Snow cover is not a permanent winter phenomenon in the central European lowlands. It may even be considered as being ephemeral in the western part of the region. However, its occurrence during the winter season or during the winter–spring transition plays a very important role from the climatological point of view (Kożuchowski and Żmudzka, 2001; Wibig and G3owicki, 2002). Owing to its physical features (high albedo, low conductivity), snow cover strongly modifies the surface–atmosphere energy budget (Robock, 1980; Robinson and Kukla, 1985). Consequently, it modifies the weather conditions mainly by lowering the air temperature (Wagner, 1973; Dewey, 1977; Walsh et al., 1982), and by changing air circulation, cloudiness and precipitation (Johnson et al., 1984; Namias, 1985; Cohen, 2001).
An important characteristic of the occurrence of snow cover in central Europe is its intra- and inter-annual variability (Falarz, 2004). Winter is the most changeable season in this region, considering the weather patterns (Niedźwiedź, 1981). Cold periods with long- or short-lasting snow cover alternate with mild periods without any snow. Over several years in central Europe, cold and snowy winters, such as in the years 1962–1963, 1969–1970, 1995–1996, were followed by mild winters with hardly any snow, such as in seasons 1972–1973, 1974–1975, 1988–1989, 1989–1990. The severity and snowiness of winters in central Europe mainly depends on the atmospheric circulation. A number of previous studies have addressed the impact of large-scale circulation patterns such as the North Atlantic Oscillation (NAO) on persistence of snow cover. The relationships between the NAO index and the annual number of days with snow cover or snow depth are statistically significant in central Europe but they become less significant towards the east (Gutzler and Rosen, 1992; Clark et al., 1999; Bednorz, 2002, 2004; Falarz, 2007). Changes in the European snow extent have also been observed during other circulation patterns, such as the Siberian (SIB), which describes the strength and position of the Siberian High and the Eurasian Type 1 (EU1), which is characterized by the centre over Scandinavia (low-pressure area in the positive phase of EU1) and two opposite centres over northern Africa and eastern Siberia (Clark et al., 1999). The most coherent signals of snow cover associated with the EU1 pattern were found in south-western Asia and in central Europe. The SIB pattern influences the appearance of snow cover mostly in eastern Europe. Some Northern Hemisphere circulation patterns, such as the NAO, East Atlantic (similar to NAO, but shifted to the southeast) and Scandinavian (describing pressure anomalies over Scandinavia and Europe) influence snow accumulation in Bulgarian mountainous regions (Brown and Petkova, 2007).
Studies on the snow cover–circulation relationships on the daily temporal scale have been carried out in different regions of Europe. The synoptic classification of snowstorms in Austria has been developed by Schalko (1949), refined by Spreitzhofer (1999b) and Spreitzhofer (1999a, 1999b, 2000). They postulated two main typical weather patterns related to intense snowfalls in Austria, one with a north-westerly flow and the other with an extended low-pressure system over the western Mediterranean. Conditions underlying the formation of severe snowstorms over the Swedish east coast have been analysed by Andersson and Nilsson (1990) and Andersson and Gustafsson (1993). They have adopted the assumption of convective bands, meaning convection developing in cold air streams over a much warmer sea surface, which is responsible for the coastal precipitation maxima. Bednorz (2006, 2008a, 2008b) found low-pressure systems from different locations to be responsible for snowfalls in the German–Polish lowlands and in the Carpathian basin. Nowosad (1992) investigated the daily changes in the depth of snow cover versus the types of atmospheric circulation in the Bieszczady Mountains (south-eastern Poland). He found that western and north-western cyclonal types of circulation cause snowfalls most frequently, while eastern anticyclonal types are favourable to the persistence of snow cover. Synoptic conditions with Mediterranean lows resulting in heavy snowfalls at several stations in the southern part of central Europe have been analysed in a case study concerning episodes in February 2003 (Bednorz, 2004). Babolcsai and Hirsch (2006) have worked out detailed characteristics and synoptic classification of heavy snowfall events in Budapest for the period 1953–2003, consisting of eight weather types. Most of them were connected with different kinds of Mediterranean cyclones and additionally with secondary lows in north-western Europe.
The aim of this study is to describe the occurrence of snow cover and its intra-annual variability briefly in the central European lowlands and to find the synoptic conditions that are responsible for heavy snowfalls and persistence of snow cover in different parts of that region. Results of the study may be helpful in recognizing the circumstances contributing to the occurrence of snow cover in the central European lowlands and forecasting thereof.
2. Data and methods
The study is based on the snow cover data from 33 ground stations, covering 40 winter seasons (November–March) from 1960–1961 to 1999–2000. The area of the study spreads throughout Poland and neighbouring countries: Germany, Austria, Czech Republic, Lithuania and Russia (Kaliningrad District), excluding mountainous regions. The elevation of the stations is generally lower than 300 m above sea level (asl); only four stations are located at 300–500 m asl. Polish snow cover data were supplied by the Institute of Meteorology and Water Management (Warsaw). Data for the German stations were derived from Deutsches Meteorologisches Jahrbuch (1960–2000) published by the Deutscher Wetterdienst (Offenbach, Potsdam). The Russian and the former Soviet Union data were obtained from the Historical Soviet Daily Snow Depth (HSDSD) Version 2 (Armstrong, 2001) and partly from the Lithuanian Hydrometeorological Service. Data for the Austrian stations were derived from Jahrbuch der Zentralstalt fur Meteorologie und Geodynamik (1960–2000) published by the Central Institute for Meteorology and Geodynamics (Vienna). Czech snow cover data were supplied by the Czech Hydrometeorological Institute and data from Budapest were supplied by the Hungarian Meteorological Service.
Observations of snow depth are taken once a day at 6:00 Universal Time Coordinated (UTC) with the precision of 1 cm at the meteorological stations. The days with a snow depth of ≥ 1 cm are considered as the days with snow cover. The basic characteristics of central European lowlands' snowiness, such as the number of days with snow cover, maximum seasonal snow depth and dates of the first and the last occurrence of snow were computed and mapped. The inter-annual variability of the number of days with snow cover was described giving the extreme values, variability coefficient and standard deviation. Furthermore, regions of occurrence of snow cover in the central European lowlands were distinguished using the principal component analysis (PCA). The PCA was applied to the S-mode data matrix, the stations being the variables and the values of mean snow cover depth in pentads of a 40-year period being the observations. The number of principal components (PCs) to be taken for further analysis was established using the Gutman criterion (eigenvalues > 1, Jolliffe, 1993; Wibig, 2001). After applying an orthogonal normalized varimax rotation, the dominant modes of the variability of snow cover were identified at each station (Horel, 1981; Richman, 1986). The analysis of the spatial distribution of five PC loadings allowed distinguishing five separate regions of occurrence of snow cover.
Finally, synoptic conditions of heavy snowfalls and persistence of snow cover in each region were determined separately. To this end, the daily mean sea level pressure (SLP) and 500-hPa geopotential height data were selected from the National Centers for Environmental Prediction (NCEP)—National Center for Atmospheric Research (NCAR) reanalysis data (Kalnay et al., 1996). The synoptic area encompasses the region 35–70°N latitude by 35°W–40°E longitude with 2.5° resolution. Composite maps of the SLP and 500-hPa geopotential heights means and anomalies were constructed separately for the days with heavy snowfalls and snow persistence. Anomalies were computed as differences between composite values and 40-year seasonal means (November–March). The days with heavy snowfalls were defined as the days during which the depth of snow cover increased by ≥ 5 cm. Changes in the depth of snow cover were calculated by subtracting the depth of snow cover of a given day from the depth of snow cover of the following day. To this end, the data for winter months from December to March were applied. Snowfalls are sometimes influenced by local conditions; therefore, only the days during which an increase of ≥ 5 cm had been observed in snow at 40–100% of the stations in a particular region were taken into consideration. Periods of persistence of snow were defined as those on which there were no changes in snow depth (zero change in snow depth) for at least 3 days, which appeared simultaneously in 40–100% of the stations in a particular region. The composite analysis and anomalies have been used previously to identify the atmospheric circulation patterns associated with heavy snowfalls (Birkeland and Mock, 1996; Bednorz and Wibig, 2008).
Snowiness is both spatially and temporally variable in central Europe. The mean annual number of days with a snow cover range from about 14 days in the west to over 90 days in the northeast, considering lowland territories only (Figure 1(A)). However, these numbers may vary from several days, or even zero, during mild winters in Germany or Hungary, to over 140 days during extremely snowy winters in eastern Poland (Table I). The variability coefficient reaches 100% in the least snowy places and is 30–40% in the most snowy regions. The values of the standard deviation prove that the length of the snow-covered period may vary by 4–8 weeks, depending on the station. The maximum seasonal snow depth also increases from the west to the east, equalling from 7 to 8 cm in the west edge of Germany to over 30 cm in the northeast (Figure 1(B)).
Table I. Mean characteristics of the occurrence of snow cover
Number of days with snow cover
Maximum seasonal snow depth (cm)
Dates of occurrence of snow cover
Variability coefficient (%)
Data are for the years from 1960–1961 to 1999–2000.
The mean data of the first appearance of snow cover falls on the the last dekad of November in the east of the studied region and in the beginning of December in western Germany (Figure 1(C)). The snow season ends between the beginning of February (the 10th of February in north-western Germany) and April (eastern Poland, Lithuania and Kaliningrad District) (Figure 1(D)).
In the central European lowlands, snowiness increases meridionally from the south to the north and zonally from the west to the east. North-eastern Poland and Lithuania are characterized by the largest snowiness in the analysed area, while in western Germany (Mozela and Rhone river valleys) snow cover appears least frequently. Considering the spatial variability in the occurrence of snow cover, different regions of snow cover were distinguished using the PCA. The PCA was applied to the S-mode data matrix, the stations being the variables and the values of mean depth of snow cover in pentads of a 40-year period being the observations. The orthogonal normalized varimax rotation was applied to the five PCs with eigenvalues > 1, which explained more than 80% of the total variance. The highest values of loadings for rotated five components show a clear spatial distribution that allows dividing the central European lowlands into five coherent regions of occurrence of snow cover (Table II, Figures 2 and 3). The first PC is most strongly correlated with 11 stations located in the Polish–German lowlands (Figure 2, panel PC1). Therefore, the first region (PC1) encompasses northern Germany and north-western and central Poland. The second region (PC2) includes five stations of the southern part of Germany, which are highly correlated with the second component (Figure 2, panel PC2). The third group of the stations, which are best correlated with the third PC, spreads in the Alpine and Carpathian foreland (Figure 2, panel PC3, and Figure 3, region PC3). The fourth region (PC4) encompasses seven most snowy stations in the north-eastern part of the studied area (Figure 2, panel PC 4). The fifth PC (not included in Figure 2) is strongly correlated with the data for Budapest only (Table II). This proves that there exist different snow conditions in the southern part of central Europe behind the barrier of the Alpine and Carpathian ranges. Therefore, the case of Budapest has been analysed separately (Bednorz, 2008b) and only four remaining regions were considered for further analysis in this study.
Table II. Loadings of five rotated PCs
The highest value for each station is marked in bold.
During the colder half of the year, from December to March, the mean SLP is the highest (>1020 hPa) southeast of the Azores and it gradually decreases towards the north (Figure 4). A low-pressure centre (<1000 hPa) is located over the north Atlantic south west of Iceland. A smaller pressure gradient is observed in eastern Europe than over the Atlantic. The 500-hPa geopotential level is often used to study the upper-level flow as it strongly relates to the surface weather. In the colder half of the year, the mean height of the 500-hPa level declines from the southwest (5700 m over the Azores) to the north (5250 m over the northern Atlantic). Such patterns cause west and southwest airflows in both the low and in the middle troposphere dominating over the studied area in winter.
Composite maps of the mean and anomalous SLP and 500-hPa geopotential heights were constructed for the days with an increase in the depth of snow cover of ≥ 5 cm in 40–100% of the stations (as defined earlier) for each region separately (number of days given in Table III). The most typical characteristics of the baric conditions favourable for heavy snowfalls in central Europe are negative SLP anomalies and depressions of 500-hPa geopotential heights over the continent: they imply the presence of low-pressure systems in these regions (Figure 5). However, the intensity and location of the lows vary, depending on the region receiving heavy snow.
Table III. Number of days with heavy persistence of snowfall/snow in each region
Number of stations
Mean number of days with heavy snowfall in one station
Number of days with heavy snowfall in 40–100% of stations
Mean number of days with snow cover persistence in one station
Number of days with snow cover persistence in 40–100% of stations
Sixty-six days with heavy snowfalls in 40–100% of the stations in the PC1 region were selected from the 40-year period. A contour map of SLP, constructed for these days, shows a low-pressure system over central Europe with its centre (<1008 hPa) located exactly over the region and south of it (Figure 5, panel PC1-A). At the same time, higher-than-normal pressure is observed over the north Atlantic (Figure 5, panel PC1-B). Contours of 500-hPa height constructed for the days with heavy snowfalls bend to the north over the Atlantic and to the south in central Europe, suggesting northerly and north-westerly flow in the middle troposphere over central Europe. A composite anomaly map shows the SLP and 500-hPa height differences between the selected weather situations (days with snowfalls resulting in increases of ≥ 5 cm in the depth of snow cover in 40–100% stations) and the 40-year seasonal means (November–March). The interpretation of the contoured composite anomalies is similar to the traditional weather anomaly maps, with clockwise (anticyclonic) flow around the positive centres and counter-clockwise (cyclonic) flow around the negative centres (Birkeland and Mock, 1996). The centre of the negative SLP anomalies occurs over the Carpathian range and the centre of the negative 500-hPa height anomalies occurs exactly over the analysed PC1 region. Consequently, circulation during the days with heavy snowfalls is characterized by a strong easterly and north-easterly flow component over the Polish–German lowlands, in contrast to the average winter circulation in central Europe.
In the least snowy PC2 region (southern Germany), 55 days with an increase of ≥ 5 cm in snow depth were recorded in 40–100% of the stations. Heavy snowfalls in this region require a vast trough of low pressure over the western part of central Europe (Figure 5, panel PC2-A). The centre of negative SLP anomalies is located over the southern part of region PC2, while the centre of 500-hPa anomalies falls exactly over Germany (Figure 5, panel PC2-B). This indicates an easterly flow component accompanying heavy snowfalls in this region. At the same time, strong positive SLP anomalies (>12 hPa) are observed over the northern Atlantic, which may be associated with the negative phase of NAO.
In the PC3 region, 119 days were recorded with heavy snowfalls in 40–100% stations during the 40-year period. In this case, a contour composite map shows a low-pressure system shifted to the south, with a centre over the Italian Peninsula (Figure 5, panel PC3-A). Contours of the 500-hPa height are bent to the south, forming a trough over Europe. They cut the cyclone meridionally, which allows distinguishing a colder part of the cyclone in the north and a warmer part in the south. In the colder part, higher air density lowers the 500-hPa geopotential level. In the southern warm part of the cyclone, the 500-hPa geopotential height is over 200 m higher than that in the northern part. The PC3 region is placed on the cold north edge of the aforementioned cyclone, and heavy snowfalls in that area probably appear due to meteorological fronts in its colder parts. The easterly flow brings polar continental air to the PC3 region, and the southern low is a reservoir of humidity from the Mediterranean. Meeting of these two elements may result in heavy snowfalls in the southern part of central Europe.
In the most snowy PC4 region (north-eastern part of the studied area), 108 days with an increase of ≥ 5 cm in snow depth were found in 40–100% of the stations. Heavy snowfalls in this area correspond to a vast trough of low pressure over Scandinavia with a small, but rather deep centre formed over the Baltic Sea and east thereof (Figure 5, panel PC4-A). Strong SLP pressure anomalies (–10 hPa) fall exactly over the PC4 region (Figure 5, panel PC4-B). The Baltic low with frequent fronts brings cold and relatively humid arctic maritime air from the northwest. The same airflow direction occurs at the 500-hPa level.
Quite different synoptic situations underlie the persistence of snow cover. To this end, the anticyclonic type of circulation is necessary in all the regions (Figure 6). High-pressure systems spread from eastern Europe through the central part of the continent encompassing Germany and south Scandinavia, as presented in the map for PC4, or they even reach the British Islands, as in the case of the least snowy PC2 region (Figure 6, panels PC4 and PC2). Usually, the centre of high pressure falls over the studied region and east thereof, except in the case of PC2, in which the higher pressures spread north of the region. The persistence of snow cover in the central European lowlands is possible under the positive SLP anomalies (up to 12 hPa) over the northern Atlantic and most of the continent, except for its south-western part, where weak negative anomalies are observed (Figure 6, panel B). Such a pattern of SLP anomalies is reminiscent of the NAO negative phase and generates eastern or south-eastern airflow, which brings cold polar continental air to central Europe preserved in the anticyclonic system over eastern Europe. Furthermore, winter anticyclones usually bring very low temperatures, particularly in the centre, mainly due to the nocturnal radiation conditioned by a cloudless sky.
4. Discussion and conclusions
The most typical characteristics of the baric conditions favourable for heavy snowfalls in central Europe are the negative SLP anomaly and depression of 500-hPa geopotential heights over the continent, i.e. the presence of low-pressure systems in these regions. However, the lows may differ in intensity and location. In the south-western part of central Europe, heavy snowfalls usually appear due to meteorological fronts in the colder parts of Mediterranean cyclones, the centres of which are often situated over the Italian Peninsula. Another typical location of the cyclones or troughs of low pressure, causing heavy snowfalls, is the Baltic Sea region, where the meteorological fronts appear frequently. Such a location of cyclonic systems causes snowfalls particularly frequently in the most snowy north-western part of central Europe. The cyclonic activity over the continent, which brings snowfalls, is simultaneous to the weakening of the Icelandic Low and significant positive SLP anomalies over the north Atlantic. It enables the Azorean High to spread towards the northeast. Weakening of the Icelandic Low corresponds to the negative NAO phase, which has been proven to contribute to a large snow cover extent in Europe (Gutzler and Rosen, 1992; Clark et al., 1999; Bednorz, 2002, 2004; Falarz, 2007).
Spreitzhofer (1999b, 2000) has postulated two main typical weather patterns related to intense snowfalls in Austria. One of these, with a low over the Italian Peninsula and a high over Scandinavia, resembles the SLP composite contour map for region PC3 identified in this study. Scherrer and Appenzeller (2006) have defined an SLP pattern responsible for snowfalls in the Swiss Alps. It is characterized by low pressure over south-eastern Europe and it slightly resembles the conditions for heavy snowfalls in the PC3 region. In the SLP pattern favourable for the occurrence of snow days in the Swiss Alps (Scherrer and Appenzeller, 2006), the southern cyclone is weaker and its centre shifts further to the south, which brings eastern flow over Europe, similar to the baric conditions of the persistence of snow cover in the PC3 region.
The persistence of snow cover in the central European lowlands is possible under the positive anomalies of SLP over most of the continent, except for its south-western part, where weak negative anomalies are observed. The centres of positive anomalies are located north of the area under study. At the same time, positive anomalies of 500-hPa geopotential heights appear over the northern Atlantic and Scandinavia. Anticyclones, which are favourable for the persistence of snow cover, usually appear as vast and strong systems with centres exactly over or close to the analysed region and are always connected with the east-European high. One of the most important factors that enables the persistence of snow is the below zero temperature. Winter anticyclones usually provide very low temperatures, mainly because of strong nocturnal radiational cooling due to a cloudless sky. Furthermore, the anticyclones over eastern Europe provide eastern circulation, which, in the winter season, brings polar continental air and temperatures well below zero. Sometimes, high-pressure centres favourable to the persistence of snow cover located northwest of the studied region cause northern airflows.
This kind of study on the relationships between climatological parameters and circulation seems to be essential for recognizing climatic conditions, particularly for the colder half of the year when the insolation weakens and the impact of circulation on the climate is dominant in the moderate zone (Wibig, 1999). Circulation strongly influences the occurrence of snow cover in the central European lowlands. Snow cover is a non-permanent climatologic element in these regions and it may appear and disappear several times during any winter season. In the case of such a diversified cycle, separate analyses of the synoptic conditions associated with snowfalls and snow persistence, as undertaken in this study, may improve forecasts of the winter phenomenon in central European lowlands.