Keywords

1 Introduction

The Dinaric Karst region lies in southern Slovenia and covers over a fourth of Slovenia’s territory. It is mainly composed of karstified plateaus and hills, and lowlands and plains strewn between them. In terms of elevation, it can be divided into the high and low Dinaric Karst region (Fig. 12.1). The high karst region covers nearly three-fifths of the entire Dinaric region, or 3140 km2, and the low karst region extends over more than two-fifths of the Dinaric region, covering 2302 km2. Because of the extensive wooded and almost completely unpopulated high Dinaric plateaus, settlement density is nearly half the Slovenian average (22 people per km2 in the high Dinaric Karst region and 93 people per km2 in the low Dinaric Karst region, compared to the 2017 Slovenian average of 109 people per km2).Footnote 1

Fig. 12.1
A map of the Dinaric Karst region is color-coded with the occasional flood area, high Dinaric Karst, and low Dinaric Karst, in which the precipitation stations, temperature stations, and gauging stations are marked.

Source authors

Locations of temperature, precipitation, and gauging stations included in the analysis.

Full size image

Slovenia’s Dinaric Karst region is part of the northwestern Dinaric Alps, which predominantly extend in a typical northwest–southeast Dinaric direction. Its main characteristics include the predominance of karstified landscape and subterranean water flow, karst poljes, intermittent lakes, high precipitation and forest density, livestock- and forestry-oriented private farming, scant settlement, sparse population concentrated in the fertile parts of poljes and plains, a traditional woodworking industry, and exceptional environmental vulnerability.Footnote 2

This chapter examines the period from 1961 to 2020 for trends in average annual temperature, annual precipitation, days with precipitation over 0.1 mm, days with snow cover, minimum annual discharges, mean annual discharges, maximum annual discharges, and absolute maximum annual discharges, as well as discharge regimes and flood hazard, using the Dinaric Karst of Slovenia as an example.

2 Methods

To determine the trend of change in selected climate and hydrological variables (Table 12.1) from 1961 to 2020, the Mann–Kendall test and Theil–Sen estimator (also known as Sen’s slope estimator) were used at selected temperature, precipitation, and gauging stations (Tables 12.2 and 12.3; Fig. 12.1). The Mann–Kendall test is a nonparametric test used to detect monotonic trends. It is insensitive to data outliers and is based on the test statistics. A positive test statistic implies an upward trend, and a negative test statistic indicates a downward trend. Sen’s slope estimator is the most frequent nonparametric test used for detecting linear time trends.Footnote 3 Compared to linear regression, Sen’s slope is more accurate for asymmetric data distribution and, with normal data distribution, it yields results that are completely comparable to those obtained by applying the least square method.Footnote 4

Table 12.1 Climate and hydrological variables examined
Full size table
Table 12.2 Weather stations with the time series analyzed
Full size table
Table 12.3 Rivers with the time series analyzed
Full size table

The free Excel template MAKESENS (Mann–Kendall test for trend and Sen’s slope estimates) was used to calculate the Mann–Kendall test and Sen’s slope estimator values.Footnote 5

In addition to the Mann–Kendall test and Sen’s slope values, the tables with climate and hydrological variables also show the confidence levels, the (initial) 1961 trend value, the (final) 2020 trend value, and absolute and relative trend differences.Footnote 6

In statistics, the confidence level indicates the probability that the confidence interval calculated includes the value of the parameter estimated. In this specific case, the higher the confidence level, the higher the probability that the upward or downward trend in the selected variable detected actually exists.Footnote 7

The initial 1961 trend value represents the value of a selected variable in 1961 read from the trend line, and the final 2020 trend value is the value of a selected variable in 2020 read from the trend line. The absolute trend difference is the difference between the final and initial trend values, and the relative trend difference is the difference between the final and initial trend values expressed as a percentage.Footnote 8

The trend value per year can be calculated using the following equationFootnote 9:

trend value per year x = Sen’s slope * (trend year x − initial trend year) \(+\) initial trend value.

3 Data

3.1 Climate Data

Climate data were obtained from the Slovenian Environment Agency.Footnote 10 The analysis included ten temperature stations and thirty precipitation stations in the Dinaric part of Slovenia (Table 12.2, Fig. 12.1) that had measurement data spanning several decades.

3.2 Hydrological Data

Hydrological data were obtained from the Slovenian Environment Agency.Footnote 11 The analysis included twenty gauging stations in the Dinaric part of Slovenia (Table 12.3, Fig. 12.1) that had measurement data spanning several decades.

4 Results

4.1 Climate Variables

The following were analyzed in terms of climate variables (Table 12.1): (1) average annual air temperature trends, (2) annual precipitation trends, (3) trends in annual days with precipitation over 0.1 mm, and (4) trends in annual days with snow cover.

4.1.1 Average Annual Air Temperature

The average annual air temperature trends from 1961 to 2020 are similar at all ten temperature stations analyzed, and they indicate a clear upward trend (Table 12.4, Fig. 12.2a). An exceptionally high confidence level is typical, amounting to 99.9% at all temperature stations.

Table 12.4 Average annual air temperature trends, 1961–2020
Full size table
Fig. 12.2
Four scatterplots of temperature in degree Celsius, precipitation in millimeters, days with precipitation and, days with snow cover from 1960 to 2020. Temperature has an increasing trend while the others have decreasing trends.

Source Authors

Climate trends at selected temperature and precipitation stations: a average annual temperature trend at the Nova vas na Blokah temperature station, 1961–2020, b annual precipitation trend at the Cerknica precipitation station, 1961–2020, c trends in annual days with precipitation at the Kočevje precipitation station, 1961–2020, and d trend in annual days with snow cover at the Predgrad precipitation station, 1961–2020.

Full size image

During the period studied, from 1961 to 2020, the temperature at the selected temperature stations rose by 0.032 to 0.050 °C a year on average, which means that over the past six decades the temperatures at these stations increased by 1.87 to 2.95 °C. Taking all the selected temperature stations into account, the temperature rose by 2.49 °C on average. The temperature difference was the smallest at Kočevje, where the temperature rose by 1.87 °C, and the greatest at Novo mesto, where it rose by 2.95 °C.

4.1.2 Annual Precipitation

In contrast to the upward temperature trends, the annual precipitation trends from 1961 to 2020 indicate a downward trend at twenty-four out of the thirty precipitation stations studied (Table 12.5, Fig. 12.2b). The confidence level is fairly modest because at as many as twenty-two stations it does not even reach 90%, which is far from statistical significance (95%). All eight stations with an at least 95% confidence level show a downward annual precipitation trend.

Table 12.5 Annual precipitation trends, 1961–2020
Full size table

Most differences in the annual precipitation are moderate. At eighteen precipitation stations out of thirty, these differences do not exceed 10%, and at twenty-eight, they do not exceed 20%. A downward annual precipitation trend is the most distinct at Podgrad (−363.5 mm or −21.7%), Hrušica pri Colu (−445.7 mm or −20.1%), and Kozina (−298.8 mm or −18.8%), and an upward trend is the most evident at Predgrad (\(+\)97.4 mm or \(+\)6.5%), Trava (\(+\)65.2 mm or \(+\)3.8%), and Nova vas na Blokah (\(+\)53.0 mm or \(+\)3.6%).

4.1.3 Annual Days with Precipitation Over 0.1 mm

From 1961 to 2020, the annual number of days with precipitation over 0.1 mm increased at twelve precipitation stations, decreased at fourteen, and remained the same at the stations in Godnje, Gorenjci pri Adlešičih, Kočevje, and Želimlje (Table 12.6, Fig. 12.2c). The confidence level is mostly very modest because at twenty-two stations, it does not even exceed 90%. The confidence level is at least 95.0% at the remaining eight precipitation stations, of which three have an upward trend and five have a downward trend in the annual days with precipitation.

Table 12.6 Trends in annual days with precipitation, 1961–2020
Full size table

Twenty-one stations show smaller negative or positive trend deviations, not reaching 10%. The decrease in the annual days with precipitation over 0.1 mm is the most distinct at Predgrad (−46.7 days or −28.9%), Trava (−36.6 days or −21.7%), and Dvor (−31.8 days or −19.9%), and the increase is the greatest at Podgrad (24.6 days or 22.1%), Otlica (19.7 days or 14.8%), and Cerknica (19.7 days or 13.8%).

4.1.4 Annual Days with Snow Cover

Between 1961 and 2020, the number of days with snow cover decreased significantly at all precipitation stations where it snows every year (Table 12.7, Fig. 12.2d). The confidence level at most of these stations is high and statistically significant.

Table 12.7 Trends in the annual days with snow cover, 1961–2020
Full size table

At the Godnje and Opatje selo stations, the snow cover is more of an exception due to their low elevation and the proximity of the sea, and therefore their trends cannot be compared with other stations. At twenty-two of the remaining twenty-eight precipitation stations, the number of days with snow cover decreased by 30 to 60%, or on average by thirty-five days a year. The only station where it decreased by less than 30% was Otlica (−20.4%), whereas a more than 60% decrease was recorded at Podgrad (−71.9%), Dvor (−63.1%), Novo mesto (−61.3%), Fužina (−61.1%), and Cerovec (−60.9%).

The duration of snow cover expressed in the number of days decreased the most at Lokve (−62.9 days), Črni Vrh (−55.6 days), and Logatec (−52.3 days).

4.2 Hydrological Variables

The following were examined in terms of hydrological variables (Table 12.1): (1) minimum annual discharge trends, (2) mean annual discharge trends, and (3) maximum annual discharge trends.

4.2.1 Minimum Annual Discharge

The minimum annual discharge trends from 1961 to 2020 were downward for all the watercourses analyzed except the Pivka and Reka (Table 12.8, Fig. 12.3a). The confidence level varies greatly: it does not exceed 90% at seven gauging stations, it reaches 95% for the Unica, and it reaches at least 99% at twelve gauging stations.

Table 12.8 Minimum annual discharge trends, 1961–2020
Full size table
Fig. 12.3
Four scatterplots of discharge in cubic meters per second from 1960 to 2020. The first four plots have a decreasing trend while the last plot has an increasing trend.

Source Authors

Discharge trends at selected gauging stations: a minimum annual discharge trend of the Malenščica River at the Malni gauging station, 1961–2020, b mean annual discharge trend of the Ljubljanica River at the Vrhnika gauging station, 1961–2020, c maximum annual discharge trend of the Kolpa River at the Petrina gauging station, 1961–2020, d absolute maximum annual discharge trend of the Krka River at the Gorenja Gomila gauging station, 1961–2020.

Full size image
Fig. 12.4
A collage of six photographs of waterfalls in a rocky terrain.

Some of the watercourses in Slovenia’s Dinaric Karst region: a the Hubelj karst spring at Ajdovščina, b the Ljubljanica karst spring at Vrhnika, c the Reka, when it disappears underground at Škocjan Caves; d the Rižana karst spring has been developed to supply drinking water, and e the source of the Unica at Planina Cave. All photos were taken when the discharge was below average (Photo Matija Zorn, GIAM ZRC SAZU Archive). f The Unica flooding the Planina Polje (Photo Miha Pavšek, GIAM ZRC SAZU Archive. Reproduced with permission.)

Full size image

The average minimum annual discharge in the period observed decreased by 0.05 to 7.48 m3/s, with an average decrease of 1.37 m3/s or 26.7%. One gauging station recorded a relative decrease of up to 10%, four stations recorded a relative decrease of 10 to 20%, five stations of 20 to 30%, and eight stations over 30%. The greatest absolute decrease was recorded for the Kolpa at Metlika, where the discharge decreased by 7.48 m3/s, and the relative difference was the greatest for the Temenica at Rožni Vrh, where the average minimum discharge decreased by 62.2%, and for the Borovniščica at Borovnica, where the decrease was 58.6%. For the Pivka and Reka, which are the only Slovenian Dinaric rivers with a positive trend, the average minimum annual discharge increased by a negligible 0.02 m3/s and 0.01 m3/s, respectively.

4.2.2 Mean Annual Discharge

The mean annual discharge trends from 1961 to 2020 were downward for all rivers and at all the gauging stations observed (Table 12.9, Fig. 12.3b). The confidence level varies from only 90% or lower at seven stations to 95% at five stations and to at least 99% for the Borovniščica, Kolpa at Petrina, Ljubljanica, Malenščica, Rižana, Temenica, Trebuščica, and Vipava.

Table 12.9 Mean annual discharge trends, 1961–2020
Full size table

During the period studied, the mean annual discharge decreased by 0.30 to 14.61 m3/s. Most rivers recorded a relative decrease of 10 to 30%, with a major decrease only observed for the Borovniščica (−51.0%), Temenica (−36.5%), and Rižana (−35.5%). The absolute trend difference in the mean annual discharge between 1961 and 2020 was the greatest for the Kolpa at Metlika, where the discharge declined by 14.61 m3/s, and the greatest relative trend difference can be observed for the Borovniščica at Borovnica, where the discharge decreased by 51.0%.

4.2.3 Maximum Annual Discharge

The maximum annual discharge trends from 1961 to 2020 were downward for all the watercourses analyzed except for the Lahinja, Prečna, and Temenica (Table 12.10, Fig. 12.3c). The confidence level varies from only 90% or even lower at six gauging stations to 95% on the Kolpa at Melika, Krka at Podbukovje, Reka, and Unica, and at least 99% at ten stations. During the period observed, the maximum annual discharge largely decreased by 10 to 40%. A decrease over 40% was recorded for the Borovniščica, Rižana, and Trebuščica. The greatest absolute decrease was recorded for the Kolpa at Metlika, where the discharge declined by 61.49 m3/s, and the relative difference was the greatest for the Trebuščica at Dolenja Trebuša, where the average maximum discharge decreased by 45.6%. The greatest increase in the maximum annual discharge was observed for the Prečna at Prečna—that is, 2.13 m3/s or 24.5%.

Table 12.10 Maximum annual discharge trends, 1961–2020
Full size table

5 Discharge Regimes

Long-term changes in temperature and precipitation not only affect the volume of the minimum, mean, and maximum discharge but also have a significant impact on changes in the discharge regime.Footnote 12 Among the climate indicators examined, changes in temperature seem especially important because they affect evapotranspiration, as do changes in the number of days with snow cover because they strongly affect all discharge regimes with a distinct snow component. In Dinaric Slovenia, the snow share is important for rivers with headwaters on the high Dinaric plateaus (Fig. 12.1), such as the Trnovo Forest Plateau (Trnovski gozd), the Javorniki Hills, and the Snežnik, Bloke, Menišija, Velika Gora, and Kočevje Rog plateaus.

In classifying the discharge regimes based on the 1961–1990 dataset,Footnote 13 rivers in Dinaric Slovenia were categorized into four groups. The Borovniščica, whose headwaters extend high up to the shady slopes of the Menišija high plateau, was categorized among rivers with an Alpine rain–snow regime typical of areas in central Slovenia with a temperate continental climate.Footnote 14 Among all the discharge regimes in Slovenia, this regime stood out in terms of modest discharge variation and unpronounced discharge maximums and minimums. The Borovniščica’s main discharge maximum in November and a secondary one in April only slightly exceeded the average annual discharge. Similarly, its secondary discharge minimums in January and February were also only slightly below the annual average. Only its main discharge minimum in August deviated more significantly from the annual average.Footnote 15

The Alpine rain–snow regime was very similar to the Dinaric–Alpine rain–snow regime, characteristic of the Cerkniščica, Krka, Ljubija, Prečna, and Temenica. This regime, too, was typical of areas in central Slovenia with a temperate continental climate,Footnote 16 but it had slightly more pronounced discharge maximums and minimums. The main discharge maximum occurred in March or April, and a secondary one in November or December.Footnote 17

The largest group of rivers, composed of the Hubelj, Idrijca, Kolpa, Lahinja, Ljubljanica, Trebuščica, Unica, and Vipava, had a Dinaric rain–snow regime typical of the temperate continental climate of western and southern Slovenia.Footnote 18 Their spring and fall discharge maximums were fairly the same, which is why some of them had their main discharge maximum in April and a secondary one in November, and the others exactly the opposite. Their main discharge minimum was in the summer, and their secondary minimums in January or February were close to the average annual discharge.

The Pivka, Reka, and Rižana, whose basins are close to the Adriatic, had a simple rain regime. Due to the impact of the sub-Mediterranean climate,Footnote 19 they had their discharge maximum in the fall, after which the volume of water did not decrease significantly up until April. They had a highly pronounced discharge minimum in the middle of the summer.Footnote 20

After three decades, the discharge regime of Dinaric rivers changed significantly, but notably less than that of Alpine rivers. The greatest changes in the discharge regime in Slovenia can be observed in rivers formerly characterized by highly pronounced snow retention in late fall and winter, and a notable increase in the discharge in the second half of spring due to the melting of the snow cover, which at that time was still ample.Footnote 21 The discharge conditions of Dinaric rivers, which only a few decades ago had been classified under three rain–snow regime versions, are now significantly more uniform. For all of them, the November or December discharge maximums significantly exceed the spring maximums, and the August minimums significantly exceed the January or February secondary minimums. Moreover, the secondary winter discharge minimums of most Dinaric rivers are already above the annual average.

Relatively small changes have been observed in the discharge conditions of rivers with a simple rain regime (i.e., the Pivka, Reka, and Rižana). Their discharge maximums have remained in November and December, and their discharge minimums have remained in July and August. Because of thinner snow cover on the Snežnik high karst plateau and its faster melting, the secondary discharge maximum for the Reka has moved from April to February. Similar conditions can be observed for the Rižana, which obtains a portion of its water from the Brkini Hills, which used to receive significantly more snow.

A comparison of the discharge regimes based on the 1961–1990 dataset with the discharge regimes based on the 1991–2020 dataset shows the following differences (Table 12.11, Fig. 12.5):

Table 12.11 1961–1990 and 1991–2020 monthly discharge coefficients (blue shading shows the main and secondary minimums, and red shading shows the main and secondary maximums); the correlation between the two datasets is based on the Pearson correlation coefficient
Full size table
Fig. 12.5
A double bar chart plots the discharge coefficients from January to December for the 1961 to 1990 and 1991 to 2020 periods. November has the highest discharge coefficient during 1991 to 2020.

Source Authors

Changes in the discharge regime of the Vipava River at the Vipava gauging station between the 1961–1990 and 1991–2020 periods.

Full size image

  • For rivers that used to have a Dinaric–Alpine and Dinaric snow–rain regime (i.e., the Cerkniščica, Hubelj, Kolpa, Krka, Lahinja, Ljubija, Ljubljanica, Malenščica, Prečna, Temenica, Trebuščica, and Vipava), the main discharge maximum has moved from the spring to the fall, which shows a significantly reduced impact of snow cover;

  • The impact of winter snow retention has significantly decreased and can now only be observed in rivers with their catchment area on high karst plateaus;

  • Summer minimums are becoming more pronounced for all rivers;

  • The November and December water levels have increased significantly, implying that winter is “running late.”

The intensity of changes in the monthly discharge coefficients for individual rivers between the 1961–1990 and 1991–2020 periods was also ascertained with the Pearson correlation coefficient (Table 12.11). By far the lowest values were calculated for the Borovniščica (0.64), where the monthly discharge coefficients between June and August decreased significantly. At the same time, the November, December, and February coefficients notably increased as a result of gradual transition in precipitation from snow to rain. At the Pokojišče precipitation station close to the upper reaches of the Borovniščica, the number of days with snow cover decreased from 70 to 30, or to less than half, between 1961 and 2020.

Pearson correlation coefficients ranging from 0.83 to 0.90 were ascertained for the Cerkniščica, Hubelj, Idrijca, Krka (at Podbukovje and Gorenja Gomila), Ljubija, Ljubljanica, Malenščica, Prečna, Temenica, Unica, and Vipava. These are mostly rivers whose basins extend onto high karst plateaus, where both the volume and duration of the snow cover have been declining significantly over the past decades.

The highest Pearson correlation coefficients, ranging between 0.91 and 0.96, are typical of the Kolpa, Lahinja, Pivka, Reka, Rižana, and Trebuščica. This small group includes very diverse rivers. For example, the Kolpa and Lahinja are karst streams into which water flows through long and branched underground karst channels, which predominantly mitigates major discharge fluctuations. In turn, the Reka and Rižana have a major portion of their basins in the Mediterranean low hills, where snow cover is relatively rare or even an exception, and therefore a smaller volume of snow has not had a significant impact on discharge changes in recent decades.

Among other things, the fact that changes in the discharge regimes of Dinaric rivers between the 1961–1990 and 1991–2020 periods has primarily been the result of stronger evapotranspiration and changes in the duration and volume of snow cover can be observed at the gauging station at the source of the Vipava River (Fig. 12.5). The November rain maximum has already strongly exceeded the spring high waters, which are primarily the consequence of snow melting on the Nanos and Hrušica high karst plateaus. Due to high temperatures and strong evapotranspiration, its main summer minimum has become even more pronounced and, due to a smaller volume of snow retained, its secondary winter minimum already somewhat exceeds the average annual discharge.

Differences in the monthly discharge coefficients between the 1961–1990 and 1991–2020 periods (Table 12.12) are largely consistent for all the rivers analyzed in Slovenia’s Dinaric Karst region. They indicate a gradual decrease in the spring and summer discharges and an increase in the fall and, to a slightly smaller extent, winter discharges. The greatest decrease in the discharge occurs in April and from June to August, and the greatest increase occurs in November, December, and February. In addition to increased late spring and summer evapotranspiration, an important reason for these developments is the gradual temperature increase and the resulting thinner and briefer snow cover. In the past, a significant share of fall precipitation was in the form of snow, which caused snow retention on the rivers, and the thicker snow cover on higher karst plateaus continued melting until late spring.

Table 12.12 Changes in the monthly discharge coefficients between the 1961–1990 and 1991–2020 periods (blue shading indicates decreasing ratios, and red shading shows increasing ratios)
Full size table

6 Floods

Because of the modest river network and the predominance of subterranean water flow, floods only pose a threat to a small portion of Dinaric Slovenia. They are the most frequent in the numerous karst poljes, where they usually build up gradually and the water then remains there for several days, weeks, or even months. They develop because the karst groundwater level rises above the surface or because the volume of water flowing into the underground drainage channels exceeds their capacity.Footnote 22 For example, flood-prone areas in the Cerknica Polje cover 2370 ha of land, in the Planina Polje (Fig. 12.4f) they cover 950 ha, and in the Lož Polje they cover 300 ha.Footnote 23 Because floods are annual and long-term in some karst poljes and in some parts of karst plateaus, the term “intermittent lakes” has been established in such cases. The best-known Dinaric intermittent lakes include Lake Cerknica in the Cerknica PoljeFootnote 24 and the Pivka Lakes along the upper reaches of the Pivka River.Footnote 25

In addition to floods in karst poljes, flashfloods also occur occasionally along individual streams. They are the most common on the Iška, Borovniščica, Cerkniščica, (Sodražica) Bistrica, and (Podbočje) Sušica.Footnote 26 Along the lower courses of major Dinaric rivers, lowland floods (Fig. 12.1) also occur, mostly posing a threat to the neighboring areas. They develop due to the difference in the speed of incoming karst high waters and the discharge capacities of the riverbeds.Footnote 27 For example, along its lower course, the Krka River threatens the Pannonian Krka Basin, and the Vipava and Hubelj rivers pose a threat to the Mediterranean Vipava Valley.Footnote 28

Floods are closely connected with the absolute maximum river discharge. In this regard, the upward trends in certain watercourses are the main cause for concern.Footnote 29 Between 1961 and 2020, the absolute annual maximum discharge trends were downward for the Borovniščica, Cerkniščica, Hubelj, Idrijca, Kolpa, Lahinja, Ljubija, Pivka, Rižana, and Trebuščica, and upward for the Krka, Ljubljanica, Malenščica, Prečna, Reka, Temenica, Unica, and Vipava (Table 12.13, Fig. 12.3d). The confidence level is low, not even reaching 90% at fifteen out of twenty gauging stations. The only exceptions are the stations at Petrina on the Kolpa (99.9%), Gorenja Gomila on the Krka (99.9%), Vrhnika on the Ljubljanica (95%), Malni on the Malenščica (99%), and Prečna on the Prečna (99.9%). At twelve gauging stations, the absolute maximum annual discharge in the period observed decreased or increased by less than 10%.

Table 12.13 Absolute maximum annual discharge trends, 1961–2020
Full size table

In terms of water share, the absolute maximum annual discharge decreased the most at Petrina on the Kolpa (i.e., by 176.34 m3/s or 34.3%), at Kubed on the Rižana (i.e., by 9.94 m3/s or 18.7%), and at Borovnica on the Borovniščica (i.e., by 4.93 m3/s or 18.3%). On the other hand, it increased the most at Prečna on the Prečna (i.e., by 6.74 m3/s or 44.4%), at Rožni Vrh on the Temenica (i.e., by 1.78 m3/s or 35.9%), and at Gornja Gomila on the Krka (i.e., by 55.49 m3/s or 27.9%).

The coefficients of variation in the maximum annual discharge between the 1961–1980, 1981–2000, and 2001–2020 periods are important for understanding flood occurrence. Specifically, the coefficients of variation show the dispersion of data: the higher the ratio, the greater the dispersion.Footnote 30 In terms of river discharge, the higher the variation ratio, the higher the probability of the occurrence of both minimum and maximum discharges.

The calculations showed that the lowest coefficients of variation were approximately evenly distributed between the 1961–1980 and 1981–2000 periods. For the Kolpa (at the Petrina gauging station), Lahinja, and Temenica, they were the highest between 1961 and 1980, for the Idrijca and Reka, they were the highest between 1981 and 2000, and for the Borovniščica, Cerkniščica, Hubelj, Kolpa (at the Metlika gauging station), Krka, Ljubija, Ljubljanica, Malenščica, Pivka, Prečna, Rižana, Trebuščica, Unica, and Vipava, they were the highest in the most recent period (i.e., from 2001 to 2020). Based on this, greater deviations in the maximum annual discharge, including occasional “catastrophic” events, can be expected for these rivers (Table 12.14).

For example, in Fig. 12.3d, which shows the absolute maximum annual discharge trend of the Krka at the Gorenja Gomila gauging station, the trend line varies between 198.9 m3/s and 254.4 m3/s, with certain annual values standing out significantly in both the downward and upward directions. For example, the 2003 absolute maximum annual discharge was only 116.2 m3/s, whereas the 2010 one reached 404.7 m3/s.

Table 12.14 Absolute maximum annual discharge coefficients of variation (DCV), 1961–2020
Full size table

With regard to the potential occurrence of floods, the increasing fall maximum discharge should also be highlighted; it already exceeds the spring maximum for all rivers because it is primarily the November and December water levels that are increasing as a result of winter “running late.” At the same time, the summer (August) minimum is decreasing increasingly (Table 12.11). The former indicates an increased probability of high water and floods in the fall months and a smaller probability in the spring. The latter indicates the occurrence of summer droughts. Other researchers also recognize the increased probability of floods in the fall and winter months.Footnote 31

However, it is not only climate change that leads to changes in the hydrological regime and the consequent occurrence of floods. One must not neglect the anthropogenic impacts, such as land-use changes, river training, dams, and urbanization.Footnote 32

7 Conclusion

The key findings about changes in the selected climate and hydrological variables between 1961 and 2020 largely agree with the trends established by other authors.Footnote 33 They can be summarized as follows:

  • The average annual air temperature from 1961 to 2020 increased by 0.032 to 0.050 °C a year at all temperature stations analyzed, which means that during that period, the average annual air temperature warmed up by approximately 2.5 °C. The temperature difference was the smallest at the Kočevje station, where the temperature increased by 1.87 °C over six decades, and the largest at the Novo mesto station, where the temperature increased by 2.95 °C.

  • From 1961 to 2020, annual precipitation decreased at twenty-four out of thirty precipitation stations. Most differences in annual precipitation are moderate, not exceeding 10% at eighteen stations and 20% at twenty-eight. The downward trend was the most pronounced at Podgrad and Hrušica pri Colu (the differences slightly exceeding −20%), and the upward trend was the most pronounced at Predgrad (a nearly \(+\)7% difference).

  • From 1961 to 2020, the annual number of days with precipitation over 0.1 mm increased at twelve precipitation stations, decreased at fourteen, and remained the same at four. At twenty-one stations, the negative or positive trend deviations were smaller, not reaching 10%. The most pronounced downward trend was recorded at the Predgrad, Trava, and Dvor stations (differences below −20%), and the upward trend was the strongest at Podgrad (differences over \(+\)20%), Otlica, and Cerknica (differences over \(+\)10%).

  • During the 1961–2020 period, the number of days with snow cover decreased significantly at all precipitation stations, where it typically snows every year. At twenty-two out of twenty-eight stations, it decreased by 30 to 60%, or by an average of thirty-five days a year. It only decreased by less than 30% at Otlica, whereas a decrease of over 60% was recorded at Podgrad, Dvor, Novo mesto, Fužina, and Cerovec. The duration of snow cover expressed in days declined the most at Lokve (−63 days), Črni Vrh (−56 days), and Logatec (−52 days).

  • During the 1961–2020 period, the average minimum annual discharge trends were downward for all the rivers analyzed, except for the Pivka and Reka. One gauging station recorded a decrease lower than 10%, four between 10 and 20%, five between 20 and 30%, and eight over 30%. On the Temenica, the average minimum annual discharge decreased by a full 62% and on the Borovniščica by 59%.

  • From 1961 to 2020, the average mean annual discharge trends were downward for all the rivers observed. The mean annual discharge of Slovenian Dinaric rivers decreased by 0.30 to 14.61 m3/s. A relative decrease of 10 to 30% was recorded for most rivers, with a more pronounced decrease only being ascertained on the Borovniščica (−51%), Temenica (−37%), and Rižana (−36%).

  • From 1961 to 2020, the average maximum annual discharge trends were downward for all the rivers analyzed, except for the Lahinja, Prečna, and Temenica, largely showing a 10 to 40% decrease. The average maximum annual discharge decreased by over 40% on the Borovniščica, Rižana, and Trebuščica. Among the rivers with a downward trend, the discharge declined the most on the Trebuščica (−46%), and among the rivers with an upward trend it increased the most on the Prečna (\(+\)25%).

  • Between 1961 and 2020, the absolute maximum annual discharge trends were downward at eleven gauging stations and upward at nine. At half of the stations, the absolute maximum annual discharge decreased or increased by less than 10%. Among the rivers with a downward trend, it decreased the most on the Kolpa at Petrina (−34%) and on the Rižana (−19%), and among the rivers with a positive trend, it increased the most on the Prečna (\(+\)44%), Temenica (\(+\)36%), and Krka at Gorenja Gomila (\(+\)28%).

  • A comparison between the 1991–2020 discharge regime dataset and the 1961–1990 dataset shows that the impact of snow cover is decreasing rapidly. At all gauging stations, the main discharge maximum due to fall rain has already strongly exceeded the spring secondary maximum, in which snow melting plays an important role. The summer minimum is growing more distinct at all stations and, with regard to the secondary winter minimum, the discharge is increasing and in many places it already exceeds the annual average, implying that winter is “running late.”

  • The increasing absolute maximum annual discharge on some rivers, the increasing discharge variability, and the increasing fall discharge, despite a decrease in the total volume of water, may indicate a greater flood hazard.