Comparative analysis of SPI and SPEI drought indices for Montenegro and the impact of teleconnections

There is a general consensus that today's global warming, along with more frequent and intense weather and climate extremes, is occurring due to the anthropogenic greenhouse gas effect, primarily resulting from excessive burning of fossil fuels. When it comes to changes in precipitation, significant differences are observed. Some regions become drier, while others become wetter, with alternating dry and wet periods in some areas (IPCC 2021). To more accurately determine precipitation changes, various methods and indices are employed. For the assessment of meteorological and hydrological drought, commonly used indices include (Mukherjee et al. 2018): the Palmer drought severity index (PDSI), the standardized precipitation index (SPI), and the standardized precipitation evapotranspiration index (SPEI), as well as the surface water supply index (SWSI) and the aggregate drought index (ADI). One widely used indicator of precipitation change, especially for characterizing meteorological drought, is SPI, recommended by the World Meteorological Organization (WMO 2012). This index quantifies precipitation for a given time unit as a standardized deviation from the average value, depending on the chosen probability distribution function – usually Gamma or a Pearson Type III distribution (Zuo et al. 2021). SPI indicates both precipitation deficit and surplus, making it suitable for identifying both dry and wet periods.

SPEI is often used to assess available water resources. The advantage of SPEI over SPI is that it considers potential evapotranspiration (PET) in addition to precipitation, making it a more representative indicator (Filipovic & Tosic 2024). Analysis of SPEI for neighboring Serbia revealed prevailing negative trends for the period 1950–2022, indicating more frequent droughts in recent decades (Djurdjević et al. 2024). Many studies analyze these two drought indices (SPI and SPEI) together (e.g., Oksal 2023; Tomasella et al. 2023).

Drought is classified among frequent climate extremes, and the remediation of its consequences can be very costly (García-León et al. 2021). Estimates indicate that over the last decades, within the EU, the number of people affected by drought has increased by 20% (Moccia et al. 2022). Drought occurs due to a prolonged lack of precipitation, lasting for months or even years. Due to climate change and rising temperatures, an increase in the frequency, intensity, and duration of drought is expected (Markonis et al. 2021). In the past, drought was typically associated with desert and warm regions, but in recent decades, it has become more common in temperate and humid climate zones. Even countries where droughts were very rare, such as Ireland, now experience them more frequently (Augustenborg et al. 2022).

The results from the trend analysis of global drought frequency, duration, and intensity during the period 1951–2010 indicate a slight overall increase in each drought component. Regionally, significant increases in drought frequency, duration, and intensity have been observed in Africa, East Asia, the Mediterranean region, and southern Australia, whereas America and Russia have shown a decrease in each drought component (Spinoni et al. 2014).

Regarding Europe, SPI and PDSI trends suggest that moderate and extreme drought (ED) conditions changed only slightly during the 20th century (1901–1999) (Lloyd-Hughes & Saunders 2002). In the early 21st century, certain regions of Europe have experienced more frequent and severe droughts than before. Particularly notable are the mega-droughts and intense heatwaves in Europe in the years 2003, 2010, 2015, 2018, and 2022. The drought and heatwave that affected much of Europe in the summer of 2022, especially the southwest and west of the continent, were unprecedented in recent history. This combination of dry and warm conditions led to water shortages and numerous wildfires across much of the European Mediterranean (Copernicus Climate Change Service 2023; Faranda et al. 2023). Future projections indicate that due to expected increases in evapotranspiration from rising temperatures, coupled with changes in seasonal precipitation patterns, droughts will become more frequent in the Mediterranean region, western Europe, and northern Scandinavia, particularly under the moderate RCP4.5 scenario. According to the more extreme RCP8.5 scenario, the entire continent is expected to experience more frequent and extreme droughts during summer and spring (Spinoni et al. 2018). Recent SSP scenarios also predict more frequent and intense droughts in Europe (IPCC 2021). Projections suggest that by 2030, extreme heatwaves and droughts in Europe could be up to 10 times more frequent compared with those at the end of the 20th century (Suarez-Gutierrez et al. 2023).

The Mediterranean region is among the more vulnerable areas to contemporary climate changes, experiencing reduced rainfall and more frequent and intense droughts in recent decades (Cook et al. 2018). For example, in most of Spain, there is a noticeable trend toward longer-lasting and more severe droughts from 1961 to 2014 (Domínguez-Castro et al. 2019). SPI calculations over a 100-year period (1919–2019) in Portugal revealed an increase in the frequency of moderate and severe droughts (Espinosa & Portela 2022). Italy witnessed more frequent drought occurrences between 2001 and 2016, with estimated economic losses ranging from 0.55 to 1.75 billion euros over that 15-year span (García-León et al. 2021). Based on SPI analysis, Greece has experienced a rise in drought frequency since 1989 (Tsesmelis et al. 2022). Nearly the entire Eastern Mediterranean region has seen an upward trend in the frequency of ED events in the last 30 years (Tsesmelis et al. 2023). Projections from various models until the end of the 21st century consistently indicate an increase in drought frequency across the Mediterranean region (Tramblay et al. 2020).

Drought occurrences have been increasing on the Balkan Peninsula, especially in the last decade. Notable events in 2018 and 2019 were characterized by significant winter precipitation deficits and record-high temperatures during the summer across much of the region (Blauhut et al. 2022). Droughts pose the greatest threat to stability and prosperity in the Balkan region, particularly affecting the agriculture and hydroenergy sectors (Barron & van Manen 2022).

Montenegro, a geographically small Mediterranean and Balkan country, has undergone significant weather and climate variations over the past two to three decades. Research indicates that weather extremes have become more intense (Mihajlović et al. 2021). Droughts and heatwaves are more frequent with varying durations from year to year, leading to significant consequences in the forestry, agriculture, and water sectors (MNDP 2020). Therefore, the aim of this study is to address the issue of drought in Montenegro for the 1961–2020 period. Through new and precise results, the goal is to draw decision-makers' attention to the occurrence of drought and highlight that this extreme event could become a serious problem in the near future. The main research questions are as follows:

• – Calculation of the trend and significance of changes in precipitation in the observed period.

• – Categorization of SPI and SPEI, i.e., trend analysis of these two indices for 1, 2, 3, 6, 9, and 12 months, in order to see if the intensity of drought/humidity increases or decreases and if the changes are statistically significant.

• – Example of data for the grid field of the main and most populous city in Montenegro (Podgorica), determination of SPEI changes during two simultaneous sub-periods within the observed time series: 1961–1990 and 1991–2020. SPEI was chosen because it is a more ‘sensitive’ drought index compared with SPI.

• – Montenegro has a Mediterranean pluviometric regime; summers are arid, even dry in most of the country. In this regard, by applying the rescaled adjusted partial sums (RAPS method) for SPEI, for the Podgorica grid field, the tipping point of sudden changes will be defined for all considered time series for the month of August (SPEI-1, SPEI-2, SPEI-3, SPEI-6, SPEI-9, and SPEI-12).

• – After a comprehensive analysis of SPI and SPEI, also on the example of the grid field of the Podgorica meteorological station (MS), the connection between changes in SPEI and 15 indicators of variations in atmospheric and oceanic oscillations (teleconnection) will be examined.

Montenegro is situated in the Mediterranean region of Southeastern Europe and the Balkan Peninsula. It spans between 41°50′–43°32′ N and 18°26′–20°21′ E, giving it a geographical width of 1.7° or 189 km, and a geographical length of 1.92° or 163 km. The country terrain, totaling 13,812 km², is morphologically divided into two main units: a larger mountainous–valley region covering the internal and northern parts, and a narrow coastal region in the south along the Adriatic Sea, along with a flat area surrounding the capital city, Podgorica. From sea level to its highest point, there is an elevation difference of 2,534 m, and the relief dissection, influenced by dominant synoptic patterns that significantly impacts Montenegro's climate differentiation. Rainfall distribution in Montenegro is highly uneven, with average annual precipitation from 1961 to 2020 ranging from 800 mm in the far north (Pljevlja Station) to as much as 4,600 mm in the southwest (Crkvice Station).

Based on data from the MS, the SPI has been calculated. SPI was formulated by McKee et al. (1993) for the purpose of defining and monitoring drought. This index measures precipitation anomalies at a given location in relation to the average precipitation for a specific time interval, i.e., a ‘backward step’ (e.g., 1, 3, 6, 12, and 24 months). For the purposes of this study, the SPI Program of the National Drought Mitigation Center (2018) has been used, which is based on the Gamma distribution function of precipitation.

To gather information for both shorter (up to 3 months) and longer (more than 3 months) durations, SPI and SPEI have been computed for each last month of the meteorological season (February, May, August, and November) across 1, 2, 3, 6, 9, and 12-month periods (SPI1, SPI2, SPI3, SPI6, SPI9, and SPI12, as well as SPEI1, SPEI2, SPEI3, SPEI6, SPEI9, and SPEI12). These calculations pertain to meteorological seasons (entire months), not calendar seasons: winter (December, January, February – DJF), spring (March, April, May – MAM), summer (June, July, August – JJA), and autumn (September, October, November – SON). For instance, SPI1 focuses solely on precipitation for the respective month (e.g., SPI1 for August 2003 considers precipitation in August 2003), while other durations involve a retrospective approach. Specifically, SPI6 for August 2003 encompasses cumulative precipitation for August and the preceding 5 months of that year, spanning two seasonal periods (summer and the prior spring of 2003). Each time unit (1, 2, 3, 6, 9, and 12 months) for a given year or two consecutive years (e.g., SPI3 for February 2003 includes cumulative precipitation for February and January 2003, and December 2002) is classified based on its dryness–wetness condition. Subsequently, a 60-year dataset (1961–2020) was utilized to calculate SPI and SPEI trends. This 60-year period suffices for evaluating general developmental trends and determining whether Montenegro is experiencing shifts toward drier or wetter conditions.

The obtained SPI and SPEI values have been categorized into specific classes (Table 1). The trend has been calculated using the Sen's method, which essentially provides a linear trend (showing a line and trend value), and its significance has been assessed using the Mann–Kendall (MK) test at significance levels of p < 0.001, 0.01, 0.05, and 0.1, corresponding to accuracy levels of 99.9, 99, 95, and 90% (Mann 1945; Kendall 1975; Salmi et al. 2002). The RAPS method has been employed to identify breakpoints, indicating the onset of abrupt changes, exemplified by SPEI for the Podgorica MS. Detailed explanation of Sen's method and the MK test was given in the study Burić & Doderović (2022), while RAPS was given in the study Šrajbek et al. (2023).

At the end of this subsection, it is important to note that there are four fundamental approaches to measuring drought (Wilhite & Glantz 1985): meteorological, hydrological, agricultural, and socio-economic. The first three approaches focus on measuring drought as a physical phenomenon, while the last one (Zhao et al. 2019) considers it from the perspective of supply and demand, monitoring the effects of water scarcity and its impact on the socio-economic system. SPI1 to SPI3 are commonly used to monitor meteorological drought, SPI6 to SPI9 for agricultural drought, and SPI12 and longer time scales for hydrological drought (Eini et al. 2023). An SPI9 value less than –1.5 is a significant indicator of drought impact in agriculture, potentially affecting other sectors and leading to additional disasters such as forest fires and heatwaves (WMO 2012). At one MS, this study examines six time scales (1, 2, 3, 6, 9, and 12 months), totaling 24 time scales for SPI and 24 for SPEI, aiming to identify general trends in meteorological, agricultural, and hydrological droughts. Thus, 432 time series for SPI (24 × 18 MS) and 240 series for SPEI (24 × 10 grids) were formed – a total of 672 series.

The association between the analyzed SPEI and indicators of atmospheric and oceanic oscillations (teleconnections) has been examined, specifically for the capital city of Montenegro (Podgorica). Table 2 lists the teleconnections used and their sources of data. Pearson's correlation coefficient has been used to calculate correlation, and its significance has been tested using the t-test at 95% (p < 0.05) and 99% (p < 0.01) confidence levels. These teleconnections have been comprehensively described in studies by Burić & Stanojević (2020). Among the 15 teleconnections considered, the Atlantic Multidecadal Oscillation (AMO) and the El Niño–Southern Oscillation (NINO3.4) have been identified as oceanic, while the others are atmospheric.

In accordance with the set research questions, this section is divided into several subsections. Since precipitation is a key meteorological variable that affects drought/humidity, i.e., SPI and SPEI index values, the trend analysis of changes in monthly precipitation will be presented first. The results of the SPI and SPEI index trends are given in the second subsection. The third and fourth subsections will deal with the analysis of SPEI for the grid field MS Podgorica in two simultaneous sub-periods (1961–1990 and 1991–2020), that is, the determination of the turning points of changes using RAPS transformations. Finally, the results of the correlation between SPEI for the grid field MS Podgorica and 15 indicators of telecommunications were presented.

It has been noted that precipitation in Montenegro is highly unevenly distributed (ranging from 800 mm in the far north to 4,600 mm annually in the southwest of the country). For this reason, precipitation trends have been calculated in percentages per decade (%/decade). The trend calculations have shown that monthly precipitation changes are generally minor. In January, May, and from September to December, precipitation trends are insignificant across all observed MS. Over the observed 60-year period (1961–2020), significant trends (p < 0.05 and p < 0.1) were observed only in February, May, June, and July at one MS each, and in August at two MS. In April, the trend is negative across all MS, with significance at 4 MS. Besides April, August is the only other month where a negative trend is observed across all observed stations. In other months, precipitation trends vary, with some being positive and others negative (Table 3). For the same period, Burić et al. (2024) found that annual precipitation trends range from –5.4 to 1.9%/decade, with only MS Krstac showing a statistically significant decrease (p < 0.01).

Table 3

Monthly precipitation trends (%/decade) in Montenegro for the period 1961–2020

   Significant changes.

Significant: *p < 0.05 and ⁺p < 0.1.

View Large

To further define the structure of changes, annual counts of days with precipitation < 1 mm (Rd < 1 mm, dry days) and Rd ≥ 20 mm (very wet days) have been isolated. Only one meteorological station (MS Krstac) shows a significant decrease in the number of very wet days (–2.1 days/decade), while changes at the other 17 MS are insignificant (ranging from –0.9 to 0.6 days/decade). Montenegro experiences a high number of dry days, averaging from 233 days annually in mountainous regions in the north to around 270 days on the coast and in Podgorica. All MS record an increasing trend in the number of dry days (from 0.2 to 3.3 days/decade), with a statistically significant positive trend observed at 6 MS (p < 0.01). The increase in annual dry days suggests intensifying and prolonged drought conditions in Montenegro, although this alone is not sufficient evidence. Closer attention will be paid to the analysis of SPI and SPEI.

The obtained results for the SPI1-F trend (F – February) have consistently shown positive trends across all observed MS, ranging from 0.01 to 0.17 per decade. The trend is statistically significant only at the Žabljak MS (p < 0.01), indicating an increase in humidity intensity in February. As for the SPEI1-F trend, the range of changes is smaller (from –0.20 to 0.03) and statistically insignificant, but a greater number of MS show a negative tendency. Except for Žabljak, where SPI shows a significant positive trend (p < 0.05), both indices show insignificant changes for the 2-month step (SPI2-F and SPEI2-F), although in this case, there is spatial dominance of the negative trend in SPEI2 compared with SPI2. For SPI and SPEI over 3 and 6 months (SPI3, SPI6, SPEI3, and SPEI6), the trend is insignificant at all stations. SPI9 shows a significant negative trend at one station (MS Krstac), while the trend for SPI12 is insignificant at all MS. However, the trend for SPEI9-F and SPEI12-F is negative (Figure 3), and significant over a larger part of Montenegro (p < 0.1).

The warmer part of the year in Montenegro experiences the least precipitation and high evapotranspiration rates. There is increased demand for water, particularly in key economic sectors such as tourism and agriculture. Considering these factors, the analysis of SPI and SPEI for spring and summer holds particular significance. SPI1-M (M – May) shows a positive trend at all observed stations (with trends ranging up to 0.10 per decade). The trend for SPEI1-M is positive in the north and northwest of Montenegro, while the majority of the country exhibits a negative trend during this month. For all other temporal scales, a negative SPI trend predominates (SPI2-M, SPI3-M, SPI6-M, SPI9-M, and SPI12-M). Only MS Tivat registers a significant negative trend for SPI9-M and SPI12-M (–0.14 and –0.15 per decade, respectively). While SPI trends show a positive sign at two to five MS, SPEI for May across these periods (SPEI2-M, SPEI3-M, SPEI6-M, SPEI9­M, and SPEI12-M) exhibits a negative trend across Montenegro (Figure 4). The central, northern, and northeastern parts of the country (7 out of 10 MS) register a significant negative trend for SPEI12-M, ranging from –0.12 to –0.17 per decade (with significance levels of p < 0.1 and 0.05).

Throughout almost the entire territory of Montenegro, summer is the driest period of the year in terms of precipitation. In the Mediterranean and similar climate regions, drought is predominantly associated with summer in human perception. As previously mentioned, it is anticipated that droughts will become more frequent and severe during summers across the Mediterranean region as a whole. Calculations of the SPI for Montenegro confirm the ongoing trend of increasing dryness during the summer season, with a clear prevalence of negative SPI trends across all time scales (Figure 5). According to SPI values, only one to three MS register a positive trend, while in all other cases, it is negative. Qualitatively similar results are obtained for SPEI for the month of August. The difference from SPI is that SPEI shows a consistently negative trend across all MS. Specifically, for SPI1-A and SPE2-A (A – August), the negative trend is significant at 10 MS and at 9 MS (p < 0.1 and p < 0.05), and for other time scales (SPI3-A, SPI6-A, SPI9-A, and SPI12-A), it is significant at 2 to 3 MS (p < 0.1).

For 1, 2, 3, 6, and 9 months, the negative trend of SPEI for August is significant at all observed MS or grid cells (significance levels: p < 0.01, p < 0.05, and p < 0.01). An increase in drought conditions for SPEI12-A (summer + spring + winter + autumn) is present across the entire territory of Montenegro, but this trend is statistically insignificant. SPI and SPEI for August across all temporal scales show the most pronounced negative trends. As a reminder, summer months (JJA) across nearly all MS exhibit a slight decreasing trend in precipitation, hence SPI1, SPI2, and SPI3 trends are predominantly negative. The decrease in summer precipitation and in April has influenced SPI trend values for longer periods, mainly showing negative trends (SPI6, SPI9, and SPI12).

The least changes are observed in the time series ending in November (N). SPI trends for time periods of 1, 2, 3, 6, and 9 months (SPI1-N, SPI2-N, SPI3-N, SPI6-N, and SPI9-N) show insignificance across all MS, with a nearly equal presence of positive and negative trends. Only SPI12-N exhibits a predominantly negative trend across most parts of Montenegro (Figure 6), with significance observed at two MS (Tivat and Herceg Novi, trend = –0.12 and –0.13 per decade, significance level p < 0.1). The quantitative and spatial distribution of the SPEI1-N and SPEI2-N trends mirror those of their corresponding SPI (SPI1-N and SPI2-N).

SPEI3-N shows a positive trend across all grid cells, indicating a decrease in drought severity for the autumn season. The three autumn months (SON) warm up the slowest throughout the year in Montenegro, resulting in minimal changes in evaporation rates. Further analysis reveals that nearly all MS experience a slight increase in precipitation trends during September and October over the period 1961–2020, while precipitation changes are insignificant in November. This explains why SPEI3-N shows a positive trend. Once summer, as well as spring and winter months, is included, negative trends emerge for other temporal scales: SPEI6-N, SPEI9-N, and SPEI12-N. For SPEI12-N, a negative trend (–0.13 per decade) is significant for the grid cells corresponding to MS Plav and Kolašin (p < 0.1).

The results show that the most unfavorable situation is in the summer season. In contrast to precipitation, monthly temperatures have significantly increased across Montenegro, with the most pronounced warming occurring during summer (Burić 2024). This significant warming has increased water evaporation, leading to negative SPEI trend values for August across all analyzed temporal scales in Montenegro. Such SPI and SPEI trends for August are highly unfavorable for agriculture, tourism, and hydroelectric power generation. Summer droughts in Montenegro are becoming more intense. For the years 1975, 1990, 1992, 1993, 2000, 2003, 2007, 2011, 2012, 2016, and 2017, SPI and SPEI values for August over 6, 9, and 12 months were mostly categorized as ED (SPI and SPEI < –2) across a large part of Montenegro. It can be concluded that there has been pronounced agricultural and hydrological drought during these years.

Mention has been made of some extremely dry years in Montenegro, with a greater number recorded in the past 2–3 decades. This section presents a comparative analysis of SPEI changes over these two 30-year periods. SPEI has been chosen because it is a more ‘sensitive’ drought index, considering both precipitation and temperature (evapotranspiration), which is a significant factor in drought assessment. For this analysis, the central MS in Montenegro, MS Podgorica, located in the capital city of the same name, is selected as it can serve as a representative for a larger part of the country. Given that changes are mostly insignificant in most cases, the focus is on the month of August, which generally registers a significant negative SPEI trend.

Similar observations are noted for SPEI-2 for August. For SPEI-3 for August (summer season), there are fewer fluctuations observed, but the year 1984 stands out clearly as a breakpoint, followed by intensified drought. For SPEI-6 for August (summer + spring), the breakpoint is the year 1980, and for SPEI-9 (summer + spring + winter), it is the year 1986. SPEI-12 for August, which accumulates values from August and the previous 11 months (summer + spring + winter + autumn), has a breakpoint in the year 1981.

The previous results indicate that the breakpoint for shorter periods is the year 1984, while for longer periods, it is 1980, 1986, and 1981 (SPEI6-A, SPEI9-A, and SPEI12-A), with several sub-series showing minor or major fluctuations. It is highly likely that the long-term component (trend) of SPEI (and also SPI) is related to global warming, while interannual variability is associated with patterns of global and regional oscillations (variations in atmospheric and oceanic oscillations).

The trend of SPI and SPEI changes has shown that drought in Montenegro is intensifying overall. The intensification of drought is likely a consequence of contemporary climate change (significant warming/increased evapotranspiration and changes in precipitation patterns), but the influence of teleconnections should not be overlooked. To better understand interannual variability and long-term trends, the relationship between SPEI for MS Podgorica and variations in global and regional atmospheric and oceanic oscillations was examined here. Correlation has been calculated with all analyzed SPEI temporal distances. Table 4 presents statistically significant correlation coefficients according to the t-test at levels of p < 0.05 and p < 0.01. Correlation coefficients with absolute values less than 0.27 are considered insignificant.

Table 4

Correlation coefficients between teleconnections and SPEI temporal series for the grid field MS Podgorica, 1961–2020^a

^aFor AMO, NINO3.4, SOI, and POLEUR, results are not shown because the correlation is not significant.

View Large

The obtained results have shown significant relationships between all analyzed SPEI temporal distances for February and various teleconnection indices such as NAO-slp, AO, MOI-1, MOI-2, NCP, SCAND, and NAO-500. The strongest correlation has been found between SPEI2-F and MOI-1 (0.74). Among the 15 teleconnection indices considered, 11 have exhibited associations with SPEI for the month of May, with the strongest connections observed with AO and MOI-1 (correlation coefficients up to 0.73). All six SPEI temporal distances for August have correlated with the SCAND pattern, while the strongest link has been found between SPEI9-A and AO (–0.65). This Arctic Oscillation (AO) pattern has demonstrated correlations with all SPEI temporal distances for November, exhibiting an inverse relationship where a negative AO phase has increased SPEI and a positive phase has decreased it.

These findings suggest that the exacerbation of drought in Montenegro, particularly evident in summer, can largely be attributed to warmer climatic conditions. This study has also shown that interannual variations in the analyzed drought indicator (SPEI) have been influenced by variations in atmospheric and oceanic oscillations.

Montenegro has a coastline along the Adriatic Sea, stretching approximately 293 km (in a straight line about 100 km). Its picturesque beaches and the advantages of a Mediterranean climate have propelled tourism, particularly beach tourism, to become the country's leading economic sector, followed by agriculture and hydroenergy production. Given these factors, the trend toward heightened summer drought conditions is particularly worrisome, as this season is critical for Montenegro's dominant economic activities. The concern is compounded by the fact that summer, already the driest season of the year, also experiences the most significant warming. In warmer climate conditions, increased evaporation from rivers, lakes, vegetation, and soil can further exacerbate all forms of drought – meteorological, agricultural, hydrological, and socio-economic.

It's important to note that out of the 10 most severe droughts in Montenegro, seven have occurred from the year 2000 onwards. The summer and autumn of 2017 were exceptionally dry. The lack of precipitation and high temperatures (resulting in high evapotranspiration) have led to a drop in river and lake levels. Using data from the Skadar Lake water levels, the largest lake in the Balkans (covering an area of 370–550 km²), for the purposes of this study, it has been determined that on 4 October 2017, the water level at the Plavnica hydrological station was measured at –12 cm. This represents the lowest water level of Skadar Lake recorded in the entire period of instrumental measurements (from 1948 to the present day). This data supports the fact that 2017 has experienced an extreme hydrological drought.

Despite Montenegro being rich in precipitation, it experiences uneven distribution throughout the year, making it susceptible to droughts. The country has faced severe droughts in 2003, 2007, 2011, 2017, 2018, and 2019. During these years, droughts affected not only meteorological aspects but also had agricultural and hydrological dimensions. The drought in 2011 escalated to the level of socio-economic drought (MNDP 2020). Projections for Montenegro suggest that it will become warmer by the end of the 21st century (Burić 2024), with an increase in the number of rainless days (Burić & Doderović 2021). This may lead to more frequent occurrences of droughts, heatwaves, forest fires, challenges in agriculture, tourism, and hydroelectric power production. Over the last two decades, river and lake levels in Montenegro have often been below average, leading to water shortages in summer, yet there has been minimal specific research and reporting on drought. If projections hold true until the end of the 21st century, drought could indeed become a serious problem in the future.

Neighboring countries on the Balkan Peninsula have also not been spared from droughts and other extreme weather and climate events. In Bosnia and Herzegovina, SPI results for the period 1961–2010 have indicated that during the last decade analyzed (2001–2010), most locations have experienced the highest number of both dry and wet summers, signaling an increase in extreme climate events during this season (Ducić et al. 2014). Similarly, in Croatia, a significant portion of the country has observed more frequent droughts and an increase in the percentage of areas affected by drought. Trends for the period 1950–2022 have shown declining trends, indicating an increase in drought intensity, particularly in northern and eastern regions, while noticeable upward trends have been observed in the southern region (Santos et al. 2023). In neighboring Serbia, from 1980 to 2010, there have been two prolonged drought intervals: 1987–1994 and 2000–2003 (Gocic & Trajkovic 2014).

The results presented in this study can assist decision-makers in developing strategies and plans to monitor droughts and mitigate the negative consequences of this phenomenon. Given that drought has affected nearly all natural and human systems, improving monitoring of this phenomenon has been crucial in agriculture, water resources (lakes, rivers, groundwater), and other sectors in Montenegro. This study has aimed to demonstrate that drought has been intensifying in Montenegro and could pose a serious problem under future warmer climate conditions. Summer has already been a dry season in Montenegro and the Mediterranean region, making the impacts of drought particularly noticeable during this season, primarily in terms of water scarcity and frequent wildfires. The main limiting aspect of this study is that it considers only two drought indices and the impact of teleconnections on SPEI. Therefore, the new research should be focused on considering additional drought indices and a comprehensive examination of the anthropogenic impact on this hazard. Also, future research should focus on investigating the impacts of drought and measures for hazard protection, especially in Montenegro's primary economic sectors such as agriculture, tourism, and hydroenergy. In the era of the Anthropocene and global climate change, humanity is facing numerous challenges (Jusup et al. 2022). Intensification and spread of drought in the Mediterranean region and climatically similar parts of the world could become a pressing problem in the near future. Therefore, drought should be understood as a potentially very dangerous threat to the environment, in general.

Two widely used drought indicators in Montenegro for the period 1961–2020 have been analyzed: the SPI and the SPEI. Trends have been calculated for 18 MS (SPI), i.e., for 10 grid fields with a resolution of 0.5° × 0.5° that cover the territory of Montenegro (SPEI), specifically for the last month of each season (February, May, August, and November), with backward steps of 1, 2, 3, 6, 9, and 12 months (SPI1, SPI2, SPI3, SPI6, SPI9, SPI12, and SPEI1, SPEI2, SPEI3, SPEI6, SPEI9, and SPEI12). Focusing on the capital city of Podgorica (MS Podgorica), the goal has been to determine breakpoints in these changes and assess the influence of atmospheric and oceanic oscillations based on SPEI values. Since SPEI has considered both precipitation and temperature (evapotranspiration), it has been a more reliable drought indicator compared with SPI, which only considers precipitation.

The obtained results have indicated the dominance of a negative trend in SPI and SPEI across almost all observed time scales, suggesting an intensification of drought in Montenegro during the period 1961–2020, overall. The intensification of drought in Montenegro has been particularly noticeable during the summer period, where the negative trends have been statistically significant. For instance, in MS Podgorica, it has been determined that the frequency of negative SPEI values in the second half of the observed period (1991–2020) has been much higher than in the first half (1961–1990). Using the RAPS method, it has been defined that the mid-1980s marked a breakpoint when SPEI sharply declined in Podgorica. The trend in monthly precipitation totals in Montenegro has mostly been insignificant, but notably negative during summer. Additionally, there has been an increase in the annual number of dry days and significant warming in Montenegro, leading to increased evapotranspiration. All these factors indicate a very likely influence of contemporary climate change, although the study has shown the importance of considering the impact of atmospheric and oceanic oscillations. Out of the 15 analyzed teleconnection indicators, results have shown that 11 have had an influence on SPEI, depending on the observed time scale.

The results of this study showed that the drought is intensifying in Montenegro. Consequently, as a warmer future is expected, it is necessary to plan and start measures to combat this phenomenon as soon as possible. Montenegro does not have a developed irrigation system, and this is one of the most important measures for mitigating the consequences of drought in agriculture. One of the measures that should be taken as soon as possible is the rational use of water resources and the collection of water in the part of the year that is rich in precipitation. On this basis, we should take into consideration the construction of artificial lakes, in order to preserve the abundant rainwater in winter and use it in the warmer part of the year, when there is very little of it. Planning and implementation of the mentioned as well as the other measures need political support (of decision-makers) and the engagement of experts of various profiles.

All authors contributed to the study conception and design. Material preparation, data collection, and analysis were performed by D.B. and J.M. D.B. wrote the first draft of the manuscript. M.D. and I.M. provided comments on previous versions of the manuscript. All authors have read and approved the final manuscript.

The authors declare that no funds, grants, or other support were received during the preparation of this manuscript.

All relevant data are included in the paper or its Supplementary Information.

The authors declare there is no conflict.