Réf. Begert & al. 2005 - A

Référence bibliographique complète
BEGERT M., SCHLEGEL T., KIRCHHOFER W. Homogenous temperature and precipitation series of Switzerland from 1864 to 2000. International Journal of Climatology, 2005, Vol. 25, p. 65-80.

Abstract: A set of 12 homogenized monthly mean temperature and precipitation series of Switzerland for the period 1864-2000 are introduced. The standardized homogenization procedure developed and implemented at MeteoSwiss during recent years has been presented. A trend analysis is performed on each temperature and precipitation series and on a mean temperature series of Switzerland. The results are compared with findings of other studies that have examined long-term temperature and precipitation trends in Switzerland and neighbouring countries. The inhomogeneities of the Swiss temperature series are up to ±1.6°C and the precipitation adjustment factors vary between 0.5 and 1.6. Each of the 12 temperature series analysed contains several inhomogeneities that cause systematic biases in the adjustment curves. The slope of a mean temperature curve derived from the original data is underestimated by 0.4°C/100 years. All precipitation series except one contain inhomogeneities, but no systematic bias is observed. The trend analysis reveals an increase in the yearly temperature series ranging from 0.9°C/100 years to 1.1°C/100 years at stations north of the alpine main crest and ~0.6°C/100 years at southern stations. Precipitation trends are observed at most sites north of the alpine main crest in winter and in some of the yearly series. The annual slopes vary between 7 and 10% and the winter slopes between 16 and 37%.

Mots-clés
Homogenization, adjustments, Switzerland, temperature, precipitation, trend analysis, alpine climate.

Organismes / Contact
Federal Office of Meteorology and Climatology (MeteoSwiss), Krähbühlstrasse 58, PO Box 514, CH-8044 Zurich, Switzerland. michael.begert@meteoschweiz.ch

(1) - Paramètre(s) atmosphérique(s) modifié(s)
(2) - Elément(s) du milieu impacté(s)
(3) - Type(s) d'aléa impacté(s)
(3) - Sous-type(s) d'aléa
Temperature, Precipitation      

Pays / Zone
Massif / Secteur
Site(s) d'étude
Exposition
Altitude
Période(s) d'observation
Switzerland   12 stations covering the whole country   273-2490 m a.s.l. 1864-2000

(1) - Modifications des paramètres atmosphériques
Reconstitutions
 
Observations
The mean adjustment curve, corresponding to the difference between the homogeneous and the original data series, shows a significant trend of 0.43°C/100 years at a confidence level of 95%. All precipitation series except one contain inhomogeneities, but no systematic bias is observed. In general, the size of a detectable inhomogeneity in this study was therefore smaller at stations of the Swiss plateau than at alpine stations and smaller in recent decades than in the 19th century. The evolution of the 12 homogeneous series is remarkably similar over the whole period. Only Segl-Maria and Lugano, which are both situated south of the alpine main crest, show a slightly different climatic signal in the 19th century and at the beginning of the 20th century.

Temperature
Most of the yearly and seasonal temperature series show a significant positive trend. The only exception occurs in summer, when no significant trend is found in the series of Lugano, Segl-Maria and Zurich. The slopes of the yearly series range from 0.9°C/100 years to 1.1°C/100 years at stations on the northern side of the Alps, whereas the two southern stations of Lugano and Segl-Maria show smaller slopes of 0.6°C/100 years. The largest increase in yearly temperature series (1.2°C/100 years) is observed in Sion (Valais). Looking at trends in different seasons, the stations with a higher elevation (Chaumont, Säntis, Segl-Maria) reveal their largest slopes in autumn, ranging from 0.8°C/100 years to 1.3°C/100 years. The largest increase for stations at lower elevations is found in winter, ranging from 0.9°C/100 years to 1.6°C/100 years. Again, the smallest slopes of all stations are observed in Lugano and Segl-Maria, independent of the season. Comparing the results with findings of previous studies (e.g. Schönwiese et al., 1994; Brunetti et al., 2000; Auer et al., 2001), there is generally good agreement.

Precipitation
For precipitation, there is no indication of a significant increase or decrease in spring, summer or autumn at the 95% confidence level. However, significant precipitation trends are observed at most sites in winter and in some of the yearly series. A significant increase in yearly values can be seen at the stations of the Swiss Plateau (Berne, Zurich, Geneva) and in the series of Chaumont, representing the Jura mountains. The slopes of the trends range from 7 to 10% per 100 years. However, looking at a progressive analysis, it becomes clear that the trends are influenced strongly by the last few years with relatively high annual precipitation amounts. In winter, positive trends of 16 to 37% per 100 years are found for the stations north of the alpine main crest. The strongest increase is observed in the western part of Switzerland (Geneva, Chaumont). The beginnings of the trends in most of the series are around 1940 to 1950. Similar to temperature, the two sites situated south of the alpine main crest clearly differ from the other stations. There is no significant increase or decrease in the yearly or seasonal precipitation series of Lugano or Segl-Maria. The comparison of these results with findings of previous studies (e.g. Auer and Böhm, 1994; Widmann and Schär, 1997; Buffoni et al., 1999; Schmidli et al., 2002) shows a qualitatively good agreement.

Swiss mean temperature trends
The main characteristics of the Swiss mean annual temperature curve are the cold period at the end of the 19th century, with a minimum around 1891, and the following more-or-less continuous increase. The increase is interrupted by a relatively warm period in the second part of the 1940s and has intensified in the last 15 years. The mean temperature increase in Switzerland amounts to 1.0°C/100 years. According to the results of the Mann-Kendall test, the overall trend becomes significant around 1950. Looking at the seasonal evolutions, a different behaviour of SON and DJF compared with MAM and JJA can be observed. The autumn and winter curves show a more continuous increase, with larger negative fluctuations around 1890 (SON, DJF) and around 1915 (SON). Both curves show significant trends beginning around 1940 in autumn and around 1970 in winter. In spite of the earlier beginning of the increase, the slope obtained from a linear least-squares fitting is smaller in autumn (1.1°C/100 years) than in winter (1.3°C/100 years). Temperature evolutions in spring and summer are subject to larger and more abrupt fluctuations compared with autumn and winter, in particular due to the warmer period 1945-55 and the remarkable sudden increase after 1980. The shift-like increase in the curves of MAM and JJA in the 1980s is responsible for the significant trends found by the Mann-Kendall test in the whole period 1864-2000. The slopes of the linear fitting are 0.8°C/100 years for spring and 0.7°C/100 years for summer. There is no continuous evolution towards warmer temperature in spring and summer.
Modélisations
 
Hypothèses
 

Informations complémentaires (données utilisées, méthode, scénarios, etc.)
Basle, Berne, Château-d'Oex, Chaumont, Davos, Engelberg, Geneva, Lugano, Säntis, Segl-Maria, Sion and Zurich stations have been selected, ranging from 273 to 2490 m a.s.l.

The homogenization procedure THOMAS (tool for homogenization of monthly data series, see Aschwanden et al., 1996; Bosshard, 1996; Baudenbacher, 1997) can be divided into two main steps: the detection of inhomogeneities and the calculation of the adjustments. The procedure allows searching and adjusting of the shifts in mean and linear trends. At least two statistical methods are used in the detection and the adjustment procedure. The detection of inhomogeneities with THOMAS is a combination of metadata analysis and the use of 12 different homogeneity tests. The results of the statistical tests and of the station history analysis serve to identify the date of an inhomogeneity and to divide the data series into segments. Each segment is then investigated separately and the iteration is repeated until each segment is determined as homogeneous at the 95% confidence level.

Reference series are used to isolate the effects of station discontinuities from regional climate change. During the adjustment procedure of THOMAS, each potential inhomogeneity detected is judged again, using different parametric and non-parametric statistical methods. The significance of the adjustments is tested using the Student's t-test for temperature and the robust Wilcoxon rank sum test for precipitation. The Kendall tau test is used as a non-parametric method to test the significance of trends.

Parameter-specific characteristics of the inhomogeneities detected were evaluated, namely their magnitude, their frequency distribution, their temporal occurrence and a possible systematic bias. The adjustments calculated describe the differences between the measuring conditions before and after an inhomogeneity, including the contributions of all different causes. The most frequent reason for a shift inhomogeneity was site relocation. Other reasons for shifts were the introduction of the ANETZ (conventional climate station converted into automated station), new types of instrument, changes in screens, changes in instruments, new observers, and changes in the time of observation and calibrations. Some inhomogeneities could not be explained by station histories.

The homogenized data set was used to study annual and seasonal characteristics of the temperature and precipitation series in the period 1864-2000. Annual values correspond to the period from December to November and are dated by the year in which January is included. Winter values refer to DJF, spring to MAM, summer to JJA and autumn to SON. All values used for trend analysis are anomalies from the 1961-90 mean value. The non-parametric Mann-Kendall test, as described in Sneyers (1990), was applied for trend analysis. A trend is considered statistically significant at a confidence level of 95% and the slopes of the trends are calculated by least squares linear fitting. The analysis was not performed for Château d'Oex (T, P), Davos (T, P) and Säntis (P) because of incomplete data series in the 19th century.

In order to give an overview of the main characteristics of climate variability in Switzerland between 1864 and 2000, the mean Swiss series were also analysed. The first principal component (PC) of a PC analysis (PCA) was used to judge whether a mean of the 12 available series adequately represents the whole Swiss territory. The PCA was performed on the standardized anomalies providing data for the whole period. Therefore, Château d'Oex (T, P), Davos (T, P) and Säntis (P) were not used because of incomplete data in the 19th century. Owing to the high percentage of explained variance (89-92%) and the high correlation, a simple averaged temperature series is a good representation for all areas of Switzerland. As expected, the precipitation field shows more local patterns and no mean series was used to analyse temporal precipitation changes in Switzerland.

(2) - Effets du changement climatique sur le milieu naturel
Reconstitutions
 
Observations
 
Modélisations
 
Hypothèses
 

Sensibilité du milieu à des paramètres climatiques
Informations complémentaires (données utilisées, méthode, scénarios, etc.)
   

(3) - Effets du changement climatique sur l'aléa
Reconstitutions
 
Observations
 
Modélisations
 
Hypothèses
 

Paramètre de l'aléa
Sensibilité du paramètres de l'aléa à des paramètres climatiques
Informations complémentaires (données utilisées, méthode, scénarios, etc.)
 
 

(4) - Remarques générales
 

(5) - Syntèses et préconisations
 

Références citées :

Aschwanden A, Beck M, Haeberli C, Haller G, Kiene M, Roesch A, Sie R, Stutz M. 1996. Bereinigte Zeitreihen — Die Ergebnisse des Projekts Klima90, Bd. 2: Methoden Klimatologie der Schweiz, Jg. 1996. Schweizerischen Meteorologischen Anstalt: Zürich.

Auer I, Böhm R. 1994. Combined temperature–precipitation variations in Austria during the instrumental period. Theoretical and Applied Climatology 49 : 161–174.

Auer I, Böhm R, Schöner W. 2001. Austrian Long-term Climate 1767–2000. Multiple Instrumental Climate Time Series from Central Europe. Oesterreichische Beiträge zu Meteorologie und Geophysik, vol. 25. Zentralanstalt für Meteorologie und Geodynamik (ZAMG): Vienna.

Baudenbacher M. 1997. Homogenisierung langer Klimareihen dargelegt am Beispiel der Lufttemperatur. Veröffentlichung der Schweizerischen Meteorologischen Anstalt, vol. 58. Schweizerische Meteorologische Anstalt: Zürich.

Bosshard W. 1996. Homogenisierung klimatologischer Zeitreihen, dargelegt am Beispiel der relativen Sonnenscheindauer. Veröffentlichung der Schweizerischen Meteorologischen Anstalt, vol. 57. Schweizerische Meteorologische Anstalt: Zürich.

Brunetti M, Maugeri M, Nanni T. 2000. Variations of temperature and precipitation in Italy from 1866 to 1995. Theoretical and Applied Climatology 65 : 165–174.

Buffoni L, Maugeri M, Nanni T. 1999. Precipitation in Italy from 1833 to 1996. Theoretical and Applied Climatology 63 : 33–40.

Schmidli J, Schmutz C, Frei C, Wanner H, Schär C. 2002. Mesoscale precipitation variability in the region of the European Alps during the 20th century. International Journal of Climatology 22 : 1049–1074. - [Fiche Biblio]

Schönwiese CD, Rapp J, Fuchs T, Denhard M. 1994. Climate Trend Atlas of Europe. Based on Observations 1891–1990 . Kluwer Academic Publishers.

Sneyers R. 1990. On the statistical analysis of series of observation. WMO, Technical Note 143. WMO, Geneva.

Widmann ML, Schär C. 1997. A principal component and long-term trend analysis of daily precipitation in Switzerland. International Journal of Climatology 17 : 1333–1356.