Atmospheric circulation patterns related to heavy snowfall days in Andorra, Pyrenees

Abstract

Heavy snowfalls over mountain regions are often a direct cause of avalanches. Specific synoptic-scale atmospheric situations are responsible for these kinds of extreme snowfall event, and this is indeed the case for Andorra, a small country located in the Pyrenees, between France and Spain. Based on days with an intensity of at least 30 cm of snow in a 24 h period, the present study uses principal component analysis (PCA) and clustering techniques to characterize the synoptic circulation patterns for these days during the winter season. The area of analysis encompasses the region 30-60°N by 30°W-15°E and the period covers the winter seasons from 1986-87 to 2000-01. The methodology proposed involves a preprocessing approach consisting of a spatial standardization of the data used for the PCA, an alternative approximation to decide the centroids and the number of groups for the K-means clustering, and the rejection of the iterations for this algorithm. This approach enables the synoptic classification of every heavy snowfall day, and composite maps were constructed for sea-level pressure, 500 hPa geopotential height, and 1000-500 m thickness (the 5270 m, 5400 m and 5520 m contour lines). The results show seven circulation patterns, most of them with an Atlantic component of the wind, and others with a clear Mediterranean advection that could be combined with cold continental air. The results, as weather charts, could be a useful tool to assist meteorological models in heavy snowfall forecasting, and the day's classification obtained opens up future possibilities for detailed meteorological and climatological analysis of the established types.

Daily atmospheric circulation catalogue for western Europe using multivariate techniques

“…As Esteban et al (2005) proposed for classifying heavy snowfall days, a PCA was conducted on the SLP data to derive the dominant patterns of variability from the raw fields (Richman, 1986;Yarnal, 1993). Later, K-means CL was used to classify the observations and to composite the circulation pattern maps (Yarnal, 1993).…”

“…Ekström et al used probably the most extended steps for the PCA and CL classifications: S-mode, correlation matrix, Varimax orthogonal rotation, and K-means CL, recently reproduced by Ekström et al (2004) for Australian dust storm analysis. Recently, in order to identify heavy snowfall day SLP patterns in the central Pyrenees, Esteban et al (2005) considered a spatial standardization of the previous data for S-mode PCA, and suggested an approximation for selecting the number of clusters and their centroids for use in the K-means method without iterations.…”

“…This is the reason why more research and classifications are needed, not only for analysing environmental problems but also for improving classifications on the short temporal scale. In this way, this paper uses the methodology proposed by Esteban et al (2005), with good results for extreme events analysis (heavy snowfalls), and applies it to the classification of a long temporal period on a synoptic scale and for the full year simultaneously.…”

“…The methodology used, previously proposed by Esteban et al (2005) to identify heavy snowfall days over Andorra (Pyrenees), involves a pre-processing approach consisting of a spatial standardization of the data used for the rotated PCA on S-mode analysis and correlation matrix, and Varimax rotation of the retained components. For the clustering process, an alternative approximation based on the PCA results (positive and negative phases) to decide the centroids and the number of groups for the K-means clustering is shown, followed by the rejection of the iterations for this non-hierarchical algorithm.…”

The influence of atmospheric circulation at different spatial scales on winter drought variability through a semi‐arid climatic gradient in Northeast Spain

“…Different studies have analysed the influence of weather-type frequency on precipitation in different Mediterranean regions, mainly for intense precipitation events (Romero et al, 1999;Kahana et al, 2002Kahana et al, , 2004Esteban et al, 2005). Nevertheless, a few studies on the influence of the frequency and persistence of weather types on droughts in this region have been published (Delitala et al, 2000;Olcina, 2001;Brunetti et al, 2002).…”