Climate change, one of the major environmental challenges facing mankind, has caused intermittent droughts in many regions resulting in reduced water resources. This study investigated the impact of climate change on the characteristics (occurrence, duration, and severity) of meteorological drought across Ankara, Turkey. To this end, the observed monthly rainfall series from five meteorology stations scattered across Ankara Province as well as dynamically downscaled outputs of three global climate models that run under RCP 4.5 and RCP 8.5 scenarios was used to attain the well-known SPI series during the reference period of 1986–2018 and the future period of 2018–2050, respectively. Analyzing drought features in two time periods generally indicated the higher probability of occurrence of drought in the future period. The results showed that the duration of mild droughts may increase, and extreme droughts will occur with longer durations and larger severities. Moreover, joint return period analysis through different copula functions revealed that the return period of mild droughts will remain the same in the near future, while it declines by 12% over extreme droughts in the near future.

Drought is a climatic disaster that heavily affects all the aspects of the ecosystem and human life. In general, drought is defined as a deficit in the amount of water, a deviation from the normal condition, while the meteorological droughts mostly depend on the decrease in the amount of the precipitation received through a long period (e.g., a season, a year, and more). Meteorological droughts have many factors that play a significant role in its occurrences such as characteristics of rainfall (i.e., duration and intensity, rainy days distribution in the growing season, onset and termination, severity and its temporal variability), temperature, low relative humidity, and hisgh winds (Mishra and Singh

The drought monitoring system is one of the fundamental essentials for drought management plans (Hao et al.

In recent years, many studies found convincing evidence that climate change impacts water resources, environment, health, and safety significantly (Danandeh Mehr and Kahya

Given that drought is a complex phenomenon in which its physical characteristics (e.g., severity and duration) are interdependent and affecting each other, multivariate analysis can provide a better description of the probabilistic characterization of droughts (Tosunoğlu and Onof

Many studies have investigated the impact of climate change on climatic events including droughts and runoff extremes (Cheng et al.

In drought frequency analysis, important drought characteristics (e.g., duration, and average severity) are generally derived from hydro-meteorological datasets using a drought index. Given that drought characteristics are typically interrelated, it is more convenient to use bivariate or multivariate models in their frequency analysis. On the other hand, given that the drought characteristics usually follow different types of univariate distributions, it is not possible to use traditional bivariate distributions in their frequency analysis. In this study, copula functions are utilized to overcome such difficulties and provide reliable estimates of multivariate drought frequencies (Hangshing and Dabral

Based on Sklar’s theorem, copula functions link the univariate distribution functions to form multivariate distribution functions. The advantage of using copula functions in forming multivariate distribution functions is that they can model the joint dependence among different variables without being dependent on their marginal distributions (Afshar and Yilmaz

After the marginal distributions of each drought characteristics are defined and their CDF’s are computed, the copula functions can be applied for joint modeling of the drought characteristics. There are multiple different copula functions available for modeling the joint behavior of different dependent univariate variables, while in this study, bivariate Archimedean (i.e., Clayton, Frank, Gumbel Hougaard, and Joe) and elliptical copulas (normal, and _{t} and C(u, v)_{e} are the theoretical and empirical copula functions, respectively, that are used for modeling the joint dependence among characteristics of n drought events. In this study, the calculation of empirical copula and validation of different theoretical copula functions are done using copula package (Yan

Equations of copula functions. Here, u and ν are two dependent univariate variables, df is the degree of freedom, _{df} is the Student

Function (family) | Joint CDF, C(u, v) | Parameter range |
---|---|---|

Clayton (Archimedean) | [max(0, u | 0 ≤ θ |

Frank (Archimedean) | θ ≠ 0 | |

Gumbel Hougaard (Archimedean) | 1 ≤ θ | |

Joe (Archimedean) | 1 ≤ θ | |

Normal (Elliptical) | −1 ≤ ρ ≤ 1 | |

Student | −1 ≤ ρ ≤ 1 df ≥ 1 |

The return period of certain drought event is associated with a specified exceedance probability. According to Shiau and Shen (_{U ≥ u} is the return period of the drought event with the characteristic of U greater than or equal to u. However, since a drought event is considered mostly a bivariate event that is characterized mostly by drought duration and severity, estimation of the joint return period of these characteristics is more helpful for the assessment and management of droughts. Therefore, in this study, the estimated joint return periods has been done by using a methodology proposed by Shiau (_{UV}(u, v); the joint probabilities of drought events with characteristics of U and V for three cases of and/or conditional can be calculated via Eqs. (12–15).

The described equations above are the main drivers of the joint probability and hence joint return period analysis. By using such formulations, the joint probability of any drought event (e.g., a drought event with duration more than 7 months and average severity of 1.25) can be calculated. Once the joint probability is being calculated, by inserting joint probability information to the equations of 7 to 8, the return period of that event can be calculated with three different scenarios of and/or conditional forms.

The SPI is one of the most commonly used drought indices which is developed based on the normalization of precipitation probabilities. Although SPI is generally calculated by using monthly precipitation data (for the different number of timescales), its values can be produced with daily or weekly precipitation data as well (WMO,

Regardless of the time interval at which precipitation values are presented, the SPI calculation method is the same for all time intervals. Based on the definition of (McKee et al.

Drought events, in general, have multiple characteristics including drought duration (the length of the dry period), drought severity (summation of SPI values during the dry period), drought average severity (average of SPI values during the dry period), and drought intensity (minimum SPI value during the dry period). Among drought characteristics, the drought severity and intensity are highly associated with drought duration and average severity (the severity of drought can be determined by multiplication of drought duration and average severity, and drought intensity is highly correlated with drought average severity; Afshar et al.

The overall procedure of drought analysis conducted in this study. D#, duration of draught event; CDF, cumulative distribution function

Given that the frequency and return period analysis require long time series of drought characteristics, in this study, the drought characteristics of events occurred over different stations are bound to generate longer drought characteristic time series and hence more robust areal average analysis. Moreover, to visualize the univariate and joint cumulative probabilities, the bound drought characteristics are fitted to different univariate probability distributions to generate synthetically continuous data of different drought characteristics. Among different univariate distribution functions (i.e., gamma, log-normal, logistic, normal, and Weibull), the best distribution is selected based on the chi-squared statistics between the cumulative distribution functions of theoretical and empirical cumulative distributions function values for each join dependence and scenario separately.

Drought events and their spatiotemporal variations are the most crucial problem to tackle over the semi and arid climatic regions like Central Anatolia (Duzenli et al.

The analysis of the reference meteorological drought characteristics is performed using monthly precipitation records obtained for the period of 1984 and 2018 over the five automated meteorological stations operated by the General Directorate of Meteorology (MGM) located in Ankara Province (Fig.

Location of the study area (Ankara Province) and the meteorological stations located in it

The drought analyses related to the future projections are performed using projected datasets between the years 2018 and 2050. The climate projections for this period are simulated by MGM using three different available GCMs (i.e., HadGEM2-ES, MPI-ESM-MR, GFDL-ESM2M), a single regional climate model (RCM; Reg4), and two different RCP scenarios (RCP 4.5, and RCP 8.5; Table

Highlight information of the GCMs used in this study

Driving GCM | Institution | Concentration scenario | GCM-RCM acronym | Spatial resolution (km) | |
---|---|---|---|---|---|

GCM | GCM-RCM | ||||

HadGEM2-ES | Hadley Center | RCP4.5–RCP8.5 | HadGEM - Reg4 | 112.5 | 20 |

MPI-ESM-MR | Max Planck | RCP4.5–RCP8.5 | MPI - Reg4 | 210 | 20 |

GFDL-ESM2M | Geophysical Fluid Dynamics Laboratory | RCP4.5–RCP8.5 | GFDL - Reg4 | 220 | 20 |

HadGEM2 (Hadley Center Global Environment Model version 2) is a second-generation global model developed by Hadley Center, a research organization affiliated with the UK Meteorological Service. There are various HadGEM2 models based on available atmospheric, hydrological, and oceanographic cycles (i.e., HadGEM2-A, HadGEM2-O, HadGEM2-AO, HadGEM2-CC, HadGEM2-CCS, HadGEM2-ES). These models have the same physical infrastructure containing different levels of detail. These models have an integrated atmospheric-ocean configuration with a vertical atmospheric expansion that provides optional better stratosphere modeling, and a surface system configuration with dynamic vegetation, ocean biology, and atmospheric chemistry (Collins et al.

The MPI-ESM (Max-Planck-Institute Earth System Model) model, developed by the Max Planck Institute of Germany, is an integrated circulation model consisting of several sub-modules. The MPI-ESM model integrates multiple models including atmospheric ECHAM6 model (Stevens et al.

The Earth System Model (GFDL-ESM2) is a global-coupled climate-carbon model which is developed at the Geophysical Fluid Dynamics Laboratory (GFDL) of the National Oceanic and Atmospheric Administration (NOAA). There are two different versions of the GFDL model (i.e., ESM2M and ESM2G). Both ESM2m and ESM25 versions use same ocean ecology and biogeochemistry but different ocean components with this difference that in the ESM2M version, the vertical coordinate is based on depth, while in the ESM2G version, the vertical coordinate is based on density (Dunne et al.

In the climatic scenarios, the different RCPs are generally distinguished from each other based on the annual changes of global greenhouse gas emissions, the socio-economic and technological development assumptions, the impact of climate-affecting gas emissions, and atmospheric particle changes. RCP scenarios used in this study (i.e., RCP 8.5 and 4.5) are defined by their total radiative forcing pathway and level by 2100. Relatively, more pessimistic RCP 8.5 scenario assumes that there will be no policy change about emission reduction in the future, while increased greenhouse gas emissions will increase the concentration of greenhouse gasses in the atmosphere. On the other hand, the relatively more optimistic RCP 4.5 scenario foresees that radiative forcing will stabilize soon after 2100 with the help of global emission reduction policies (van Vuuren et al.

To illustrate the impact of climate change on drought characteristics, the number of drought events at each meteorological station and overall average over Ankara Province are derived both for the reference and the projected periods using the approach described in Section

The number of drought events compared among the reference and future scenarios over stations located in Ankara Province. 4.5 = RCP 4.5; 8.5 = RCP 8.5; HG HadGEM

The accuracy statistics and the parameters of the best fitting functions for the estimation of the univariate CDF of drought characteristics are given in Table

The best univariate distribution and copula function that has been selected based on the comparison of chi-squared and RMSE values between theoretical and empirical cumulative probabilities.

Time period | Dataset | Univariate | Bivariate | |||
---|---|---|---|---|---|---|

Distribution | Parameter | Chi-squared | Copula (P) | RMSE | ||

Reference | Duration | Logistic | l = 17.372; s = 4.671 | 0.295 | Normal (0.47) | 0.031 |

Ave. Sev. | Logistic | l = 0.886; s = 0.114 | 0.373 | |||

HadGEM 4.5 | Duration | Lognormal | ml = 2.881; sl = 0.669 | 0.165 | Frank (1.93) | 0.036 |

Ave. Sev. | Logistic | l = 0.799; s = 0.14 | 0.195 | |||

HadGEM 8.5 | Duration | Logistic | l = 15.326; s = 3.156 | 0.076 | Frank (0.18) | 0.039 |

Ave. Sev. | Gamma | sh = 9.429; r = 10.642 | 0.056 | |||

MPI 4.5 | Duration | Gamma | sh = 3.385; r = 0.193 | 0.087 | Frank (2.95) | 0.024 |

Ave. Sev. | Gamma | sh = 8.878; r = 9.825 | 0.127 | |||

MPI 8.5 | Duration | Normal | mean = 15.935; sd = 7.23 | 0.338 | Frank (0.15) | 0.03 |

Ave. Sev. | Logistic | l = 0.918; s = 0.168 | 0.152 | |||

GFDL 4.5 | Duration | Gamma | sh = 5.633; rate = 0.288 | 0.184 | Frank (2.93) | 0.031 |

Ave. Sev. | Lognormal | ml = −0.115; sl = 0.271 | 0.065 | |||

GFDL 8.5 | Duration | Gamma | sh = 3.459; r = 0.217 | 0.074 | Normal (0.47) | 0.032 |

Ave. Sev. | Lognormal | ml = −0.152; sl = 0.292 | 0.123 |

CDF and PDF of different drought characteristics (duration and average severity) over Ankara Province generated by using reference and future scenario simulations

The joint behavior of drought characteristics for different occurrence probabilities simulated with different copula functions with respect to their performance in capturing the joint dependence among drought duration and average severity is given in Table

The comparison of the relationship between drought characteristics by considering the same joint probability for both reference and future projections over Ankara Province

Average drought duration and severity values corresponding to various univariate return periods (5, 11, 17, 24, and 30) and joint return periods reflecting various levels of drought events are given in Table

Comparison of the impact of climate change on univariate and joint return periods of droughts with different duration and average severity amounts. The reference refers to the 1986–2018 period, and RCP 4.5 and RCP8.5 are two climate projections up to 2050 (results are averaged across three GCMs and five stations within Ankara Province)

Univariate return period (year) | Drought characteristics according to reference dataset | Dataset | Joint return period (year) | |||
---|---|---|---|---|---|---|

Duration (month) | Average severity | And | Or | Conditional | ||

5 | 12 | 0.76 | Reference | 7 | 4 | 5 |

RCP 4.5 | 8 | 5 | 5 | |||

RCP 8.5 | 8 | 4 | 5 | |||

11 | 21 | 0.94 | Reference | 24 | 8 | 9 |

RCP 4.5 | 22 | 7 | 8 | |||

RCP 8.5 | 28 | 7 | 10 | |||

17 | 25 | 1.12 | Reference | 86 | 18 | 9 |

RCP 4.5 | 52 | 12 | 8 | |||

RCP 8.5 | 81 | 11 | 13 | |||

24 | 27 | 1.18 | Reference | 131 | 25 | 9 |

RCP 4.5 | 70 | 14 | 8 | |||

RCP 8.5 | 118 | 13 | 14 | |||

30 | 28 | 1.23 | Reference | 189 | 33 | 9 |

RCP 4.5 | 93 | 17 | 8 | |||

RCP 8.5 | 169 | 16 | 16 |

Moreover, the comparison of the return periods of drought events shows that the joint return period of mild droughts (a drought event with a return period of 5 years in case of univariate probability) does not vary much within the reference and future projection time periods (and, or, conditional columns under joint return period). In contrast, the joint return period of extreme drought (a drought with a return period of 30 years in case of univariate probability) decreases within future projection periods (e.g., joint return period decreases from 189 years to 93 and 169 years for RCP 4.5 and 8.5 scenarios, respectively, for and case) implying that extreme droughts will be more frequent within future projections while the frequency of mild droughts will remain the same.

Drought is one of the extreme events that has inverse effects on various sectors including agriculture and water resources. Investigation of the effects of climate change on drought is important for the planning and the management of water resources. In this study, the impact of climate change on univariate and bivariate drought characteristics (duration and average severity) is investigated using long-term station-based historical observations and climate projections (three GCMs and two emission scenarios) over Ankara Province.

The comparison of the drought characteristics shows that while over extreme droughts the average severity of drought events increases, the mild drought events will be expected to happen with longer duration and milder average severities (Fig.

Results also show that the return periods of events with low and high drought characteristic (both duration and severity) values decrease (i.e., such events will occur more frequently), while the return periods of events with medium drought duration and severity increase (i.e., such events will occur less frequently). Given that events with high drought duration or severity values are more significant from the drought planning and management perspective (i.e., such events yield more significant economic/hydrological/social results), these results imply that Ankara Province will need to prepare and adapt for worse drought conditions in the future than the past.

Overall, the joint return periods determined using copula formulations for three possible probability combinations (i.e., and, or, conditional) showed that the frequency of mild drought events obtained using different GCMs will be close to those experienced in the reference period, while the return period of extreme events (i.e., drought events with return period more than 30 years) will decrease by 30%. These results imply that the extreme droughts will occur more frequently in comparison with the reference time period. The results found here should be repeated over larger regions (e.g., country scale) to make a general conclusion and determine the impacts of climate changes on mild and extreme droughts.

We thank the Turkish State Meteorological Service for providing precipitation data of both reference and future scenarios, the ESA CCI Land Cover project for the global land cover classification map, and the Land Processes Distributed Active Archive Center (LP DAAC; “

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.