*4.1. Air Temperature and Precipitation Variability*

For all stations included in the study, an increase in mean annual air temperature has been observed; the regression analysis showed that it was statistically significant at *p* < 0.05. The rate of increase varied from 0.5 to 0.7 ◦C per 10 years (R<sup>2</sup> from 0.34 to 0.47). The Mann-Kendall test confirmed statistically significant positive trends for all stations, and Sen's slope values confirmed the rate of increase described above (Appendix B). The variability coefficient for all series is <25%, which means a relatively low variability of mean annual air temperature in the study area. Figure 2 shows that the mean decadal air temperature has been gradually increasing at all stations, too. The most striking increase is observed for Kasprowy Wierch, one of the highest peaks of the Tatra Mts, where mean annual air temperature has exceeded 0 ◦C which means a shift from the "moderately cold" vertical climatic zone (i.e., from mean annual air temperature from −2 to 0 ◦C) to "very cool" (0 to 2 ◦C). Such a shift can also be observed for Zakopane (from "moderately cool" to "moderately warm"). Lesko and Limanowa shifted from "moderately warm" to a mean annual air temperatures above 8 ◦C, not included in the original scheme described in [29].

**Figure 2.** Mean annual air temperature (◦C, black horizontal marks inside the boxes) in specific decades at the stations studied. Boxes mark the first and the third quartile, and the whiskers show the highest and the lowest value in a certain decade. Standard errors for the mean values are provided in Appendix C. Stations are ordered following the concept of TPZ explained in Section 4.1.

In the case of precipitation, there are no statistically significant changes for annual totals (according to regression analysis and the Mann–Kendall test; see Appendix B). The comparison of mean annual totals in specific decades confirms this fact (Figure 3); the highest values were noted in the second decade. The values of the variability coefficient for annual totals are below 25% which means low variability. However, the values for particular months reveal that for Kasprowy Wierch, Bielsko-Biała, Limanowa and Kraków, for all months the coefficient values are >45%, which means high variability. For Zakopane, Katowice, Łazy, Nowy S ˛acz, Gaik-Brezowa and Krynica, from 1 to 3 months show mean variability (25–45%) but for all other months, the coefficient exceeds 45%. For Lesko and Koma ´ncza, 4–5 months show mean variability while all other months have high variability. Additionally, there is no clear dependency between precipitation and altitude (except Kasprowy Wierch) or precipitation and longitude, and this is linked to the strong local impacts of landforms in the mountains on spatial patterns of precipitation.

Both air temperature and precipitation are key factors controlling drought occurrence and Figure 4 shows their combination for the stations included in the study. The stations can be then assigned to the following temperature-precipitation zones (TPZ):


Kraków and Katowice, where the mean annual air temperature is the highest in the vertical profile (8.9–9.0 ◦C), and annual precipitation totals are very diversified, from 670 to 1000 mm, so the potential drought risk is high.

**Figure 4.** Comparison of mean annual air temperature and precipitation totals for the period 1991–2020 for the stations studied. The value for Nowy S ˛acz is not visible in the figure as it is almost identical to the value for Katowice and the symbols overlap each other.

#### *4.2. Drought Frequency and Trends in the Polish Carpathians*

Drought occurrence was determined by SPI, SPEI, RPI and the Selianinov index. For SPI and SPEI, the percentages of dry months (i.e., with SPI ≤ –1.00, SPEI ≤ −0.8) for the whole year and for the subperiods May–October and November–April were calculated. All data series show very high variability, that is, the values of Vc exceed 100%.

SPI values for the 1-month time scale vary from 3.74 to −4.28, for the 3-month scale from 3.81 to −3.13, and for the 6-month scale from 3.62 to −2.78. None of the SPI 1- and 3-monthly series shows any statistically significant trend; the results of the Mann–Kendall test are presented in Appendix B. In the case of SPI 6-monthly series for Zakopane, Krynica, Koma ´ncza (TPZ 2) and Gaik-Brzezowa (TPZ 4), the *p*-values indicate statistically significant increasing trends, but Sen's slope values are as low as 0.001, and low tau values indicate that the trends are weak (Appendix B). A 1-month SPI reflects short-term conditions, related closely to meteorological drought along with short-term soil moisture and crop stress, while a 3-month SPI reflects short- and medium-term moisture conditions and provides a seasonal estimate of precipitation, and a 6-month SPI indicates seasonal to medium-term trends in precipitation [39]. Figure 5 presents the data for the decades and it shows that in a short-term perspective, a clear increase in the frequency of dry months in the cold subperiod can be seen in the last decade in comparison to the previous ones (except Gaik-Brzezowa; Figure 5c). This is also the reason for the increase of dry-month frequency at an annual scale (Figure 5a). The increase is observed throughout the whole study area. In the decade 2011–2020, the frequency of dry months according to SPI reached about 15% on an annual scale at all stations. For medium-term and seasonal perspectives (Figure 5, data for 3- and 6-monthly timescales), there are no clear spatial or temporal patterns.

**Figure 5.** Percentages of dry months according to SPI (SPI ≤ −1.0) for 1-, 3- and 6-monthly timescales, for the whole year (**a**) and for the subperiods May–October (**b**) and November–April (**c**) in the decades of the study period. Stations are ordered following the concept of TPZ explained in Section 4.1. Explanation of numbers on axis x: 1—1991–2000, 1 month; 2—2001–2010, 1 month; 3—2011–2020, 1 month; 4—1991–2000, 3 months; 5—2001–2010, 3 months; 6—2011–2020, 3 months; 7—1991–2000, 6 months; 8—2001–2010, 6 months; 9—2011–2020, 6 months.

For annual values of RPI, the Mann–Kendall test by definition gives the same results as for annual precipitation totals (Appendix B). For the test's results for particular months and particular stations, all *p*-values exceeded 0.05, so none of the series shows any statistically significant trend. Tau values varied from −0.264 to 0.209. Figure 6 shows that in the decade 2011–2020, for most of the stations, a large increase in the number of dry months per year can be seen which is mainly the effect of the increase in the frequency of such months in the cold half-year. In the last decade of the study period, according to RPI, from 35 to 45% of months per year were dry.

**Figure 6.** Percentage of dry months according to RPI (RPI ≤ 75%-for the whole year) and for the subperiods May-October and November-April in the decades of the study period. Stations are ordered following the concept of TPZ explained in Section 4.1. Explanation of numbers on axis x: 1—1991–2000, year; 2—2001–2010, year; 3—2011–2020, year; 4—1991–2000, May–Oct; 5—2001–2010, May–Oct; 6—2011–2020, May–Oct; 7—1991–2000, Nov–Apr; 8—2001–2010, Nov–Apr; 9—2011–2020, Nov–Apr.

The Selianinov index could be calculated for all stations (except Kasprowy Wierch) and for each year only for June, July and August; for other months the values could be calculated only sporadically. The Mann–Kendall test for those three months showed no statistically significant trend at any station. All *p*-values exceeded 0.05, tau values varied from −0.159 to 0.062. However, the Selianinov index, unlike SPI and RPI, showed a large spatial variability of drought occurrence in the warm part of the year (Figure 7), with a clear increase in drought risk with decrease in altitude (i.e., from less than 10% of dry months at Kasprowy Wierch (TPZ 1) to over 50% in Kraków, TPZ 4). Additionally, the data show that for most stations, the share of dry months is greater in the last decade than in the two previous ones. An increase is especially visible in the highest parts of the Carpathians. Until 2015, the Selianinov index could be calculated for Kasprowy Wierch only once over several years, and only for July, while later it could be calculated also for June and August, as the index is calculated only for the months when the mean daily air temperature exceeds 10 ◦C. There is no significant difference in the W-E profile concerning drought risk, but the data for Bielsko-Biała are worth attention as the risk is much lower than in other stations of similar locations. Bielsko-Biała and Katowice (TPZ 4) are the westernmost points of the study area, and both of them are exposed to moist oceanic air masses coming from the west; however, Bielsko-Biała is located in the Carpathian foothills, that is, close to an orographic barrier which enhances precipitation. Figure 3 shows that Bielsko-Biała has higher precipitation sums than neighboring stations.

**Figure 7.** Percentage of dry months according to the Selianinov index for the subperiod April–October in the decades of the study period. Stations are ordered following the concept of TPZ explained in Section 4.1.

SPEI values for the 1-monthly time scale vary from 2.54 to −3.15, for the 3-monthly scale from 2.60 to −2.83, and for the 6-monthly scale from 2.72 to −3.44. For the 1-monthly series, statistically significant trends were found with the Mann–Kendall test for Katowice, Nowy S ˛acz and Kraków (TPZ 4), and for 3-monthly and 6-monthly series for Kasprowy Wierch (TPZ 1), Katowice, Nowy S ˛acz and Kraków (TPZ 4) (Appendix D). The trends indicate an increase in drought risk although Sen's slope values are as low as 0.001–0.002 which indicates that the trends are weak. However, most tau values are much higher than for SPI (Appendix B), which shows that those trends, although weak, should be considered important signals of the increasing drought risk at least in some areas of the Polish Carpathians, mainly in foreland areas. SPEI presents combined effects of precipitation and air temperature and concerning the results presented above, it is the increasing temperature that is mainly contributing to those trends. Figure 8 presents the percentage of the dry months (SPEI ≤ −0.8) for the decades and it shows that in the last decade of the study period, there were much more dry months observed at most of the stations than previously. Such tendency is more clear for the cold half-year than for the warm one, especially for the 1-monthly time scale.
