Lorenzo et al. (2008) show that in Western
Europe on a seasonal scale, circulation may have an influence
on the long-term development of sunshine duration, which
can be used as proxy for surface solar radiation (Stanhill and
Cohen, 2005). Still for the overall annual mean long-term
trend in sunshine duration, they identified changes in sur-
face solar radiation from anthropogenic aerosol emissions as
a more likely explanation.
To sum up, factors influencing surface solar radiation and
factors that are influenced by surface solar radiation seem
to account for the most of the changes in DTR in Europe.
Consequently, we consider changing surface solar radiation
as a major cause for the different types of DTR development.
Since solar radiation incident at the top of atmosphere
has not changed substantially during the investigated period
(Beer et al., 2000), two different candidates are likely to have
influenced surface shortwave radiation, namely clouds and
aerosols (first and secondary effect). Norris and Wild (2007)
showed that by removing the cloud cover influence from
surface solar radiation data, the reversal from dimming to
brightening becomes even more pronounced for most of Eu-
rope. Consequently cloud coverage changes acted as a dis-
guise rather than a cause for the variations in surface solar
radiation.
A much more likely candidate for the varying surface so-
lar radiation and DTR trend types and their time shifted
reversals are different patterns of emissions, leading to re-
gionally differentiated backscattering of solar radiation by
aerosols. A reduction of incoming radiation has been re-
ported by Liepert and Kukla (1997), Gilgen et al. (1998)
and Abakumova et al. (1996). Wild et al. (2005) reported
a reversal from global dimming to brightening in mid to late
1980s at widespread locations throughout the world. From
Abakumova et al. (1996), a reduction in incoming shortwave
radiation until at least 1990 is evident for the specific region
of Russia. These results indicate that changes in surface so-
lar radiation were found in many regions though they do not
have to be necessarily simultaneous. The global background
signal and forcing from aerosol as presented by Mishchenko
et al. (2007), showing a general decrease during the 1990s,
can be dominated by local influence as described by Alpert
et al. (2005). Publications from Stern (2006) and Lefohn et
al. (1996) assume that a reversal from increase to decrease
of Eastern European emissions (dominated by Russia) takes
place in the late 1980s or early 1990s. In contrast, West-
ern European emissions are peaking already during the early
1970s according to Smith et al. (2004), Streets et al. (2006)
and Stern (2006). This is confirmed in Mylona (1996) and
Vestreng et al. (2007) who have shown that the maximum in
SO2emissions from fossil fuel for early industrialized coun-
tries, such as the UK or the former Federal Republic of Ger-
many, can be as early as the second half of the 1960s.
4.1 DTR, radiation and emissions – the biggest European
emitters
In the following section we discuss the qualitative connection
between trends in SO2emissions, sunshine duration, radia-
tion and DTR in several examples. We included the annual
sunshine duration in this section as a proxy for surface solar
radiation (Stanhill and Cohen, 2005), which is only available
for a sufficient number of sites since the 1960s. Sanchez-
Lorenzo et al. (2007 and 2008) provide data for sunshine
duration in Western Europe between 1938 and 2004. SO2
emissions at land level are available from Mylona (1996) and
Vestreng et al. (2007). Vestreng et al. (2007) is partially a
succeeding work of Mylona (1996) and both provide SO2
emission estimates for every fifth year. Therefore, we com-
bined the estimates to one time series if they cover the same
source region. Data from Mylona (1996) are available be-
tween 1880 and 1990, Vestreng et al. (2007) provide data be-
tween 1980 and 2004. For the overlapping period between
1980 and 1990 the data from Vestreng et al. (2007) were
favored for the following analysis. Further SO2estimates
can be derived from Lefohn et al. (1996). They provide an-
nual estimates of sulfur emissions between 1850 and 1990.
Obviously not all sulfurous emissions will be in the form of
SO2but still we converted all sulfur into SO2for the gain of
a much simpler comparison against the Mylona (1996) and
Vestreng et al. (2007) data. Valuable information on trans-
boundary fluxes and trend development of SO2since 1980
can be obtained from Klein and Benedictow (2006). Data
from long-term surface solar radiation measurement can be
found in Ohmura (2006). A recently submitted paper from
Gilgen et al. (2008) provides additional information on rever-
sal years and overall trends from gridded data of the Global
Energy Balance Archive. To make the characteristics of the
DTR time series more easily comparable against the rather
slow changes of emissions and the low pass filter data of sun-
shine duration, we calculated the eleven year running mean
of the annual DTR time series. The miss weighting at the
edges was indicated by using dashed lines (Figs. 5 and 6).
The emissions of SO2were plotted upside down to indicate
the potential sulfate aerosol forcing on DTR.
According to Berge et al. (1999) the 10 biggest emitters
(in total t/a) in Europe are: Bulgaria, France, Germany, Italy,
Poland, Spain, Great Britain, Ukraine, Russia and the Czech
Republic (or FCZS). Except for Bulgaria and Italy all coun-
tries above were analyzed with respect to their DTR trends
and will be discussed subsequently in more detail beginning
with the Eastern European countries, from south to north.
The best fitting trend for the Ukrainian DTR time series
is a linear trend with a slope of −0.014◦C/a, emphasizing a
continuous decrease (Fig. 2l). A more detailed inspection of
the running mean in Fig. 2l and Fig. 5a reveals, however, a
tendency towards an increase around 1978 which is reverses
to a continued decrease from 1987 onward. Data for emis-
sions are available from 1980 (Mylona 1996 and Lefohn et
Atmos. Chem. Phys., 8, 6483–6498, 2008 www.atmos-chem-phys.net/8/6483/2008/
K. Makowski et al.: Diurnal temperature range over Europe between 1950 and 2005 6491
Figure 5. Sketch of mean diurnal temperature (T) cycle under a) weak anthropogenic
radiative influence, dominant radiative processes are denoted by arrows, b) enhanced
shortwave radiative cooling – “global dimming” (represented by the black arrows) and
long-wave radiative warming (represented by the grey arrows), and c) weakening
shortwave radiative cooling – “global brightening” (thinner black arrows) and continued
long-wave radiative warming.
a)
d)
b)
c)
Fig. 5. Time series of annual mean DTR and SO2emissions. Eleven-year running means of DTR, expressed as relative (rel.) deviations from
the 1971–2000 mean, are plotted as solid black line. Differently weighted first and last five years of the time series are denoted as dashed
black lines. SO2emissions from Mylona (1996) (2) and Vestreng et al. (2007) (2) were plotted up side down, to indicate the presumed
forcing. Estimated sulfur emissions from Lefohn et al. (1996) (3) were converted to SO2 equivalent and also plotted upside down. All SO2
estimates are expressed in megatons per year.
al., 1996 estimated their values for the Union of Soviet So-
cialist Republics – USSR), showing a distinct decrease since
1990 which is not reflected in the DTR data. However, the
previously described short increase and decrease of DTR is
reflected in a short decreasing and increasing period of SO2
emissions. Surface solar radiation measurements for Odessa
show a decrease from 1960 until 1987 (end of data). The
described short increase in DTR between 1978 and 1987 is
mirrored as well in the radiation time series plot (Abaku-
mova et al., 1996 (Fig. 4), most evident between 1977 and
1983. To summarize, the continued decrease in DTR since
1980 cannot be explained by a continued increase of national
emissions. However, the findings are not contradictory with
respect to the connection between DTR and radiation (for
further details on the Ukraine, as an example for linear de-
creasing trends, see Sect. 4.3).
For the FCZS, data of emissions are available for the
whole period from Mylona (1996) and Vestreng et al. (2007).
The highest values occur around 1980 in line with the re-
versal of DTR from decrease to increase which is calculated
around 1977 (Fig. 4 and Fig. 5b). In Gilgen et al. (2008) the
reversal of surface solar radiation from dimming to brighten-
ing is estimated between 1978 and 1983.
Daily maximum and minimum temperature for Poland is
available since 1966. The reversal from decrease to increase
is calculated at 1977 in the seven year running mean. The
running mean of the Polish time series (Fig. 2g and Fig 5c)
shows a short increasing and subsequently decreasing period
between 1975 and 1986. The reversal of the second order
polynomial fit to DTR which omits the described hump is
between 1980 and 1985, this is close to the peaking of emis-
sions in 1985. Consequently, both reversals are consistent
with the reversal in incoming shortwave radiation in 1980.
Russian emissions peak at 1975 according to Vestreng
et al. (2007). Other emission estimates as e.g. from Stern
(2006) and Lefohn et al. (1996) suggest that the decrease
of emissions started much later namely in the late 1980s
(Fig. 5d) with the breakdown of the former Soviet Union.
However, the significant decrease of emissions after the col-
lapse of the Eastern Bloc is reflected in all cited emis-
sion estimates. The DTR decrease for Russia lasts until
1990 consistent with the decrease in surface solar radiation
www.atmos-chem-phys.net/8/6483/2008/ Atmos. Chem. Phys., 8, 6483–6498, 2008
6492 K. Makowski et al.: Diurnal temperature range over Europe between 1950 and 2005
Figure 6. Sketch of mean diurnal temperature (T) cycle under a) weak anthropogenic
radiative influence, dominant radiative processes are denoted by arrows, b) enhanced
shortwave radiative cooling – “global dimming” (represented by the black arrows) and
long-wave radiative warming (represented by the grey arrows), and c) weakening
shortwave radiative cooling – “global brightening” (thinner black arrows) and continued
long-wave radiative warming.
a)
d)
b)
c)
Fig. 6. As Fig. 5. In addition, sunshine duration series from Sanchez-Lorenzo et al. (2008) (4) and Sanchez-Lorenzo et al. (2007) (5) are
plotted as solid grey line. Plotted data are low-pass filtered values (11-year window 3 year σGaussian low-pass filter) expresses as relative
deviations from the 1961–1990 mean.
(Abakumova et al., 1996) measured at Moscow. Likewise,
the observed increase of surface solar radiation at Moscow
(Wild et al., 2005) is mirrored in an increasing DTR. Both
are potentially caused by the strong decrease of SO2emis-
sions reported from various estimates.
In Western Europe the biggest emitters during 1985 and
1995 are France, (West) Germany, Great Britain and Spain
(Berge et al., 1999). The DTR for all four regions is best
described with a second order polynomial trend, significant
above 95% except for Spain with p-values of 0.35, 1st coef-
ficient) and 0.21 (2nd coefficient) (not shown). For the com-
parison of DTR against SO2emissions and sunshine dura-
tion we used in addition the low-pass filter time-series from
Sanchez-Lorenzo et al. (2007) for Iberia to compare it to our
DTR data of Spain. In Sanchez-Lorenzo et al. (2008) the
same authors provided annual means of sunshine duration for
most of Western Europe, split into six regions. Here we also
used the low pass filtered data of sunshine duration for the
regions NC (north central), CW (central west) and CE (cen-
tral east). We compared NC to Great Britain, CW to France
and CE to West Germany according to their spatial coverage
given in Fig. 1 of Sanchez-Lorenzo et al. (2008).
For Great Britain a reversal in DTR is apparent around
1965 simultaneously to the emissions from Mylona (1996)
of SO2which are peaking in 1965, according to Lefohn et
al. (1996) emissions reach their maximum already around
1960. The annual sunshine duration from Sanchez-Lorenzo
et al. (2008), which covers the central to south-eastern part of
the UK, shows a reversal from decrease to increase in the late
1960s. Subsequently, sunshine duration increases along with
DTR and the decreasing emissions until the present (Fig. 6a).
In the former Federal Republic of Germany, the DTR re-
versal is calculated at 1967 by the 7-year-running mean.
Reversal of SO2emissions is in 1965 according to My-
lona (1996). The most dominant increase of DTR and de-
crease of SO2respectively, however begins during the 1980s
which is in line with the end of the decrease in surface solar
radiation in Germany (Liepert and Kukla, 1997 (Fig. 2) and
a strong increase in sunshine duration (Fig. 6b). Notably the
horizontal visibility increased already since the second half
of the 1960s in most of the Western German stations investi-
gated by Liepert and Kukla (1997). These results point to a
decrease of turbidity and thus a reduction of aerosol burden
of the troposphere.
In France the DTR reversal is calculated at 1980 while SO2
emissions estimated by Mylona (1996) start to decrease from
Atmos. Chem. Phys., 8, 6483–6498, 2008 www.atmos-chem-phys.net/8/6483/2008/
K. Makowski et al.: Diurnal temperature range over Europe between 1950 and 2005 6493
1975, Lefohn et al. (1996) in contrast determine the highest
value of emissions not until 1980 (Fig. 6c). The reversal in
shortwave incoming radiation is between 1980 and 1986 ac-
cording to Gilgen et al. (2008). Sunshine duration reverses
between 1980 and 1985. The maximum lag of about one
decade between radiation and SO2emission (form Mylona
1996) reversals can be explained by the method used to de-
termine the year of reversal. Gilgen et al. (2008) used a sec-
ond order polynomial fit to define the period of reversal. The
reversal in the DTR retrieved from the fitted second order
polynomial would be similar, namely around 1980 to 1982
(Fig. 2s).
Spain as the most southern representative of the largest
emitters in Europe has reduced its emissions remarkably
since 1980 (Mylona, 1996). The reversal of DTR derived
from the 7 year-running mean is at 1977 (Fig. 2r and Fig. 4).
According to Sanchez-Lorenzo et al. (2007) the reversal for
sunshine duration for the whole Iberia Peninsula is in 1982
most evident during spring and summer, with mostly clear
sky situations. The period where all three independently in-
vestigated measures (namely emissions, DTR and sunshine
duration) show a reversal in their long-term behavior lies
consequently within 5 years (Fig. 6d).
To summarize on the section with the biggest emitters in
Europe we would like to point to the good qualitative agree-
ment between the low frequency trends in DTR and SO2
emission estimates in six (four in Western, two in Eastern
Europe) of the eight regions analyzed.
The long-term evolution of sunshine duration from
Sanchez-Lorenzo et al. (2007 and 2008) supports our con-
clusions for Western Europe. The biggest disagreement be-
tween DTR and SO2emissions can be found for Eastern Eu-
rope, namely in Russia and the Ukraine. However, trends
in surface solar radiation from Abakumova et al. (1996) for
these regions support our results for long-term DTR devel-
opment. To conclude that though we found good agreement,
more work involving chemistry climate models with an ap-
propriate input and transport of pollutants is required in order
to improve our understanding of the relation between DTR
and air pollution in the future.
4.2 Long range effects on DTR and radiation
According to the previous section a further feature which has
to be discussed is the inconsistency in the DTR reversal com-
pared to the reversal in emissions in a number of regions such
as Finland. The running-mean curve for the DTR in Finland
(Fig. 2c) shows a reversal in the early 1990s, in line with
the surface solar radiation measurements, taken in Sodankyla
in the north of Finland (Ohmura, 2006, Fig. 9). Accord-
ing to Gilgen et al. (2008) 1990 is the year of reversal from
dimming to brightening, for the mean of seven stations in
southern Finland. Emissions, however, peaked around 1975
(Vestreng et al., 2007). It can be seen from the EMEP (Co-
operative program for monitoring and evaluation of the long-
range transmission of air pollutants in Europe) Report 1/06
(Klein and Benedictow, 2006) for Finland that for 2004 about
80% of the oxidized sulphur deposition originates from out-
side Finland. Biggest contributor is Russia with as much as
23% for overall Finland. This implies that the influence on
DTR especially for stations in the North and East of Finland
is likely to be dominated by Russian emissions, thus giving
a possible explanation for the reversal in DTR and surface
solar radiation as late as 1989 (Fig. 4).
Similar to Finland, other countries in Northern Europe,
such as Sweden, Norway, Iceland, Latvia, Lithonia and Den-
mark contribute no more than 10% to their total of oxi-
dized sulphur deposition, leaving these regions as dependent
on neighboring countries such as Great Britain, Germany,
Poland, Estonia, Ukraine and Russia and their patterns in
matters of emissions (Klein and Benedictow, 2006).
4.3 Linear downward trends of DTR
Another interesting feature is the linear downward trend of
the DTR in Sweden, the Baltic States and Ukraine. It is note-
worthy that they seem to build a north-west, south-east ori-
entated zone between Eastern and Western Europe (Fig. 3).
The linear decreasing DTR trend is not explainable by the na-
tional emission trends of the corresponding regions, since the
emissions for all above mentioned countries have declined at
least since 1990. Surface solar radiation for southern Swe-
den and the Baltic States started to increase in the late 1980s
(Ohmura, 2006; Gilgen et al., 2008), subsequently DTR in
both regions levels-off or increases slightly as well. How-
ever, during the 1990s DTR stopped increasing which re-
sulted in a significant decreasing linear trend for the whole
period. The continuous decrease of DTR in Ukraine cannot
be explained by a continued increase of emissions. Also, no
radiation data is available for further interpretation. Support
for the findings on an overall decreasing DTR can be found
from soil moisture measurements. Robock and Li (2006)
have shown that between 1958 until the mid 1990s soil mois-
ture increased significantly for the Ukraine. They state that
precipitation and temperature alone could not have caused
this development. Using a land surface model they show
that a reduction in downward shortwave radiation could have
caused the observed increase in soil moisture, which is in line
with the DTR decrease noted above.
4.4 Early increase in DTR and radiation
The final feature we want to discuss in detail is the early in-
crease of DTR during the 1950s and 1960s, visible in dif-
ferent regions all over Europe but mainly in the northern,
eastern and the periphery regions, namely Norway, Russia,
the Baltic States, FCZS, FRY as well as Iceland, Algeria
and Turkey. The early increase visible from the different
DTR time series might be due to an “earlier brightening”
during this period. The hypothesis of an earlier brightening
www.atmos-chem-phys.net/8/6483/2008/ Atmos. Chem. Phys., 8, 6483–6498, 2008
6494 K. Makowski et al.: Diurnal temperature range over Europe between 1950 and 2005
in Eastern Europe during the 1950s and 1960s is again sup-
ported by soil moisture measurement. Figure 3 of Robock
and Li (2006) suggests that soil moisture in Russia decreased
slightly between 1958 and 1970, indicating an increase in ra-
diation.
No radiation data is available from the above mentioned
regions prior to 1960. Still we can assume a similar techno-
logical development for Eastern Europe as for Western Eu-
rope but with a time shift of one to two decades as indicated
by a later reversal in emissions and radiation during the pe-
riod since 1960.
By investigating radiation and radiation related measure-
ments for Western Europe prior to 1950 we can consequently
find potential explanations for the observed early increase of
the DTR. Evidence for an early brightening period (increase
of incoming shortwave radiation) in Western Europe is pre-
sented in Ohmura (2006). The surface solar radiation data in
Fig. 1 of Ohmura (2006) for Wageningen, Stockholm, Davos
and Potsdam increases until the 1950s. For Wageningen
this is supported from De Bruin et al. (1995). In Sanchez-
Lorenzo et al. (2008) sunshine duration for Western Europe
increases from 1938 (earliest value of plot) until 1950.
5 Conclusions
We investigated annual mean DTR for the period 1950 until
2005 for 23 different countries and regions in and around Eu-
rope as well as Europe as a whole. A total of 16 out of these
23 regions as well as the European mean show a statistically
significant period of decrease and a subsequent increase in
DTR. Two additional regions (BeNeLux, Spain) show an in-
crease, which however is not statistically significant in the
multiple regression analysis. Of the remaining five regions,
two (East Germany, Portugal) show no specific trend and
three (Sweden, Baltic States, Ukraine) regions show a con-
tinuation of the decreasing trend. The trend analysis is lim-
ited by the lack of a standard homogeneity procedure and by
the limited number of available measurement sites and their
spatial distribution.
The connection between DTR, shortwave radiation and
SO2emissions has been qualitatively discussed with respect
to a common trend reversal. The period of reversal of DTR
from decrease to increase is in most cases in line with social
and economic development as indicated by SO2-emissions or
deposition, respectively. All reversals of DTR were shown
to take place between 1965 and 1990. This is consistent
with the change from decrease to increase of incoming short-
wave radiation (“Global Dimming” to “Global Brighten-
ing”). Consequently, we conclude that the long-term trends
in DTR are strongly affected by changes in incoming short-
wave radiation, presumably largely influenced by direct and
indirect effects of aerosol from sulphurous emissions.
This may suggest that in more regions around the globe
DTR will increase if the surface solar radiation continues to
increase on a widespread basis.
Appendix A
A1 Illustrative example how to read Table 1
The example of Denmark, (Table 1, line 5) reads as follows:
the first column contains the name of the region. The second
column contains the R2between the time series and the best
fitted trend of the form:
f (x) =f1∗x+f o (A1)
Following the R2a small “o” indicates that the linear coef-
ficient is statistically significant above the 90% level or in
more common words: it is 90% likely that the linear coeffi-
cient cannot be zero if the time series should be represented
by the given equation. Column three contains again the value
for the R2. However, now the comparison is performed be-
tween the time series of annual mean DTR of Denmark and
the best fit of the type:
f (x) =f2∗x2+f1∗x+f o (A2)
Following this R2two lines of coding symbols contain the
information that the linear coefficient (f1)is now 99% sig-
nificant (two “x”) and the quadratic trend is different from
zero at the 95% significance level (one “x”). For the third or-
der polynomial, shown in the fourth column, the R2increases
again to now 0.3. The three lines of symbols following the
R2indicate that the cubic coefficient is now significant at the
99% level, the quadratic at the 95% level but the linear co-
efficient misses the 90% level and is marked consequently
with a small “–”. In the 5th column the R2increases to 0.33
thus explaining already 33% of the given annual mean time
series. However the polynomial of the form,
f (x) =f4∗x4+f3∗x3+f2∗x2+f1∗x+f o (A3)
overestimates for the given time period.
A2 Explanatory example Fig. 4
The method underlying Fig. 4 can be illustrated comparing
Figs. 2f (Denmark) and 2o. The highest value for the seven
year running mean during the given period for FRY appears
in 1991 with 9.8◦C, the lowest is 8.98◦in 1977. The absolute
difference is 0.82◦; 10% of 0.82◦is 0.082◦. Consequently all
years with a seven year running mean value of the time series
of FRY within the range of 8.98◦and 9.06◦have been marked
with a dash in Fig. 4, line 6 (1975, 1976, 1978). These dashes
consequently give a sort of error bar for the calculated year
of reversal. For the reversal in the annual mean time series of
Denmark, a much bigger uncertainty range is given, namely
between 1981 and 1987. The highest value in the period 1965
Atmos. Chem. Phys., 8, 6483–6498, 2008 www.atmos-chem-phys.net/8/6483/2008/
K. Makowski et al.: Diurnal temperature range over Europe between 1950 and 2005 6495
to 1995 of the seven year running mean of the annual means
from Denmark is 5.75◦(1970) the lowest is 5.21◦(1981). So
the difference between the two extremes is 0.54◦which is
only two thirds of FRY difference. This fact gives credit to
the different overall variability of the investigated time series.
After adding 10% of 0.54◦to the minimum of 5.21◦all years
within the range of 5.21◦to 5.26◦(1982–1987) are marked
with a dash (Fig. 4, line 14).
Appendix B
Addititonal, detailed information on the
regional annual means
B1 Western Europe
Norway. The mean DTR, predominantly governed by
stations around 60◦N, shows an increase during the 1950s
followed by a significant decrease until the late 1980s.
Starting in 1987 DTR increases, but is then interrupted by a
dip around 2000 (Fig. 2a). This dip reduces the significance
of the fitted polynomial, still the third and forth order
polynomials are significant above 90%.
Sweden. The averaged time series for Sweden shows
a highly significant negative linear trend (Fig. 2b). The
selected stations are all located south of 64◦N, representing
southern Sweden (Fig. 3). The DTR appears to level off
since the late 1980s. However, when reducing the Swedish
data to cover only stations for the period until 2005, a
tendency to an increase became apparent, this trend was
not significant in any model. Also the selection would have
given even more weight to the most southern part of Sweden.
Finland. The data for Finland consists of three stations,
evenly divided from north to south, namely Helsinki,
Jyvaskyla and Sodankyla. The national trend is best repre-
sented by a second order polynomial trend significant above
the 99% level (Fig. 2c).
Denmark. Though one of the smaller countries, Denmark
contributes 5 equally distributed stations to the dataset
(Fig. 3). The best fitting trend model is the third order
polynomial (Fig. 2f). The second and third order coefficients
are significant at 95% level, whereas the linear term shows
only p-value of 0.18 corresponding to approximately 80%
confidence level.
Great Britain. Only three stations met the demanded
quality requirements of temporal coverage up to 2003,
namely Oxford, Wick and Waddington. The former two are
located in the industrialized southern area of the UK and
show a distinct DTR reversal from decrease to increase.
Wick is situated at the northern tip of the British mainland
showing a general decrease. Despite this the fitted second
order polynomial is significant at the 95% level, indicating a
trend from decrease to increase (Fig. 2e). An early reversal
around 1965 is visible from the 7-year running-mean.
BeNeLux. Belgium and Luxembourg each contribute only
one station to the selected dataset, hence they were analyzed
together with the seven stations from the Netherlands. The
analysis of the BeNeLux region showed no significant trend.
The best fit however is a second order polynomial with
p-values around 0.23 (Fig. 2m). The seven year running
mean trend shows an overall increase since 1980.
East Germany. No significant trend is apparent. Best
fit is the second order polynomial (Fig. 2j). P-values are
in general above 0.7 (confidence level, below 30%) in all
models and coefficients.
West Germany. For the mean of the 13 stations a distinct
reversal from decrease to increase is visible in the national
mean time series. Consequently the second order polyno-
mial trend is significant at the 95% level in both coefficients
(Fig. 2i).
France. The 25 selected stations are distributed equally
over France (Fig. 3). Similar to West Germany the second
order polynomial is significant at the 95% level emphasizing
the DTR development form decrease to increase with the
reversal period between 1965 and 1985.
Alpine Region. There are only two stations, one from
Austria and one from Switzerland. Most Swiss stations had
to be rejected due to homogeneity issues. Problems were
caused by change of location and instrumentation. The only
Swiss station that met the quality requirements is Basel-
Binningen. For Austria only one station (Kremsmuenster)
with complete data coverage from 1950 to 2005 is provided
in ECA&D-P. The mean trend derived from the two stations
is best described by a polynomial of the second order
(Fig. 2n). The main contribution to this shape is given by
the Basel-Binningen station which shows a distinct decrease
and increase.
Italy. The mean trend for Italy is calculated from four sta-
tions. The best fitting polynomial is second order (Fig. 2t).
Overall a strong decrease and subsequent increase is visible.
Spain. For Spain a slight decrease in the seven year running
mean up to 1977 is visible. Thereafter an equally slightly
visible increase in DTR can be seen (Fig. 2r). However, the
statistical analysis shows no significant trend on the 90%
confidence level. The p-values for Spain are, 0.36 for the
linear and 0.22 for the cubic coefficient.
www.atmos-chem-phys.net/8/6483/2008/ Atmos. Chem. Phys., 8, 6483–6498, 2008
6496 K. Makowski et al.: Diurnal temperature range over Europe between 1950 and 2005
Portugal. A total of three stations are sufficient according
to the demanded quality requirements .Braganca shows an
overall increase for the whole period while Lisboa (Lis-
bon) and Porto show a continuous decrease in DTR (Fig. 2q).
B2 Eastern Europe
Former Republic of Yugoslavia. The reliable period for
the FRY region is from 1956 to 2004. The best fitting trend
model for that time series is a fourth order polynomial trend
with p-values below 0.05 for all coefficients. Consequently
the development shows more than a single period of decrease
and increase. From 1956 to about 1965 the DTR increases
this is followed by a distinct decrease up to around 1980.
From 1980 until 1990 a second and more emphasized
increase is dominant with a subsequent phase of more or
less constant development until 2004 (Fig. 2o). However,
the most pronounced feature in this period is the decrease
and then subsequent increase in DTR from 1965 to 1991.
Romania. The best fit is a third order polynomial with
p-values of about 0.03 for the first and second coefficient,
the p-value for the third coefficient is slightly higher with
0.051 and therefore misses the 95% confidence boundary.
From the seven year running mean a period with a distinct
decrease from 1961 to 1971 is visible, followed by a longer
period of increasing DTR lasting until 1990, subsequently
the running mean shows a constant development (Fig. 2p).
Czechoslovakia. As for the area of the FRY, the former
Czechoslovakian states are best fitted by a fourth order
polynomial (Fig. 2k). All tested coefficient of the fourth
order polynomial are above the 95% significance level
(p-values <0.05). The decreasing period lasts until 1977
according to the seven year running mean, then the DTR
increases until it stops around 1992. This is followed by a
stable to slightly decreasing period until 2004.
Poland. Only two stations with data from 1966 to 2005 are
available, Leba and Siedlce these both show very similar
long-term trends. The best fit is a second order polynomial.
p-values for the coefficients are 0.083 and 0.046. For Poland
a decreasing period is visible from 1966 to 1980 and an
increasing period from 1986 to 2005 (Fig. 2g).
Baltic Region. A strong increase is visible in the seven year
running mean up to 1966, followed by a decrease, leveling
off in 1991. A short increase starting in 1990 come to an end
by 1996 and then becomes a continued decrease (Fig. 2h).
The result of this is an overall linear decrease in the fitted
trend model significant at 95% level.
Ukraine. The nearly monotonic drop of the Ukrainian
mean DTR lasts over the whole period from 1951 to 2005
(Fig. 2l). The linear trend is significant above 99%. The
two most westerly located stations, L’Vov (Lwiw/Lemberg)
and Uzhgorod show a dominant increase since the middle of
the 1970s. For the two biggest cities of Ukraine, Khrakov
and Kyiv (Kiev) a decrease in DTR until the mid-1990s is
dominant follow by a leveling off or increase thereafter.
Russia. The largest region of the so called Eastern-European
section is the European part of Russia with Brest (Brestzon-
alnaya) as only representative station for Belarus included.
The mean DTR development for the overall 36 stations is
best described by a third order polynomial. p-values are
around 0.001 the R2is 0.22. Assuming that none of the
high frequency is caught by the polynomial this is a quite
high value. The seven year running mean describes an in-
crease until 1966 followed by a continuous decrease until
1992 and thereafter an equally uninterrupted increase until
2003 (Fig. 2d).
B3 Surrounding regions
Iceland. The best fit for the DTR time series is the third
order polynomial (Fig. 2u). All coefficients are above
95% significant. Equal to Denmark and Finland, Iceland
is considered to be a mixture of the Western and Eastern
European trend type.
Turkey. The shape of the mean data series of the three
stations is best described by a third order polynomial. The
seven year running mean describes a distinct increase from
1950 until 1963, then a subsequent decrease is disturbed by
a short period of increase between 1974 and 1984, thereafter
the long-term decrease is continued until 1990. Finally
an increase until 2004 is visible from the smoothed 7 year
running-mean curve. The described interruption causes a
reduction in the significance of the trend model, p-values are
0.150 (1st), 0.053 (2nd) and 0.027 (3rd). When smoothing
the described period the p-values are: 0.03 (2nd) and 0.009
(3rd). The linear coefficient never becomes statistically
significant since there is no overall decrease or increase in
the series.
Algeria. Three stations are available which are distributed
roughly evenly along a north south transect (Fig. 3). All sta-
tions, namely Alger-Dar el Beida, El Golea and Tamanras-
set, present a similar trend best described with a third order
polynomial. The significance of coefficient is above the 99%
level. The seven year running mean is dominated by an in-
crease from 1950 to 1963, followed by a decrease lasting
until 1986. Finally, an increase can be noted up to 2004. The
peek in 2001 is a prominent feature of the mean and can be
equally found in each of the contributing stations.
Atmos. Chem. Phys., 8, 6483–6498, 2008 www.atmos-chem-phys.net/8/6483/2008/
K. Makowski et al.: Diurnal temperature range over Europe between 1950 and 2005 6497
Acknowledgements. We would like to thank Royal Netherlands
Meteorological Institute for access to the ECAD-P data set and
the meteorological services of Romania (Sorin Cheval), Nor-
way (Elin Lundstad, Eirik Forland, Knut A. Iden) and Iceland
(Trausti Jonsson) for providing additional data. Discussions with
Thomas Peter, J¨
org M¨
ader and Andreas Roesch were highly
appreciated. Paul Southern’s proofreading is very gratefully
acknowledged. The work was funded by ETH Zurich, Polyproject:
“Variability of the sun and global climate” – Phase II. MW and AO
acknowledge NCCR Climate funded by the Swiss National Science
Foundation.
Edited by: J. Quaas
References
Abakumova, G. M., Feigelson, E. M., Russak, V., and Stadnik, V.
V.: Evaluation of long-term changes in radiation, cloudiness, and
surface temperature on the territory of the former Soviet Union,
J. Climate, 9, 1319–1327, 1996.
Alpert, P., Kishcha, P., Kaufman, Y. J., and Schwarzbard, R.:
Global dimming or local dimming? Effect of urbanization
on sunlight availability, Geophys. Res. Lett., 32, L17802,
doi:10.1029/2005GL023320, 2005.
Beer, J., Mende, W., and Stellmacher, R.: The role of the sun in
climate forcing, Quaternary Sci. Rev., 19, 403–415, 2000.
Berge, E., Bartnicki, J., Olendrzynski, K., and Tsyro, S. G.: Long-
term trends in emissions and transboundary transport of acidi-
fying air pollution in Europe, J. Environ. Manage., 57, 31–50,
1999.
Braganza, K., Karoly, D. J., and Arblaster, J. M.: Diurnal
temperature range as an index of global climate change dur-
ing the twentieth century, Geophys. Res. Lett., 31, L13217,
doi:10.1029/2004gl019998, 2004.
Dai, A., DelGenio, A. D., and Fung, I. Y.: Clouds, precipitation and
temperature range, Nature, 386, 6626, 665–666, 1997.
Dai, A., Trenberth, E., and Karl, T. R.: Effects of clouds, soil
moisture, precipitation, and water vapour on diurnal temperature
range, J. Climate, 12, 2451–2473, 1999.
De Bruin, H. A. R., van den Hurk, B. J. J. M., and Welgraven, D.:
A series of global radiation at Wageningen for 1928–1992, Int. J.
Climatol., 15, 1253–1272, 1995.
Easterling, D. R., Horton, B., Jones, P. D., Peterson, T. C., Karl,
T. R., Parker, D. E., Salinger, M. J., Razuvayev, V., Plummer,
N., Jamason, P., and Folland, C. K.: Maximum and minimum
temperature trends for the globe, Science, 277, 364–367, 1997.
Englehart, P. J. and Douglas, A. V.: Changing behavior in the di-
urnal range of surface air temperatures over Mexico, Geophys.
Res. Lett., 32, L01701, doi:10.1029/2004gl021139, 2005.
Gilgen, H., Wild, M., and Ohmura, A.: Means and Trends of short-
wave Irradiance at the surface estimated from global energy bal-
ance archive data, J. Climate, 11, 2042–2061, 1998.
Gilgen, H., Roesch, A., Wild, M., and Ohmura, A.: Decadel
changes in shortwave irradiance at the surface estimated from
Global Energy Balance Archive data, in review, 2008.
Karl, T. R., Kukla, B., and Gavin, J.: Decreasing diurnal temper-
ature range in the United States and Canada from 1941 through
1980, J. Clim. Appl. Meteorol., 23, 1489–1504, 1984.
Karl, T. R., Jones, P. D., Knight, R. W., Kukla, G., Plummer, N.,
Razuvayev, V., Gallo, K. P., Lindseay, J., Charlson, R. J., and
Peterson, T. C.: Asymmetric trends of daily maximum and mini-
mum temperature, B. Am. Meteorol. Soc., 74, 1007–1023, 1993.
Klein Tank, A. M. G., Wijngaard, J. B., K¨
onnen, G. P., B¨
ohm, R.,
Demar´
ee, G., Gocheva, A., Mileta, M., Pashiardis, S., Hejkr-
lik, L., Kern-Hansen, C., Heino, R., Bessemoulin, P., M¨
uller-
Westermeier, G., Tzanakou, M., Szalai, S., P´
alsd´
ottir, T., Fitzger-
ald, D., Rubin, S., Capaldo, M., Maugeri, M., Leitass, A., Bukan-
tis, A., Aberfeld, R., van Engelen, A. F. V., Forland, E., Mietus,
M., Coelho, F., Mares, C., Razuvaev, V., Nieplova, E., Cegnar,
T., Antonio L´
opez, J., Dahlstr¨
om, B., Moberg, A., Kirchhofer,
W., Ceylan, A., Pachaliuk, O., Alexander, L. V., and Petrovic, P.:
Daily dataset of 20th-century surface air temperature and precip-
itation series for the European Climate Assessment, Int. J. Cli-
matol., 22, 1441–1453, doi:10.1002/Joc.773, 2002.
Klein, H. and Benedictow, A.: EMEP MSC-W (Meteorological
Synthesizing Centre-West) Data Note 1/2006: Transboundary
data by main pollutants (S, N, O3)and PM, ISSN 1890-0003,
2006.
Kukla, G. and Karl, T.R.: Nighttime warming and the Greenhouse-
Effect, Environ. Sci. Technol., 27, 8, 1468–1474, 1993.
Lefohn, A. S., Husar, J. D., Husar, R. B., and Brimblecombe, P.:
Assessing historical global sulfur emission patterns for the period
1850–1990, Report number: DOE/ER/30234-1, prepared for the
US department of energy, 190 pp. [NTIS DE96014790], 1996.
Liepert, B. G. and Kukla, G. J.: Decline in global solar radiation
with increased horizontal visibility in Germany between 1964
and 1990, J. Climate, 10, 2391–2401, 1997.
Liu, B., Xu, M., Henderson, M., and Gong, W.: A spatial analysis
of pan evaporation trends in China 1955-2000, J. Geophys. Res.,
109, D15102, doi:10.1029/2004jd004511, 2004.
Mishchenko, M. I., Geogdzhayev, I. V., Rossow, W. B., Cairns, B.,
Carlson, B. E., Lacis, A. A., Liu, L., and Travis, L. D.: Long-
term satellite record reveals likely recent aerosol trend, Science,
315, 1543, doi:10.1126/science.1136709, 2007.
Mylona, S.: Sulphur dioxide emissions in Europe 1880–1991 and
their effect on sulphur concentrations and depositions, Tellus, 48,
662–689, 1996.
Norris, J. R. and Wild, M.: Trends in aerosol radiative effects over
Europe inferred from observed cloud cover, solar “dimming”
and solar “brightening”, Geophys. Res. Lett., 112, D08214,
doi:10.1029/2006JD007794, 2007.
Ohmura, A. and Lang, H.: Secular variations of global radiation
in Europe, IRS’88: Current problems in atmospheric radiation,
edited by: Leonoble, J. and Geleyn, J. F., A. Deepak Publ.,
Hampton, VA, 298–301, 1989.
Ohmura, A.: Observed long-term variations of solar irradi-
ance at the earth’s surface, Space Sci. Rev., 125, 111–128,
doi:10.1007/s11214-006-9050-9, 2006.
Pinker, R. T., Zhang, B., and Dutton, E. G.: Do satellites de-
tect trends in surface solar radiation?, Science, 308, 850–854,
doi:10.1126/science.1103159, 2005.
Robock, A. and Li, H.: Solar dimming and CO2effects
on soil moisture trends, Geophys. Res. Lett., 33, L20708,
doi:10.1029/2006gl027585, 2006.
Roderick, M. L. and Farquhar, G. D.: The cause of decreased pan
evaporation over the past 50 years, Science, 298, 1410–1411,
2002.
www.atmos-chem-phys.net/8/6483/2008/ Atmos. Chem. Phys., 8, 6483–6498, 2008
6498 K. Makowski et al.: Diurnal temperature range over Europe between 1950 and 2005
Sanchez-Lorenzo, A., Brunetti, M., Calbo, J., and Martin-Vide, J.:
Recent spatial and temporal variability and trends of sunshine du-
ration over the Iberian Peninsula from a homogenized dataset, J.
Geophys. Res., 112, D20115, doi:10.1029/2007jd008677, 2007.
Sanchez-Lorenzo, A., Calb´
o, J., and Martin-Vide, M.: Spatial
and temporal trends in sunshine duration over Western Europe
(1938–2004), J. Climate, doi:10.1175/2008JCLI2442.1, in press,
2008.
Stanhill, G. and Cohen, S.: Global dimming: a review of the ev-
idence for widespread and significant reduction in global radia-
tion with discussion of its probable cause and possible agricul-
tural consequences, Agr. Forest Meteorol., 107, 255–278, 2001.
Stanhill, G. and Cohen, S.: Solar radiation changes in the United
States during the Twentieth Century: Evidence from sunshine
measurements, J. Climate, 18, 1503–1512, 2005.
Stechikov, G. L. and Robock, A. : Diurnal asymmetry of climatic
response to increased CO2and aerosol: forcings and feedbacks,
J. Geophys. Res., 100(D12), 26211–26 227, 1995.
Stern, D. I.: Reversal of the trend in global anthropogenic
sulphur emissions, Global Environ. Change, 16, 207–220,
doi:10.1016/j.gloenvcha.2006.01.001, 2006.
Stone, D. A. and Weaver, A. J.: Factors contributing to diurnal tem-
perature range trends in twentieth and twenty-first century sim-
ulations of the CCCma coupled model, Clim. Dynam., 20, 435–
445, doi:10.1007/s00382-002-0288-y, 2002.
Smith, S. J., Pitcher, H., and Wigley, T. M. L.: Future sulphur diox-
ide emissions, Clim. Change, 73, 267–318, doi:10.1007/s10584-
005-6887-y, 2005.
Streets, D. G., Wu, Y., and Chin, M.: Two-decadel
aerosol trends as a likely explanation of the global dim-
ming/brightening transition, Geophys. Res. Lett., 33, L15806,
doi:10.1029/2006GL026471, 2006.
Trenberth, K. E., Jones, P. D., Ambenje, P., Bojariu, R., Easterling,
D., Klein Tank, A., Parker, D., Rahimzadeh, F., Renwick, J. A.,
Rusticucci, M., Soden, B., and Zhai, P.: Observations: Surface
and Atmospheric Climate Change. The Physical Science Basis,
Contribution of Working Group I to the Fourth Assessment Re-
port of the Intergovernmental Panel on Climate Change, edited
by: Solomon, S., Qin, D., Manning, M., Chen, Z., Marquis, M.,
Averyt, K. B., Tignor, M., and Miller, H. L., 987 pp., 2007.
Vestreng, V., Myhre, G., Fagerli, H., Reis, S., and Tarras´
on, L.:
Twenty-five years of continuous sulphur dioxide emission reduc-
tion in Europe, Atmos. Chem. Phys., 7, 3663–3681, 2007,
http://www.atmos-chem-phys.net/7/3663/2007/.
Vose, R. S., Easterling, D. R., and Gleason, B.: Maximum and min-
imum temperature for the globe: An update through 2004, Geo-
phys. Res. Lett., 32, L23822, doi:10.1029/2005GL024379, 2005.
Wild, M., Gilgen, H., Roesch, A., Ohmura, A., Long, C. N., Dut-
ton, E. G., Forgan, B., Kallis, A., Russak, V., and Tsvetkov,
A.: From dimming to brightening, Science, 308, 741–908,
doi:10.1126/science.1103215, 2005.
Wild, M., Ohmura, A., and Makowski, K.: Impact of global dim-
ming and brightening on global warming, Geophys. Res. Lett.,
34, L04702, doi:10.1029/2006GL028031, 2007.
Atmos. Chem. Phys., 8, 6483–6498, 2008 www.atmos-chem-phys.net/8/6483/2008/
