Long-term occurrence, variability and tracks of deep cyclones over Krakow (Central Europe) during the period 1900–2010

Authors


Abstract

This article discusses patterns in the long-term and seasonal occurrence of deep cyclones over Krakow. This study analysed the frequency of occurrence of air pressure values equal to or lower than the 1st percentile (equivalent to ≤995.3 hPa) of all air pressure values recorded at 12:00 UTC over a period of 110 years (1900/1901–2009/2010). Special attention was devoted to the tracks of deep cyclones. No distinct changes were found in the frequency of occurrence of deep cyclones during the study period. Overall the frequency peaked in December, but in recent years there has been an increase in frequency towards the end of winter and beginning of spring. A similar general lack of noticeable change in the number of days with deep cyclones can also be found in specific tracks. There were minor increases in the frequency of occurrence of cyclones from the Norwegian Sea (T1), the Atlantic (T3), Bay of Biscay (T6) and the Mediterranean (T7) after 1950. The study also found confirmation of the theory that cyclone tracks had shortened at their northeastern extremities. Copyright © 2012 Royal Meteorological Society

1. Introduction

The occurrence and movement of pressure systems are important components of atmospheric circulation. At a continental scale it is low pressure areas, especially those travelling from west to east with their associated systems of atmospheric fronts that generally have a significant influence on European weather. The passing of such a system is often accompanied by meteorological phenomena of a violent nature such as sudden changes of pressure and temperature, strong winds, heavy precipitation including hail, and electrical discharges. Very often these phenomena cause considerable damage to the environment and the economy and may adversely influence human health and well being. Very strong winds have been known to cause major damage and are classified second after floods as causes of natural disaster according to a report of the World Meteorological Organization (Cornford, 2002; Berz, 2005). In Europe, these hurricane-strength winds have been increasingly frequent and relatively regular in their occurrence over the last 10 years or more, including in countries such as the UK, France, Denmark, Germany and the Benelux countries (Leckebusch and Ulbrich, 2004; Leckebusch et al., 2008). Some of the natural disasters caused by these winds include coastal floods in Britain (Zong and Tooley, 2003) and the occurrence of sea-level extremes in the Gulf of Finland (Averkiev and Klevannyy, 2010). A series of three violent storms in December 1999 (named Anatol, Lothar and Martin), which particularly affected Central Europe, are also worth mentioning in this regard (Ulbricht et al., 2001; Berz, 2005; Dupont and Brunet, 2006). They contributed to the death of more than 130 people and caused 13 billion Euros worth of damage (Ulbricht et al., 2001; Berz, 2005). Other similar events include the Erwin/Gudrun cyclone of January 2005 (Soomere et al., 2008; Averkiev and Klevannyy, 2010) and an extremely strong wind that occurred in the Carpathian Mountains in November 2004 as a result of a deep cyclone over Central Europe (Unton-Pyziołek, 2005; Widawski and Łakomiak, 2008). This latter wind destroyed ca. 12 000 hectares of forests in the Slovakian Tatras and the Slovakian nickname ‘Vel'ká Kalamita,’ or the Great Calamity, describing the effects of the event has stuck even in professional literature.

At a time of ongoing debate about climate change and the impact of human activities, questions have been asked whether a further increase in the frequency and intensity of similar events might be expected in the near future. According to some scenarios, there will be an increase in both the speed (by 7–10%) and incidence (by ca. 20%) of winds in comparison with the last decades of the 20th century (Knippertz et al., 2000; Leckebusch and Ulbrich, 2004). Rockel and Worth (2007) claim that there will be a particularly noticeable increase in wind speed in Central Europe in January and February. IPCC (2007) researchers associated this potential effect with a change in the tracks of extratropical cyclones, including the strongest winter cyclones (with core pressures below 970 hPa), which have been noted to increase in frequency over the Northern Hemisphere in recent years (Schinke, 1993; Haak and Ulbrich, 1996; Lambert, 1996). In the Euro-Atlantic sector, the number and intensity of cyclones have increased particularly strongly in its northern regions (Wernli et al., 2003) and their tracks have moved further north, which has resulted in a noticeable decrease in the occurrence of cyclones below 55°N (Bartholy et al., 2006; Trigo 2006). Scenarios of future changes that take into account an increase in greenhouse gas (GHG) concentration suggest that the current changes are likely to continue. It should be expected that the incidence of shallower cyclones will decrease (Leckebusch and Ulbrich, 2004; Bengtsson et al., 2006), while deep cyclones will be even more frequent and their tracks will move further away (Carnell and Senior, 1998; Knippertz et al., 2000; Geng and Sugi, 2003; Leckebusch and Ulbrich, 2004; Lambert and Fyfe, 2006). According to Knippertz et al. (2000), cyclone tracks will move further north and east while Carnell and Senior (1998) anticipate the possibility of the shortening of the northeastern extremities of these tracks. In their study, Leckebusch and Ulbrich (2004) suggest that cyclonic activity will weaken above 60°N, but will strengthen below that line of latitude, particularly over the western part of Central Europe and the northeastern Atlantic. The difference between these scenarios for change in cyclone track activity may stem from a number of causes, such as the choice of model used, its spatial resolution, forecasting period and the GHG scenarios assumed. Additionally, it must be borne in mind that mechanisms leading to cyclone development and movement are very complicated and that the climate is constantly changing because of natural causes (Brayshaw et al., 2005). For this reason, forecasting changes in cyclonic activity is very difficult and assessing anthropogenic impacts is even harder.

The above considerations have led to the following questions: Are deep cyclones also increasing in frequency in Poland? Are their tracks changing in frequency? If these changes are real, is the likelihood of very strong winds increasing in Poland?

To find answers to these questions, the authors attempted to determine the long-term variability of frequency of those extremely deep cyclones which have an impact on the weather in Poland, particularly in its southern area around the city of Krakow. A special focus was placed on seasonal variability in the occurrence of these systems and on analysing the tracks of deep cyclones during the last 110 years (1900/1901–2009/2010).

2. Source material and methods

The study was based on the values of atmospheric pressure recorded at 12:00 UTC and reduced to sea level which were recorded at the Historic Weather Station of the Jagiellonian University in Krakow (50°04′N, 19°58′E, 220 m a.s.l.) during the period 1900–2010. The station is one of the few in Europe to possess a long and homogeneous record of air pressures (Trepińska, 1997) and has been demonstrated in available research to be representative of the climatic conditions in this European region (Kożuchowski et al., 1994). In the section devoted to the tracks of deep cyclones the authors used tracks identified by Bielec-Bakowska (2010) as well as sea-level air pressure maps and maps of 500 hPa geopotential height at 12:00 UTC (www.wetterzentrale.de). Specific cases of deep cyclones were described using data from observations made at the Historical Station in Krakow and synoptic maps.

In the climatological literature there are a number of ways to define a deep depression. For example, Schinke (1993) and Chen and Zhang (1996) define deep cyclones as systems with the core pressures equal to or lower than 990 hPa. Also Kłysik (1995) classified cyclones as deep ones if their core pressure was lower than (but not equal to) 990 hPa. The most frequent approach, however, is to classify low pressure areas as deep cyclones if their core pressures are equal or lower than 970 hPa (Lambert, 1996; Knippertz et al., 2000; Leckebusch and Ulbrich, 2004). In Poland, such values are extremely rare. Indeed, during the study period, there was only one instance of a pressure lower than 970 hPa (969.6 hPa on 26 February 1989) and even pressure values lower than or equal to 990 hPa were only recorded 102 times. For this reason, the definition of a day with a deep cyclone adopted for the purposes of this study requires that the pressure value recorded at the Krakow station at 12:00 UTC (reduced to sea-level) be equal or lower than the 1st percentile of all days in the period. In local terms, this corresponds to 995.3 hPa, which is not far off the 990 hPa criterion. Also the 1st percentile criterion is compatible with the universally adopted assumptions for identifying extreme phenomena (IPCC, 2007).

Particularly deep cyclones occur mostly during the cool half of the year (ca. 89% cases in Krakow) and especially in winter (ca. 56% in December–February). This fact combined with the treatment of the atmospheric circulation conditions of this half of the year as a ‘circulation season’ has led to the adopting of a non-calendar year as the basic year for the purpose of this study. Numbers of days with deep cyclones were calculated for the years starting on 1 July and ending on 30 June. Therefore, the first year of this study (1900/1901) is a period starting in July 1900 and ending in June 1901 and the last year (2009/2010) starts in July 2009 and ends in June 2010. Consequently there were 110 ‘years’ taken into account.

The second section, devoted to cyclone tracks, is based on the classification by Bielec-Bakowska (2010) which identifies seven tracks and a group of unclassified cyclones (Figure 1). This subjective (manual) classification was based on data from NCEP/NCAR reanalyses and on an analysis of sea-level and 500 hPa geopotential maps. To identify tracks of deep cyclones, the study identified days when the height of the 1000 hPa surface was lower than or equal to the 1st percentile of all cases at five grid points representing the area of Poland during the period 1971–2000 (Bielec-Bakowska, 2010). Finally, source areas and tracks of relevant cyclones were identified using synoptic maps and maps of the pressure field at 500 hPa. The tracks were then included to one of seven groups (subjectively identified by the author) characterized by a similar source area and path. Naturally, tracks identified in this way constitute a generalization of tracks of all cyclones included in a given group, which means that the actual tracks of some of them may differ from the ones marked on the figure.

Figure 1.

The tracks of deep cyclones over Poland (1971–2000) (according to Bielec-Bakowska, 2010)

Each of the cyclone days (≤995.3 hPa) studied in the period (1900/1901–2009/2010) was included to one of the previously identified seven tracks or to an unclassified group (Figure 1). This was achieved by analysing maps at sea-level pressure and maps at 500 hPa. In the next step, an analysis was made of the annual and long-term variability of occurrence of cyclones along each track.

A couple of reservations are in order before the results can be read meaningfully. The values recorded at the Krakow weather station are likely to be higher than the core pressures of the relevant cyclones. This applies specifically to such cyclones, which were deep enough to exert a powerful influence on the region's weather, but at the time of measurement only reached Krakow with their peripheries yielding pressure values higher than the cut-off point adopted in the study. In such cases, the particular days would have been excluded from the analysis altogether.

3. Days with the lowest air pressure—the annual and long-term perspectives

During the period 1900–2010, the values of air pressure recorded in Krakow at 12:00 UTC ranged from 969.6 hPa (26 February 1989) to 1058.6 hPa (23 January 1907). Most of the values remained within the range 1000–1040 hPa and the most frequent range of air pressures was 1010–1020 hPa (from about 30% in December to about 70% in July). The average monthly values showed little variation in Krakow, but air pressure was slightly higher in winter than in the warm season. During the study period, the differences range from 1016.4 hPa in March to 1018.9 hPa in January while summer values stayed around 1014–1015 hPa (Figure 2). The highest and the lowest pressure values were typically recorded in the cool season. This is confirmed by long-term monthly pressure amplitudes, which were at their highest in the winter months of January (82.1 hPa), February (76.0 hPa) and December (74.3 hPa) and at their lowest in summer months from June to August (36.8 hPa, 32.2 hPa and 37.0 hPa, respectively).

Figure 2.

Atmospheric pressure in Krakow during the period 1900/1901–2009/2010 at 12:00 UTC

Taking this into account, the study only looked at the days with an air pressure lower than or equal to the 1st percentile of all air pressure values recorded at 12:00 UTC, i.e. ≤995.3 hPa. These days are referred to below as days with deep cyclones.

During the study period (1900/1901–2009/2010), in Krakow days with deep cyclones occurred virtually exclusively during the cool season (October–April) when 96.1% of such days were recorded (Figure 3). The frequency peaked in winter (December–February) at 55.9% and December was the month with the highest frequency of 23.7%. Spring (March–May) was noteworthy for its higher frequency of deep cyclones, at 24.9%, than autumn (September–November), at 18.0%, a fact linked to a known pattern involving a higher frequency of anticyclonic systems in early autumn in Poland (Buchert, 1992; Woś, 2010). During the study period, there were only 5 days with deep cyclones in summer (June and August), and none of these occurred in July (Figure 3).

Figure 3.

Annual course of the frequency of the occurrence of deep cyclones (≤995.3 hPa) in Krakow in the period 1900/1901–2009/2010

The annual pattern of days with deep cyclones varies widely from year to year. Typically, these systems were only recorded on 1 or 2 d a month during 2–3 months in a given year. There were only 23 months altogether with 3 days with deep cyclones and 10 months with four such days. Eight (March 2008) or six such days (March 1981) in a month must be regarded as extreme. Also unusual were years, when deep cyclones were recorded during 5 months (1918/1919, 1969/1970 and 1998/1999). In these years, there was a significant increase in the frequency of deep cyclones in spring. Often, however, monthly numbers of days with very low pressures were so low that it was difficult to identify a month with the maximum frequency of occurrence. Where this was possible the maximum identified fell most typically in December, especially during the second half of the 20th century (Figure 4). A trend was also identified in which the number of days with deep cyclones increased in the last few decades of the study period and there was also a growing frequency of deep cyclones occurring in late winter and early spring which pushed the annual maximum towards springtime.

Figure 4.

Number of days with deep cyclones (≤995.3 hPa) in particular months in Krakow in the period 1900/1901–2009/2010

During the entire study period, there were on average 3.7 days per year with deep cyclones recorded in Krakow. There were also 8 years, when no such days were recorded, including 4 before 1952 and the remaining 4 no later than 1976. At the other end of the scale there were years with 11 days with deep cyclones (1954/1955) followed by 9 days each in 1998/1999, 2007/2008 and 2008/2009 (Figure 5).

Figure 5.

Long-term variability of the number of days with deep cyclones (≤995.3 hPa) in Krakow in the period 1900/1901–2009/2010

In the study period, there is a pattern with roughly 10-year periods that appear regularly and that are characterized by higher numbers of days with deep cyclones (more than 40 occurrences per 10 years). Conversely, a particularly low number of such days (27) was recorded in the 1940s, while the average of calculations for 10-year periods was 37.3. Despite the fact, however, that the average annual number of days with deep cyclones started picking up slightly in 1951 (3.9 vs 3.5 days in the first half of the 20th century) no significant trend was detected for the entire study period. It is worth noting that a certain decrease in the number of days was recorded in autumn, while an increase was recorded in winter and spring.

4. Tracks

The study also investigated the tracks of deep cyclones that influenced the weather in Krakow during the study period. A classification of deep cyclone tracks passing through Poland involving seven groups of tracks (Figure 1) was used for this purpose (see Section 2). The original classification also involved between two and four sub-groups of tracks in three of the main groups (T1–T3), but the slim sample size in this study allowed for just the main groups of tracks to be used (T1–T7).

Results clearly show that one-half of the deep cyclones (50.7%) arrived from the Atlantic (T1–T3). This large group was dominated by systems passing over Iceland towards the east and northeast (T2, 24.1% of all) and over the British Isles (T3, 18.0%) moving eastwards (Figure 6). Depressions moving along these tracks pass over the North Atlantic Current and from it they receive energy (Bąkowski and Piotrowicz, 2007), which often boosts their activity. This was, for example, the route followed by the Erwin/Gudrun cyclone, mentioned above for its significant impact in increasing the level of the Baltic Sea, especially in the Gulf of Finland (Averkiev and Klevannyy, 2010). This cyclone, however, only brushed Krakow and recorded higher values than those adopted for this study (1018.4 hPa).

Figure 6.

Frequency of days with occurrence of deep cyclones tracks over Krakow (1900/1901–2009/2010)

Quite a large group of cyclones arrived in Krakow from over the Mediterranean moving northeastward (T7, 12.0%), while depressions originating over the mid-Atlantic, west of the British Isles and moving over Germany to southern Europe, were the rarest group influencing the weather in Krakow (T5, 2.4%). Cyclones from groups T4 and T6, which travelled from the Mediterranean and the Bay of Biscay across Poland into Northeastern Europe, occurred with the frequency of 6.3 and 9.5%, respectively (Figure 6). Finally, the fact that there was a relatively large group of low pressure systems grouped as unclassified (X, 19.0%) would confirm a considerable degree of complication which is characteristic of circulation mechanisms and their apparent ‘capriciousness.’ There was a shortening of some of the tracks observed towards the end of the period and especially during its last decade. This was particularly true of those deep cyclones travelling from the Atlantic along one of the tracks identified that would ‘suddenly’ lose power upon arrival over the Scandinavia or the Baltic Sea. This may be evidence confirming the theory about a shortening of cyclone tracks proposed by Carnell and Senior (1998).

Each of the groups of deep cyclones defined had a slightly different annual occurrence pattern. The cyclones are mostly concentrated between November and March with the exception of systems arriving from the south, i.e. from the Bay of Biscay and the Mediterranean (Figure 7). The largest group of cyclones, passing from Iceland (T2) and the central Atlantic (T3), showed a very significant degree of concentration in wintertime (accounting for 91.9 and 94.6% of all occurrences). The annual pattern shown by the T2 group involves a sharp increase in occurrence between October and January/February and a rapid decline in March. Group T3 stands out with a strong concentration in December (34.3%) that is two to three times greater than in all other months in the period between November and March. Also cyclones arriving from the Norwegian Sea (T1) and from the North Sea (T4) predominantly occur in the cool season (88.6 and 96.2% respectively), but they are fewer overall and are evenly distributed between all the months of that season. The depressions arriving from the Bay of Biscay (T6) and the Mediterranean (T7) have different patterns. The former group is mostly active in November–December and March–April (87.2% of all cases). More than half of the Mediterranean depressions (63.3%) affect the weather in southern Poland in Spring, and especially in March and April (55.1%).

Figure 7.

Number of days with deep cyclones (≤995.3 hPa) over Krakow from particular groups of tracks (T1–T7) in the period 1900/1901–2009/2010—annual pattern

There is also a considerable difference in long-term variability of occurrence between deep cyclones belonging to different track groups. As it was mentioned above, during the 8 years of the period considered no such deep systems were recorded. Still, depending on the track of cyclones the number of years when there were noticed varied from 9 (T5) to 57 instances (T2) (Figure 8), which means that even the most frequent group of Icelandic depressions (T2) was absent in as many as 53 years of the study period.

Figure 8.

Number of days with deep cyclones (≤995.3 hPa) over Krakow from particular groups of tracks (T1–T7) in the period 1900/1901–2009/2010—long-term variability

A more detailed analysis revealed that some of the track groups became slightly more frequent from the mid-20th century. Groups T1, T3, T6 and T7 stand out particularly, as close to 60% of their occurrences were recorded after 1950 (between 59.0% for T6 to 62.9% for T1) (Figure 8). Group T2 was slightly more frequent at the beginning of the study period, but it also occurred frequently in the second half of the 20th century.

5. Deep cyclones—special cases

The occurrence of deep cyclones is associated with rapid and radical weather changes. A drop in pressure causes strong winds that can cause considerable material damage and adversely influence the human body. These phenomena are normally also accompanied by thunderstorms, heavy precipitation and rapid temperature changes. The scale and intensity of their impact depends on the cyclone's depth and the distance of its centre from the affected area. Even if some cyclones meeting the pressure criteria result in only minor weather changes, but very deep systems that persist for days invariably cause dramatic weather changes and significant damage. Examples of such impacts in Krakow are given below.

5.1. March 2008 - 8 days with deep cyclones

Most of the instances of deep cyclones during the study period involved either isolated days (60.5% of all days) or two-day sequences (30.2%). Three-day events were much less frequent (5.1%), while 4 days with a deep cyclone happened only three times (1922, 1976 and 1989) accounting for just 2.9% of all days. A five-days event recorded in March 2008 may be regarded as particularly extreme, especially given that it was accompanied by three more days with deep cyclones, including a two-day sequence, in the same month.

The two-day sequence opened the series (1–2 March). It was a Greenland-born cyclone that followed the T2 track over Iceland, the Baltic Sea and Finland. Its lowest pressure values at the centre did not exceed 960 hPa. During the night of 29 February/1 March, Krakow experienced a sudden pressure drop (by 22.3 hPa in 12 h), which lasted until 2 March when 981.2 hPa was recorded. Accompanying fronts brought thunderstorms and hailstorms and produced 13.1 mm of precipitation. Strong winds, which reached a top speed of 9 m·s−1 and gusts up to 20 m·s−1 at the evening measurement time, broke trees and damaged buildings. After the passage of the front, the wind abated to 6 m·s−1 (gusts up to 13 m·s−1), the temperature ranged from 4.5 to 6.9 °C and the precipitation totalled 10.7 mm on the second day (Figure 9).

Figure 9.

Changes of sea-level pressure (at 12:00 UTC), average daily temperature and daily precipitation totals during selected days of deep cyclones that influenced the weather in the Krakow region

Meanwhile, another Atlantic cyclone was following the T3 track eastwards over the British Isles and the Baltic. Its centre was as low as the previous one (<960 hPa) and it's lowest values were recorded in Krakow on 12 March. On that day, the lowest of the three values measured was observed at noon (992.9 hPa), when the highest wind speed was also recorded (5 m·s−1 with gusts up to 13 m·s−1 during the day). Interestingly, the calm occurred both in the morning and in the evening.

The final cyclone event of the month was also the most severe. Between 21 and 25 March, a group of Atlantic-borne cyclones (21–23 March) and of a Mediterranean system (24–25 March) passed over Poland and produced five solid days with a pressure equal or lower than 995.3 hPa measured at 12:00 UTC.

The Atlantic cyclone group was connected with a powerful and vast anticyclone over the North Atlantic that forced a cyclone from Iceland to move over Poland (with the centre values ranging from 975 to 980 hPa). The cyclone then began to fill in and retired towards Finland. In Krakow, this was accompanied by a considerable drop in pressure (22 hPa within 24 h starting on 20 March at 18:00 UTC) and strong winds (6 m·s−1 with gusts up to 14 m·s−1). Accompanying snowfall left a cover of 1 cm (Figure 9).

In the meantime, another cyclone developed over the Mediterranean and reached Poland via the T7 track on 24 March. The lowest value linked to that system in Krakow was 990.0 hPa and it produced much weaker winds (4 m·s−1, gusts < 10 m·s−1). Positive temperatures were recorded throughout the 24 h period (up to 6.8 °C) and accompanying precipitation ranged from rain to sleet and snow (!).

5.2. Winter of 1954/1955 - 11 days with deep cyclones

During the three winter months of 1954/1955, 11 days with a deep cyclone were recorded: 7–8 and 22–24 December; 11, 14 and 17 January; and 17–18 and 20 February. Most of the systems came from Iceland (T2) and the Norwegian Sea (T1) with some from the Atlantic (T3). The cyclones were not particularly deep (centre values around 980 hPa) and produced pressures ranging from 987.0 to 995.3 hPa in Krakow. One exception was a cyclone passing in late December, which had a pressure of less than 965 hPa at the centre and 977.8 hPa in Krakow. The passage of such powerful systems caused a considerable temperature increase, intense precipitation and strong winds. The December events provide a good example with air temperatures exceeding 5 °C on 50% of all days and peaking at 14.3 °C (10 December)—a very high value for a winter month. January and February were cooler, but the deep cyclones always caused thaws on the day of their passage (Figure 9).

5.3. Minimum pressure in Krakow on 26 February 1989

Exceptional event was the recording of the absolute minimum pressure of the study period. The value of 969.6 hPa was observed on 26 February 1989, at the time of a very deep cyclone (with less than 955 hPa at its centre), which developed 2 days earlier over the British Isles and the North Sea. The system covered virtually the whole of Europe and stagnated for more than a week, after which it filled in and moved towards the Baltic Sea. This coincided with the development of two anticyclones over the Atlantic and Russia, which were connected by a higher pressure zone over southern Europe. The systems were also reflected in higher pressure levels and their relative location caused powerful zonal air flow over Europe.

In February 1989, the weather in Krakow was exceptionally warm. On 25 and 26 March, it reached 17.2 and 15.6 °C, respectively. The minimum air temperature at night never dropped below 9 °C, while the wind peaked at 4 m·s−1, at midday on 25 February (Figure 9).

6. Discussion and conclusions

European climate is strongly determined by the occurrence of mobile cyclones, most of which come from the Atlantic along habitual tracks. For this reason, even a minor, but systematic, change in their frequency of occurrence, intensity or track course, especially in the case of deep cyclones, may have a long-term impact on the continent's climate, including on local climates (Bengtsson et al., 2006). Therefore, an attempt was made to analyse the long-term variability of deep cyclones that influenced the weather in the Krakow region, which was viewed as providing a good representation of Central European circulation conditions.

The study found that the frequency of deep cyclones in Poland, both overall and in each of the track groups, failed to change significantly. A certain increase in the number of days with deep cyclones that started at the beginning of the second half of the 20th century, fell short of statistical significance. The highest numbers of days with deep cyclones were recorded in 1950 and the 1960s and after 1998 (9–11 per year). During the whole study period, the frequency peak fell typically in December, but in a recent trend, noticed in the last few decades, the frequency increased in late winter and early spring. This would suggest that the annual maximum frequency of deep cyclone occurrence is shifting towards spring.

More than half of all deep cyclones developed over the Atlantic and travelled over or near Iceland via the Baltic Sea and/or the Scandinavian Peninsula and further on eastwards or northeastwards (T1 and T2) and across the British Isles eastwards or northeastwards (T3). Towards the end of the study period, it was observed that deep cyclones following these tracks shortened their journeys considerably. As they moved over the Scandinavian Peninsula or the Baltic Sea they ‘suddenly’ weakened and filled-up. This would confirm a theory by Carnell and Senior (1998) according to which cyclone tracks are shrinking at their northeastern ends.

These results were then compared with track-specific frequencies of deep cyclones recorded during the period 1971–2000 and influencing the weather in Poland (Bielec-Bakowska, 2010). Because of differences between circulation conditions in various regions of Poland the study focused on the grid point 50°00′N 25°00′E (southeastern Poland, PLSE). This point represents the southeastern region of the country in which Krakow is very often included. The overall pattern remained unchanged, but certain differences appeared (Table I).

Table I. Frequency (%) of days with occurrence of deep cyclones tracks in the period 1971–2000
Grid points/stationTracks
 T1T2T3T4T5T6T7x
  1. x, unclassified cases; PLNW, PLNE, PLC, grid points representing particular regions of Poland: NW, northwestern; NE, northeastern; C, central. Data concerns grid points and Poland according to Bielec-Bakowska (2010).

PLNW7.242.331.55.40.95.40.07.2
PLNE12.645.021.65.40.04.50.99.9
PLC8.837.225.78.00.98.03.58.0
PLSW7.020.225.45.34.410.56.121.1
PLSE6.222.124.86.20.96.223.010.6
Poland8.429.326.83.82.46.310.512.5
Krakow10.525.721.98.61.910.510.510.5

In most cases, the frequency of deep cyclones that influenced Krakow was greater than that in the farthest southeastern corner of Poland. The greatest increases were found in deep cyclones from the Norwegian Sea (T1, up by 4%) and the Bay of Biscay (T6, up by more than 4%). Decreases were recorded in cyclones from the Mediterranean (T7 by more than 12%) and the mid-Atlantic (T3 by 2.9%). It is also interesting to note that sometimes deep cyclone frequencies in Krakow are more similar to those recorded in south-western Poland than in the easternmost sections of the southeastern region (tracks T1, T6, T7; Table I). The comparison of frequencies of occurrence along the identified tracks confirms the existence of quite evident differences in circulation conditions in various regions of Poland. This is due to Poland's location in Central Europe at a contact point between maritime and continental air masses, visited by pressure systems arriving from the Atlantic, Arctic, Asia, and the Mediterranean.

The study also attempted to assess potential similarities between the frequency of deep cyclones in Krakow and elsewhere in Europe. This proved difficult due to a want of equally long pressure records with uniform observation times and daily minimums. These were replaced by average daily pressures at sea level at nine European stations for the period 1901–2009. The data was sourced from the European Climate Assessment & Dataset (ECA&D) (http://eca.knmi.nl) and the Krakow station. The stations selected (Bodo, Haparanda, Bergen, Visby, Hammer, Berlin, Krakow, Vienna and Lugano) are located at altitudes ranging from 5 to 273 m a.s.l. along a belt 20° in longitude (05°19′–24°07′E) and a similar latitude (46°00′–67°16′N) (Figure 10). The data selection implies that the comparison involves particularly low daily average pressures rather than individual deep cyclones. Having admitted that, however, it seems that this approximation reflects the trends of the changes studied quite well.

Figure 10.

Location of meteorological stations

The analysis began by looking at the degree of similarity between pressure changes at selected European stations and in Krakow. The results (Table II) confirm a high degree of similarity of changes observed in atmospheric circulation in Central Europe and even between Krakow and Scandinavia and the Baltic Sea (correlation coefficients between 0.49 and 0.95).

Table II. Correlation coefficient (R) between average daily values of atmospheric pressure recorded in Krakow and at selected stations in Europe in the period 1901–2009
StationBodoHaparandaBergenVisbyHammerBerlinViennaLugano
  • a

    Confidence level 0.01.

Correlation coefficient (R)0.1820.198a0.486a0.679a0.813a0.898a0.946a0.816a

The next step involved particularly low pressure values. At each station the number of such days per year was calculated and the year was defined from 1 July to 30 June. Three methods of identification of the days were adopted. The first permitted the assessment of pressure changes vis-a-vis a particular place using pressure values equal or lower than the 1st percentile of all measured values. These values were found to vary from 975.0 hPa at Bodo to 997.0 hPa at Lugano (Table III). This means that the value of 990 hPa regarded in literature as the threshold of deep cyclones is ‘too high’ for areas above 60°N and slightly ‘too low’ for Central Europe. Two other methods involved days with air pressure equal or lower than either 990 or 995 hPa. In this way, it is possible to compare atmospheric circulation over larger areas and to assess to what degree the occurrence of very low pressure values (≤990 hPa) reflected very short episodes or true circulation conditions at the time.

Table III. The average daily air pressure values equal to the 1st percentile of all cases at selected stations in Europe in the period 1901–2009
StationBodoBergenHaparandaVisbyHammerBerlinViennaKrakowLugano
Value of the 1st percentile975.0977.5977.9982.5987.2991.8996.0996.4997.0

Regardless, however, of the criterion adopted no significant long-term trends were found in the number of days identified in this way (Figure 11). Stations to the north of Poland did show an increased frequency of low pressure values in the second half of the period, while in the remaining stations these values were more frequent early in the study period, but none of these tendencies were statistically significant during the entire period. It was also difficult to identify periods with a clear increase or decrease in the number of days with low pressure that would coincide at all of the stations. There were, however, certain features of changes in the pressure values that permitted grouping at regional level. In most cases the similarities did not span the entire study period, but covered up to several years and the strength of the similarity (of an increase or decrease) declined with the distance from a station. It is also worth noting that further northwards was a station, the greater the change in the number of days with particularly low pressures and the more difficult it was to determine any specific period of such increased frequency.

Figure 11.

Number of days with extremely low average daily values of atmospheric pressure at selected station in Europe in the period 1901/1902–2008/2009: (a) values calculated as air pressure values equal to or lower than the 1st percentile of all cases, (b) pressure values equal to or lower than 995 hPa (grey) and 990 hPa (dark grey)

The passage of cyclones over Europe is strongly related to the general atmospheric circulation in this area. The development of cyclones over the North Atlantic and their eastwards moving across the continent is of particular importance here. The Mediterranean is another important area of cyclogenesis, from which cyclones travel into eastern and northeastern Europe. For this reason, the authors decided to examine the extent, to which changes in the number of deep cyclones over Krakow were linked with changes in the larger scale circulation. To accomplish that, the study attempted to correlate the number of days with deep cyclones over Krakow with a circulation indices calculated for southeastern Poland (Niedźwiedź, 2011) and with NAO (http://www.cru.uea.ac.uk).

Due to the small number of the category of days under consideration it was not possible to determine their significant relationships with the selected circulation indices. The only statistical significance, albeit weak, was identified in a relationship between the number of days with deep cyclones and the cyclonicity index (C) in southeastern Poland (Table IV). The existence of this relationship was to be expected, as it reflects the number of days with the types of circulation that are known to be linked with cyclones. It is the lack of any direct relationships between the number of deep cyclones and other types of circulation situations, which suggests that causes to the changes in the occurrence of deep cyclones must be sought elsewhere. Perhaps changes in ocean temperatures and ocean circulation patterns may be the next areas to consider here.

Table IV. Correlation coefficient (R) between number of days with deep cyclones over Krakow and atmospheric circulation indices: NAO, zonal (W), meridional (S) and cyclonic (C) in the period 1900/1901–2009/2010
IndicesNumber of days with deep cyclones
 YearAutumnWinterSpringSummer
  • a

    P, S, C—indices according to Niedźwiedź (2011). 0.369, value statistically significant at p < 0.05.

NAO0.008− 0.100− 0.002− 0.0800.044
Wa0.1830.1530.1020.082− 0.023
Sa− 0.133− 0.0320.019− 0.101− 0.133
Ca0.1230.1510.3690.1490.168

The results presented here may suggest that in Central Europe deep cyclones should be defined as ones with a central pressure no higher than 995 hPa, while systems with pressures lower than or equal to 990 hPa should be defined as particularly deep. The study failed to clearly confirm any increase in the frequency of particularly deep cyclones, which means that forecasts envisaging higher frequencies of strong winds accompanying deep cyclones must be treated with caution. Having said that, strong wind is not just the result of the pressure value but also of its gradient and therefore future research should take this aspect into consideration as well. It may also be that a greater change in the occurrence of deep cyclones and accompanying phenomena should be expected on an annual basis than in the long-term. Such change could be a result of an overall change in atmospheric circulation that is reflected in such phenomena as shifts in the timing and duration of certain seasons of the year.

Acknowledgements

This study was supported by a grant from the Ministry of Science and Higher Education (N306 0479 39).

Ancillary