The main objective of this study was to analyse the spatial and temporal characteristics of drought in Serbia using the standardised precipitation index (SPI) and the standardised precipitation and evapotranspiration index (SPEI). The duration and severity of drought were determined, and the trend of the indices for different accumulation periods, for 3 months (SPI3 and SPEI3), 6 months (SPI6 and SPEI6) and 12 months (SPI12 and SPEI12) were analysed for 10 meteorological stations in Serbia in the period 1961–2020. There was a high correlation between the SPI and SPEI on all time scales. Droughts were observed in Serbia in the following periods: 1961–1963, 1971–1972, 1987–1993, 2000–2003 and after 2011. The drought observed in 2000–2001 was recorded with both indices for all time scales at all stations. The longest drought was measured from June 2011 to February 2013, with a duration of 21 months and a severity of −37.5 with the SPEI12 in Loznica. The 1972 drought is observed in almost all of Serbia, although it was more pronounced in northern Serbia. A statistically significant negative trend is observed for SPEI12 at 6 stations. The SPEI tends to show longer and more severe dry periods than the SPI for the later part of the period under consideration, especially after 2000. As there is a significant increase in temperature in Serbia, the SPEI is better suited than the SPI for analysing drought characteristics under warming conditions.

This paper examines extreme precipitation events (EXPEs) and their trends based on daily precipitation values observed at 14 stations in Serbia for the period 1961–2020. The following EXPEs were investigated: RR10mm (heavy precipitation days), RR20mm (very heavy precipitation days), Rx1day (highest 1-day precipitation amount), Rx3day (highest 3-day precipitation amount), Rx5day (highest 5-day precipitation amount), R95p (very wet days) and R99p (extremely wet days). A positive trend for all EXPEs was dominant in Serbia from 1961 to 2020. All annual Rx1day time series show a positive trend, which is significant at 12 out of 14 stations. The highest values of all EXPEs were observed in 2014, when the annual precipitation totals were the highest at almost all stations in Serbia. To examine the potential influence of global warming, the mean values of the EXPEs were calculated for two periods: 1961–1990 and 1991–2020. In the second period, higher values were determined for all EXPEs than in the first period. The large-scale variability modes, such as the North Atlantic Oscillation (NAO), the East Atlantic Oscillation (EA), and the East Atlantic–West Russia (EAWR) pattern, were correlated with the EXPEs. A negative correlation was found between the EXPEs and the NAO and the EAWR, and a positive correlation between the EXPEs and the EA pattern. For future research, the contribution of high-resolution data will be examined.

A better understanding of climate change impact on dry/wet conditions is crucial for agricultural planning and the use of renewable energy, in terms of sustainable development and preservation of natural resources for future generations. The objective of this study was to investigate the impact of climate change on temporal fluctuations of dry/wet conditions in Serbia on multiple temporal scales through multifractal analysis of the standardized precipitation evapotranspiration index (SPEI). We used the well-known method of multifractal detrended fluctuation analysis (MFDFA), which is suitable for the analysis of scaling properties of nonstationary temporal series. The complexity of the underlying stochastic process was evaluated through the parameters of the multifractal spectrum: position of maximum ${\alpha }_0$ (persistence), spectrum width $W$ (degree of multifractality) and skew parameter $r$ dominance of large/small fluctuations). MFDFA was applied on SPEI time series for the accumulation time scale of 1, 3, 6 and 12 months that were calculated using the high-resolution meteorological gridded dataset E-OBS for the period from 1961 to 2020. The impact of climate change was investigated by comparing two standard climatic periods (1961–1990 and 1991–2020). We found that all the SPEI series show multifractal properties with the dominant contribution of small fluctuations. The short and medium dry/wet conditions described by SPEI-1, SPEI-3, and SPEI-6 are persistent ($0.5<{\alpha }_0<1$); stronger persistence is found at higher accumulation time scales, while the SPEI-12 time series is antipersistent ($0<{\alpha }_0-1<0.5$). The degree of multifractality increases from SPEI-1 to SPEI-6 and decreases for SPEI-12. In the second period, the SPEI-1, SPEI-3, and SPEI-6 series become more persistent with weaker multifractality, indicating that short and medium dry/wet conditions (which are related to soil moisture and crop stress) become easier to predict, while SPEI-12 changed toward a more random regime and stronger multifractality in the eastern and central parts of the country, indicating that long-term dry/wet conditions (related to streamflow, reservoir levels, and groundwater levels) become more difficult for modeling and prediction. These results indicate that the complexity of dry/wet conditions, in this case described by the multifractal properties of the SPEI temporal series, is affected by climate change.

In this study the impact of climate change on drought conditions in Serbia is investigated trough the multifractal analysis of standardized precipitation evapotranspiration index (SPEI). The SPEI time series for accumulation time scale of 1, 3, 6 and 12 months were calculated using the high resolution meteorological gridded dataset E-OBS with a resolution of 0.10 in regular latitude/longitude coordinates and time span from 1961 to 2020. The multifractal analysis was performed for two standard climatic periods (1961−1990 and 1991−2020) using multifractal detrended fluctuation analysis method (MFDFA) which is suitable for nonstationary temporal series. In each of 1278 grid cells that cover Serbian territory we analyzed 4 SPEI time series for two 30 years periods (which totals 10224 MFDFA runs) and calculated multifractal spectrum parameters that describe persistence, degree of multifractality and the influence of small/large fluctuations. We found that all the analyzed SPEI series show multifractal properties with the dominance of small fluctuations. The short and medium drought conditions described by SPEI-1, SPEI-3 and SPEI-6 are persistent (the position of maximum of multifractal spectrum is in the range 0.5<α0<1), stronger persistence is found at higher accumulation time scales, while SPEI-12 time series are antipersistent (0<α0-10) indicating that small SPEI temporal fluctuations contribute more to multifractality than large fluctuations. In the second period SPEI-1 SPEI-3 and SPEI-6 series become more persistent with weaker multifractality, indicating that short and medium drought conditions (which are related to soil moisture and crop stress) become easier to predict, while SPEI-12 changed toward random regime and stronger multifractality in eastern and central part of country, indicating that long term drought conditions (related to streamflow, reservoir levels, and groundwater levels) become more difficult for modeling and prediction.

Quantifying the intensity of a drought is not an easy task, as one has to take intoaccount multiple parameters, for example rainfall, temperature, soil water content, etc.Numerous indices have been developed to tackle this issue, and some of the most widelyused are standard precipitation index (SPI) and standard precipitation evapotranspiration index(SPEI), which, as the name implies, standardize precipitation and evapotranspiration over apoint for a certain time period and use that as a reference for an estimate of available water.The goal of this work is to analyze these drought indices for a long-term time period,take a look at their correlation and compare them with available records. The theoreticaldifference between the two indices is in the fact that SPEI takes (potential) evapotranspirationinto account, while SPI does not. Since sufficient measured evapotranspiration data was notavailable, potential evapotranspiration was used, calculated by the Thorntwaite method(taking into account temperature and geographical position). Precipitation and temperaturedata was acquired from the station network ran by the Republic Hydrometeorological Serviceof Serbia, for the time period of 1961-2020. Since the SPI and SPEI can be computed fordifferent timescales, in this study the indices for 3-, 6- and 12-months were compared.The results show that despite high correlation between SPI and SPEI (r > 0.9), theirbehavior in regards to drought events shows an important difference. This is visible in thetemporal spread of the dry months. SPEI depicts the last two decades as much drier than SPIdoes, and inversely, the sixties and the seventies as much wetter. Also, for the whole timeperiod, SPEI shows a trend towards more dry conditions, while SPI shows no clear trend.This is a consequence of rising potential evapotranspiration, which in itself is a consequenceof rising average temperatures.The conclusion is that the indices should not be used interchangeably and withtemperature data being as available as precipitation data, SPEI is more representative ofwater budget in the area in question.

Extreme precipitation events (EXPEs) were analyzed based on daily precipitation data from 1961 to 2020 in Novi Sad, Serbia. The temporal characteristics of the following EXPEs were investigated: Heavy precipitation days (R10mm), Very heavy precipitation days (R20mm), Highest 1-day precipitation amount (Rx1day), Highest 3-day precipitation amount (Rx3day), Highest 5-day precipitation amount (Rx5day), Very wet days (R95p), Extremely wet days (R99p), Precipitation fraction due to very wet days (R95pTot) and Precipitation fraction due to extremely wet days (R99pTot). The EXPEs were analyzed on an annual and seasonal basis and for two reference periods 1961-1990 and 1991-2020. Positive trends were found for both annual and seasonal values for all indices, except for R20mm and R99pTot in winter. A significant increase in Rx1day, Rx3day and Rx5day was observed in all seasons (except for Rx1day and Rx5day in winter) and on an annual basis. The high value of Rx1day (116.6 mm) was recorded in the summer of 2018 in Novi Sad, caused by convective precipitation that led to urban flooding. The possible influence of large-scale circulation patterns was investigated. A strong positive and negative influence of the East Atlantic pattern and the East Atlantic Western Russia pattern on the EXPEs was found. The results of this work support the growing evidence that the impact of extreme conditions is likely to become even stronger due to changes in their intensity and frequency.