Case I.1
Links between circulation changes and climatic trends in European regions
(Monika Cahynová) 1
Introduction
In recent decades the awareness of anthropogenic climate change has boosted research in the field of
synoptic climatology. More and more studies are trying to attribute the climatic trends and variability to
atmospheric circulation, taking into account either various characteristics of the pressure field (e.g.
circulation indices) or synoptic types classifications.
Generally, the links are found to be more pronounced in the winter season (e.g. Chen, 2000; Huth,
2001; Beck et al., 2007; Kostopoulou and Jones, 2007). The well-documented strengthening of zonal
flow over Europe connected with a positive trend in the North Atlantic Oscillation in the decades
preceding the 1990s is considered to be the main cause for the observed wintertime warming (Hurrell,
1995). When analyzing century-long or longer time series, several authors point out that the
relationships between circulation and climate are changing on decadal time scales, thus strongly
limiting the usefulness of statistical downscaling models (e.g. Hanssen-Bauer and Førland, 1998;
Beranová and Huth, 2008). Other causes than circulation changes are blamed for at least part of the
local climate variability and trends (e.g. Huth, 2001; Goodess and Jones, 2002).
In this study we quantify the links between large-scale circulation changes over Europe and local
surface climatic trends in the period 1961-2000. The main aim is to compare the results of 24
circulation classifications from the COST733 catalogue 1.2. To our knowledge, such a comparative
approach has not been used so far. We further compare the classifications according to their skill to
stratify daily climatic data into types. In both the analyses we focus on the effect of the number of
types and the size of the spatial domain on the results.
Data & Methods
For the description of atmospheric circulation we have used 8 classification methods from the
COST733 cat. 1.2 (CKMEANS, GWT, LITADVE-LITC18-LITTC, LUND, P27, PETISCO, SANDRA,
TPCA), each applied on sea level pressure fields with a pre-defined number of circulation types (9, 18,
and 27). The classifications were computed across the whole Europe and 11 European regions. The
subjective Hess-Brezowsky catalogue (HBGWL, HBGWT) was used for comparison in the Czech
Republic.
Meteorological data from the Czech Republic are represented by daily values of eleven variables at 21
stations in the period 1961-1998. The variables include daily maximum, minimum, and mean
temperature, precipitation amount, the occurrence of precipitation, relative humidity, cloudiness (in
tenths), and sunshine duration.
More information can be found in Cahynová M., Huth R. (2009): Changes of atmospheric circulation in central
Europe and their influence on climatic trends in the Czech Republic. Theor. Appl. Climatol. 96: 57-68.
Cahynová M., Huth R.: Circulation vs. climatic changes over the Czech Republic: A comprehensive study based
on the COST733 database of atmospheric circulation classifications. Phys. Chem. Earth,
doi:10.1016/j.pce.2009.11.002.
1
At the scale of Europe, we have used daily data of maximum (Tmax) and minimum (Tmin) air
temperature and precipitation amount from 29 stations of the European Climate Assessment &
Dataset in the period 1961-2000.
For the assessment of “skill” of circulation classifications to stratify daily climatic data into circulation
types, we have applied the explained variance index (EV index).
Seasonal circulation and climatic trends were estimated using linear least-squares regression applied
to the seasonal occurrence of days with a specific circulation type (CT), and to the seasonal average
of the given climatic variable. The common t-test was performed to evaluate the statistical significance
of both circulation and climatic trends, and only the stations with climatic trends significant at the 95%
level were further studied.
For the detection of relationships between changes in atmospheric circulation and trends in surface
climatic variables, two methods were used: first the method of “hypothetical” (circulation-induced)
linear trends (see e.g. Huth, 2001), and second the decomposition of climatic change that occurred
between the 1st and the 2nd half of the study period (e.g. Beck et al., 2007).
Results
The EV index is the highest for Tmax, lower for Tmin, and the lowest for precipitation. The highest
values are reached at the Icelandic, Norwegian, and Alpine stations in winter. Classifications defined
on small domains produce higher EV index than those defined on D00.
Using the first method of attribution, circulation changes in the classifications defined over small
domains are responsible for about one fourth of the spring trends of Tmax and precipitation, and less
for Tmin. In summer the circulation changes have virtually no influence on the observed trends, the
ratio of circulation-induced and observed trends being close to zero in most cases (and sometimes
even negative values occur). In autumn very few climatic trends are significant, and the influence of
circulation changes on these trends is again very small for all the three variables (usually between 0
and 30%). It means that the within-type trends – i.e. changes in the internal climatic properties of
individual circulation types – are the main driver of observed climatic trends in spring, summer, and
autumn. This pattern occurs systematically over all stations and circulation classifications.
Winter is the only season when circulation changes play a major role in the observed climatic trends;
however, there are some stations in the Balkans and the Mediterranean where the within-type
changes are still more important even in this season. Circulation changes have a major influence
(around or more than 50%) on the massive recent warming over the British Isles and Central Europe.
Precipitation trends in winter are best resolved by circulation changes at one of the easternmost
stations – Kyiv – that underwent substantial desiccation in 1961-2000, and this is the only case that
the ratio of circulation-induced and observed trend reaches 1 (although only by one classification).
The individual classifications produce very different results (see Figure Error! No text of specified
style in document.-1; note that average of all the stations is shown). The results are usually higher
for circulation defined on small domains compared to D00, but in Iceland and Scandinavia the
observed trends are better resolved by the large-scale circulation changes (not shown). The results
are correlated with the number of circulation types.
The second method of attribution of climatic changes (Beck et al., 2007) gives fairly similar results,
only in winter in Central Europe the frequency-related part of climatic change between the 1st and the
2nd half of the period is higher and reaches 50 to 100%.
In the Czech Republic the same analysis was performed at 21 stations in 1961-1998. The results are
consistent to those obtained for other Central European stations in the period 1961-2000. In winter the
circulation influence on climate variables trends is the highest for temperature and lower for
precipitation occurrence, relative humidity, and sunshine, while in spring and summer the results for
temperature are lower than for the other variables (in autumn the values are about the same). The
variability within results from individual stations is comparable to the variability produced by the 24
objective circulation classifications. There is no clear dependence of the results on the geographical
position of the stations.
The subjective Hess-Brezowsky catalogue with 29 types (HBGWL) and 10 types (HBGWT) produce
substantially higher proportion of circulation-induced climatic changes in the Czech Republic than the
objective catalogues for both methods of assessment in winter for temperature, and in autumn and
spring for nearly all climatic variables. The reason might be that in the HBGWL and HBGWT
catalogues, more circulation types bear significant seasonal trends in their frequency compared to the
objective catalogues.
SON-Dxx
SON-D00
DJF-Dxx
DJF-D00
TN
0.6
0.5
0.5
0.4
0.4
0.3
0.3
ratio
0.6
0.2
0.1
0.2
0.1
0
0
-0.1
-0.1
-0.2
-0.2
-0.3
-0.3
CKMEANSC09
CKMEANSC18
CKMEANSC27
GWTC10
GWTC18
GWTC26
LITADVE
LITC18
LITTC
LUNDC09
LUNDC18
LUNDC27
P27C08
P27C16
P27C27
PETISCOC09
PETISCOC18
PETISCOC27
SANDRAC09
SANDRAC18
SANDRAC27
TPCAC09
TPCAC18
TPCAC27
ratio
JJA-Dxx
JJA-D00
CKMEANSC09
CKMEANSC18
CKMEANSC27
GWTC10
GWTC18
GWTC26
LITADVE
LITC18
LITTC
LUNDC09
LUNDC18
LUNDC27
P27C08
P27C16
P27C27
PETISCOC09
PETISCOC18
PETISCOC27
SANDRAC09
SANDRAC18
SANDRAC27
TPCAC09
TPCAC18
TPCAC27
MAM-Dxx
MAM-D00
TX
Figure Error! No text of specified style in document.-1. Ratio of circulation-induced and observed
climatic trends (i.e. the proportion of observed trends directly linked to changing frequency of
CTs) in 1961-2000. Average of European stations where seasonal observed trends are
significant at the 95% level.
Summary & Conclusions
We evaluate and compare 24 circulation classifications from the COST733 cat. 1.2 according to their
ability to stratify daily station climatic data into circulation types, and assess the magnitude of seasonal
temperature and precipitation trends in Europe in the period 1961-2000 (and in the Czech Republic in
1961-1998) that can be linked to changing frequency of circulation types in these classifications.
The skill to stratify climatic data into types (measured by Explained Variance index) is the highest for
maximum temperature, lower for minimum temperature, and the lowest for precipitation. The highest
values are generally obtained in winter for classifications with 27 types that are computed on the small
spatial domains.
Seasonal climatic trends can be only partly explained by the changing frequency of circulation types,
the link being the strongest in winter. In the other seasons, within-type climatic trends are responsible
for a major part of the observed trends. Circulation changes in the small domains are usually more
tightly connected with climatic trends than those in the large domain except for Icelandic and
Scandinavian stations where circulation over the whole Europe explains a larger part of the observed
trends. There are large differences between results obtained with individual classifications, which
suggests that a comparative approach is highly desirable in such synoptic-climatological studies.
Recommendations
1. Family of algorithms
No clear result: the ranking varies for different seasons and climatic variables.
2. Domain size and locations
Small scale is usually more tightly linked to climatic variability and trends. Exception: winter and spring
climatic trends in Iceland and Scandinavia are better described by circulation in the large domain.
3. Number of classes
More classes usually means better separability of daily climatic variables into types (this results from
the definition of EV index), and also larger proportion of climatic trends that can be attributed to
circulation changes. Exceptions occur.
4. Input variables,
NOT tested
5. Seasonal
NOT tested
6. Sequencing
NOT tested
References
Beck C, Jacobeit J, Jones PD (2007) Frequency and within-type variations of large-scale circulation
types and their effects on low-frequency climate variability in central Europe since 1780. Int J
Climatol 27: 473–491
Beranová R, Huth R (2008) Time variations of the effects of circulation variability modes on European
temperature and precipitation in winter. Int J Climatol 28: 139–158
Chen D (2000) A monthly circulation climatology for Sweden and its application to a winter temperature case
study. Int J Climatol 20: 1067–1076
Goodess CM, Jones PD (2002) Links between circulation and changes in the characteristics of Iberian
rainfall. Int J Climatol 22: 1593–1615
Hanssen-Bauer I, Førland EJ (1998) Long-term trends in precipitation and temperature in the
Norwegian Arctic: can they be explained by changes in atmospheric circulation patterns? Clim
Res 10: 143–153
Hurrell JW (1995) Decadal trends in the North Atlantic Oscillation: regional temperatures and
precipitation. Science 269: 676–679
Huth R (2001) Disaggregating climatic trends by classification of circulation patterns. Int J Climatol 21:
135–153
Kostopoulou E, Jones PD (2007) Comprehensive analysis of the climate variability in the eastern
Mediterranean. Part II: relationships between atmospheric circulation patterns and surface
climatic elements. Int J Climatol 27: 1351–1371