Comparative analysis of SPI, SPEI, and RDI ındices for assessing spatio-temporal variation of drought in Türkiye

Abstract

This research presents a comprehensive drought analysis using climate data obtained from 219 homogeneously distributed meteorological stations in Türkiye between 1991 and 2022. In this context, Standard Precipitation Index (SPI), Standardized Precipitation Evapotranspiration Index (SPEI) and Reconnaissance Drought Index (RDI) drought indices were used and comparative analysis was made. Türkiye. The study demonstrates that below-normal precipitation over extended periods and increasing temperatures have contributed to the increased frequency of meteorological drought events. Türkiye's topographic conditions, particularly its location in the Mediterranean basin, significantly influence drought occurrences. It is noted that over the past 20 years, Türkiye has been trending towards drier conditions, with rising temperatures reinforcing this trend. The study observes that the moderate drought class range is the most frequently recurring in the SPI, SPEI, and RDI methods utilized. Regarding atmospheric conditions affecting the climate in Türkiye, it is observed that increased drought severity stands out prominently in years when the North Atlantic Oscillation is positive. During these years, increased drought severity is evident in the SPI, SPEI, and RDI indices, particularly in winter and autumn, while a wide area experiences drought effects in the summer months. Long-term analyses emphasize that drought periods occur less frequently but have more prolonged impacts, attributed to variations in precipitation patterns from year to year and the influence of rising temperatures due to global climate change. The potential future increase in drought in the Mediterranean basin due to global climate change and Türkiye's vulnerability to this situation could have adverse effects on water resources, food security, energy sources, and ecosystems.

Introduction

Since the mid-twentieth century, the increasing impact of anthropogenic global climate change has brought about a rise in the frequency, intensity, and impact of extreme meteorological events, with drought undoubtedly being the most devastating when considering its long-term effects (Olsen et al. 2023).

Drought is a water deficit situation that occurs when natural water resources, utilized by various systems, fall below the long-term average or normal for a specific period and at a regional scale (Türkes 2007). Due to reduced precipitation and increasing temperatures, drought extending over one or more seasons and even spanning years is a meteorological disaster. Among the natural disasters worldwide, 28 out of 31 are meteorologically-related, with drought being the most hazardous (Partigöc and Sogancı 2019). Drought possesses the characteristic of affecting larger populations negatively compared to other disasters, owing to its slow onset and extensive coverage. Beginning with precipitation deficiency, drought encompasses a four-stage process (Mishra and Singh 2010) The first stage is meteorological drought, where precipitation falls below normal levels (Wilhite and Glantz 1985; Mishra and Singh 2010). The second stage is agricultural drought, when there is not enough water to meet the needs of crops and occurs after moisture loss and scarcity in water resources (Wilhite and Glantz 1985; Dai et al. 2020; Guo et al. 2020) The third stage is hydrological drought, where a prolonged lack of rainfall causes a reduction in water resources (Wilhite and Glantz 1985; Guo et al. 2020). The final stage is socio-economic drought, which encompasses the impact of drought on agriculture, living organisms, water resources, and sectors reliant on these resources, as well as the inability to meet water demand as an economic input, negatively affecting overall societal well-being in terms of quality of life, health, employment, and livability (Mishra and Singh 2010; Moriondo et al. 2011; Rousta et al. 2023).

The socio-economic impacts of drought, closely associated with various fields such as ecosystems, agriculture, and livestock, often lead to significant problems. Furthermore, the extensive and prolonged influence of drought on large areas increases the importance of management and adaptation efforts compared to other meteorological disasters (Sahana and Mondal 2023) Therefore, studies examining the effects of drought and regional drought analyses hold a significant share in the literature, prioritizing academic language and commonly used geographical terms (Mishra and Singh 2010; Ali et al. 2011; Sobral et al. 2018; Tigkas et al. 2019; Lim Kam Sian et al. 2023). For this reason, many drought indices have been developed to determine the temporal and spatial changes of meteorological drought in the world and in our country. Drought indices can be divided into two types: parametric indices and non-parametric indices. While parametric drought indices consider a certain statistical distribution based on long-term data sets in drought assessments, non-parametric drought indices are based on calculations that are independent of the data distribution (Soľáková et al. 2013). As parametric methods are based on statistical theory, they are more reliable and allow comparisons to be made between different geographical regions. This is why they are the preferred method in the literature (Soľáková et al. 2013). The main parametric indices are Standardised Precipitation Index (SPI) (McKee et al. 1993), Standardized Precipitation Evapotranspiration Index (SPEI) (Vicente-Serrano et al. 2010) Palmer Drought Severity Index (PDSI) (Palmer 1965), De Martonne Drought Index (DAI) (De Martonne 1926), non-parametric drought indices are Reconnaissance Drought Index (RDI) (Tsakiris et al. 2007), Effective Drought Index (EDI) (Byun and Wilhite 1999) and Percentage of Normal Index (PNI) (Willeke et al. 1994) (Zargar et al. 2011; Zarei et al. 2021; Ndayiragije and Li 2022).

In this study, the most preferred indices for investigating meteorological drought are SPI, SPEI, and RDI. Among these indices, the Standardized Precipitation Index (SPI) is particularly favored due to its ease of calculation using only precipitation data, comparability across different climatic conditions, and facilitation of representing wet/dry periods (WMO 2012; Ndayiragije and Li 2022). However, the SPI method's focus on temporal changes in precipitation can overlook variations in other climatic variables (Ndayiragije and Li 2022; Kapluhan 2023; Pekpostalcı et al. 2023). Numerous studies worldwide have utilized SPI to examine wet/dry periods, contributing to the generation of insights into future drought development (Das et al. 2020; Malik and Kumar 2021; Isia et al. 2023; Miro et al. 2023; Yang et al. 2023). Another drought index preferred in the study is the Standardized Precipitation Evapotranspiration Index (SPEI), which is calculated similarly to the SPI formula. However, unlike SPI, SPEI takes into account both precipitation and evapotranspiration (ET). ET, reflecting the impact of the relationship between temperature and evaporation on drought, (Mehr et al. 2020), makes SPEI a more sophisticated method compared to SPI. In various studies in the literature, SPI, SPEI, and RDI indices have been compared to determine meteorological drought in different climatic regions and assess its impact on agricultural production (Xu et al. 2015; Pathak and Dodamani 2020; Zarei et al. 2021). Similar to the SPEI index, another preferred index used in the study is the Reconnaissance Drought Index (RDI), which incorporates both precipitation and ET (Jabbi et al. 2021). The RDI and SPEI indices provide similar climatic information. However, while RDI adopts a more general approach to rainfall distribution and assessment, SPEI is advantageous for determining agricultural drought and evaluating the impact of the temperature-evaporation relationship on drought (Tsakiris et al. 2007; Sobral et al. 2018; Zarei et al. 2021; Abbes et al. 2023). The similarity in the climatic parameters used to calculate SPI, RDI, and SPEI indices enables comparisons, and their utilization in the study is favored due to their ability to depict spatial variations and similarities in describing them.

In studies utilizing drought indices to analyze the spatial–temporal variation of drought, interpolation methods developed to generate data in locations without data are commonly used. Stochastic methods such as Kriging, which consider spatial autocorrelation, have been successfully used in analyzing the spatial and temporal distribution of drought (Moreira et al. 2008; Akhtari et al. 2009; Gocic and Trajkovic 2013). For Türkiye, which has a wide and varied geography, studies involving the interpolation of climatological data have determined that Ordinary Kriging is the method that best represents spatial distribution (Yavuz and Erdoğan 2012; Aydın, 2014; Ozelkan et al. 2015; Cağlıyan and Ayhan 2018; Khorrami and Gündüz 2021). In our study, the Ordinary Kriging method was chosen for the spatial interpolation of drought data. The implementation of drought methods and the creation of maps were conducted in a Geographic Information Systems (GIS) environment.

The Mediterranean basin is referred to as a 'hotspot of climate change' according to climate models (Lionella and Scarascia 2018; Lange 2020). Due to its location in a dry and semi-drought climate region and the combination of increasing temperatures with decreasing rainfall, it is anticipated that the impact of climate change will intensify, leading to more severe droughts in the Mediterranean basin (Türkes 2020a; Erlat and Güler 2023). Particularly in the Mediterranean basin, an increase in annual temperatures by 25% above the global average and 40% more during summer months is expected by the mid to late twenty-first century (Lionella and Scarascia 2018). This situation contributes to an increased vulnerability of Türkiye to drought events and other potential impacts of global climate change. Therefore, rising temperatures and decreasing precipitation values may lead to adverse effects on factors such as water resources, food security, energy sources, terrestrial and marine ecosystems, and human health and safety. In recent times, the direct impact of drought on water resources and water management, especially in extensive geographic areas with pronounced effects of climate change such as Türkiye with its dry and semi-drought climate characteristics, is among the most intensively researched topics. The severity of drought in Türkiye and its accurate prediction are of vital importance for the sustainable management of water resources. The effects of climate change have been increasing recently. Therefore, climate change and the reasons for it and situations causing serious problems such as drought need to be updated. Previous studies have looked at specific drought years and specific areas. This study covers the current data of the last few years and at the same time constitutes a whole of the studies carried out in the past periods. For this reason, the aim of this study is to reveal the temporal and spatial changes of drought in Türkiye using different drought indices and to evaluate the impact of drought on atmospheric circulation systems with current data. In this context, the main objectives of this study are as follows:

• Between 1991 and 2022, the comparison of the severity of drought in Türkiye using SPI/SPEI and RDI indices for 1, 3, 6, 9, and 12-month periods across stations in the study area, and determination of the driest/moistest years by month,

• Determination and comparison of average frequency values of drought classes over years for SPI/SPEI and RDI methods between 1991 and 2022,

• Spatial distribution of the driest years identified for each index type will be mapped using Geographic Information Systems (GIS) to depict changes in drought severity in Türkiye for those years.

The subsequent sections will describe the general characteristics of the study area, the meteorological data used, and the methods employed in processing this data and analyzing the temporal and spatial changes in drought.

Data and methods

This section will sequentially describe the general characteristics of the study area, the meteorological data used, and the processing of this data, as well as the methods employed in analyzing the temporal and spatial changes in drought. In the study, drought indices of SPI, SPEI, and RDI for 1, 3, 6, 9, and 12-month periods were calculated for 219 stations between 1991 and 2022. These calculations were examined in four stages:

• 1. Average frequency values for each year were generated based on drought classes to compare which drought class occurred most frequently in each year according to the drought indices.

• 2. The driest and wettest years for each month between 1991 and 2022 were determined in all drought indices, along with the corresponding regions where these values were observed in the study area.

• 3. The driest years for each month were identified in each drought method, and spatial distribution maps were created for these years.

• 4. The results of the conducted studies were evaluated.

The data and methods to be used within this scope are explained as follows:

Study area

The study area covers an area of 783,562 km² between 36°-42° North latitude and 26°-45° East longitude (Fig. 1). The average elevation is 1132 m and the region is located in the Mediterranean major climatic zone of the Subtropical belt. However, the diversity of landforms and especially the presence of mountains running parallel to the coast cause climatic conditions to vary from region to region. While the coastal areas have milder climatic conditions, the inland parts of the Anatolian plateau are characterized by hot summers, cold winters and low rainfall. According to the Thornthwaite climate classification, arid regions are in the interior, east and southeast of the study area (Fig. 2) (Sensoy et al. 2008). Humid regions are located in the northern parts, around Bitlis city (located in the south of Türkiye) and in the southwest (Fig. 2). The other regions of Turkey are generally characterised by semi-arid and semi-humid climate zones. The Mediterranean climate zone, where Turkey is located, is highly affected by global climate change due to being under the influence of mobile mid-latitude and Mediterranean depressions (Olgen 2010). The frequency and effectiveness of mobile depressions, high atmospheric pressure, westerly winds and polar jet stream caused by the Mediterranean front and the Polar front control the precipitation conditions, which causes the occurrence of dry periods (Kömüscü, 2001). Apart from the effect of depressions due to its location, precipitation regimes and variability in Türkiye are also affected by atmospheric controls such as the North Atlantic Oscillation (NAO). In particular, periods with positive NAO anomalies correspond to dry periods over Türkiye (Kömüscü, 2001). At the same time, the fact that Türkiye is surrounded by seas on three sides, the increase in the average elevation from west to east and the direction of extension of the mountains cause differentiation of climate characteristics in short distances (Sensoy et al. 2008). These geographical features of Türkiye bring about spatial and temporal variations in precipitation and temperature due to atmospheric controls and microclimatic zones. This change also causes regional rainfall irregularity, water shortage and drought effects. While the areal precipitation value for 2022 is 503.8 mm in Türkiye, the average areal precipitation between 1991 and 2020 is 573.4 mm. The average temperature value is 13.9 °C (MGM 2022). According to the 2022 climate assessment report, the highest temperature was 47.9 °C in July in Şırnak/Silopi and the lowest temperature was -34.4 °C in January in Van/Özalp (MGM 2023). According to 2022 data, the highest precipitation was 87 mm in January and the lowest precipitation was 9 mm in July (MGM 2023). Climatic differences in the study area cause changes in the natural vegetation cover in a short distance, forming various formations such as forest, shrubs and grasses (Avcı, 2005; Günal 2013).

Fig. 1

Meteorological stations in the study area

Fig. 2

Climate classification of Türkiye according to Thorntwaite precipitation index (Sensoy et al. 2008)

Meteorological data

Monthly total precipitation data and monthly maximum, minimum and average maximum, minimum and average temperature data measured between 1991 and 2022 at 219 meteorological observation stations across Türkiye (Fig. 1) belonging to the General Directorate of Meteorology of the Ministry of Environment, Urbanization and Climate Change of the Republic of Türkiye were used to calculate the SPI, RDI and SPEI drought indices. The main reason for selecting this period is the lack of missing data in the measurements made for all stations during the observation period between 1991 and 2022.

Determination of average frequency values

Seven drought classes used in the evaluation of drought indices were determined. The recurrence rates of these drought classes for each year were determined and graphs were created according to their average values. As a result, interpretations were made according to the change in the average recurrence values of the drought classes in which year.

Drought ındices

Standard precipitation ındex

SPI is considered as a statistically correlated index calculated based on the probability of precipitation in a given time period (Mlenga et al. 2019) (1). The first period when the index falls below zero is considered the beginning of drought, and the month when the index rises to a positive value is considered the end of drought (McKee et al. 1993). In the classical approach of SPI, the cumulative distribution function of rainfall totals is fitted to the appropriate frequency distribution (McKee et al. 1993). In SPI, precipitation is converted into normalized common units based on numerical measurements (Türkeş and Tatlı, 2009; Badji et al. 2023). The number of standard deviations refers to the deviation of standard measurements from the long-term average (Mlenga et al. 2019). As a result, SPI can be used for normal, log-normal and gamma distributions of precipitation (Zeybekoglu et al. 2023). In this study, drought analyses were calculated with the R-Studio package program and it is assumed that the intensity function of precipitation in a given period fits the following gamma distribution (1):

$$\text{g}\left(\text{x}\right)=\frac{1}{{\beta }^{\alpha }\Gamma \left(\alpha \right)}{x}^{a-1}{e}^{-x/\beta }x>0$$

(1)

The continuation of the formulas given in accordance with the Gaussian distribution can be found in the sources. In this section, only the first and the last part of the formula are given (see e.g. following references for a similar approach: (Lim Kam Sian et al. 2023). SPI is the average of the difference of the precipitation total for a studied time period from the long-term average for that time period, again obtained from the long-term data for that time period.

$$\text{SPI}=\frac{\left({\text{X}}_{\mathrm i}-X_{\mathrm{ort}}\right)}\sigma$$

(2)

where ${\text{X}}_{\text{i}}$ represents the precipitation total within the time period, $\overline{X }$ ̅ represents the long-term average of the precipitation totals within the studied time period, and σ represents the long-term standard deviation of the relevant time period (2) (Tatlı et al. 2004; Tatlı and Türkes 2008; Ozelkan and Karaman 2018). The obtained SPI values can be calculated for different periods (such as 1, 3, 6, 9, 12, 24, 48 months) and these values can be subjected to the classification given in Table 1.

Table 1 Drought analysis classification values (Zarei et al. 2021)

Standardized precipitation evapotranspiration ındex

SPEI is used to assess the severity of agricultural drought, taking into account plant evapotranspiration and meteorological drought. SPI is calculation-based and expressed as the difference (D) between potential evapotranspiration (PET) and precipitation (P) (3). This difference (D) constitutes the water surplus or deficit for the analyzed month (i) (Bakanogulları 2020; Badji et al. 2023).

$${\text{D}}_{i}={P}_{i}-{PET}_{i}$$

(3)

In SPEI calculation, unlike SPI analysis, the difference between precipitation and PET (D) is calculated as water deficit. After the accumulated precipitation is converted into probabilities, these probabilities are transformed into a standard distribution to form drought index values. In this way, SPEI values represent the number of standard deviations from total precipitation or climatic water deficit for a given location and year (Stagge et al. 2015). The classification of the value ranges used for the evaluation of the analyses is shown in Table 1.

Reconnaissance drought ındex

RDI (Reconnaissance Drought Index), which is one of the most widely used indices in drought analysis, is advantageous for calculating PET like SPEI. Indices such as SPI and SPEI are calculated based on precipitation and temperature factors and are important in evaluating PET values and agricultural productivity (Tsakiris et al. 2007). The RDI was presented at the MEDROPLAN coordination meeting as a new drought analysis method and has been described in more detail in other publications (Tsakiris 2004; Tsakiris and Vangelis 2005).

A detailed calculation of the formula can be found in the references. The last step of the formula is given in this section (see e.g. following references for a similar approach: (Tsakiris and Vangelis 2005; Tsakiris et al. 2007; Zarei et al. 2021).

$${RDI}_{st}=\frac{{\sum }_{i=1}^{N}\left({a}_{0}^{\left(i\right)}-\overline{{a }_{0}}\right)}{\sqrt{\frac{1}{N}} \sum_{i=1}^{N}({a}_{0}^{\left(i\right)}-\overline{{a }_{0 }}{)}^{2}}$$

(4)

In this formula, ${RDI}_{st}$ represents the standardized RDI; N represents the total number of years; ${a}_{0}^{\left(i\right)}$ ${a}_{0}$ parameter of year i; $\overline{{a }_{0}}$ represents the arithmetic mean of ${a}_{0}$ parameter calculated for N years of data (4) (Tsakiris et al. 2007). The standardized RDI includes the standard deviations of each year's a_0 parameter relative to the mean values, thus providing a standardized measure in drought analysis. This index is used to compare and interpret meteorological drought conditions (Zarei et al. 2021).

Spatial ınterpolation of climate data

Spatial interpolation methods provide convenience for estimating and reproducing values at other unobserved points by using the data of points where observations of a variable are made. For this reason, Kriging method is the most widely used among geostatistical methods that are more preferred in studies. In Kriging method, interpolation surfaces are generated from observations with limited sample (Isaaks and Srivastava 1989). In order to apply these geostatistical methods, it is an important prerequisite to determine the different spatial conditions of observations close to each other (Khorrami and Gündüz 2019). One of the most preferred geostatistical methods to show the spatial distribution of climatic data is Ordinary Kriging (OK) method (Wang et al. 2014). The OK method estimates the values of points with no data by weighted averages of points with nearby data (Hadi and Tombul 2018). The OK method is expressed by the following formula (15) (Hadi and Tombul 2018). The OK method is formulated as follows:

$$\widehat{Z }\left({X}_{0}\right)= \sum\nolimits_{j=1}^{N}{\lambda }_{i}Z({X}_{i})$$

(5)

Here, the value of the predicted point $\widehat{Z}$ at location ̂ ${X}_{0}$ is equal to the sum of the values of each sampled point at location ${X}_{i}$, multiplied by its weight ${\lambda }_{i}$ (Khorrami and Gündüz 2019).

Semivariogram models form the basis of geostatistics and are used to model the spatial dependence of variables (Aksu 2023). The semivariogram model is a technique that measures the half-mean square difference between values against the lag distance or separation of points in spatial autocorrelation (Biswas and Si 2013). There are different types of semivariograms such as spherical, exponential and gaussian. Among these types, the most preferred gaussian method is used to reduce the effect of outliers in climatic parameters by normalizing the data distribution and was also preferred in this study (Hadi and Tombul 2018; Varouchakis 2021).

Results

Precipitation and temperature assessment

The monthly variation of the total annual precipitation between 1991–2022 used in the study is shown in Fig. 3. The average annual precipitation value between 1991–2022 is 1610.9 mm. The months with precipitation above the annual average belong to the months other than July, August and September (Fig. 3). According to the averages of 1991–2022, the highest precipitation in our country is in the Eastern Black Sea region, while the lowest precipitation is in the central parts of Central Anatolia and around Şanlıurfa and Igdir (Fig. 4) (MGM 2022).

Fig. 3

Graph of monthly average total precipitation, average temperature and average PET values in Türkiye between 1991–2022

Fig. 4

Spatial variation of annual total precipitation 1991–2022

The variation of average temperature values by months between 1991–2022 is shown in Fig. 2. The months with the lowest precipitation values in June, July and August correspond to both the highest temperature values and the highest evaporation values. The annual average temperature value is 14 °C and the annual average potential evapotranspiration value is 137.5 mm. Above-average temperature values were experienced between March and October between 1991–2022. In the average potential evapotranspiration values, the months above the average are between April and September, when temperature values and evaporation increase (Fig. 3). According to the average temperature values of 31 years used in the study, the highest temperature values extend from the Marmara coast to Southeastern Anatolia (Fig. 5). The lowest mean temperature values are observed in the north–south direction in eastern Central Anatolia and Eastern Anatolia (Fig. 5).

Fig. 5

Average temperature distribution between 1991–2022

Average frequency values of drought methods by year

In drought analyses, there are class intervals determined to evaluate the time periods when drought periods begin and end and to express drought severity. These class intervals, which are given in Table 1, enable the drought severity to be expressed on a certain scale and facilitate the interpretations made for the region. The high topographic and climatic diversity of the study area causes drought severity to vary. In this context, the average frequency values of drought classes between 1991–2022 were determined. Graphs were created in order to understand the change of the average frequency values calculated for the evaluation of drought classes according to years for all index types. The years with the highest average frequency values in the graphs were evaluated in detail by comparing the drought analyzes.

In monthly assessments, the highest average frequency values for extremely wet conditions (2.0 and above) and very wet conditions (1.50 to 1.99) in SPEI₁ and RDI₁ methods belong to 2022 (Fig. 6). In the monthly SPI₁, extremely wet conditions were observed in 2001, while very wet conditions were identified in 2009 (Fig. 6). These findings emphasize the consistency of different drought indicators in assessing wet conditions and their ability to identify wet periods that vary by year. Moderately wet conditions (1.49 to 1.00), as in other drought classes, have the highest average frequency values in SPEI₁ method in 2022, while RDI₁ and SPI₁ indices show the highest frequency value in 2009 (Fig. 6).

Fig. 6

Representation of average frequency values of drought classes for 1991–2022 for 1-month assessment results

While the average frequency values of mild drought conditions (-0.99 to 0.99) were high in 2005 in all three index types, it was determined that the average frequency value was high in 2017 in RDI₁ and SPI₁ indices (Fig. 6). For moderate drought conditions (-1.00 to -1.49), the highest mean frequency values for SPEI₁ were in 2016, while for RDI₁ and SPI₁ indices they were in 2008 (Fig. 6). The highest average frequency values of severe dry periods (-1.50 to -1.99) in all indices were observed in 2021. The average frequency value of exceptionally dry conditions (-2.00 and below) was highest in 2001 for all index types (Fig. 6).

In the 3-month analysis evaluations, when analyzed according to drought classes, extremely wet conditions (2 and above) reached the highest average frequency values in SPEI₃ and RDI₃ methods in 2022 (Fig. 7). However, in the SPI₃ index, these conditions had the highest recurrence rate in 2002 (Fig. 7). The very wet (1.50 to 1.99) drought class range reached the highest values in 2022 in the SPEI₃ and RDI₃ methods, as in the case of extremely wet conditions, while the highest average frequency value was determined in 2009 in the SPI₃ method (Fig. 7). In addition, moderately wet conditions (1.49–1.00) had the highest average frequency value in 2009 for all index types (Fig. 7).

Fig. 7

Representation of the average frequency values of drought classes for the period 1991–2022 for the 3-month assessment results

The highest mean frequency value for near normal drought conditions (0.99 to -0.99) belongs to 2005 for all three index species (Fig. 7). The drought class range for moderately drought conditions (-1.00 to -1.49) has a high average frequency in similar years in all three methods. It was determined that moderate drought conditions were more recurrent in 2007, 2008 and 2021 in all methods (Fig. 7). During severe dry periods (-1.50 to -1.99), the highest average frequency value was observed in 2007 and 2020 in all three methods (Fig. 7). Exceptionally drought (-2.00 and below) value intervals were repeated most frequently in 2007 and 2021, similar to each other in the three methods (Fig. 7).

In the 6-month analysis evaluations, the extremely wet (2.00 and above) and very wet (1.50 to 1.99) conditions with the highest average frequency values belong to 2022 for SPEI₆ and RDI₆ indices (Fig. 8). As in the other assessments, 2002 has the highest average frequency value for the SPI₆ index. The year with the highest mean frequency value for moderately wet (1.00 to 1.49) conditions is 2009, which is the same year for all three indices (Fig. 8).

Fig. 8

Representation of the average frequency values of drought classes for the period 1991–2022 for the 6-month assessment results

Near normal drought (0.99 to -0.99) conditions had the highest mean frequency value in 1996 for all index types (Fig. 8). In moderately drought (-1.00 to -1.49) conditions, the years 2001 to 2021 indicate high values in all indices (Fig. 8). In severely drought (-1.50 to -1.99) conditions, 2021 is a remarkable year with high average frequency values in all three index types (Fig. 8). In exceptionally drought (-2.00 and below) conditions, the years 2021 and 2007 indicate high average frequency values in all three methods (Fig. 8).

According to the 9-month analysis results, the years with the highest recurrence of extremely wet (2.00 and above) conditions were determined as 2022 for SPEI₉ and RDI₉ indices (Fig. 9). Moreover, 2015 is the period with the highest average frequency value for SPI₉ (Fig. 9). The average frequency values for the very wet (1.50 to 1.99) value range were highest in 1998 for RDI9 and SPI₉ indices (Fig. 9). For SPEI₉, 2022 is the period in which the average frequency values increase compared to other years (Fig. 9). Moderately wet (1.49 to 1.00) values are observed as conditions repeated in 1998 in all three index types (Fig. 9).

Fig. 9

Representation of the mean frequency values of drought classes for the period 1991–2022 for the 9-month assessment results

Near normal drought (0.99 to -0.99) conditions were most repeated in 1997 for all indices (Fig. 9). The years in which moderately drought (-1.00 to -1.49) conditions were repeated the most in all three index types were 2001 and 2008, while the extreme drought (-1.50 to -1.99) value ranges were in 2001 and 2021 (Fig. 9). Exceptionally drought (-2.00 and below) conditions were repeated most frequently in 2007, 2014 and 2021 in all three methods (Fig. 9).

In the 12-month analysis of extremely wet (2.00 and above) conditions, the highest average frequency values were reached in 2015 for the SPI₁₂ and RDI₁₂ indices and in 2022 for the SPEI₁₂ index (Fig. 10). One of the years in which very wet (1.50 to 1.99) conditions persisted, 1998, was observed in all three index types (Fig. 10). The highest average frequency values occurred in 2015 for the RDI12 and SPI₁₂ methods and in 2022 for the SPEI₁₂ index (Fig. 10). Moderately wet (1.49 to 1.00) conditions were repeated in 1998 for all three index types (Fig. 10).

Fig. 10

Representation of the average frequency values of drought classes for the period 1991–2022 for the 12-month assessment results

Unlike the wetity classifications, near normal drought (0.9 to -0.99) conditions reached the highest mean frequency value in all methods in 1997 (Fig. 10). Moderately drought conditions (-1.00 to -1.49) had the highest frequency in 2001 in all indices (Fig. 10). In 2014 and 2021, it was observed that severe drought (-1.50 to -1.99) value ranges had the highest average frequency values in all methods (Fig. 10). It was determined that extreme conditions (-2.00 and below) had high average frequency values in all methods in 2007, 2014 and 2021 (Fig. 10).

The years of effective and severe water deficit in most of Türkiye are noteworthy in the recurrence of drought severity in average frequency values. In 1 and 3-monthly evaluations, the average frequency value of severe drought values in 2001, 2005, 2007, 2008, 2016, 2017, 2020 and 2021 is high. In the short-term drought indices, the effect of meteorological drought, which was combined with the summer drought in 2013, which increased in severity after 2012, and the drought of 2001, 2007–2008, which was the longest drought in Türkiye (Türkes and Erlat 2003; Türkes 2020b) . In the 6, 9 and 12-month indices, it was determined that the average frequency values in 1996–1997, 2001, 200–2008, 2014 and 2021 were high in severe droughts. This indicates that long-term drought indices reflect the years corresponding to prolonged severe drought in Türkiye.

SPI, SPEI and RDI methods 1-month evaluation analyses

In drought indices, 1-month periodic assessments are based on short-term monthly precipitation distribution. Short-term soil moisture, especially during the growing season of plants, is closely related to plant stress and meteorological drought (WMO 2012). In 1-month periodic evaluations, the years 1994, 2003, 2006, 2007, 2010, 2014, 2015, 2017 and 2020 were the years with the highest drought values in SPI₁ method. In the RDI₁ method, the highest drought values belong to the years 1994, 2006, 2010, 2012, 2015, 2016 and 2017. In the SPEI₁ method, the years 1999, 2001, 2007, 2008, 2009, 2016, 2017 and 2019 are the years with the highest drought values. The spatial variation of these years in the study area was mapped for each month. In addition, by comparing all indices, the driest/moist years for each month and their spatial distribution in the study area are briefly summarized below with the reasons.

The years and areas with the highest drought values in January, February, March and April are shown in Fig. 11. In January, SPI₁ (-6.08 to -1.29) and RDI₁ (-4.72 to -1.28) indices show high drought values in the southeastern part of the study area, while SPEI₁ (-4.54 to -1.28) index shows high drought values in the northern part of the study area. In the winter season of 2010, when the highest drought values were observed in January in SPI₁ and RDI₁ indices, the above-average temperature values caused the drought severity in the southeastern part to increase (Güler and Erlat 2023). The fact that 2010 was the warmest winter season explains the high drought severity in the stations in this region for both index types. In the SPEI₁ index, especially the winter precipitation in the form of snow in 2014 caused exceptionally drought values in the northern parts of the study area (Fig. 11) (Simsek et al. 2013). While drought values in February were high in SPI₁ (-6.05 to -0.83) and RDI₁ (-4.83 to -0.62) methods in almost all of the study area, high drought values in SPEI₁ (-3.02 to -0.63) index were observed in the inland areas (Fig. 11). The increasing trends in drought observed in SPI₁ and RDI₁ indices in February 2017 were detected during the period from December to September, which also affected inland Türkiye, eastern and western Iran, northern Iraq and Syria, western Jordan, and Israel (Sahour et al. 2020). The increasing trend in drought is particularly associated with reduced precipitation and increased potential evapotranspiration.

Fig. 11

Graphs of the years (January-April) in which the highest droughts were determined for each month of the 1-month analysis in SPI, SPEI and RDI methods

While the drought severity in March was high in the western part of the study area in SPI₁ (-5.64 to -1.12), it increased in the eastern part of the study area in RDI₁ (-4.37 to -0.86) and SPEI₁ (-3.22 to -0.38) indices. In the SPI₁ index of March, it is seen that the drought severity in 2007 increased especially with the below normal fall and winter precipitation starting from 2006 until December 2008 (Türkes 2012). Consistent with the results of the study, this aridification trend was most effective in the Aegean Mediterranean, Marmara and Southeastern Anatolia in Türkiye (Türkes 2007). In the spring season, March, especially the temperature increase trends observed since 2001 explain the increasing drought severity in SPI₁ and RDI₁ indices in 2012, 2015 and 2018 (Fig. 11). For SPI₁ (-4.82 to -0.94), the drought severity in April was higher in the western part of the country and for SPEI₁ (-3.70 to 0.00) in the eastern part of the country, while for RDI₁ (-4.82 to 0.00), the drought severity increased from the western part to the northeast (Fig. 11). For the SPEI₁ index, drought severity was high from February to April 2008 (Fig. 11). The year 2008 was a period of high drought severity in Türkiye with increasing temperatures and decreasing precipitation. During this period, especially in agricultural production, the country experienced serious production losses in crops such as wheat, corn, paddy, sunflower and potato (Taskın et al. 2022).

In January, exceptionally wet conditions with wet conditions of 2 and above were observed in RDI₁ (4.51) in 2008 in the eastern part of the study area, SPI₁ (3.42) in 2015 in the south of the Marmara Sea and SPEI₁ (2.58) in 2016 in Central Anatolia. The southward expansion of the wet northern Black Sea belt, especially between 2015 and 2020, caused exceptional wet conditions (Kömüscü et al. 2022). In February assessments, the highest wet conditions were observed in RDI₁ (3.89) in the south of the Black Sea in 2011, SPI₁ (3.10) in the east of the study area in 2004 and SPEI₁ (2.73) in the south of the study area in 2022. According to the index types, the highest precipitation values in 2011 and 2022 are observed on the Black Sea coast and in the southern part of the study area as a result of the topographic effect due to the mountains extending parallel to the coastal area. March wet conditions are highest in Central Anatolia in 2015 with RDI₁ (4.37) and SPI₁ (3.91) and in the east of the study area in 2003 with SPEI₁ (2.60). The highest April wet conditions in RDI₁ (5.42) were in the northwest of the study area in 2013. SPI₁ (3.81) in 2012 in the western part of the study area and SPEI₁ (2.62) in 2015 in Central Anatolia. The fact that the rainy period of Central Anatolia coincides with the months in the spring season causes high wet conditions.

The years and areas with the highest drought values in May, June, July and August are shown in Fig. 12. In the drought assessments in May, SPI₁ (-4.40 to -0.89) and SPEI₁ (-3.54 to 0.00) indices show moderate drought conditions across Türkiye, while the highest drought value in RDI₁ (-5.44 to -0.75) index is observed in the northeast (Fig. 12). May is the month with the most severe warming after April 2008, and the increase in drought severity with increasing temperatures, especially after 2003, is remarkable (Güler and Erlat 2023). The SPI₁ method, in which 2014 is represented by severe drought, is in parallel with other studies (Simsek et al. 2013). RDI₁ 2016 in May and SPEI₁ 2017 in May show that the impact area of drought has expanded with the effect of increasing warming after 2008. In June, SPI₁ (-4.99 to -1.27) shows high drought severity in the south, while SPEI₁ (-3.92 to 0.00) shows an increase in drought severity in the coastal area from north to south (Fig. 12). Unlike the other indices, RDI₁ (between -3.90 and -1.02) indicates normal drought conditions in the eastern parts and wet conditions in the interior.

Fig. 12

Graphs of the years (May–August) in which the highest droughts were determined for each month of the 1-month analysis in SPI, SPEI and RDI methods

SPI₁ in July (-3.42 to -1.80) and August (-4.63 to -1.78) shows that the study area is generally dominated by normal drought values. However, the RDI₁ index shows an increase in drought severity in the northeast in July (-3.44 to -0.87) and in the north in August (-4.18 to -1.11) (Fig. 12). In the SPEI₁ index type, drought severity increases in the inland areas in July (-3.22 to 0.00), while this situation shifts towards the southeast in August (-3.83 to -1.05) (Fig. 12). The positive anomaly values observed in summer season temperatures after 2010 are associated with the mega-temperatures affecting Türkiye, with Russia in the center. The effect of this temperature increase on aridity is seen in RDI₁ index in June and July 2015 (Fig. 12). In SPEI₁, July 2012 and August 2019 show the impact of increasing temperatures on drought. In June, SPI₁ 2003 and SPEI₁ 2007 are the years when the highest drought severity is repeated. SPI₁ 1991 in July and 2006 in August are the years with the highest drought severity in SPI₁ and RDI₁ indices. It was determined that the drought severity, which was especially effective starting from 2006 until 2008, caused the highest drought values to be seen in these indices in the summer months (Türkes 2012; Akbas 2014).

In 2022, the highest wet conditions for all methods were determined in the Black Sea coastal region. RDI₁ (5.55) was measured in the Eastern Black Sea station, while SPEI₁ (2.94) in the Central Black Sea and SPI₁ (3.00) in the Western Black Sea were the highest. When ranked according to the wet conditions in July, the highest value was obtained in the RDI₁ (4.46) method measured in the Eastern Black Sea region in 2016. Then, the second highest value was SPI₁ (3.56) calculated in the eastern part of the study area in 2009. The lowest wet conditions were observed in 2022 with a value of 3.41 obtained with the SPEI₁ method applied in the coastal part of the Aegean Sea. The highest wet conditions in August in 2022 were RDI₁ (3.64) in the west of the study area and SPEI₁ (3.08) in the southwestern coastal area of the Aegean Sea. In the SPI₁ (2.75) method, wet conditions were high in the south of the Marmara Sea in 1997.

The years and areas with the highest drought values in September, October, November and December are shown in Fig. 13. In September, drought values in SPI₁ (between -4.46 and -1.38) and RDI₁ (between -3.35 and -1.01) indices are high in the northern and northwestern parts. In SPEI₁ (-3.09 to 0.00), the highest drought values are observed in the northern and eastern parts (Fig. 13). The warming trend that started in the first half of the 1990s has gradually increased since the twenty-first century. The highest drought values were determined in SPI₁ and RDI₁ indices in September 1991 and in SPEI₁ index in December 1998. In Türkiye, the period when a significant warming trend is observed belongs to the fall months. The warming trend seen especially since 2010 causes drought to recur more frequently. With the effect of increasing temperature and evaporation, the highest wet conditions in the SPEI₁ index is in 2017 as in other studies (Güler and Erlat 2023). In October, it is observed that the drought severity increases from south to north and east in the SPI₁ (-6.10 to -0.39) and RDI₁ (-4.62 to -0.38) indices, while the drought severity increases in the western parts of the SPEI₁ (-3.04 to -0.59) index (Fig. 13). In October, SPI₁ and RDI₁ show the highest drought severity in Türkiye in the fall season of 2020.

Fig. 13

Graphs of the years (September-December) in which the highest droughts were determined for each month of the 1-month analysis in SPI, SPEI and RDI methods

In November, drought severity in SPI₁ (-4.89 to -0.75) index has the highest values from south to north. RDI₁ (-4.25 to -0.81) has the highest drought values from inland to the east (Fig. 12). SPEI₁ (-2.53 to -0.99) has high drought values only in inland areas. In December, SPI₁ (-5.56 to -0.19) and RDI₁ (-4.61 to -0.23) indices have similarly high drought values in the western and southern parts, while SPEI₁ (-2.39 to 0.00) also has high drought values in the eastern parts (Fig. 13). In November, drought intensity is high for SPI1 and RDI1 in 2010. In December, SPI1 and RDI1 indices show an increase in drought values for 2015. The stations in the east and southeast, where drought severity was the highest during the analysis period, are located in both topographic conditions and continental climate zone. In addition, the fact that different types of air masses are effective according to the months causes the drought severity to increase. The stations in the continental climate regions are closed basins and local differences in the climate in mountainous areas and basins and the variability in the distribution and regime of precipitation increase the vulnerability to drought (Bahadır 2013; Kale 2021).

In September 2022, high wet conditions in RDI₁ (4.48) and SPEI₁ (2.70) were determined in the east of the study area. In the SPI₁ (2.81) method, the wet conditions is high in the northern part of the study area in 2014. The highest October wet conditions at the stations in the east of the study area belong to RDI₁ (4.55) in 2022 and SPI₁ (3.11) in 2015. In the SPEI₁ (2.55) method, the highest wet conditions in Central Anatolia in 2010 are noteworthy. RDI₁ (5.10), where wet conditions were highest in November, was determined in the east of the study area in 2022. In 2001, the highest wet conditions for both SPEI₁ (2.72) and SPI₁ (3.33) belong to the stations in the southwestern and southern parts. The highest wet conditions in December were determined as RDI₁ (4.00) in 1994 and SPI₁ (3.06) in 2013 at the stations in the east of the study area. However, in the SPEI₁ (2.57) method, the wet conditions were highest in 2014 in the southwest of the study area.

SPI, SPEI and RDI methods 3-month evaluation analyses

In 3-month assessments, it enables the comparison of precipitation in certain 3-month periods with the same 3-month precipitation for all years in the historical record (WMO 2012). In the SPI method, the highest drought values were determined in 2006, 2007, 2008, 2011, 2012, 2014, 2015 and 2022 in the 3-month analysis. In the RDI ₃ method, the highest drought values are found in 2006, 2008, 2012, 2013, 2015 and 2021. SPEI₃ method shows the highest drought values in 1998, 2001, 2007, 2011, 2017, 2018, 2020 and 2021. The months of these years are mapped and the spatial variation of 3-month drought is shown. By comparing all indices, the driest/moist year and stations for each month are briefly summarized below.

The years and areas with the highest drought values in January, February, March and April are shown in Fig. 14. In January, it is seen that normal drought values are dominant in SPI₃ (-5.72 to -0.94) and RDI₃ (-5.47 to -1.07) indices throughout Türkiye. However, in SPEI₃ (-3.55 to -0.88), drought severity is shown to be high in the eastern part (Fig. 14). There are significant decreasing trends in annual precipitation and especially in winter precipitation in the Eastern Mediterranean basin and in general. This situation is observed to increase the severity of drought with the decrease in the frequency of mid-latitude and Mediterranean low pressures, which are especially effective in the winter season, and the increase in high pressure conditions in 2015 (Topcu 2022). This is supported by the results of SPI₃ and RDI₃ in January in Fig. 14. Especially in SPEI₃ index, the drought values that intensified in 2011 can be interpreted with the decrease in precipitation in the winter season and the increase in the effect of high pressure conditions in the region. In February, SPI₃ (-5.71 to -0.75) shows increased drought severity in the eastern part of the study area, while RDI₃ (-5.06 to -0.80) shows normal drought values throughout the study area. In SPEI₃ (-3.14 to -093), high drought values are observed in the southeastern part in February. In March, the areas with high drought severity differ in all index types (Fig. 14). Drought intensity increased in the northwestern and eastern parts in SPI₃ (-5.62 to -0.91), in the south in RDI₃ (-5.08 to -4.87) and inland in SPEI₃ (-3.67 to -0.62). In April, drought intensity was high in the eastern parts in SPI₃ (-5.38 to -0.75) and SPEI₃ (-3.00 to -0.31), while drought values increased in the Western Black Sea in RDI₃ (-5.10 to -0.54).

Fig. 14

Graphs of the years (January-April) in which the highest droughts were determined for each month of the 3-month analysis in SPI, SPEI and RDI methods

The increase in the drought values of SPI₃ in February and April 2012 in the eastern parts of the country is due to the effect of continental climate and at the same time, precipitation values below normal in this period (Simsek et al. 2012). With the effect of this drought, losses were observed in crops such as wheat, barley and sugar beet (Simsek et al. 2012). RDI₃ The drought in February and March 2013 can be evaluated with the increase in seasonal average temperatures since 2010 (Güler and Erlat 2023). The 2014 severe drought SPI₃ was observed in March. The drought covering the 2013–2014 agricultural year was severe, resulting in losses in agricultural yields (Simsek et al. 2013; Ozelkan et al. 2016). The 23.6% decrease in precipitation in February 2021 compared to normal SPEI₃ represents a severe drought period effective in the whole country (Taskın et al. 2022). The highest increase in annual average temperatures was in 2018, and the increase in evaporation intensity due to this situation is represented by the increased drought severity in the terrestrial regions and the entire study area in April SPEI₃ (Fig. 15). After 2000, increasing temperature and decreasing precipitation values affect severe drought in March SPEI₃ and April RDI₃.

Fig. 15

Graphs of the years (May–August) in which the highest droughts were determined for each month of the 3-month analysis in SPI, SPEI and RDI methods

In January, 2002 was the year with the highest wet conditions in SPI₃ (3.70) and SPEI₃ (2.94) methods and wet conditions was high at stations in Central Anatolia. In the RDI₃ (3.52) method, it was determined that the wet was the highest in the east of the study area in 2014. In RDI₃ (4.42) and SPEI₃ (2.87) methods, the highest wet conditions in February were in the Black Sea coastal area in 2022. In SPI₃ (3.58) method, the wet conditions is high in the south of the Marmara Sea in 2015. According to the wet conditions, the highest wet conditions was obtained in March in 2022 with RDI₃ (5.54) in the east of the study area and SPEI₃ (2.67) in the western Black Sea coastal area. The lowest wet conditions were observed in Central Anatolia in 2015 with SPI₃ (3.85). April wet conditions in the northwest of the study area in 2013 were high in SPI₃ (5.03) and RDI₃ (5.56) methods. The SPEI₃ (2.68) method shows the highest wet conditions in Central Anatolia in 2015.

The years and areas with the highest drought values in May, June, July and August are shown in Fig. 15. The effects of the 2007–2008 period of severe drought in Türkiye are seen in SPI₃ in May, RDI₃ in June and SPI₃ and RDI₃ indices in July. In the May assessments, it is observed that normal dry conditions prevailed across Türkiye in all three index types (Fig. 15). In the May map of SPI₃ (-4.93 to -0.85), drought increased in the western part and in the SPEI₃ (-3.20 to -0.47) index in the west coast. However, in the RDI₃ (-6.06 to -1.09) May map, drought is higher in the southeastern and northeastern parts. The drought effect seen in the entire study area in the RDI₃ and SPI₃ indices in May in 2016–2017 was due to the effect of the heat wave centered in Russia. At the same time, 2017 was the year with the highest heat waves in Türkiye (Erlat et al. 2021).

In the SPI₃ maps for the month of July (-4.70 to -1.03), the normal drought values seen in the southern part in June (-5.96 to -1.02) are seen to increase in drought severity from southwest to north (Fig. 15). In the SPEI₃ index, the drought values that intensified in the eastern parts in June and July in 2021 are related to the effect of decreasing precipitation compared to normal and increasing temperature evaporation intensity (Taskın et al. 2022). RDI₃ drought severity in June (-5.24 to -1.28) and July (-4.15 to -1.06) is high in the inland and southeastern parts. In SPEI₃ June (-3.33 to -0.64) and July (-4.05 to -0.10), drought intensity increased at stations in the east of the study area (Fig. 15). In August, the effects of 2007–2008 drought intensity are observed in SPI₃ and SPEI₃ indices (Fig. 15). The fact that the drought intensified during the 2007–2008 periods is predominantly concentrated in the western, southern and eastern parts of the study area is related to the above-average temperature values in response to the decreasing precipitation values during this period (Simsek and Cakmak 2010; Türkes 2020b; Erlat and Güler 2023).

Similar to April, wet conditions in May are also high in 2013. In SPEI₃ (2.90), it was determined that the wet conditions increased in the northern part of Central Anatolia in 2022. In June, as in other months in 2013, RDI₃ (6.06) and SPI₃ (4.82) show an increase in wet conditions in the northwestern part. In the SPEI₃ (2.97) method, wet conditions are higher in the northeastern part of Central Anatolia in 2022. In SPEI₃ (3.03) and RDI₃ (5.72) for July in 2022, the highest values were found in the stations in the eastern areas. In SPI₃ (4.03), there is a high relative wet conditions in the east of the study area in 2015. In August, similar to July, increasing wet conditions are observed in SPEI₃ and RDI₃ methods in 2022. The eastern part for RDI₃ (5.77) and the southwestern part for SPEI₃ (3.03) have the highest wet conditions. In SPI₃ (3.31), the highest wet conditions were determined in the eastern part in 2016.

The years and areas with the highest drought values in September, October, November and December are shown in Fig. 16. In the SPI₃ maps for September (between -4.97 and -1.15), while normal arid conditions are generally dominant in the study area, there is an increase in drought severity in some stations in the eastern part. The increased drought severity in the eastern part in the SPI₃ index coincides with the hottest fall season in 2022 (Fig. 16). In RDI₃ (-4.04 to -1.20), there are no stations with high drought intensity outside of normal conditions, while in SPEI₃ (-3.19 to -0.94), drought intensity is high in the western coastal areas (Fig. 16). In 2006, the drought intensity observed in April is remarkable in RDI₃ September and SPI₃ October. The atmospheric conditions affecting the occurrence of heat waves, which have intensified especially since 2005, cause south sector winds, which increases the severity of drought (Erlat and Türkeş, 2015). The 2007–2008 drought severity observed in August is also observed in September and October in the SPEI₃ index (Fig. 16).

Fig. 16

Graphs of the years (September-December) in which the highest droughts were determined for each month of the 3-month analysis in SPI, SPEI and RDI methods

SPEI₃ shows normal arid conditions in October (-3.52 to -0.51), November (-3.46 to -0.88) and December (-4.39 to -1.03), with higher drought severity in the stations in the western and eastern parts of the Black Sea coast, respectively (Fig. 16). In the RDI₃ index (-4.33 to -0.33), the drought intensity, which increases only in October in the south, is represented by normal dry conditions in the other months. In the SPEI₃ index, drought intensity values are high in the western stations in September, inland in October, eastern in November and along the western Black Sea coast in December (Fig. 16). In SPI₃ and SPEI₃ indices, the effect of the increasing drought severity in the winter months of 2011–2012 is observed in SPI₃ and RDI₃ indices in November and December in the eastern parts. It is seen that the effect of winter drought is high with the high pressure conditions prevailing in continental regions and the effect of decreasing precipitation. The years 2020, when annual average temperatures increased, and 2021, when the most severe drought was experienced, are seen in RDI₃ November and SPEI₃ December (Fig. 16).

In September, the highest wet conditions belong to SPEI₃ (2.67) and RDI₃ (5.09) methods in 2022, while SPI₃ (3.15) method has the highest wet conditions in 2009. In SPEI₃ and SPI₃ methods, stations with high wet conditions were determined in the Black Sea coastal area and in the eastern part of the Black Sea coast in RDI₃ method. In October, RDI₃ (5.28) method in 2022, SPI₃ (3.29) and SPEI₃ (2.73) methods in 2015, the highest wet conditions are present in the stations in the eastern part of the study area. In 2022, RDI₃ (6.03) method in the east of the study area and SPEI₃ (3.33) in the Black Sea coastal area have high wet conditions in November. In the SPI₃ (3,56) method, wet conditions increased in the south of the Marmara Sea in 2013. December wet conditions in RDI₃ (5,79) method are the highest in the east of the study area in 2022. In 2001, the wet conditions increased in SPEI₃ (2,64) in the south of the study area and SPI₃ (3,25) in the western part of the station.

SPI, SPEI and RDI methods 6-month evaluation analyses

It provides seasonal precipitation forecasts by reflecting short- and medium-term wet conditions. Indices for 6, 9 and 12-month periods basically express medium and long-term trends in precipitation (WMO 2012). In the 6-month index results, the highest drought values were determined in 2011, 2012, 2013, 2014, 2016 and 2022 in the SPI method. In the RDI method, the highest drought values belong to the years 2006, 2007, 2012, 2013, 2017 and 2020. For the SPEI method, the years 1993, 2003, 2007, 2008, 2012, 2014, 2020 and 2021 are the years with the highest drought values. These years were mapped and the driest/wettest year and stations for each month are briefly summarized below.

The years and areas with the highest drought values in January, February, March and April are shown in Fig. 17. In the SPI₆ maps for January (-5.61 to -0.78), February (-5.94 to -0.96) and March (-5.28 to -0.85), it is seen that drought severity increases in the east of the study area (Fig. 17). The severe drought observed in January 2011 in SPEI₃ index is also observed in January in SPI₆ index. This shows that the effect of the decrease in winter precipitation seen in the 3-month analysis is also noticeable in the 6-month analysis (Fig. 17). Similar to the SPI₃ indices, the drought, which intensified in the eastern part of the country where the continental climate is observed in winter, continues to affect SPI₆ in February and March. In the April SPI₆ maps (between -5.32 and -1.04), high drought values are observed in the stations in the northeastern part of the study area (Fig. 17). In RDI₆ index, it was determined that the normal drought severity in January (-5.70 to -0.92) in the whole study area increased in February (-4.74 to -1.10) in the inland areas. The RDI₆ index was higher in March (-4.97 to -1.15) at the stations in the northern and western parts of the study area and in April (-5.40 to -1.07) at the stations in the east (Fig. 17). During these periods, topographic conditions and the effect of decreasing precipitation conditions cause an increasing effect on drought severity. In 2007, the period of severe drought was observed in March in RDI₆. The most severe drought period of 2013–2014 also shows its effect on the 6-month indices (Fig. 17).

Fig. 17

Graphs of the years (January-April) in which the highest droughts were determined for each month of the 6-month analysis in SPI, SPEI and RDI methods

SPI₆ April, RDI₆ January and SPI₆ indices show severe dry periods from February to April. These periods support the yield decreases in agricultural products expressed in other studies (Akbas 2014; Taskın et al. 2022). In the maps prepared for the SPEI₆ index, drought severity increases from the north to the interior in January (between -3.01 and -0.88) and February (between -2.84 and -1.00). In March (-2.96 to -0.96) and April (-2.81 to -0.94), it is observed that it is effective from north to south in the eastern part of the study area (Fig. 17). The year 2021, when extreme drought was experienced, draws attention with its high drought severity in the SPEI₆ index throughout the study area (Erlat and Güler 2023).

The highest wet conditions in 2015 are observed in all three methods in the 6-month analysis in January. The highest wet conditions in RDI₆ (4.03) and SPI₆ (3.47) are in the southern part of the Marmara Sea. In SPEI₆ (2.64), the highest wet conditions is observed in Central Anatolia. In February 2015, similar to January, RDI6 (4.06) and SPI₆ (3.48) have high wet conditions in the southern part of the Marmara Sea. In SPEI₆ (2.60) method, the highest wet conditions in 2009 were highest in Central Anatolia. The highest wet conditions for March belongs to the Western Black Sea in 2022 for the RDI₆ (4.35) method. In 2019, SPI₆ (3.40) method had the highest wet in the southeast, while SPEI₆ (3.23) method had the highest wet conditions in the northern part of the Marmara Sea in 1998. In April 2013, SPI₆ (4.56) and RDI₆ (4.89) methods determined the highest wet conditions in the northwest of the study area. In the SPEI₆ (2.82) method, 2022 has the highest wet conditions in the Black Sea coastal area.

The years and areas with the highest drought values in May, June, July and August are shown in Fig. 18. It is seen that the drought values of the three index types in Türkiye increased between 2000 and 2014 with the effect of the record air temperatures observed between these years (Fig. 18). SPI₆ increases in the north–south direction in the east of the study area in May (-5.61 to -1.09), June (-5.39 to -1.20) and July (-5.92 to -1.05) (Fig. 18). In August (-4.77 to -1.03), there are high drought values in the southeastern part of the study area in SPI₆ index. In RDI₆ months, drought intensity increased from May (-5.64 to -1.07) to August (-5.29 to -1.11) in the western Black Sea coastal area (Fig. 18). In the SPEI₆ monthly SPEI₆ maps, the drought intensity increased only in the easternmost part of the study area in May (-3.44 to -0.90) and was replaced by normal dry conditions in July (-3.26 to -0.35), moving from west to east (Fig. 18). The SPEI₆ maps for June (-3.27 to -0.88) and August (-3.94 to -0.50) similarly show a general increase in drought severity from west to southeast (Fig. 18). In 2012, the impact of high droughts is evident in SPI₆ and SPEI₆ in May. In 2014, severe drought with severe agricultural yield losses is observed in the eastern part between July and August in the SPI₆ index. The 2007–2008 SPEI₆, another period of severe drought and agricultural yield decline, was observed throughout the study area in July and August. While the yield decrease in wheat, which is an important food for the world, was 14% in 2008, this yield decrease was recorded as 13.8% in 2014 (Taskın et al. 2022). In the RDI₆ index, drought severity increased in the Western Black Sea coasts with increasing temperature values between May–November 2006.

Fig. 18

Graphs of the years (May–August) in which the highest droughts were determined for each month of the 6-month analysis in SPI, SPEI and RDI methods

Similar to April, in 2013, RDI₆ (5.86) and SPI₆ (5.28) methods show high wet conditions in the northwest of the study area. SPEI₆ (3.18) also shows high wet conditions on the south coast of the Marmara Sea in 2022. In June 2013, the highest wet conditions in RDI₆ (5.99) and SPI₆ (5.54) methods are the highest in similar areas with April and May. SPEI₆ (2.83) determined the highest wet conditions on the north coast of the Marmara Sea in 2022 as in other months. The high wet conditions observed in the northwest of the study area since April 2013 also belong to RDI₆ (6.01) and SPI₆ (5.20) methods in July. Similarly, the SPEI₆ (2.85) method shows the highest wet conditions in 2022, with high values in the western part. In August 2013, high wet conditions recur in the southwest of the study area in the RDI₆ (5.86) and SPI₆ (5.03) methods. In 2022, SPEI₆ (2.93) shows high wet conditions in the western Black Sea coastal area.

The years and areas with the highest drought values in September, October, November and December are shown in Fig. 19. In September, SPI₆ (-5.42 to -0.52) and SPEI₆ (-4.47 to -0.80) indices show high drought severity in the southwestern and southeastern parts of the study area, while RDI₆ (-5.22 to -1.33) shows high values in the western Black Sea region (Fig. 19). SPI₆ shows high drought intensity in the west, east and south in October (-5.66 to -0.93), expanding its area of influence in November (-5.74 to -0.84), while in December (-6.07 to -1.02) it is only observed at stations in the northeast (Fig. 19). RDI₆ October (-5.04 to -0.94) is characterized by high drought intensity in the western Black Sea region, while RDI₆ (-5.53 to -0.94) and SPEI₆ (-3.17 to -0.70) have normal dry conditions in November. In SPEI₆ October (-4.12 to -1.01) and December RDI₆ (-6.07 to -1.02) and SPEI₆ (-2.92 to -0.77) indices, drought intensity was high in the western parts and the area of influence expanded (Fig. 19). In 2012, drought severity increased in November and December in SPI₆ index (Fig. 19). However, while the increasing drought intensity in 2020 is in the west of the study area in SPEI₆ and RDI₆ indices, the drought intensity in 2021 is high in the west of the study area in September in SPI₆ and SPEI₆ indices. It is thought that the increasing temperatures in the winter season in 2020 may increase the evaporation intensity and cause an increase in drought severity in the western part (Fig. 19). In 2021, severe drought in the eastern part of the study area indicates that the effect of decreasing precipitation throughout the study area has increased in semi-arid areas such as the continental climate zone (Fig. 19).

Fig. 19

Graphs of the years (September-December) in which the highest droughts were determined for each month of the 6-month analysis in SPI, SPEI and RDI methods

In the September assessments, RDI₆ (5.68) and SPI₆ (4.72) have the highest wet conditions in 2013 in the northwest of the study area. The highest wet conditions in 2022 with SPEI6 (2.90) value is in the Mediterranean coastal area. The highest wet conditions in October is in the east of the study area with RDI₆ (6.16) in 2022 and SPI₆ (4.32) in 2015. In SPEI₆ (2.91) method, the highest wet conditions were determined in the Black Sea coastal area in 2022. The highest wet conditions in November in 2022 was in the east of the study area in RDI₆ (6.17) and in the Black Sea coastal area in SPEI6 (2.73). In SPI₆ (3.44) method, the highest wet conditions were determined in the south of the Marmara Sea in 2013. The highest wet conditions in December belong to the east of the study area in RDI₆ (6.23) and the Black Sea coastal area in SPEI₆ (2.77) in 2022. SPI₆ (3.20) in the Mediterranean coastal area has the highest wet conditions in 2001.

SPI, SPEI and RDI methods 9-month evaluation analyses

In drought indices, the driest/wettest years and stations were evaluated by months. In the SPI₉ method, 2007, 2012, 2013 and 2014, and in the RDI₉ method, 2006, 2007 and 2014 were the years with the highest drought values. In the SPEI₉ method, the highest drought values were determined in 1994, 2003, 2007, 2014 and 2021. The months of these years were mapped and the spatial distribution of drought was shown.

The years and areas with the highest drought values in January, February, March and April are shown in Fig. 20. The effect of record temperatures between 2000 and 2014 caused an increase in drought severity in the 9-month index maps (Fig. 20). Especially after 1980, the Siberian anticyclone weakened and the Azores anticyclone became effective in the eastern Mediterranean, which led to an increase in winter drought (Topcu 2022). This would increase the impact of the North Atlantic Oscillation and lead to more recurrence of severe dry periods in Türkiye (Türkes 2020a; Topcu 2022). In the SPI₉ maps for January (-5.50 to -0.98) and February (-4.56 to -0.96), it was determined that normal dry conditions prevailed in general, although high drought values were observed at a few stations in the interior. SPI₉ The drought increase in January and February 2013, which was effective throughout the study area, caused a decrease in agricultural production. It affected cereal production and especially hazelnut production in the Black Sea region (Yılmaz 2023). The SPI₉ index shows a widening of the area of influence of drought severity in March, while high drought values are observed in April (between -4.96 and -0.83) in the eastern and inland areas (Fig. 20). It was determined that drought severity in 2012 increased in these continental climate areas due to the effect of decreasing precipitation due to topographic conditions in SPI₉ April.

Fig. 20

Graphs of the years (January-April) in which the highest droughts were determined according to each month of the 9-month analysis in SPI, SPEI and RDI methods

In RDI₉ values, drought severity is high in January (-5.86 to -0.99), February (-5.87 to -1.04) and March (-5.77 to -0.91) in the coastal area from the western Black Sea to the south (Fig. 20). The fact that the highest drought in RDI₉ monthly indices between January-April 2007 was observed in the western coasts is related to the increased effect of the North Atlantic oscillation in this year (Türkes 2020b). As seen in Fig. 20, the effect of severe drought in 2014 is remarkable for the entire study area. In 2014, the increasing temperature-evaporation trend combined with decreasing precipitation was effective in the severity of drought between January and April, especially in SPEI₉ months. Drought intensity extending from north to south to coastal areas is observed in SPEI₉ January (between -3.15 and -1.04). Apart from these assessments, drought intensity is high from inland to eastward in SPEI₉ February (-2.92 to -0.64), March (-3.08 to -0.78) and April (-3.00 to -0.79) and RDI₉ April (-5.52 to -0.74) (Fig. 20).

The highest wet conditions in January, February and March were in the southern part of the Marmara Sea in 2015.In RDI₉ method, the highest wet conditions were determined in January (4.48), February (4.30) and March (3.88).In SPI₉ method, the highest wet conditions were determined in January (3.98) and February (3.61) in the southern part of the Marmara Sea in 2015.The highest wet conditions in February SPEI₉ (2.63) were in Central Anatolia in 2019. In March, SPI₉ (3.30) in 2013 in the southeast and SPEI₉ (2.58) in 2022 in the Black Sea coast. In 2013, the highest wet conditions for April, RDI₉ (4.59) and SPI₉ (4.02) were in the northwest of the study area.In SPEI₉ (2.73) method, wet conditions show an increase in Central Anatolia in 2009.

The years and areas with the highest drought values in May, June, July and August are shown in Fig. 21. In SPI₉, the drought intensity increases from the western coastal areas towards the central Anatolia in May (-5.97 to -0.71), June (-5.80 to -0.82) and August (-5.84 to -1.06), and in July it is observed only in the northeastern coastal area (Fig. 21). In RDI₉ maps, unlike SPI₉, drought intensity increases towards the east of the study area in May (-5.51 to -0.90), June (-6.09 to -0.92) and July (-5.87 to -0.97) (Fig. 21). SPEI₉ drought intensity spread over the entire study area, centered in Central Anatolia in May (-3.03 to -0.83) and was high in the northern and western parts in June (-3.75 to -0.85) and July (-2.84 to -0.86) (Fig. 21). SPEI₉ August (-3.17 to -1.02) is dominated by normal drought values. The 2007 drought intensity, which is effective in SPI₉ May, June, August and SPEI₉ July, is observed in the western part (Fig. 21). Consistent with the other indices, the effect of the North Atlantic Oscillation led to an increase in summer drought. In 2014, the severe drought effect was determined to be severe in the eastern part between RDI₉ May–July. The effect of 2021, another severe drought year, is prominent throughout the study area in SPEI₉ May with increasing temperature and evaporation (Fig. 21).

Fig. 21

Graphs of the years (May–August) in which the highest droughts were determined for each month of the 9-month analysis in SPI, SPEI and RDI methods

May wet conditions in RDI₉ (5.50) and SPI₉ (4.84) indices are the highest in 2013 in the northwestern part. An increase is observed in SPEI₉ (2.77) values in the West Black Sea coastal region in 2022. Similarly, in June, it was determined that the increase in wet conditions continued in the northwest station in RDI₉ (5.50) and SPI₉ (4.99) indices in 2013. SPEI₉ (2,92) has the highest wet conditions in the northern part of the Marmara Sea in 1998. It is observed that the wet conditions increased in 2013 in the northwest of the study area in July. RDI₉ (5,60), SPI₉ (5,05) and SPEI₉ (2,95) wet conditions increased in 2022. In August, RDI₉ (5.72), SPI₉ (4.92) in the northwest in 2022 and SPEI₉ (3.56) in the south of the Marmara Sea with the highest wet conditions.

The years and areas with the highest drought values in September, October, November and December are shown in Fig. 22. The SPI₉ is higher in the east and south inland in November, extending from north to south in the east of the study area in September (-5.54 to -1.30). Normal dry conditions are present in October (-6.00 to -0.87) and November (-4.46 to -0.94) (Fig. 22). SPI₉ September–November is characterized by the impact of the 2014 severe drought, while November and December are characterized by the 2012 severe drought in the eastern part (Fig. 22). RDI₉ in September (-5.77 to -1.31), the effect of the increasing drought severity in the western Black Sea coastal area continues until December (-5.83 to -1.12). In addition, drought intensity increases in the coastal area from the Western Black Sea to the south in December (Fig. 22). In the RDI₉ index, the severe drought that was effective in the northern coastal region from August to December 2006 caused an increase in the drought effect of the lack of precipitation, especially in winter months (Partal and Yavuz 2020).

Fig. 22

Graphs of the years (September-December) in which the highest droughts were determined by each month of the 9-month analysis in SPI, SPEI and RDI methods

Normal dry conditions in the study area give way to high drought values in the north in September SPEI₉ (-3.82 to -0.82) (Fig. 22). In October (-3.34 to -0.71), drought intensity is also high in the western part, while in November (-3.12 to -0.70) and December (-3.82 to -1.07), drought intensity increases from the southeast (Fig. 22). In 2021, it is observed that drought severity increased in November–December in the SPEI₉ index with low precipitation at normal precipitation values and increasing temperature evaporation values in the southern part (Fig. 22).

The highest wet conditions with the values determined in RDI₉ (5.57) and SPI₉ (4.78) indices in September are in the northwestern part of the study area in 2013. With the value of SPEI₉ (3.20), wet conditions increase in the south of Marmara in 2022. In the northwest of the study area, there is high wet in RDI₉ (5.50) and SPI₉ (4.74) methods in October 2013. The highest wet with SPEI9 (2.94) value is in the south of Marmara in 2022. In November 2022, RDI₉ (5.34) in the western Black Sea coastal area and SPEI₉ (2.95) in the eastern part of the study area have increased wet conditions. For SPI₉ (4.66), high wet conditions were determined in the northwestern part of the study area in 2013. In December 2022, RDI₉ (5,12) in the western Black Sea coastal area and SPEI₉ (2,73) in the north of the study area, and SPI₉ (4,30) in the northwestern part of the study area in 2013, wet conditions increase.

SPI, SPEI and RDI methods 12-month evaluation analyses

In the 12-month analyses where long-term drought is expressed, the highest drought values were determined in 2006, 2007, 2012, 2013 and 2014 in the SPI method and in 2006, 2007, 2013 and 2014 in the RDI method. In SPEI management, the highest drought values are found in 1994, 2008, 2012, 2014 and 2015. By mapping the months of these years, the spatial distribution of drought is visually presented.

The years and areas with the highest drought values in January, February, March and April are shown in Fig. 23. It is seen that the year 2014, which was a severe drought year for Türkiye, caused high drought severity in March and April across the study area in SPI₁₂/RDI₁₂ and SPEI₁₂ indices (Fig. 23). In the SPI₁₂ maps, the area of influence of drought severity is large in January (-5.15 to -0.98), March (-4.42 to -0.88) and April (-5.20 to -0.85) (Fig. 23). In February (-5.40 to -0.86), drought intensity increased in the northeastern coastal area. RDI₁₂ showed high drought intensity in January (-6.16 to -0.94) and February (-6.03 to -1.10) in the west coast and western Black Sea, while in March (-6.40 to -0.97), increased drought values were determined in the northeast and a few stations in the south (Fig. 23). Similar to SPI₁₂, drought intensity was high throughout the study area in RDI₁₂. SPEI₁₂ increased in January (-3.21 to -0.71), March (-2.96 to -0.87) and April (-3.06 to -0.99), extending southward along the eastern Black Sea coast, and increased in February (-3.35 to -0.81) in the western part (Fig. 23). In January SPI₁₂ and SPEI₁₂, the severe drought values of 2014 are also noteworthy. The increased agricultural loss due to the winter drought in this period and the frost in March support the results of this study (Taskın et al. 2022). Since 2008, the severe drought in 2013 and 2014, which was affected by increasing temperature and decreasing precipitation, became evident in March RDI₁₂ and February SPI₁₂ (Fig. 23). Another severe dry period and agricultural losses in 2007–2008 were observed in January–February RDI₁₂ and February SPI₁₂ in the western part.

Fig. 23

Graphs of the years (January-April) in which the highest droughts were determined for each month of the 12-month analysis in SPI, SPEI and RDI methods

The highest January wet conditions for RDI₁₂ (5.18) and SPI₁₂ (4.52) were in 2014 in the northwest of the study area. For SPEI₁₂ (2.66), wet increased in the Black Sea coastal area in 2010, while the highest wet conditions were determined for RDI₁₂ (4.92), SPI₁₂ (4.25) and SPIE₁₂ (2.62) in 2014 in the northwest of the study area in February. In March, similar to the other months, RDI₁₂ (4.86), SPI₁₂ (4.10) and SPEI₁₂ (2.66) had the highest wet conditions in the northwest of the study area in 2014. In April, similar to the other months in 2013, RDI12 (4.13) in the northwest has the highest wet conditions. SPI₁₂ (3.72) in Central Anatolia in 2015 and SPEI₁₂ (3.15) in 2019 have the highest wet conditions in the southeast.

The years and areas with the highest drought values in May, June, July and August are shown in Fig. 24. SPI₁₂ shows high drought severity on the west coast in May (-5.99 to -0.58), July (-5.76 to -0.77) and August (-5.94 to -0.76), while dry conditions prevail throughout the study area in July (Fig. 24). RDI₁₂ monthly assessments show high drought intensity in the east of the study area from May (-5.56 to -0.82) to August (-6.07 to -1.11). In SPEI₁₂ May (-2.94 to -0.72), June (-2.93 to -0.78) and July (-2.88 to -0.92), drought intensity was high in the east of the study area and increased in August (-3.24 to -0.86) in the northern part (Fig. 24). While the severe drought values of the 2006–2007 period were high in the western part of the study area in SPI₁₂ May–August, it is seen that the effect of drought severity in 2014 increased in the continental climate zone in the eastern part of the study area in RDI₁₂ and SPEI₁₂ indices (Fig. 24).

Fig. 24

Graphs of the years (May–August) in which the highest droughts were determined for each month of the 12-month analysis in SPI, SPEI and RDI methods

The highest wet conditions for May belong to RDI₁₂ (5.18), SPI₁₂ (4.57), and SPEI₁₂ (2.70) indices in 2013. Similar to May, wet increases in the northwest of the study area in RDI₁₂ (5.38) and SPI₁₂ (4.64) methods in 2013. The Eastern Black Sea coastal area has the highest wet conditions determined by the SPEI₁₂ (2.81) method in 2022. Similar to June, SPEI₁₂ (2.94) value is high in the Eastern Black Sea coastal area in 2022. RDI₁₂ (5.36) and SPI₁₂ (4.65) methods show high wet conditions for the northwest of the study area in 2013. In SPEI₁₂ (3.00) method, wet conditions increase in the northern part of the Marmara Sea in 2022.

The years and areas with the highest drought values in September, October, November and December are shown in Fig. 25. From September to December, there is an increasing drought severity in the eastern and coastal areas in SPI₁₂ (-5.76 to -0.64) and in the western Black Sea coastal area in RDI₁₂ (-5.84 to -1.09) (Fig. 25). In SPEI₁₂ maps, drought values are high in the northern part in September (-3.00 to -0.45), while normal drought values dominate in October (-2.69 to -0.49) and November (-3.07 to -0.66). In SPEI₁₂ December (-3.02 to -0.56), drought intensity was high in the west coast and southeast stations (Fig. 25). Drought severity increases in the east of the study area during the high drought periods of 2012–2013. This can be explained by both topographic conditions and decreasing precipitation due to the effect of increasing temperature in continental regions. This drought effect between September-December in SPI₁₂ was effective in November in SPEI₁₂ (Fig. 25). In 2006, the drought effect intensified in the northern part of the region between September and December in RDI₁₂ monthly. The effect of decreasing precipitation in this region causes the increase in drought severity in 2006 (Partal and Yavuz 2020). The increasing atmospheric temperature and heavy precipitation in the Black Sea region support the increasing drought severity (Donava 2017).

Fig. 25

Graphs of the years (September-December) in which the highest droughts were determined for each month of the 12-month analysis in SPI, SPEI and RDI methods

In September, RDI₁₂ (5.16) and SPI₁₂ (4.45) show the highest values in the northwestern part, while SPEI₁₂ (2.85) shows extremely humid values in the western Black Sea coastal region in 2022. October RDI₁₂ (5.37) in 2022 shows an increase in extremely humid in the Western Black Sea coastal region and SPEI₁₂ (2.85) in Central Anatolia. SPI₁₂ (4.62) values in 2013, indicating extremely humid conditions, are in the northwest of the study area. In November and December in 2022, RDI₁₂ shows an increase in extremely humid in the western Black Sea coastal area and SPI₁₂ in the northwest of the study area. For SPEI₁₂ in 2022, in November (3.41) Mediterranean coastal area and in December (2.91), extremely humid conditions prevail in the south of Central Anatolia.

Conclusion and assessment

This study includes a comprehensive drought analysis for Türkiye covering the period 1991–2022. In this analysis, SPI, SPEI and RDI methods are used to present the spatial and temporal variation of drought, the impact of geographical factors on this variation and its agricultural consequences in detail. In the study area, meteorological drought events are experienced more frequently than in the past years due to the long-term below-normal precipitation and increasing temperature. In addition, Türkiye's topographic conditions and its location in the Mediterranean basin, which is a sensitive area to climate change, are among the most important factors affecting the temporal and spatial variation of the rule. The results of the study are given below:

• In the SPI, SPEI and RDI methods applied by using the data of 219 meteorological stations between 1991 and 2022 covering the study period, it was observed that the slightly arid class range (0.99 to -0.99) had the highest recurrence rate. One reason for Türkiye's tendency to shift to more arid conditions in the last 20 years is the increasing trend in annual average temperatures in the Mediterranean basin where Türkiye is located, increasing the frequency and duration of heat waves and the areas affected by them (Güler and Erlat 2023).

• Among the three indices SPI, SPEI and RDI, it is observed that drought severity generally increases in the whole study area. In short-term drought analyses, dry periods are more recurring, whereas in long-term analyses, dry periods are less recurring. One of the main reasons for this situation is the variability of precipitation periods from year to year and the increase in heavy rainfall due to increasing temperatures with global climate change. With the increase in temperature, atmospheric evaporation rates cause more intense and intense precipitation. Although this situation shows the absence of precipitation in short-term drought analyzes, it has a misleading effect. The most important feature of climate change is that wet periods are wetter and dry periods are more severe droughts. Especially in Türkiye, the increase in surface runoff with the effect of heavy precipitation from 2010–2021 increases the severity of drought by reducing the amount of water stored in the soil.

• In the 1-month indices, the seasonal variability of precipitation caused the areas where drought severity increased to differ according to the indices. Especially in the winter months, SPI, SPEI and RDI reflect the drought severity in the western coasts and in the whole study area, while SPI and RDI show the drought severity in the northern and eastern parts best in other months. In the SPEI index, drought variation is high in the eastern parts.

• In the 3-month indices, the SPI index draws attention to increased drought severity in the eastern and western parts, while the RDI index draws attention to increased drought severity in the northern parts. In SPEI indices, periodically changing drought severity is effective in the northern and western coastal and continental regions.

• In the 6-month indices, drought severity increases in the eastern parts for the SPI index. RDI 6-month indices show high drought severity in the northern coastal areas, while SPEI index type shows high drought severity in the western coasts and eastern parts.

• Although long-term severe drought varies periodically in the 9- and 12-month indices, it is observed that the stations in the northern and western coasts and in the eastern part have similarities in examined indexes.

• In the long-term drought indices, fewer but long-lasting dry periods are noteworthy. This situation is the result of the variability of the mid-latitude wandering depressions, which are among the atmospheric conditions that control the climate of Türkiye, according to the North Atlantic oscillation, and the year-to-year and seasonal changes in precipitation (Türkes 2020b). In periods when the North Atlantic oscillation is positive, especially the winter season is colder and less rainy. The relationship between the severe dry periods in Türkiye and the periods when the North Atlantic oscillation is positive, which is also included in the results of the study, has a wide place in the literature (Kömüscü, 2001; Tatlı and Türkes 2008; Türkes 2020b; Topcu 2022). Long-term droughts with high severity of heat waves and decreased precipitation have an impact on agricultural production (Sen et al. 2012; Ozelkan et al. 2016; Taskın et al. 2022).

This study can form a basis for building resilient and sustainable societies against climate change and drought. The future goal of this study, which was carried out in Türkiye, one of the regions where climate change is felt the most, is to reveal the effects of different types of drought (meteorological, hydrological, agricultural and socioeconomic) with quantitative and qualitative data obtained from field studies.