1. Introduction
Climate change is one of the major contemporary environmental issues that can increase the risk of exposure to disasters by changing the dynamics of the frequency and intensity of climate-related hazards, affecting vulnerability, and altering exposure patterns [
1]. The detection of climate-induced changes is only possible with an adequate monitoring network with sufficiently long observation series and with the analysis of the spatiotemporal trend of key meteorological variables. The analysis of temperatures and precipitation is therefore the first step in understanding the progression of climate change and its impact on the hydrological cycle.
The high sensitivity of the hydrological cycle to climate change is particularly evident in the Mediterranean region, which is considered one of the major hotspots of climate change [
2,
3,
4,
5]. High sensitivity is a consequence of the geographical position in the transition zone between the temperate climate of the mid-latitudes and the hotter North African climate [
2,
5,
6]. Local factors resulting from the topography of the terrain, orography, and distance from the sea further increase the sensitivity and local and regional differences. Analyses of observation-based data indicate that in the past half-century, the Mediterranean region has experienced a trend toward warmer and drier conditions, with a corresponding rise in the frequency of extreme events [
2,
5,
7,
8]. Statistically significant warming trends are found in most of the subcontinental regions studied by Giorgi [
9], including the Mediterranean, where the regional temperature trends are characterized by pronounced interdecadal variability. The analysis of five pilot sites, namely, Portugal, Spain, Tunisia, Greece, and Turkey, also revealed positive temperature trends in all the areas studied [
8]. Except for Greece, all the observed temperature trends were statistically significant. Thus, the Mediterranean region is currently experiencing temperature increases above the global average, with the largest changes occurring in the warm temperature extremes [
5,
10,
11]. As a result of the increase in the air temperatures and evaporation, and the decrease in the runoff, droughts have become more frequent and intense.
However, the trend in precipitation in the Mediterranean region varies greatly depending on the location of the observation site within the basin [
7]. An analysis of 40 stations [
12] showed a decrease in the annual precipitation in the western, central, and eastern Mediterranean, and an increase in northern Africa, southern Italy, and the western Iberian Peninsula, consistent with the conclusions of Todaro et al. [
8]. According to Mariotti et al. [
13], the trend of the mean annual precipitation has decreased by 0.6 × 10
−2 mm/day/decade in the last century throughout the Mediterranean region. Increased interannual variability and large spatial variability of precipitation are observed in most Mediterranean regions [
5,
9,
11]. Predictions also suggest a decrease of 4–22% in most areas, depending on the emission scenario [
7]. In addition to the expected decrease in precipitation, heat waves and droughts are also expected to increase, as is precipitation variability [
5,
14].
As in the Mediterranean region, three predominant processes of climate change occur in Croatia, including the increase in average temperature, the decrease in precipitation, and more frequent and severe extreme weather events. A more pronounced increase in air temperature occurred in the 20th century [
15,
16,
17], and according to Bonacci [
18,
19], a sudden increase was recorded already after 1988. The increase in air temperature is more pronounced on the coast than inland [
15]. The decadal increase in air temperature was in the range from 0.04 °C to 0.08 °C until 2004, and between 0.05 °C and 0.10 °C until 2008 [
17]. The observed trends of the increase in mean annual air temperatures were statistically significant at all analyzed stations except at the Osijek station [
17]. The largest contribution to the increase in mean annual air temperatures is the result of changes in the warm-season temperatures on the coast and changes in the winter temperatures inland [
15,
17].
The variability of annual precipitation shows a decrease in northwestern Croatia, mountainous regions, and northern coastal areas, and an increase in the eastern lowlands and Dalmatian islands [
20,
21,
22]. The negative trend is most pronounced on the northeast coast of the Adriatic [
20]. The analysis of changes in the precipitation regime at four Croatian meteorological stations (Split, Hvar, Lastovo, and Zagreb) has shown a decrease in the number of days per year with precipitation, and an increase in the number of days with intense precipitation at all analyzed stations, as well as a redistribution of precipitation throughout the year [
21]. This is evidence of the great variability of precipitation at the regional level, which is further supported by the division of Croatia into thirteen precipitation zones based on the vertical gradient of precipitation and the base potential of precipitation [
23]. Despite the downward trend in annual precipitation, according to Gajić-Čapka et al. [
23], there is no evidence of major secular changes in precipitation extremes.
This paper studies the changes in temperature and precipitation in the karst region of Imotski. Since climatological data have not been previously analyzed in this area, this analysis will contribute to a better understanding of local climate patterns by providing valuable information on the long-term trends and variations in climate factors affecting the local environment. The analyses performed over the period of 40 years (1981–2021) and 28 years (1993–2021) aim to detect non-stationarity (trends, jumps, and seasonality) in the temperature and precipitation regime on two time scales (month and year). The goal is to identify climate trends that can assist in addressing the significant challenge of managing karst aquifers and developing effective management strategies in light of climate change. Climate change has the potential to affect climate patterns at various scales, including local, regional, and global, and may result in climate-related hazards such as droughts, floods, reduced water availability, and damage to various sectors. Considering the importance of Imotski and its surroundings due to their rich karst phenomena, such as the Blue and Red Lakes, the identification of climatic changes could play a crucial role in the hydrological cycle of the area. Among other things, water from the nearby Opačac Spring serves as a water supply for the entire region, and the availability and abundance of the spring may be threatened by climate change. In addition, there are also hydropower plants in the area where a reduction in river discharge may result in a reduced power supply. The decrease in precipitation and water scarcity may also have a negative impact on agriculture. This study will provide an initial assessment of the changes taking place and serve as a starting point for assessing the risk of climate change impacts at the local and regional scales as there is a lack of previous literature on climate trends in the area studied.
2. Materials and Methods
2.1. Study Area
The study is based on the analysis of data from the Croatian Meteorological and Hydrological Service for the two climatological stations, Imotski and Ričice (
Figure 1). The climatological station Imotski is located in the center of the small town of Imotski (LAT: 43°26′41″, LON: 17°13′17″) at an altitude of 399 m above sea level. The other station, Ričice, is about 920 m air distance from Imotski station and is located near the Ričice reservoir (LAT: 43°29′48″, LON: 17°8′2″) at an altitude of 402 m above sea level. These two stations are the only climate observatories in the wider area of Imotsko Polje, our area of research interest. Imotsko Polje is an example of a polje in the karst, surrounded by the Biokovo Mountains to the southwest and the Dinaric Alps to the northeast. This particular geographical position and relief provide unique climatic characteristics with often strong diurnal air temperature fluctuations, as well as the occurrence of temperature inversions, which are particularly evident in valleys such as Imotsko Polje. In the scientific literature published so far, no long-term analysis of climatological parameters in the area has been carried out.
The study area can be divided into a higher NE karst plateau consisting of permeable carbonate rocks, mainly Upper Cretaceous limestone, and an area of Imotsko Polje with mainly alluvial deposits, ending with impermeable rocks in the south of the Sija and Matica rivers [
24,
25,
26]. Due to the highly karstified carbonate rocks, the area is rich in karst features—dolines, collapsed doline, caves, etc. There are no surface streams in the karst plateau, and due to the high degree of karstification, most precipitation infiltrates, feeding the karst aquifer. The boundaries of the catchment area and the directions of groundwater circulation are unknown [
25,
27], but it is known that the karst aquifer is fed from wider areas that also include streams and reservoirs in Bosnia and Herzegovina [
25,
28,
29]. Tracer tests have proven the connection between the Ričice reservoir and the spring in the Polje [
25]. Thus, the water from the higher karst plateau and also from the Ričice reservoir occurs in Polje in the springs or in the surface stream Vrljika [
25]. This is evidence of the extraordinary interdependence of surface and subsurface flow and, finally, of the extremely complex water circulation in karst, especially in this area of highly karstified rocks.
The climate is Mediterranean with continental influence, and according to the Köppen–Geiger [
30] climate classification, it belongs to class Cfb, a temperate climate without dry seasons and with warm summers. The most famous phenomena are the Blue and Red Lakes, collapsed dolines, but also natural piezometers, as the oscillations of the water level follow the fluctuations of the groundwater level in the karst aquifer [
25]. Opačac, which supplies water to the Imotski region, is the most important spring.
2.2. Data
This paper analyzes monthly and annual temperatures, and monthly and annual total precipitation observed at the Imotski and Ričice climatological stations of the Croatian Meteorological and Hydrological Service (DHMZ). All data used in this paper are official data from Croatian Meteorological and Hydrological Service (DHMZ) and Federal Hydro-Meteorological Institute (FHMZ) in Bosnia and Herzegovina, for which we express our gratitude. The primary focus is on the analysis of temperatures and precipitation at the Imotski and Ričice stations, but data from nearby stations were also considered to further discuss spatial precipitation variability and interdependence between stations.
The observation period for the Imotski station is the period from 1981 to 2021, while the observation period for the Ričice station is somewhat shorter, covering 1993 to 2021. For the Imotski station, the annual mean temperatures for 2004 and 2020, and the annual minimum temperatures for 2008 and 2020 are missing, while the annual maximum temperatures are complete. The series of annual minimum temperatures for the Ričice station is complete, while the annual mean and maximum temperatures for 1993 are missing. Methods of measuring daily mean temperatures and temperature extremes vary, resulting in inconsistent gaps in the data.
The mean annual temperature is the average of the mean monthly temperatures, while the mean monthly temperatures are the average of the mean daily temperatures. The mean daily temperature is calculated according to the expression [
31,
32,
33,
34,
35] used in some central and eastern European countries (e.g., Croatia, Slovenia, Serbia) as follows:
where
T7,
T14, and
T21 are temperatures measured at 7, 14, and 21 h (local time). The equation in use today is over a century old and was formulated based on the assumptions and understanding of that time [
32,
36].
Daily minimum and maximum temperatures are read using minimum and maximum thermometers in the meteorological station at 2 m above the ground. The readings are taken at 9 h (local time) and refer to the period from 9 p.m. of the previous day to 9 p.m. of the day on which they are recorded, i.e., the last 24 h.
2.3. Homogeneity of Data and Differences between Two Time Series
The homogeneity of the data was tested using the Standard Normal Homogeneity Test (SNHT) [
37]. The SNHT tests the null hypothesis that the tested values are independent and identically distributed, and the alternative hypothesis that there is a break in the time series, i.e., that a stepwise shift in the mean occurs [
38]. The test can determine the probable year of break and is therefore called a location-specific test [
39].
If there is a mean shift between two subseries, the significance of the mean shift is tested using the F-test and the t-test. The F-test is a statistical test used to test the null hypothesis that two normal populations have the same variance. The t-test, on the other hand, is used to determine if there is a significant difference between the means of the two groups. In both tests, the value of p < 0.05 was chosen as the level of the significance of differences.
2.4. Trend and Variability Analysis
A variety of techniques were used in the data analysis to identify trends, analyze variability, and determine the frequency of occurrence of extreme values.
Trends are represented by the direction of linear regression. Each direction is defined by the following expression:
where
y is variable of interest (air temperature in °C or precipitation in mm),
t is a time label,
a is the slope of the line, and
b is the y-intercept. The sign of the coefficient
a indicates the trend (falling or rising). The strength of the linear relationship between the variables
t and
y is indicated by the square of the Pearson correlation coefficient R
2.
Two nonparametric tests for trend detection were used to analyze potential trends in time series data; the Theil–Sen (TS) estimator [
40] was used to determine the slopes of trends, and the Mann–Kendall (MK) test [
41,
42] was used to assess statistical significance. The advantage of using nonparametric tests is that they do not require distributional fitting and the estimators are not susceptible to the influence of extreme values (outliners) that are common in meteorological data [
43,
44]. The MK test compares the null hypothesis that there is no trend in time series with the alternative hypothesis that there is a trend. In this work, a modified Mann–Kendall test for Matlab [
45] was used, which takes into account the modification of Hamed and Ramachandra Rao [
46] regarding the presence of autocorrelation in the data. The statistical significance of the trends was examined at a 95% confidence level. If a linear trend is present, the magnitude of the monotonic trend in the time series is estimated using the Theil–Sen estimator. The magnitude of the trend (
β) is equal to the median of the slope values between values
xi and
xj at time steps
i and
j as follows:
The value of the magnitude of the trend indicates its steepness, while its sign indicates whether the trend is increasing or decreasing.
Rainfall variability was analyzed using the coefficient of variability (
CV), the standardized rainfall anomaly (
SRA), and the precipitation concentration index (
PCI). The coefficient of variation (
CV) shows the extent of variability in relation to the mean and is defined as the ratio of the standard deviation to the mean. It is applicable to data measured on a ratio scale, i.e., continuous data with a meaningful zero. Otherwise, data with an interval scale will yield different coefficient values depending on the scale used. For this reason, the coefficients of variation were not calculated for the temperature data given in Celsius. The coefficient of variation of precipitation data defines the variability of precipitation in the observed time interval at a certain location of interest. A higher value of
CV is an indicator of greater variability and vice versa, calculated as follows:
where
σ is the standard deviation and
μ is the mean precipitation.
CV is used to classify the degree of variability of rainfall events as low (
CV < 20), moderate (20 <
CV < 30), and high (
CV > 30) [
47,
48].
Standardized rainfall anomaly (or standardized rainfall anomaly index, Standardized Anomaly Index [
44,
49,
50,
51]) defines the nature of trends and provides information about the frequency and severity of drought. Thus, drought severity can be classified as extreme drought (
SRA < −1.65), severe drought (−1.28 >
SRA > −1.65), moderate drought (−0.84 >
SRA > −1.28), and no drought (
SRA > −0.84) [
52].
SRA is defined as follows:
where
Pi is the annual precipitation of the particular year (mm),
is the mean annual precipitation over a period of observation (mm), and
σ is the standard deviation of annual precipitation over the period of observation.
The variability and distribution of rainfall at different scales is defined by the precipitation concentration index (
PCI). This is a dimensionless value that examines the heterogeneity pattern of rainfall [
47]. The most commonly used scale is the annual scale, and
PCI is then defined as follows:
where
Pi is the monthly precipitation in month
i (mm). In addition to the annual scale,
PCI can also be calculated on the seasonal or even supra-seasonal scale, which includes the division into dry and wet periods [
53]. According to Oliver [
54], a
PCI of less than 10 indicates a uniform distribution, and a value of 11 to 20 suggests a seasonal distribution. At the same time, an index of more than 20 represents marked seasonal differences with increasing monthly concentration. Depending on the value of
PCI, monthly precipitation concentration can be classified as low (for uniform distributed precipitation), moderate for values from 11 to 15, high for values from 16 to 20, and very high for values above 21 [
47,
49,
50,
51,
55,
56].
4. Conclusions
This paper examines air temperature and precipitation dynamics as two of the most important factors affecting karst hydrology that can be extremely useful for comprehending both the movement of hydrologic parameters and the hydrologic state of the study area. In addition, identifying trends in air temperature and precipitation allows for more efficient management of regional water resources and associated hazards.
The analysis of temperature dynamics reveals a statistically significant increase in the mean and maximum annual air temperature at the Imotski station and a statistically significant increase in the mean annual temperatures at the Ričice station. The minimum annual temperatures decrease at both stations, but neither trend is statistically significant. The results of the SNHT test show that both the Imotski and Ričice meteorological stations experienced statistically significant shifts in temperature trends over time. At the Imotski station, there was a statistically significant shift in the annual mean temperatures after 2007 and annual maximum temperatures after 1998, while at the Ričice station, there was a statistically significant shift in the annual mean temperatures after 2011 and annual maximum temperatures after 1998. These results suggest climate variability rather than inhomogeneity, considering there was no relocation of the stations or measurement instruments at these years in the history of the two stations.
The precipitation trend at the Ričice station is negative, while at the Imotski station, it is positive. The differences in temperature and precipitation at the two stations are due to the different lengths of the time series and the influence of local factors. The influence of local factors is especially reflected in the amount of precipitation, which depends on the humidity of the air mass, the intensity and direction of the airflow, and also on the vertical component of motion, which can be significantly modified by local influences [
6,
23]. The significant spatial variability of precipitation and varying trends observed at different stations make it difficult to draw a single conclusion about the actual trend of the precipitation regime in this area [
23].
The precipitation variability is high at the monthly level, but low at the annual level. The PCI values at both stations yielded the same results, indicating a seasonal distribution of precipitation. Although the trend of the PCI at the Imotski station is decreasing, both stations have a high or very high value of the PCI, indicating a pronounced seasonality, i.e., large differences between the precipitation and non-precipitation periods. The number of dry periods increases at the Ričice station, while it decreases at the Imotski station, but their intensity increases. Both stations also show an increase in wet years. The SRA values indicate the presence of extreme events, either drought or extremely wet years, affecting the water flow between karst features.
The analysis of climatological data of long time series carried out for the first time in this area brought valuable and significant conclusions about local climate patterns that can be significant for predicting future climate trends and assessing the potential impacts of climate change on the local environment and communities. Because this is a highly karstified area with a very complex circulation of water in karst, understanding the climate trends and variability at these two stations is particularly important, as extreme phenomena associated with rising temperatures and seasonally poorly distributed precipitation can have negative impacts on karst hydrology and water availability. To mitigate climate-related risks, novel irrigation technologies, water storage, and efficient water supply infrastructure, coupled with the use of real-time monitoring systems and advanced prediction tools, can play a crucial role in ensuring the long-term sustainability of communities and ecosystems in karst areas. It is also important to emphasize the importance of the availability of high-resolution data, which is possible with a sufficiently dense network of meteorological stations, which unfortunately does not exist in this part of Croatia.
The authors hope that this study will contribute to a better understanding of both local and regional climate, and that the results presented in this paper will provide valuable insights into climate change in the wider Mediterranean region and will be considered in the further consideration of the hydrology of this karst area.