**1. Introduction**

Spatial and temporal variability of precipitation amounts in Poland is very high. The diversity of relief means that the areas with the lowest precipitation covering the central part of the country receive less than 500 mm of precipitation annually. On the other hand, on the upper border of the moderately warm story in the Western Carpathians, precipitation of 1000 mm should be expected, while on the upper border of the moderately cool story, which is the limit of agricultural use, it was 1400 mm [1–3].

The location of Poland in moderate latitudes of Central Europe determines high variability of weather in particular years. The values of the lowest and highest precipitation in Warsaw may vary from 60% to 150% of the standard multi-year average for the year, 27% to 250% for seasons, and even 5% and 505% for October [4,5]. In the summer, the maximum daily sums may exceed the multi-year average monthly sums. Such high variability of precipitation with simultaneous variability of values of other meteorological elements results in different meteorological conditions of crops in particular years manifested by the occurrence of dry periods and periods with an excess of precipitation.

In recent years, a growing interest among climatologists, hydrologists, environmentalists, farmers, and political activists has been the observed and projected climate change [6]. The trends of precipitation changes in a warming climate have not yet found a clear assessment. In the area of Poland lying between northern Europe, with marked and predicted increasing trends of annual precipitation sums and southern Europe where decreasing annual precipitation sums are observed and predicted, no major changes in precipitation are observed and predicted, with only a very slight increasing trend in some areas [7,8]. The largest increases in precipitation with different significance of changes were observed in the northern part of Poland with different levels of significance and in a small area of south-eastern Poland [9]. In the light of the analysis of observational series averaged for the

area of Poland, only the increasing variability of precipitation sums is undeniable [10,11]. Rising air temperatures, especially since the second decade of the 20th century, have increased evapotranspiration and may be a significant cause of increased drought caused by insufficient precipitation among other things [12–14].

In contrast to floods, the effects of droughts are not immediate. The phenomenon increases slowly and its consequences become visible over a longer period of time. Furthermore, they are initially less visible and extend over a larger area than in the case of other extreme weather events [15–17]. The impact on the economy in a drought-affected region depends not only on the duration, intensity, and spatial extent of the phenomenon, but also on the vulnerability of the environment to the negative effects of droughts. On soils with deep groundwater levels and low useful retention, which prevail in Poland, with relatively low and variable precipitation and the observed increase in air temperature as well as the signaled increase in the frequency of meteorological extreme events, an increase in both the frequency and intensity of drought phenomena should be expected [18,19].

In the extensive literature on droughts, and particularly on droughts in Poland, one can distinguish several research trends or thematic sections covering the phenomenon in question. Drought as a meteorological phenomenon unfavorable for agriculture does not occur suddenly but it shows a specific cycle of development. D ˛ebski [20] breaks down this cycle into four phases: distinguishing atmospheric drought, soil drying out transforming into soil drought, lowering of groundwater level, occurrence of deep lows in rivers and drying out of springs and small watercourses–hydrological drought, and long-term lowering of groundwater resources defined as hydrogeological drought. Extended atmospheric drought may develop into soil drought, often referred to as agricultural drought. It occurs when a lack of precipitation, usually combined with high air temperatures, causes the soil to dry out, severely restricting the growth and development of crops and resulting in a significant decrease in crop yields. The time scale over which soil drought can occur is 1–3 months [15]. The prolonged period of low precipitation, often in combination with increased air temperature, leads to hydrological drought manifested by decreased water flow in rivers and water level in lakes, and at a further stage to hydrogeological drought manifested by decreased groundwater resources. The phases distinguished by the author correspond to the divisions into atmospheric, soil, and hydrological droughts often used in the literature [21,22].

Each of the mentioned phases is characterized by a different course and requires different research methods. One of them is the method of rain-free sequences. In Poland, Schmuck [23] was the first to analyze droughts on the basis of rain-free sequences. The author together with Ko´zmi ´nski [24] presented a spatial distribution of frequencies of droughts lasting over 8 and 17 days in Poland. Drought monitoring uses many indices based on precipitation alone or precipitation and other meteorological elements and indices, often taking into account evapotranspiration of plants and soil water reserves, as well as groundwater. A review of this is provided by Przedpełska [21] and later by Łab ˛edzki [15,25]. The simplest and the most widely used are indices using precipitation in such modifications as the relative precipitation index RPI [4] and the standardized precipitation index SPI [26–29].

Drought monitoring in agricultural areas should take into account meteorological evaporation conditions in addition to precipitation. For example, we can mention the Sielianinov hydrothermal index, the De Martonne dryness index, index evapotranspiration, or standardized climatic water balance [30–33]. The Institute of Soil Science and Plant Cultivation–State Research Institute (IUNG-PIB) in Puławy, at the request of the Ministry of Agriculture and Rural Development, developed and launched an agricultural drought monitoring system (PL abbr. SMSR). The importance of the problem of drought monitoring is also emphasized by the fact that it has found appropriate legal empowerment [34]. Based on the Act, the Ministry of Agriculture and Rural Development carries out the task of drought monitoring by specifying that the IUNG-PIB determines the current values of climatic water balance (CWB) "in the period from 1 June to 20 October, within 10 days

after the end of the six-day period, indicators of climatic water balance for individual crop species and soils, broken down by voivodeship, on the basis of data provided by the Institute of Soil Science and Plant Cultivation–State Research Institute" [35].

Drought is a relative phenomenon and its assessment should be related to the current agricultural area and a specific crop. Therefore, by analyzing the current state of research, a number of studies on drought can be listed for particular regions, such as Schmuck [36] for Lower Silesia, Prawdzic and Ko´zmi ´nski [37] for the Szczecin province, Konopka [38] for the Bydgoszcz region, and Łab ˛edzki [39] for the Bydgoszcz-Kujawy region, as well as publications on selected crops, e.g., winter wheat [40], medium–late and late potatoes [41], or spring cereals [42].

Characteristics of particular droughts or drought periods, such as the 1959 drought [43], the drought of 1969 [44], the dry period of 1982–1992 by Bobi ´nski and Meyer [45], the drought of 1992 [46], or the drought of 2005 [47], have been developed. A few sentences on the course and effects of droughts in 2003 and 2005 may be found in the monograph by Łab ˛edzki [15] devoted to agricultural droughts in Poland.

As for drought problems discussed directly in this paper and concerning precipitation deficits and drought cases in Poland in the first years of the 20th century, it is worth mentioning the first study by Hohendorf [48] on precipitation deficits and excesses for the period 1891–1930 and a more recent one by Dziezyc et al. [ ˙ 49] for the realities of the period from 1952–1980, Farat et al. [50] for the 40-year period from 1951–1990, Ziernicka-Wojtaszek and Zawora [51] for the 30-year period from 1971–2000, Doroszewski et al. [14] for the period from 1961–2010, and Przybylak et al. [52] for the period of over 1000 years from the establishment of the Polish state in the year 1996 to 2015.

In recent years, studies from the stream of contemporary climate change taking into account the impact of rising temperature on the occurrence of droughts have become increasingly frequent [53–56]. Drought is increasingly monitored using satellite data [57,58].

In the warmest year in the history of observation, 2019, especially in the summer, another intense drought occurred in Poland. This study uses preliminary documentation and characterization, and is a continuation of such studies made in the 20th and 21st centuries for dry years, such as 1959, 1969, 1982, 1983, 1989, 1992, 2000, 2003, 2005, 2008, and 2013. The study includes analysis of the causes, the course, and the consequences of the summer drought of 2019, and characterizes consecutive drought-meteorological, agricultural, hydrological, and hydrogeological phases. It shows, by comparison with extreme thermal droughts of 2003, 2018, and 2019 in Central Europe, its uniqueness. It proves that, in terms of climatic water balance values in summer, it is a case of maximum maximorum, which means that there is no other value in the series that is above or at least at the level of this maximum. It is a typical case of thermal drought caused not so much by precipitation deficit as by intensive evapotranspiration, following extremely high air temperatures. Hence, the methods which were atypical or modified for drought analysis, such as the analysis of months with precipitation frequency lower than the average rather than the analysis of classical sequences of days without precipitation, or the additional analysis of the simultaneous influence of precipitation and air temperature on the value of climatic water balance using the method of stepwise multiple regression, were used to show the leading role of temperature in generating the drought phenomenon.

#### **2. Materials and Methods**

The research material consists of verified, homogeneous monthly mean values of insolation, air temperature, and precipitation totals from 47 meteorological stations evenly distributed across Poland for the periods June–August 2019 and June–August 1981–2010 (Figure 1). Due to the insufficient number of stations, mountainous, and especially highmountainous, areas are poorly represented. Data on sunshine (hours), temperature (◦C), and precipitation (mm) were taken from the database of the Institute of Meteorology and Water Management. The study material was evaluated in terms of the degree of data homogeneity [59].

**Figure 1.** Distribution of weather stations in Poland: (**a**) geographical regions; (**b**) voivodeships. PL-EN: Beskidy Wschodnie-Eastern Beskids, Centralne Karpaty Zachodnie-Central Western Carpathians, Niziny Sasko-Łuzyckie- ˙ Lowlands Sasko-Łuzyckie, Niziny ˙ Srodkowopolskie-Central Polish Lowlands, Pobrze ´ za Południowobałtyckie-South ˙ Baltic Coast, Pobrzeza Wschodniobałtyckie-Eastern Baltic Coast, Podkarpacie Wschodnie-Eastern Subcarpathia, Pojezierza ˙ Południowobałtyckie-South Baltic Lakes, Pojezierza Wschodniobałtyckie-Eastern Baltic Lakes, Polesie–Polesie, Północne Podkarpacie-Northern Subcarpathia, Sudety z Przedgórzem Sudeckim-Sudety Mountains with the Sudeten Foreland, Wysoczyzny Podlasko-Białoruskie-Podlasie-Byelarus Uplands, Wyzyna Lubelsko-Lwowska-Lublin-Lviv Upland, Wy ˙ zyna ˙ Małopolska-Małopolska Upland, Wyzyna ˙ Sl ˛ ´ asko-Krakowska-Silesia-Cracow Upland, Wyzyna Woły ´ ˙ nsko-Podolska-Volyn-Podolia Upland, Zewn ˛etrzne Karpaty Zachodnie-Outer West Carpathians.

In addition to meteorological data, the following were used: maps published within the agricultural drought monitoring system of IUNG–CWB, maps with comments from 14 reporting periods, summaries of drought-prone areas for the mentioned 14 reporting periods, all 14 studied plants and 16 voivodeships, and summaries of drought occurrence for selected communes of the Lubuskie voivodeship on particular soil categories for the year 2019.

The characteristics of weather patterns, i.e., temperature, precipitation, and synoptic situations against multi-year averages contained in the Bulletin of Climate Monitoring of Poland of the Institute of Meteorology and Water Management and the Bulletin of the National Hydrological and Meteorological Service for 2019; data on the outflow from the Bulletin of the National Hydrological and Meteorological Service for 2019; and data on the groundwater table level from the Hydrogeological Annual Report Polish Hydrogeological Survey for 2019 were also used.

The scope of the work included: (1) Characteristics of the weather pattern during the 2019 drought: (a) pluviothermic December 2018 to October 2019 including days with maximum temperature equal to and above 25 ◦C on the background of the multiyear (1981–2010); (b) characteristics of the occurrence of probability of extreme values of temperature and precipitation on the background of the multi-year (1951–2018); (c) frequency of synoptic situations on the background of the multi-year (1951–2018); (d) characteristics of the driest months June to August 2019 on the background of the multiyear (1951–2019). (2) The characteristics of meteorological drought: (a) the low rainfall frequency method for the entire summer drought period; and (b) the relative precipitation method RPI for the period of June–August. (3) The characteristics of agricultural drought: (a) on the basis of the climatic water balance (CWB) values for the following months

June, July, and August 2019 and in the period 1981–2010; (b) the temporal dynamics of drought (during the growing season); (c) a ranking of the provinces affected by the drought; and (d) the vulnerability of particular crops to drought. (4) Elements of hydrological drought. (5) Elements of hydrogeological drought.

Due to the extraordinary nature of the phenomenon, very different methods were used. The method for periods of low precipitation frequency.

It was not possible to distinguish classical sequences of rain-free days, even in the areas regarded as the driest, comprising south-western and central-western parts of Poland. Precipitation-free sequences of several days at the most, often lasting several days, were separated by one, two, or even several days of local precipitation of thunderstorm character. Attempts were made to separate such drought sequences interrupted by rainfall in June– August by the predominance of rain-free days in the selected periods. At the same time, meteorological stations were found at which precipitation dominated separated by rainless days and which could not be regarded as drought periods. These were high-mountain stations and the Lesko station and stations in north-central and north-eastern Poland, such as L ˛ebork, Elbl ˛ag, Olsztyn, K ˛etrzyn, and Suwałki. The limiting value was the average multi-year number of days with precipitation amounting to 41% of the duration of the summer period from June–August [60]. On the basis of this criterion, drought periods with rare precipitations were distinguished, in which the average number of days with precipitation amounted to 24%. In the remaining periods, not included in the drought periods, the number of days with precipitation amounted to 43%, which slightly exceeded the average number of days with precipitation—41%.

The relative precipitation method *RPI*.

The relative precipitation index RPI [4] is defined as the ratio of the total precipitation in a given period to the average multi-year sum, taken as the norm:

$$RPI = \frac{P}{P} \cdot 100\% \tag{1}$$

where *P*—the total precipitation in the study period (mm); and *P*—the average precipitation in the studied multi-year period (mm). The 1981–2010 norm was assumed.

The method of climatic water balance (CWB).

The climatic water balance (*CWB*) was calculated as the difference between precipitation (*P*) and potential evapotranspiration (*PET*):

$$CWB = P - PET \tag{2}$$

where *P*—precipitation (mm); and *PET*—potential evapotranspiration (mm).

To calculate the potential evapotranspiration, the simplified formula developed by Doroszewski and Górski [61], based on Penmann's algorithm [62], was used:

$$PET = -89.6 + 0.0621 \cdot t^2 + 0.00448 \cdot h^{1.66} + 9.1 \cdot f \tag{3}$$

where: *PET*—the monthly potential evapotranspiration (mm · [month]ˆ(−1));

*f*—the length of the middle day of the month (hour);

*h*—the monthly insolation (hour);

*t*—the average monthly air temperature 2 m above the ground surface (◦C).

The Institute of Soil Science and Plant Cultivation (IUNG) method of determining climatic water balance (CWB) in 6-decade periods with a step every decade was modified to characterize the CWB in monthly periods. The calculated values of climatic water balance for the months of June, July, and August were compared with the same months from the multi-year period from 1981–2010 and with the corresponding IUNG reporting periods for the estimated monthly periods.

The temporal dynamics of drought.

The dynamics of drought over time in the growing season were characterized by presenting—for the entire area of Poland without regional differentiation (discussed in

other chapters)—the values of climatic water balance (mm) and the area at risk of drought (%) in the following 14 reporting periods from 21 March to 30 September 2019: reporting period 1 (from 21 March to 20 May), reporting period 2 (from 1 April to 31 May), reporting period 3 (from 11 April to 10 June), reporting period 4 (from 21 April to 20 June), reporting period 5 (from 1 May to 30 June), reporting period 6 (from 11 May to 10 July), reporting period 7 (from 21 May to 20 July), reporting period 8 (from 1 June to 31 July), reporting period 9 (from 11 June to 10 August), reporting period 10 (from 21 June to 20 August), reporting period 11 (from 1 July to 31 August), reporting period 12 (from 11 July to 10 September), reporting period 13 (from 21 July to 20 September), and reporting period 14 (from 1 August to 30 September).

Ranking of voivodeships by area at risk of drought.

A ranking of voivodeships by drought risk, illustrating spatial diversification of drought phenomenon, in addition to values of climatic water balance, was presented for two time periods: the whole reporting period covering the vegetation period from 21 March to 30 September, and the reporting period 8 of the highest drought intensity covering the months of June and July. The corresponding summaries for voivodeships published in IUNG reports for the individual 14 reporting periods were used. Voivodeships were ranked in terms of the area at risk of drought based on the average value of the 14 reporting periods (as above) and the eighth period with the highest drought intensity.

Sensitivity of crops to drought.

Sensitivity of crops to drought was determined by the percentage of cases of drought occurrence in all 14 reporting periods for the entire Poland, in all reporting periods for the three voivodeships where drought was the most intense (i.e., lubuskie, wielkopolskie, and łódzkie voivodeships), and for three voivodeships for the driest reporting period covering June and July. The 14 crops studied by IUNG were ranked from the most drought-sensitive to the least drought-sensitive on the basis of the average risk for the whole of Poland, within the three voivodeships most threatened by drought and within those voivodeships in the eighth period with the highest drought intensity.

Hydrological aspects of drought.

Hydrological aspects of drought were characterized by the acceleration of the timing of minimum flows and comparison of minimum flows with multi-year averages.

Hydrogeological aspects of drought.

Hydrogeological aspects of drought were characterized on the basis of the difference in the level of the groundwater table with respect to multi-year average values in individual months.

The results on the background of drought patterns in different European countries are presented against a historical reconstruction of a 254-year climate database for Europe and drought projections in Europe for the period 2041–2070 compared to 1981–2010 for two emission scenarios: RCP4.5 and RCP8.5.

The uniqueness of the 2019 drought phenomenon against the background of (compared to) the hottest and driest years of 2003 and 2018 was demonstrated by comparing the values of the climatic water balance in individual months of the summer (June–August) and for the whole summer period in the stated three years. The results of the comparison are included in the final "Discussion" chapter.

The maps of spatial distribution of drought were made in the Surfer 10 program. The kriging method using spherical function fitting was used for their interpolation. Taking into consideration small scale of the maps, it was evident that they had illustrative character. To describe spatial variability of precipitation, the names of mesoregions according to physico-geographical regionalization by Kondracki [63] were used.

#### **3. Results**

*3.1. Synthesis of Weather Patterns December 2018–October 2019 for Air Temperature, Precipitation, Atmospheric Circulation*

The months leading up to the 2019 growing season were treated more generally, more specifically the driest months of June, July, and August.

The winter period and the April–October 2019 growing season in question can be characterized as follows:

December 2018 was anomalously warm with a deviation of 2.0 ◦C from the 1981–2010 norm and very wet with precipitation of 126–150% of normal with snow cover lingering for several to a dozen days.

January 2019 was slightly cool with a temperature deviation from normal of −0.5 ◦C on the borderline between humid and very humid with precipitation of the order of 125% of normal with snow cover persisting throughout the month, in warmer regions a few to a dozen days.

February was very warm with temperature deviation from the norm of 3.0–4.0 ◦C and was dry with precipitation of 80% of the norm with snow cover lasting several days, especially in colder regions.

March was anomalously warm with temperature deviation from the norm of 2.0–3.0 ◦C and was slightly dry with precipitation of 80% of the norm.

April, in terms of temperature, was anomalously warm with temperature deviations from the norm exceeding 2.0 ◦C in the central-western part of Poland, in the Mazowiecka Basin and near Suwałki. Only the coast, i.e., the Sudety Mountains and the Carpathians, saw very warm conditions. In terms of precipitation, the month was normal in the south of the country and, locally, it was humid and very humid there. In the remaining part of Poland, April was mostly extremely dry. Anticyclonic situations prevailed: 57% over cyclonic 20% and zero 23%. Advections from the north-east direction prevailed for 30% of days, followed by advections from the east, the south-east, and the south for 20% each.

May was the only month of the year with temperatures lower than normal within the limits of −0.7 ◦C and was very cold in the prevailing area of the country. In the prevailing area of Poland, it exceeded the precipitation norm and was the month with the highest relative precipitation in the year within 145% of the norm. Anticyclonic situations prevailed at 54%, and the prevailing north-west direction of advection was 28%. This frequency was more than twice the frequency of the 1951–2018 multi-year average of 12%.

June was extremely warm over most of Poland, with temperatures within the range of 5.0–6.0 ◦C. Across about 80 percent of the country, there were days with a maximum temperature of >25.0 ◦C every day. In 18 days on almost half of the Polish territory, the maximum temperature exceeded 30.0 ◦C. The probability of such a warm June can be estimated at less than 1%. June was extremely dry on most of the country, and only Eastern Pomerania and Warmia and Mazury received normal or even wet precipitation. The probability of occurrence of such low precipitation at representative meteorological stations can be estimated as 1% at the Kraków station, 27% at the Słubice station, 31% at the Toru ´n station, 6% at the Warszawa station, and 11% at the Wrocław station. Anticyclonic situations prevailed, accounting for 59%. The dominant direction of advection was southerly. Advection from the south direction was 2.5 times higher than the average of 8%. The frequency of circulation from the south-east direction was twice as high as the average for the 1951–2018 period.

July was normal in terms of temperature, only slightly warm in the Carpathians, and very warm in the Sudetes. The probability of such a warm July can be estimated at 40%. July in the prevailing area of Poland was dry or extremely dry. After June and April, it was the third relatively driest month of the year. The probability of occurrence of such low precipitation at representative meteorological stations can be estimated as 59% at the Kraków station, 45% at the Słubice station, 16% at the Toru ´n station, 20% at the Warszawa station, and 22% at the Wrocław station. Cyclonic situations accounted for almost half of the cases. Advection from the north-east direction prevailed strongly, accounting for 33% of all cases in this month.

August was extremely warm over the prevailing area, with anomalies exceeding 2.0 ◦C in the warmest places. The probability of such a warm August can be estimated at 5%. August was humid in the east and south-east of Poland, while the remaining areas were dry and very dry, and, in the Lubuskie Land, it was extremely dry. The probability

of occurrence of such a dry August can be estimated as 59% at the Kraków station, 6% at the Słubice station, 2% in Toru ´n, 20% in Warszawa, and 30% in Wrocław. Anticyclonic situations prevailed with a frequency of 44%. South-west-oriented advection prevailed with a frequency of 29%, followed by south-oriented advection with a frequency of 16%.

September was warm and sometimes slightly warm. September on the prevailing area of Poland was in the range between humid and extremely humid, only in the south-east, and, locally, in Silesia and the Sudeten basins, September was dry or very dry. Advection from the north-west direction prevailed with a frequency of 17%, which was 8% higher than the norm. Anticyclonic, cyclonic, and null situations occurred with an equal frequency of 33%.

October was generally very warm. The month was dry and very dry over the prevailing area of Poland, and, locally, in the center of the country, in the east, and in the south, it was even extremely dry. Cyclonic systems prevailed, accounting for 41%, with zero circulation 38%. Advection from a southernly direction with a frequency of 20% was dominant, which was 7% higher than normal. Advection from west direction was frequent—19%, also 7% more frequent than the norm, north-east direction—16% which was 9% more frequent than the norm.

Driest months (June, July, and August).

June with a temperature deviation of 5.0 ◦C for the whole Poland was the warmest month in the period since 1951. Although in the period 1951–2019 there were extreme years with very high temperatures as in 1964 with a deviation of 2.2 ◦C and in 1979 with a deviation of 1.9 ◦C but they were within 2.0 ◦C or below.

July was no longer so warm but normal. Years with temperature deviations above 2.0 ◦C occurred in the period since 1951 in 2006 with a deviation of 3.5 ◦C, with a deviation of 2.6 ◦C in 1994 and with a deviation of 2.4 ◦C in 2010.

In August, which was an extremely warm month in the mentioned period 1951–2019, there were years even warmer, such as 2015 with a deviation of 3.5 ◦C, 2.7 ◦C in 1992, 2.6 ◦C in 2018, and 2.1 ◦C in 2002.

As far as the moisture characteristics of the year 2019 are concerned, in the light of the flows for the whole area of Poland for the period 1951–2019, it ranks fifth after the driest years of 1954, 2015, 2016, and 1952 [Bulletin of the National Hydrological and Meteorological Service No. 13/215 2019].
