Large day-to-day changes in extreme temperature in Poland and atmospheric circulation

The primary purpose of the study was the determination of the spatial day-to-day variability of extreme temperatures in Poland and the dependency of large temperature changes on atmospheric circulation in accordance with the Grosswetterlagen (GWL) classification. The goal was implemented based on data from 1966 to 2015, made available by the Institute of Meteorology and Water Management – National Research Institute. Day-to-day changes in maximum and minimum temperatures were designated with a rate of ≥ 6 °C (large) and with a rate of ≥ 12 °C (very large) and their spatial distribution was presented. Finally, the analysis of the dependency of considerable temperature changes on atmospheric circulation in accordance with the Grosswetterlagen (GWL) classification was conducted. The obtained results showed that in Poland in the period 1966-2015 the number of large changes in both Tmax and Tmin slightly increased, although the fluctuations show spatial variability. Large changes in Tmax occur more frequently in spring, and in Tmin in winter. Large changes in Tmax and Tmin are mainly recorded during cyclonic circulation, however, the anticyclonic circulation types favour especially large decreases in Tmin.


Introduction
Day-to-day temperature variability is of significance not only as an important characteristic of climate dynamics. It is also interesting from a practical point of view. Changes in mean daily temperature are on average lower than those in the case of extreme temperatures [1,2]. The decrease in the diurnal temperature range is approximately equal to the increase in mean temperature [3]. Knowledge of the probability and conditions of occurrence of considerable temperature changes are of particular importance. Substantial temperature fluctuations are unfavourable for the well-being of man, showing an irritating effect [4], and can have a negative impact on many areas of the economy, e.g.: rapid cooling in spring can cause freeze damage of plants, and rapid warming can lead to fast snowmelt resulting in flood threat [4,6,7].
The observed contemporary climate warming is visible in the increase in the average global air temperature [8]. In the majority of cases, changes in mean daily temperature are relatively low, rarely interrupted with high temperature increases [9]. Together with climate change, the values of minimum and maximum temperatures also increase. This results in day-to-day variability of maximum and minimum temperatures. It is also worth emphasizing that variations in minimum temperatures are usually greater than those in maximum temperatures [10]. The variability of such changes in a year is probably uneven. Air temperature is largely generated by circulation factors [9,11]. High increases and decreases in extreme temperatures have been observed during air inflow from the north and west [12]. The greatest day-to-day changes in temperature are related to the North Atlantic Oscillation (NAO) [12]. Higher values of the NAO index are usually related not only to higher temperatures but also to lower day-to-day variability [10]. It was pointed out that extreme values of air temperature occur with anticyclonic circulation and strong highs blocking the zonal circulation, which in recent decades have been increasingly occurring over Europe, are of particular importance [14,15,16].
Analyses of trends in the period  show that the variability of the spatial distribution of day-to-day temperature changes was subject to the following changes: a 5% increase occurred in south-western Europe, a change of 0 to -5% in the north-west, and a decrease by 5 to 10% in north-eastern Europe. Very large day-to-day changes in temperature are forecasted to occur increasingly frequently in summer [17]. The effects of climate change can depend more on the change in the minimum (Tmin) or maximum temperature (Tmax) than mean temperature [18]. Otherwise, the effect of precipitation deficit on an increase in extreme values of daily temperature amplitude has been observed [19].
Fewer studies in the climatological literature are devoted to the day-to-day variability of maximum and/or minimum temperature [e.g. :2,12,13,20-24] comparing to the number of studies concerning the mean temperature interdiurnal variations (ITV) or diurnal temperature range (DTR) [e.g. :7,10,17,25,26,27]. Additionally, the occurrence and changes of extreme air temperature values should be discussed in the context of circulation conditions because atmospheric circulation is considered one of the most important factors influencing weather and climate conditions in moderate latitudes [28,29].
Bearing the aforementioned in mind, the main objective of the research was the determination of the spatial day-to-day variability of extreme temperatures in Poland and its circulation conditions. The implementation of the primary study objective covers the determination of the spatial and multiannual variability of changes in extreme temperature: maximum (Tmax) and minimum (Tmin) from day to day in the period 1966-2015, and the designation of cases of large (≥ 6 °C) and very large (≥ 12 °C) increases and decreases in extreme temperatures in particular months, years, and throughout the multiannual period. Particular attention was paid to the analysis of the impact of atmospheric circulation on the occurrence of large changes in extreme temperature (Tmax and Tmin).

Materials and Methods
The paper employed meteorological data from 18 stations in Poland ( Figure 1, Table 1   The calculations were based on daily values of maximum and minimum air temperature. A day-to-day change in maximum temperature (Tmax) was calculated as the difference of maximum temperature on a given day and the preceding day. A day-to-day change in minimum temperature was determined analogically (Tmin). Day-to-day changes in maximum and minimum temperatures were designated with a rate of ≥ 6 °C, in the paper described as large, and these with a rate of ≥ 12 °C, described as very large. Day-to-day changes in temperature of more than 6 °C have an irritating effect on the human organism (classification by Bajbakova et al. [4]), and changes of more than 12 °C are described as severe [4]. The number of positive and negative changes were determined for Tmax (Tmax↑, Tmax↓) and Tmin (Tmin↑, Tmin↓), as well as their occurrence in particular months and years in the analysed period. Trends of multiannual changes in temperature increases and decreases were determined, as well as their statistical significance. The determination of the trend of values averaged for the territory of Poland employed a 5-day moving average. The statistical assessment of changes applied a nonparametric Mann-Kendall test, and their statistical significance was assessed by means of the Sen method [30]. The spatial distribution of large changes in extreme temperatures was presented in maps. The next stage involved the analysis of the dependency of considerable temperature changes on the atmospheric circulation in accordance with the Grosswetterlagen (GWL) classification [31,32,33] (Table 2). To eliminate individual cases occurring in the few stations, only these cases were analysed when large changes in extreme air temperature values occurred at least in 20% of the stations. The frequency of occurrence of GWL circulation types in the years 1966-2015 was calculated, as well as the probability of occurrence of large day-to-day changes in Tmax and Tmin in particular circulation types.

Mean annual day-to-day changes of Tmax and Tmin
In the analysed period 1966-2015, the mean annual day-to-day changes in Tmax and Tmin in Poland were 2.3 °C and 2.5 °C, respectively, and showed spatial variability (Table 2, Figure 2). Mean day-to-day changes in Tmax increased from the north to the south of the country and ranged from 2.1 °C in Łeba, Świnoujście, and Chojnice to 2.5 °C in Kraków. Mean changes in Tmin were lowest in the western part of Poland and increased eastwards generally except for the central-eastern area which is characterised by slightly lower values. They varied from 2.0 °C in Świnoujście and Zielona Góra to 2.7 °C in Wrocław, Toruń, Kielce, Rzeszów, and Białystok. The dynamics of mean annual changes from year to year in temperature, Tmax and Tmin, in the majority of the territory of Poland is similar. Mean changes for the entire country showed a statistically significant increase in changes in extreme temperature ( Figure 3). 72% of the analysed stations showed an increase in mean changes in both Tmax and Tmin ( Figure 4). Only in Świnoujście and in Warsaw, this trend is statistically significant in the case of Tmax. In the case of Tmin, a statistically significant increasing trend is typical of Białystok, Chojnice, Łódź, Terespol, Toruń, and Wrocław. In Rzeszów, a decrease in these changes in the range of Tmax and a lack of changes in the case of changes in Tmin are observed. The stations in Toruń, Poznań, and Wrocław are characterised by a negative tendency of changes in Tmax and positive in Tmin. Stations in Świnoujście and Zielona Góra deserve particular attention, where year-to-year changes in Tmax and Tmin show an increasing tendency, although the mean annual changes in Tmax are higher than Tmin, unlike in the remaining stations.

Large day-to-day changes of Tmax and Tmin
The total annual number of large day-to-day changes in maximum temperature (Tmax > 6 °C) in the analysed stations in Poland in the research period is 340 cases with Tmax > 6 °C. The number varied in particular stations from 16/year in Chojnice and Elbląg to 25 cases/year in Kraków. In the discussed period, just an average of 6 cases of Tmax> 12°C occurred annually, from 0.2 cases/year in Gorzów Wielkopolski to 0.8 cases/year in Świnoujście.
The total number of large interdiurnal changes in minimum temperature (Tmin > 6 °C) in the selected stations in Poland in the analysed period averaged 445 cases annually, varying from 35 in Kielce to 9 in Świnoujście. Very large changes in minimum temperature (Tmin > 12 °C) in Poland occurred on average 13 times in a year in the period 1966-2015 -from 0.1 cases in Gorzów Wielkopolski and Zielona Góra to 1.6 in Suwałki.
In the years 1966-2015, the average annual number of cases of increases in large changes in Tmax (Tmax↑ > 6 °C) in Poland reached 165 and showed spatial variability. A high increase in Tmax usually occurred at the coast and near Kraków (more than 12) (Figure 5a). The least cases of such changes occurred in the belt of lakelands (below 8). Decreases in large changes in Tmax (Tmax↓ > 6 °C) were recorded in Poland on average 175 times in a year, and their number increased from the north-east to the south-west of the country. They most commonly occurred in the foothills of the Carpathians (more than 12), and most seldom in north-western Poland (below 8) (Figure 5b). Very large increases in maximum temperature (Tmax↑ > 12 °C) are very rare: up to 0,5 cases annually in Świnoujście and occasional occurrences in central Poland (Figure 5c). Very large decreases in Tmax (Tmax↓ > 12 °C) were also infrequent: up to 0,3 cases/year in Kraków and Lublin (Figure 5d). Large increases in Tmin (Tmin↑ > 6 °C) in Poland occurred 255 times a year. Their number generally increased from the west to the east of the country, varying from 6 cases in Zielona Góra and Świnoujście to 19 cases annually in Rzeszów (Figure 6a). Decreases in large changes in Tmin (Tmin↓ > 6 °C) in Poland were recorded on average 190 times in a year, and the spatial distribution of their number was variable (Figure 6b). They occurred most seldom in the north-western part of the country with the lowest number of cases recorded in Świnoujście (3 times in a year). The highest number of large decreases in Tmin was recorded in stations in Kielce, Rzeszów, and Białystok (more than 15). Very large increases in minimum temperature (Tmin↑ > 12 °C) were the most common of all the very large variations within the range of 0,1 cases/year in northeastern Poland to 1,1 cases annually in Białystok (Figure 6c). Very large decreases in Tmin (Tmin↓ > 12 °C) did not exceed 0.6 cases a year in Suwałki (Figure 6d).   Considering the annual course of the number of large changes in Tmin, the stations in Świnoujście, Zielona Góra, Białystok, and Suwałki stand out. Świnoujście is a station characterised by the lowest monthly number of decreases in Tmin, particularly in July and in September (7 cases at each). Zielona Góra is characterised by the lowest number of decreases in Tmin, particularly in July (9). In Suwałki, the highest monthly number of increases in Tmin was observed in February, and the decreases in January (123 and 94 cases, respectively), and in Białystok a high increase in Tmin usually occurred in January, and a decrease in February (115 and 105 cases, respectively). The variability of the number of increases in Tmin of more than 12 °C also deserves attention -their highest number was recorded in February in Suwałki (19) and Białystok (17) and in January in Kielce (18) and Terespol (16). Very high decreases in Tmin usually occurred in January in Rzeszów (13) and Suwałki in January and December (13 and 11, respectively).
During the period considered in all the analysed stations in Poland, the highest daily decrease in Tmax exceeded 14 °C, and the increase exceeded 13 °C (Figure 9). The highest decreases in particular stations varied from -14.1 °C in Toruń (19.10.1967)   The share of all decreases and increases in temperature in the case of Tmax is more even in comparison to Tmin (Table 3). On average in Poland, in 49.4% of cases, an increase in Tmax occurred, and in 48.9% of cases a Tmax decrease.
Interdiurnal Tmax increases occur slightly more frequently in the north-west of Poland. The stations in the north-western part of the country are characterised by the advantage of an increase in Tmax over a decrease. The remaining area is dominated by interdiurnal declines in Tmax.
Interdiurnal Tmax increases occur slightly more frequently in the north-west of Poland. The stations in the north-western part of the country are characterised by the advantage of an increase in Tmax over a decrease. The remaining area is dominated by interdiurnal declines in Tmax. In Świnoujście, the variation in the share is the greatest (Tmax↑ = 46.8%, Tmax↓=50.8%), and lack of it in Toruń (Tmax↑ and Tmax↓ = 49.1%).
Throughout the country, a day-to-day decrease in minimum temperature occurs more frequently, averaging 50.7% of cases. The share of increases in Tmin averages 47.6% in Poland. The stations in the north of the country and located in the area from Elbląg to Poznań are characterised by the greatest advantage of Tmin decreases over increases. In Elbląg, the range of the share is the biggest (Tmin↑ = 46,7%, Tmin↓ = 51,7%), and in Lublin the smallest (Tmin↑ = 48,5%, Tmin↓ = 50,0%). Days with no changes either in Tmax or Tmin constitute 1.7% of cases. Considering the share of large changes in Tmax and Tmin, a predominant share of large interdiurnal changes in Tmin over Tmax (56.7% and 43.3% respectively) in Poland is found. When considering the share of large increases and decreases in the case of Tmax, spatial differentiation is noticeable ( Table 2). Tmax most often increases day-by-day on the coast and in the western part of the country. The maximum occurs in Świnoujście of 56.8%, which is also characterised by the greatest variation in the share of increases and decreases in Tmax (Tmax↑ = 56.8%, Tmax↓ = 43.2 %). The smallest share of Tmax increases was recorded in southeastern and eastern Poland, with the minimum in Rzeszów (43.7%). In this region also, Tmax declines are the highest with the maximum in Rzeszów (56.3%). The lowest share of decreases is in Świnoujście (43.2%). On the other hand, in the case of Tmin, the day-to-day increases are more frequent than decreases. The highest share of increases is recorded at the stations in Świnoujście and Elbląg. Świnoujście has the highest variation in share (Tmin↑ = 6.7%, Tmin↓ = 33.3%), and the lowest -in Kielce (Tmin↑ = 54.5% and Tmin↓ = 45.5%).

Influence of atmospheric circulation
In the years 1966-2015, circulation types WZ (15.5%) and BM (10.8%) occurred the most frequently ( Figure 10). The following types occurred the most seldom, with a frequency below 2%: NA (0.6%) and SZ (0.8%), as well as HFZ (1.  Explanations of the GWL types: see Table 2. Large day-to-day changes in Tmax primarily occurred during cyclonic circulation (62.8%). The highest probability of occurrence of a large increase in maximum temperature was recorded during SWZ (8.0%) (Figure 11a). High probabilities of Tmax > 6 °C also occurred during types TRW (7.1%), SZ (7.1%), and TB (7.0%). The lowest probability of a large increase in Tmax was related to SEA and TRM (1.6% each). A large decrease in Tmax was the most probable during NEA and HNZ circulation (5.4% each) (Figure 11b). Such changes also occurred relatively frequently with types TRM (4.5%) and TRW (4.3%). They were recorded most seldom during types TB (1.0%) and HFA (1.4%).  Table 2.
Large day-to-day changes in Tmin usually occurred during cyclonic circulation types, reaching 83.5% on the whole. The highest probability of occurrence of a large increase in minimum temperature was recorded during SWZ (10.7%) (Figure 11c). Moreover, circulation types WA (9.3%), NWZ (8.8%), and SZ (8.5%) favoured high increases in Tmin. The lowest probability of a high increase in Tmin was usually related to circulation types SA (1.4%) and HFZ (1.6%). The highest probability of occurrence of a high decrease in Tmin was recorded in type HNA (10.6%) ( Figure  11d). A higher number of large day-to-day changes in Tmin was also probable during circulation NZ (6.7%), WA (6.2%), as well as SEA and BM (6.1% each). The lowest probability of a high day-to-day decrease in Tmin occurred in types TM and NEZ (1.9% each).
A very high increase in Tmax over the predominant territory of Poland, i.e., a day-to-day increase of more than 12 ⁰C, occurred in the study period sporadically. It occurred one time during the following circulation types: NZ, HNZ, SWZ, and SEZ. A rapid day-to-day decrease in maximum temperature, i.e., a decrease by more than 12 ⁰C, occurred seven times in the study period in the majority of the analysed stations during the circulation types: HNFZ, WS, WZ, HNFA, NZ, HM, WA.
Very large changes in Tmin were recorded more frequently over a prevalent area of Poland. Day-to-day changes in Tmin > 12 °C occurred 36 times in the period 1966-2015, and the highest probability of their occurrence was determined during SZ, NZ, HB, and SWZ types. A rapid decrease in Tmin was recorded 6 times: twice during type WS and one occurrence each during types: WZ, NZ, NWZ, and BM. All these cases of rapid changes in Tmin occurred in winter.

Discussion and Summary
In the analysed period 1966-2015, the mean annual day-to-day changes in Tmax and Tmin in Poland reached 2.3 °C and 2.5 °C, respectively, confirming earlier results [11,12,21,22,23,24]. The changes showed spatial variability reflecting the range of the effect of marine and continental properties on the climate of Poland. In the coastal region, day-to-day temperature changes are the least considerable as a manifestation of the mitigating effect of the sea on the climate [21]. According to Kossowska-Cezak [1], a very rapid change in thermal conditions in the lowland area of Poland is caused by the advection of air masses and the inflow of air from over a thermally contrasting surface.
Considerably more day-to-day changes in minimum than maximum temperature increase occurred in the analysed period. This is in accordance with the results of Panfil [22]. Similarly, Fortuniak et al. [7] determined that at the scale of the entire year, temperature increases are characterised by a lower mean value, but higher frequency than temperature decreases. This pattern is particularly evident in the warm season of the year. Regarding day-to-day variations, Piotrowicz et al. [24] pointed out that significant decreases in Tmax were more numerous than in Tmin, whereas the opposite was true for significant increases -they were more numerous in the case of Tmin than for Tmax.
In the years 1966-2015, the number of considerable changes in both Tmax and Tmin increased in Poland, although the variability shows spatial differentiation, and the changes are statistically significant only in some of the stations. A similar trend of changes was also found in north-eastern Poland -the direction of trends in the analysed stations was variable as Panfil [22] stated. Trends of multiannual changes in Olsztyn showed statistically significant increases and decreases only for Tmax [12]. In the major area of the territory of Poland, based on this research, higher ranges of large changes in maximum than minimum temperatures were recorded. Less increases in Tmax exceeding 10 degrees than Tmin were recorded in north-eastern Poland, but in the case of decreases, the number is approximate [22]. As Ciaranek [23] indicated, the rapid changes of Tmax (> 10 °C) in winter occurred in Poland more frequently in the first half of the analysed period 1961-2010. Therefore, if this tendency of changes is maintained, a decrease in the frequency of occurrence of rapid changes in maximum air temperature can be expected in the following years in Poland in the cold season.
The obtained results showed that the number of large Tmax fluctuations in Poland is variable latitudinally: it is the highest in the coastal area and increases from the north to the south of the country. The number of large changes in Tmin in Poland is the lowest in the northwestern part and increases east and southward. The presented results are consistent with the results of the research of Panfil and Dragańska [11], who within the analysed area of north-eastern Poland, designated areas with large (≥ 6 °C) day-to-day changes in temperature that occurred most frequently in the east.
The annual course of the number of both increases and decreases in Tmax reached culmination in spring, in accordance with the results of Panfil [22]. Changes in Tmin -both increase and decrease, are more considerable in winter than in summer, similarly proved by Wibig [21]. Fortuniak et al. [7] confirm that such considerable changes in temperature are usually observed in the winter period as a result of the advection of both cool and warm air masses. The temperature swings which are most strongly felt by the human body -not prepared to rapidly adapt to them -are mainly those occurring in winter and spring as a result of the advection of thermally different air masses, Kossowska-Cezak [9] explained. Kossowski [34] pointed out that significant short-term variations in temperature most frequently occur in the cooler half of the year when the air masses over Poland are the most thermally diverse and weather fronts are most frequent. In the summer months, such shifts are smaller and happen less frequently, while changes in cloud cover, which influence night-time radiation cooling, become a contributing factor [24,26].
The share of temperature decreases and increases in the case of Tmax is more aligned in comparison to Tmin in the study period. On average in Poland, in 49.4% of cases, an increase in Tmax occurred, and in 48.9% of cases its decrease. A day-to-day decrease in Tmin occurs more frequently, averaging 50.7% of cases. The share of increases in Tmin averages 47.6% in Poland. While according to Ciaranek [23], in the years 1961-2010 in Poland, the share of increases in large changes in Tmax was lower than in the case of decreases and showed spatial variability.
Temperature changes at moderate latitudes remain in close relation to the variability of atmospheric circulation, particularly in winter when insolation plays a secondary role in forming weather conditions [21]. The research carried out showed that in Poland, day-to-day changes in Tmin exceeding 6 degrees are considerably more frequent in winter. Large changes in Tmax, however, show a higher frequency in spring. The majority of large day-to-day changes in Tmax, and particularly Tmin in accordance with the above analysis are related to cyclonic circulation types. However, the anticyclonic circulation types favour especially large decreases in Tmin. While Kuziemska [35] pointed to the important role of advection in shaping considerable temperature anomalies in winter, both positive and negative, whereas the most high temperature decreases occur when the inflow of cold air is accompanied by strong effective radiation, i.e. in the case of anticyclonic circulation types. In summer, the dominant role in shaping thermal conditions is played by insolation, and secondary, although still important, by advection. Kossowska-Cezak [9] evidenced that day-to-day temperature changes also depend on the sequence of circulation types. Day-to-day changes in minimum and maximum temperatures are usually the highest between the second and fourth day of occurrence of a given circulation, and then they decrease or even change the sign to the opposite one [21].
The study showed the variability of the direction of atmospheric advection favouring changes in Tmax and Tmin. Large increases in Tmax were favoured by the southern and southwestern circulation types and decreases -by the north and northeastern. High increases in Tmin were primarily favoured by the northern and western types and decreases by the northeastern types. The obtained results coincide with earlier studies. This is confirmed by research by Wibig [21], according to which western types are characterised by an increase in temperature, but maintain shorter for 4-5 days, and day-to-day changes are lower than in the case of eastern types. Eastern types are characterised by a day-to-day decrease in temperature maintained for 5-6 days. Panfil [12], analysing the dependencies of large changes in extreme temperatures on the atmospheric circulation in Olsztyn, proved a particular impact of the northern and westerly component. In winter, high decreases in temperature correspond to circulation types of the easterly component, particularly from the northern sector [1]. Moreover, Wibig and Głowicki [12] stated that day-to-day temperature changes are well correlated with cloudiness, and extreme temperature changes are related to the North Atlantic Oscillation, particularly in winter and spring. However, besides the weather situation and the advection of various air masses, the scale of short-term temperature changes is influenced by local factors, such as relief and land cover [36]. This suggests that the daily minimum temperature is less associated with the regional circulation patterns than the daily maximum temperature and more influenced by local factors instead. [36]. Piotrowicz et al. [24] suppose that the short-term variation in Tmax and Tmin decrease, which has been gradual over the entire analysed long-term period, without clear and sudden changes to the overall trend, was largely influenced by anthropogenic factors. The decrease in the daily temperature range is partially related to increases in cloud cover. Furthermore, a large number of atmospheric and surface boundary conditions are shown to differentially affect the maximum and minimum temperatures. For rural sites, day-to-day temperature variation was typically greater for Tmin than Tmax. The opposite was found for urban locations, with statistically significant stronger signals for larger cities [27].
Linkages of the observed changes in the diurnal temperature range to large-scale climate forcings, such as anthropogenic increases in sulfate aerosols, greenhouse gases, or biomass burning remain tentative [3,27]. Future research on temporal and spatial variability and the influence of local conditions on the occurrence of large changes in extreme temperatures would be not only valuable but also necessary.

Conclusions
The research on large day-to-day changes in extreme temperature in Poland in the period 1966-2015 and the influence of atmospheric circulation leads to the following conclusions: • the number of large changes in both Tmax and Tmin slightly increased, although the fluctuations show spatial variability, • higher fluctuations of maximum than minimum temperature were recorded over the majority of the territory of the country, • the number of large changes in Tmax increases from north to south, and the number of large changes in Tmin from north to east and south of Poland, • annual course of increases and decreases in both Tmax and Tmin is analogical, and the annual course of changes in Tmax and Tmin differs: large changes in Tmax occur more frequently in spring and Tmin in winter, • the share of temperature increases in the case of Tmax is higher than that of decreases, and it is the opposite for Tmin, • large changes in Tmax and Tmin are mainly recorded during cyclonic circulation, however, the anticyclonic circulation types favour especially large decreases in Tmin.
• large increases in Tmax were favoured by southern and southwestern circulation types, and decreases -by north and northeastern, • large increases in Tmin were primarily favoured by northern and western types and decreases by northeastern types.

Conflicts of Interest:
The author declares no conflict of interest.