Réf. Schmidli & al. 2001 - A

Référence bibliographique complète
SCHMIDLI J., SCGMUTZ C., FREI C., WANNER H. and SCHÄR C. Mesoscale precipitation variability in the Alpine region during the 20th century. International Journal of Climatology, 2001, vol. 22, 1049-1074.

Abstract: For the period 1901-90, an increase of winter precipitation by 20-30% per 100 years in the western part of the Alps, and a decrease of autumn precipitation by 20-40% to the south of the main ridge are detected. The correlation analysis reveals weak and highly intermittent correlations with the North Atlantic Oscillation Index to the north and more robust correlations to the south of the main Alpine crest. It implies that the increase of winter precipitation cannot be explained by the observed trend of the NAOI, at least in the framework of a simple linear regression model. The observation that precipitation changes in winter are primarily due to increasing precipitation activity rather than changes in the frequency of weather types and that the changes are associated with an increasing frequency of intense precipitation events lend some support to the hypothesis of an intensified water cycle.


Precipitation variations, data reconstruction, North Atlantic Oscillation, Alps, subdomains, 20th century

Organismes / Contact
Climate Research, ETH Zürich, schmidli@geo.umnw.ethz.ch
University of Berne

(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



Pays / Zone
Massif / Secteur
Site(s) d'étude
Période(s) d'observation

Austria, Switzerland, Italy, southern Germany, France.

European Alps (3.2°-16.2° E, 43.2°-48.8° N)




(1) - Modifications des paramètres atmosphériques

Mean annual values of Alpine precipitation vary between 1.5 and 8 mm day-1 according to ANALYSIS. The main features of the spatial distribution are two high-precipitation bands extending along the northern and southern rim of the Alpine ridge, and drier conditions in the interior of the mountain range and over the adjacent atland. The maximum along the southern rim is divided into two major wet zones, centered over southern Switzerland/northern Italy and over eastern Italy/Slovenia, respectively. The monthly precipitation patterns exhibit pronounced seasonal and interannual variations, as reected in the standard deviation. Regions with high standard deviation coincide with regions of high annual means.

Time reconstructions of seasonal precipitation means of the 5 leading RPCs for the period 1901-90 have shown a large variability on annual to multi-decadal time-scales. The magnitude of the interannual variability varies considerably between RPCs and between seasons. Various of the decadal-scale extrema and also some of the annual peaks are common between the corresponding principal components (PCs).

Linear precipitation trends for the 1901-90 period have shown that only winter and autumn precipitation exhibit well-established areas of significant (at 10% level) trends. In winter, significant positive trends of 20-30% per 100 years are found for a region extending from southern Germany, through Switzerland, to the south-western tip of the Alps; and non-significant negative trends of 20-30% are observed in the south-eastern part of the Alps. In autumn, significant negative trends of 20-40% per 100 years occur in southern France, parts of northern Italy, and in the eastern Alps. While the negative trend along the Mediterranean coast should be considered with caution (only a few stations), the negative trend in the eastern Alps is trustworthy as it is supported by a few dozen stations. The western Alps PC shows a highly significant increase (at the 5% level) by 35% per 100 years in winter, and the Mediterranean and south-eastern Alps PCs show a significant decrease by -38% and -30% per 100 years, respectively, in autumn.

Linear trends of seasonal precipitation means for the period 1961-90 have shown significant trends in winter, spring, and autumn. In winter, positive trends of more than 20% per 30 years are observed for most regions north of the Alpine crest, and also for the central Alps. Although the trends are three times larger than for 1901-90 (20-40% per 30 years), they are only significant in southern Germany (small sample size). In spring, positive trends of 20-40% are observed in the Ticino and Piedmont region, south-western Alps, and along the western boundary of the domain; significant negative trends of 20-30% per 30 years are found in a small region along the northern Alpine rim. In autumn, a precipitation decrease of up to 40% per 30 years and more is found south of the Alps, in southern France and in most of northern Italy.

Interannual precipitation variations in the Alps exhibit marginal linear relationships with lower tropospheric circulation variations as cast by the univariate NAOI.


Informations complémentaires (données utilisées, méthode, scénarios, etc.)

The reconstruction of the mesoscale precipitation fields is based on the reduced-space optimal interpolation method of Kaplan et al. The reconstruction and analysis of centennial precipitation variations proceeds from two datasets: a high-resolution gridded monthly precipitation analysis for the recent decades, and a comparatively sparse sample of homogenized station records extending back to the beginning of the 20th century. The statistical combination of the two datasets involves a decomposition of the dense component into a set of dominant modes of space-time variability and the estimation of the coefficients of the modes from the sparse component.

The high-resolution gridded dataset (ANALYSIS) consists of monthly precipitation fields for the period 1971-90 with a grid spacing of 25 km (0.22° lat * 0.3° long). It was derived from 6800 station records of the operational high-resolution rain-gauge networks in the Alpine region. The spatial analysis represents area-average precipitation in the surrounding of each of the 928 land grid points. The original rain-gauge network exhibits a typical inter-station distance of 10-15km and hence the effective resolution of the ANALYSIS is close to 25km over most parts of the study domain.
The sparse dataset comprises 140 long-term station records covering the period 1901-90. The network exhibits great difference in the density of the stations. Long-term station records can be affected by many changes, so the dataset was subjected to a rigorous homogenization procedure (Alexandersson test).
The performance of the reconstruction has been assessed by crossvalidation with ANALYSIS, to be validated. The large month-to-month fluctuations, observed in all regions, have been accurately reproduced by the reconstruction, which explains between 88% and 99.9% of the variance in subdomain-mean precipitation. The final reconstruction was obtained by calibration over the period 1971-90 and using truncation at L = 28.

The rotated principal components (RPCs) represent estimates of area mean precipitation anomalies. Loading patterns of the five leading RPCs of monthly precipitation for the years 1971-90 have been presented. The five reconstructed RPCs explain 70% of the total variance of the ANALYSIS. Thus the leading five modes explain 90% of the total variance of the reconstruction.
The spatial distribution of linear trends of seasonal precipitation has been presented for the complete reconstruction (28 PCs), and compared to linear trends in the 5 leading RPCs. Linear precipitation trends for the period 1901-90 have been obtained from linear regressions for each of the grid points and using the full reconstruction dataset. The trends are given as percentage change over 100 years of the seasonal mean.

Values of monthly NAOI have been taken from Hurrell (2000). The correlation study has been undertaken for several independent time periods to examine the stationarity of the NAOI-Alpine precipitation link.

(2) - Impacts du changement climatique sur le milieu naturel



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

(3) - Impacts du changement climatique sur l'aléa

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