1. Introduction
One of the main goals of synoptic climatology is not just to classify atmospheric circulation types according to certain rules assessing baric fields but also to compare the frequencies of the types with a great variety of meteorological elements and to study the dependencies between atmospheric circulation patterns and meteorological phenomena. Numerous studies have been dedicated to investigating extreme events during specific synoptic conditions and their trends over recent decades, which is an important task, especially in the context of climate change. For example, hot spells or heat waves and their relationship with atmospheric circulation are explored in [
1,
2,
3,
4] for different regions of Europe and in [
5,
6,
7,
8,
9,
10] for other parts of the world. Synoptic analyses of cold spells have been performed for many regions and focus either on climatological aspects [
11,
12,
13] or on concrete extreme cold periods [
14,
15,
16]. Cases of extreme precipitation with the potential to cause flood hazards and other dangerous events such as landslides are also depicted in a lot of research [
17,
18,
19,
20,
21,
22,
23,
24,
25]. There are several approaches to composing circulation types in atmospheric circulation classifications. Some of them use the direction of the air masses near the land surface as a main variable for naming and numbering the baric formations of circulation types, such as the classification by Lamb [
26,
27] for the British Isles.
The other main approach to determining circulation types is to use the position of the centre of the baric formation (cyclone/anticyclone) towards the territory of interest. Such classifications include the widely known “Grosswetterlagen”, composed by Hess and Brezowsky for the central parts of Europe [
28,
29,
30], Maheras’s classification composed for Greece [
31,
32,
33], and the classification for Bulgaria [
34,
35,
36], used in the present work and explained in detail in [
34]. The latter classification is made for both SLP (sea level pressure) and 500 hPa geopotential height at approximately 5.5 km above sea level in the middle troposphere. A great dependence on the atmospheric circulation between both baric levels occurs. In the front part of the high trough, an airflow divergence of 500 hPa exists, which causes a decrease in the air pressure at that level. Meanwhile, this is a reason for upward movements between the surface level and the 500 hPa level, i.e., the convergence of the airflow near the surface and, respectively, the cyclonic circulation there. The opposite processes are valid in the back parts of the high trough, where a convergence of the airflow exists, followed by downward movements, as the result is anticyclonic circulation near the surface, below this part of the high trough. This is one of the reasons why research dedicated to atmospheric circulation classifications at 500 hPa makes sense.
In the northern hemisphere, the airflow at the front of the high trough at 500 hPa comes from the southern quarter, bringing warmer air masses to the investigated area. For the Balkans, these air masses originate from the Mediterranean or North Africa. In the back parts of the high trough, the winds are from the northern quarter and cause cold penetration in the middle troposphere. Many studies investigate the dependencies between the circulation at 500 hPa and various meteorological manifestations. The connection between synoptic types at 500 hPa and fires has been examined for Canada [
37,
38], Alaska [
39], the western Mediterranean region [
40,
41], and Australia [
42]. In the field of air quality, atmospheric and synoptic conditions are examined for continuous and severe weather pollution in many regions like China [
43,
44,
45,
46,
47,
48], other parts of Asia [
49,
50], the Arctic [
51], and the Iberian Peninsula [
52]. Other articles assess the influence of circulation at 500 hPa on climatological seasons [
53,
54] or concentrate on specific cases or consequences of some meteorological phenomenon [
55,
56].
Considering temperature as a meteorological element, studies mostly concentrate on their minimum or maximum values and, to a lesser extent, on their average values. This is associated with the fact that the minimum and maximum temperatures are extreme values, which are of interest if the purpose of the study is to investigate extreme meteorological phenomena. The average daily or monthly temperatures could also be used to explore extreme events because the minimums and maximums might be considered as a function of the average values, but the choice depends on every concrete research goal. Regardless of which temperature parameter is chosen, a predefined rule is required to define an extreme temperature event. One of the solutions is to choose a constant number as a threshold, but the problem with using this approach is that it is defined for a particular geographic region, as the value varies and is dependent on latitude and climate characteristics, altitude, and the season of the year [
57,
58]. This is the main reason why relative thresholds of the empirical distribution of the temperature, or so-called percentiles, take part and are widely used. This method allows for comparing data gathered from meteorological stations in different geographical areas and with different climates and was proposed in [
59,
60]. Therefore, it is ensured that a given part of the temperature observations are extreme by definition, regardless of their range or impact. [
61,
62]. Temperature extremes are most often determined using percentiles ranging from the 1st to the 10th for cold days and from the 90th to the 99th for warm or hot days and are presented and commonly used in various climatological indices [
63,
64,
65]. Percentile indices could be defined for different time frames. For example, [
66,
67] calculates the percentile for the entire calendar year, for the entire summer season (June, July, August) [
68], and separately for every summer month [
69].
The main goal of the current work is to investigate the relation between the anticyclonic circulation types at 500 hPa, formed according to the rules in [
34], and summer seasonal temperature, respectively, for every summer month in the period 1961–2020, which includes the last two climatological periods 1961–1990 and 1991–2020. A comparison between extremely warm and cold seasons for these two periods is also performed. The thresholds of the 10th and 90th percentiles of the examined 60-year period are used for determining the extreme cold and warm temperature months and seasons. This approach is applied not only to the temperature time series but also to the frequencies of the circulation types. Conclusions are drawn for each type based on its correlation with seasonal and monthly temperatures.
2. Materials and Methods
The studied area includes climatological data from 15 meteorological stations (
Table 1 and
Figure 1), which are relatively evenly distributed and situated in different climate zones. Thirteen of them are within the territory of Bulgaria, and two, Nis (Serbia) and Calaras (Romania), are situated near the Bulgarian border. The Musala and Murgash stations represent high mountain areas above 2000 m a.s.l. and mid-mountain areas between 1000 and 2000 m a.s.l. According to the Köppen classification [
70], they have a typical mountain climate, corresponding to E and Dfc zones. The stations Sofia, Kjustendil, and Razgrad, at an altitude between 346 and 586 m a.s.l., correspond to the climate zone Cfb. It is characterized by a hot summer and cool winter with relatively evenly distributed annual precipitation. The southernmost stations in the studied area are Sandanski and Kardzhali, which fall into the Csa climate zone, characterized by hot and dry summers, mild winters, and maximum precipitation in the cold half of the year. The remaining stations are in the Cfa zone.
The period of the study was 1961–2020, and the monthly and correspondingly summer seasonal temperature values were calculated according to the climatological methodology [
71,
72], based on the daily average values derived from three fixed-in-time daily observations at 7, 14, and 21 local standard time, as the evening observation was weighted twice (1).
Summer temperatures were obtained by averaging the monthly values for June, July, and August.
The atmospheric circulation classification used in the study was carried out using a subjective (manual) approach to the circulation types. This means that every day in the examined period was classified manually, according to some predefined rules, without using an automatic execution of a numerical algorithm. The classification was based on the data and isolines for a geopotential height at 500 hPa, derived from 20th-century reanalyses, NCEP CFSR/GFS reanalyses, and NCEP/NCAR reanalyses [
73,
74,
75]. The visualisations from wetter3.de were mainly used for this purpose, where the data for these three reanalyses were united, covering different periods [
76]. We considered the atmospheric circulation to be of an anticyclonic type in cases where a certain anticyclone covers the territory of Bulgaria, regardless of where its centre is located. According to the position of the centre of the anticyclone relative to the territory of Bulgaria, five types of anticyclonic circulation at 500 hPa were determined. The centre of anticyclone type A1 is situated northwest of Bulgaria, north of the 45th parallel, and west of the 25th meridian, and A2 is north of the 40th parallel and east of the 25th meridian. Type A3 is over the country between the 40th and 45th parallel and the 20th and 30th meridian. The anticyclonic type A4 is often a ridge from the Azores High, considered as such if its centre is west of the 20th meridian and south of the 45th parallel. The centre of type A5 is east of the 20th meridian and south of the 40th parallel (
Figure 2).
The summer (June, July, August) frequencies of the anticyclonic types at 500 hPa over the 60 years (1961–2020) are presented in
Table 2. The second column represents the frequencies concerning all circulation types (anticyclonic and cyclonic) at that baric level, while the third column shows the frequencies only amongst anticyclonic types.
Types A1, A2, and A3 have very low frequencies compared to the A4 and A5 types; hence, their climatological influence is relatively small, and that is why more attention is paid to the last two. The presence of the A1 type at the 500 hPa level leads to a northern direction of airflow at that baric level, and often, there is the presence of a high trough east of the anticyclone. In the presence of of A2, the wind direction at the 500 hPa level over Bulgaria comes from the eastern quarter, and it depends on the exact position of the isolines of the baric field. The A3 type’s centre is over the territory of Bulgaria, so it is the reason for there being almost no wind in this area (
Figure 2). A4 has the highest frequency, not only among anticyclonic types but also among all circulation types at 500 hPa, during the summer season. It has a leading climate-forming role, especially in the summer. Its frequency is 28.7% for all types and 71.9% for the anticyclonic types. The A4 type causes westerly and south-westerly winds and thus causes relatively warmer air masses, predominantly from the Mediterranean or Northern Africa, to reach the middle levels of the troposphere above central parts of the Balkan peninsula and especially Bulgaria. The frequency of type A5 is second in value, although it is much smaller than that of A4. Depending on the exact position of the baric formation, the winds have a southwestern or southern direction. In almost all cases with A5, there is a trough situated west of it, over a territory most often above the middle of the Mediterranean. This fact means that over this region, relatively colder air masses are already intruding, through the counterclockwise directions of the airflow of the system of this high trough. As a consequence, this south–western direction of the winds in the neighbouring A5 anticyclone, in most cases, has, to some degree, a cooling effect in the middle troposphere over the territories of the Balkan peninsula, and because the circulation in the western back part of the high anticyclone is similar to that in the front part of the high trough and some kind of cyclonic circulation at lower atmospheric levels, respectively, cloudiness or even precipitation can occur.
The seasonal distribution of anticyclonic circulation types is shown in
Figure 3. The domination of type A4 over A5 is so substantial only in the summer season. In the transitional seasons (spring and autumn), A4 still prevails, but in the winter, the frequency of A5 is even greater than that of A4. Only A4 and A3 have their maximums in the summer, A2 in the autumn, and A5 and A1 in the winter.
To find a relation between the frequencies of the circulation types and the summer seasonal temperature, Pearson’s correlation coefficient was used, with a standard level of significance (α = 0.05) determined by
t-tests. According to the chosen significance level and the length of the row of 60 values concerning the period (1961–2020), the calculated statistically significant correlation coefficients have values greater than 0.26. For assessing the tendencies of the anticyclonic types at 500 hPa, the non-parametric Mann–Kendall test [
77] and Sen’s slope estimator [
78] were applied. The calculations of correlations and trends on a monthly and seasonal basis were carried out through MS Excel’s RealStatistics software ver.8.7 [
79].
4. Conclusions
Comparing the last two 30-year climatological periods (1961–1990) and (1991–2020), an undoubtedly rapid decrease in extremely cold summer seasons and summer months in the last 30 years has occurred, and at the same time, a rapid increase in extremely warm summer seasons and months was found, according to the accepted thresholds of the 10th and 90th percentiles. The highest number of extremely warm seasons was established during 1991–2020 along the seaside, and, at the same time, extremely cold seasons were not registered there at all. The reason for this fact is the rise in the frequency of type A4 on one side and the increase in the sea surface water temperature on the other. The correlation between circulation types and the temperature shows the warming effect at least on the middle part of the Balkans of Azores anticyclone at the 500 hPa baric level in the middle troposphere (type A4). August has the highest and most statistically significant correlation coefficient between type A4 and the temperature of all summer months. In most cases, anticyclones or ridges with a centre south and southeast of Bulgaria (type A5) have a cooling effect, although the direction of the air masses during the presence of this type is most often from the southwest. This fact is probably related to the neighbouring high troughs or cyclones situated west of the anticyclone over the middle Mediterranean, which means relatively colder air there. Such a kind of circulation at 500 hPa could be a reason for cyclogenesis in the lower levels and eventually cloudiness, possible precipitations, and, therefore, relatively lower temperatures, especially in the summer. The increasing trend in the frequencies of warming type A4 and the decrease in cooling type A5 prove that the change in the atmospheric circulation in the middle troposphere, concerning at least anticyclones, is one of the major reasons for the warming summers over the central Balkans in recent decades.