Decreasing precipitation in the Upper Euphrates Basin and the negative impact of climate change directly affect water resources and hydroelectricity generation in the basin. This basin, which contains the largest dams in terms of hydroelectricity generation potential, requires research studies to assess and characterize drought for risk prevention and mitigation applicable to water resources management. To better assess drought in the upper Euphrates Basin due to recent warming, FEHEM is developed, a hydro-economic optimization model of the integrated reservoir system of the Upper Euphrates Basin. Using a historical hydrological dataset, water management and hydroelectric operations are evaluated with a linear programming model at monthly time steps. This paper uses two different drought indices: (1) the standardized precipitation index, which is based on precipitation alone; and (2) the reconnaissance drought index, which takes into account both evaporation and precipitation. These indices were used to evaluate the impact of temporal drought characteristics in the Upper Euphrates Basin on the hydropower generation of 10 dams with a total installed capacity of over 3255 MW in the basin, based on 45 years of precipitation data from more than a hundred measuring stations in the basin.

  • Study participants included Turkey's main hydroelectricity-producing region, the Upper Euphrates Basin, and the effects of the region's 50 years of drought.

  • A direct correlation between the index curves and energy production values is found when examining the temporal fluctuations of the SPI–RDI index and the energy production graphs.

AKL

Asagi Kalekoy Dam

BGT

Bagistas Dam

BYH

Beyhan Dam

CALVIN

California Value Integrated Network

DAP

Eastern Anatolia Project

DMI

Turkish State Meteorological Service

DSI

General Directorate of State Hydraulic Works

EPDK

Turkiye Energy Market Regulatory Authority

EPIAS

Energy Markets Operation Incorporated Company

FEHEM

Upper Euphrates Basin Hydro-economic Model

GAP

Southeastern Anatolia Project

GCM

General circulation models

IPCC

Intergovernmental Panel on Climate Change

JDI

Joint deficit index

KBN

Keban Dam

KGI

Kigi Dam

M$

Million Dollar (Revenue)

OZL

Ozluce Dam

PMB

Pembelik Dam

RDI

Reconaissance drought index

SPI

Standardized precipitation index

SYR

Seyrantepe Dam

TEDAS

Turkiye Electricity Distribution Company

TTR

Tatar Dam

UZN

Uzuncayir Dam

Recent studies confirm that droughts are becoming more frequent around the world due to factors such as population growth, improved living standards, and increased agricultural and industrial activity (Xiao et al. 2024). According to the Fifth Assessment Report of the Intergovernmental Panel on Climate Change, the Earth has warmed over the past 100 years (IPCC 2013). Droughts have become a global trend, with Eurasia and Africa experiencing the most severe droughts (IPCC 2013; Zarch et al. 2015; Zhang et al. 2015).

Drought is characterized by low rainfall, high temperatures, and intense evaporation over an extended period (Zhao et al. 2020). The significant variation in water resources across space and time has often led to water shortages in various regions and at different times. Three types of drought are recognized: meteorological or climatic drought; hydrological drought; and agricultural or crop drought (Tallaksen & van Lanen 2004). Meteorological drought, defined as a lack of precipitation leading to drying relative to the average situation, is the root cause of various types of drought. Adequate precipitation is crucial to avoiding the ramifications of this phenomenon. The impacts of meteorological drought are widespread throughout the hydrological cycle. Extended periods of meteorological drought often result in hydrological drought or a reduced flow of streams, leading to a deficiency in water supply. Agricultural or crop drought arises from insufficient soil moisture, reducing the moisture supply available for crops (Ozturk et al. 2015; Lelieveld et al. 2016; Waha et al. 2017; Ajjur & Al-Ghamdi 2021; Paltán et al. 2021).

Regarding the existence of different types of drought, several drought identification and assessment indices have been introduced in the literature, including the following: the copula-based joint deficit index (Kao & Govindaraju 2010); the standardized precipitation index (SPI) (McKee et al. 1993); the standardized evapotranspiration index (Vicente-Serrano et al. 2018); the reconnaissance drought index (RDI) (Tsakiris& Vangelis 2005); the comprehensive meteorological drought index (Zhao et al. 2017); and the impact of drought on large reservoirs (Wu et al. 2017). Among them, SPI, RDI, and Palmer Drought Severity Index (PDSI) have been widely used for the assessment and monitoring of meteorological droughts in recent years (Li et al. 2017). Another study for the United States summarizes a multivariate and multi-index drought monitoring approach (Hao & AghaKouchak 2013).

Numerous studies have demonstrated that drought frequency will likely rise amidst extreme global climatic conditions, particularly in warm climates (Jahani & Mohammadi 2019; Tam et al. 2019). Future changes in meteorological drought were projected using three emission scenarios across 29 general circulation models, and it is anticipated that interannual changes will be more pronounced in the latter half of the 21st century (Zhang et al. 2019).

The FEHEM is a hydro-economic optimization model for the interconnected water supply system of the Upper Euphrates Basin. The FEHEM represents approximately 8.1% of Turkey's total drainage area. Using 45 years of historical hydrological data to represent hydrologic variability, the model determines the optimum hydropower generation decisions for the modeled reservoirs. With the developed model, it will be possible to prepare water management plans under different policies and future climate scenarios to determine adaptation strategies in advance (Aytac et al. 2023).

This article assesses the temporal patterns of drought occurrences in the Upper Euphrates Basin. The evaluation employs two different drought indices, namely the SPI and the RDI. The SPI, which is more common, solely relies on rainfall and quantifies the departure of precipitation from historically established norms for a certain period. The RDI considers evaporation, in addition to rainfall, as a vital measure for evaluating the magnitude of drought.

This study enhances quantitative comprehension of the integrated water distribution and economic system of the Upper Euphrates Basin, where Turkey's biggest installed capacity power dams are found. The study area was deliberately chosen to depict the water infrastructure of the region, which encompasses every surface reservoir, over 100 groundwater reservoirs, hundreds of transport connections, large-scale agricultural irrigation projects, such as GAP and DAP, as well as numerous urban demand areas.

The purpose of this paper is to forecast possible water demand shortages in the Upper Euphrates Basin water supply system caused by drought, examine them with management options, and offer engineering and management solutions. The developed indicators were used to determine the direct effects of drought on hydroelectric power generation. The resulting influence on annual energy revenues was computed using average unit energy prices.

Hydrology

The study area comprises 45 years of hydrological data ranging from 1971 to 2016, obtained from the primary tributaries of the Karasu River, Munzur River, Peri River, and Murat River, where Upper Euphrates Basin's largest dams are sited, as well as from the primary tributary of the Euphrates River that forms at their confluence. This 45-year dataset characterizes the hydrological conditions of the Upper Euphrates Basin. FEHEM's hydrological data consist of runoff, precipitation, temperature, evaporation, and groundwater inflows. The calculation of net evaporation rates and evaporation losses for reservoirs was also performed. The accuracy of the results obtained is directly influenced by the length and reliability of the data used in statistical studies and modeling. Therefore, precise hydrological and meteorological data, physical characteristics (including length, size, slope, cultivation, and soil structure), operating rules, accumulation structures, characteristics of the basin, and past operating results were meticulously obtained from relevant institutions and organizations. Figure 1 illustrates the map of flow observation stations situated in and around the study area. The DMI (Turkish State Meteorological Service) and DSI (General Directorate of State Hydraulic Works) operate these stations, where they typically conduct measurements of precipitation, flow, and temperature. Some stations also record observations of snow depths, humidity, and wind.
Figure 1

Observation station map of the study area (Aytac et al. 2024).

Figure 1

Observation station map of the study area (Aytac et al. 2024).

Close modal
In the study, net evaporation was calculated separately for each reservoir (Figure 2). Evaporation values taken from DMI were used to represent the relevant reservoirs.
Figure 2

Euphrates basin long-years monthly average evaporation (mm) (Aytac 2023).

Figure 2

Euphrates basin long-years monthly average evaporation (mm) (Aytac 2023).

Close modal
The hydrological, water demand and other water allocation data used in this study were obtained from the respective government agencies. However, considering the measurement time intervals, there may be erroneous, contradictory, or missing data, which limits the quality of the datasets. To address this, numerous correlation and regression analyses were conducted in the basin to complete the missing and outlier data (Figure 3). Significant correlation equations and coefficients were discovered through operations performed during the period shared by both stations. Missing or erroneous values were completed using the correlation equations found, and missing or contradictory flow values on the same branch were determined by adjusting them with the area ratio instead of the relevant stream gauging station. Figure 3 contains an example of this process.
Figure 3

Correlation scatter plots (Aytac 2023).

Figure 3

Correlation scatter plots (Aytac 2023).

Close modal

Rainfall

The Upper Euphrates Basin, including its sub-basins (covering a total of 63,118.0 km2), spans the Murat River Basin (25,779 km2), Karasu Basin (20,687.6 km2), Munzur Basin (3,940 km2), and Peri Basin (5,770 km2). The study area's elevation ranges from 835 to 4,050 m. Meteorological stations within the study area provide precipitation and snow data, with approximately 100 stations measuring these variables throughout the basin. All stations mentioned are situated within the basin drainage area (Figure 4). The average monthly precipitation for the relevant stations during the 1971–2016 observation interval is 596.73 mm (Figure 5).
Figure 4

Study area gauging stations (Aytac 2023).

Figure 4

Study area gauging stations (Aytac 2023).

Close modal
Figure 5

Average total precipitation of the stations representing the basin (Aytac 2023).

Figure 5

Average total precipitation of the stations representing the basin (Aytac 2023).

Close modal

Model development

Hydrological and economic data are used in hydro-economic optimization models. Hydrological data consist of reservoir and tributary flows and evaporation rates. Reservoir storage capacities and hydroelectric turbine capacities also constitute model inputs. Economic data consist of unit operating costs and unit energy prices. The main model outputs are reservoir water storage and releases, hydroelectric energy production, and revenue.

The FEHEM model, a deterministic hydro-economic optimization model of the Upper Euphrates Basin, is adopted from the California Value Integrated Network (CALVIN) model. CALVIN is a hydro-economic optimization model for California's interconnected water supply system (Draper et al. 2003; Dogan et al. 2018). FEHEM uses a network-flow structure and is modeled using Pyomo, a Python-based high-level optimization modeling language (Hart et al. 2017). In network-flow problems, the physical system is represented by a set or matrix of nodes (N) and a set of links (A). In hydro-economic optimization models, where the objective is to minimize operating costs and maximize hydroelectric generation revenue, this process can be expressed mathematically by the following objective function:
(1)
In Equation (1), z represents the cumulative cost. For each link, index i represents the start node, and index j represents the end node. In the piecewise linear programming technique, k represents each linear segment. The k part results from the linearization of nonlinear cost functions and c represents the linear unit cost. In this equation, the independent variable is X and represents the flow from node i to node j. All functions used in the model must be convex in minimization problems to guarantee a global optimum. For hydroelectric modeling, the inverse of the benefit curve (penalty) is taken in order to maximize hydroelectric generation:
(2)
(3)
(4)

Equation (1) is subject to three constraints. The first limiter represents the lower flow limit (l) for each link (Equation (2)). This lower limit can also be used to represent minimum flow requirements. The second limiting function (Equation (3)) represents the upper limit (u) on a given link. This function can also be used to represent the capacity of reservoirs, canals, and turbines. The last limiter (Equation (4)) represents the mass balance. For each connection, the incoming flow must be equal to the outgoing flow. In this function, a is used to represent loss factors such as evaporation. All parameters (c, l, u, and a) are predetermined and fixed.

The objective of the FEHEM model is to minimize costs and maximize hydroelectric generation. All equations expressed mathematically in Equations (1)–(4) are defined in the format Pyomo uses. The Pyomo solves the optimization problem through predefined solvers (such as GNU Linear Programming Kit (GLPK)) and the results are organized and analyzed as time series through postprocessors. The primary results are reservoir storage and turbine release time series. Post-processed results include hydroelectric generation and revenue.

The general network representation, which includes all nodes and reservoirs used in the model, is shown in Figure 6.
Figure 6

A general network diagram of nodes and reservoirs used in the model.

Figure 6

A general network diagram of nodes and reservoirs used in the model.

Close modal

For the applicable dams, multiple datasets were gathered from different institutions, with a focus on DSI as the primary source of hydroelectric data. Table 1 presents a selection of these data.

Table 1
Dam IDInstalled power (MW)Turbine flow (m3/s)Max. water level volume (hm3)Min. water level volume (hm3)Lake area (km2)Max. water level (m)Min. operating level (m)Net water drop (m)Tailwater elevation (feet)
BGT 140.63 351.57 250.00 114.75 15.58 915.00 892.53 50.00 2,828.00 
UZN 81.99 157.67 307.93 123.30 13.43 903.00 885.00 65.00 2,772.31 
KGI 138.00 104.54 507.55 153.31 8.40 1,365.00 1,307.00 165.00 4,019.03 
OZL 170.00 147.56 1,081.5 391.30 25.80 1,141.00 1,105.00 144.00 3,346.46 
PMB 127.34 212.23 358.38 135.63 15.63 1,026.40 1,005.00 75.00 3,129.92 
SYR 56.84 197.36 24.01 6.38 1.50 964.40 945.00 36.00 3,018.37 
TTR 128.22 228.96 299.57 100.00 13.07 915.00 909.50 70.00 2,782.15 
BYH 582.10 868.70 590.98 190.46 19.15 990.00 970.00 76.00 2,963.91 
AKL 500.00 628.73 596.06 233.03 15.90 1,105.00 1,080.00 87.55 3,320.21 
KBN 1,330.00 1,146.55 31,001.60 17,074.20 687.00 845.00 820.00 145.00 2,272.00 
Dam IDInstalled power (MW)Turbine flow (m3/s)Max. water level volume (hm3)Min. water level volume (hm3)Lake area (km2)Max. water level (m)Min. operating level (m)Net water drop (m)Tailwater elevation (feet)
BGT 140.63 351.57 250.00 114.75 15.58 915.00 892.53 50.00 2,828.00 
UZN 81.99 157.67 307.93 123.30 13.43 903.00 885.00 65.00 2,772.31 
KGI 138.00 104.54 507.55 153.31 8.40 1,365.00 1,307.00 165.00 4,019.03 
OZL 170.00 147.56 1,081.5 391.30 25.80 1,141.00 1,105.00 144.00 3,346.46 
PMB 127.34 212.23 358.38 135.63 15.63 1,026.40 1,005.00 75.00 3,129.92 
SYR 56.84 197.36 24.01 6.38 1.50 964.40 945.00 36.00 3,018.37 
TTR 128.22 228.96 299.57 100.00 13.07 915.00 909.50 70.00 2,782.15 
BYH 582.10 868.70 590.98 190.46 19.15 990.00 970.00 76.00 2,963.91 
AKL 500.00 628.73 596.06 233.03 15.90 1,105.00 1,080.00 87.55 3,320.21 
KBN 1,330.00 1,146.55 31,001.60 17,074.20 687.00 845.00 820.00 145.00 2,272.00 

Energy prices represent the economic value of producing energy units in the energy market. The Turkish energy market includes several participants, such as a system operator (EPDK – Turkiye Energy Market Regulatory Authority), scheduling coordinators, an energy exchange (EPIAS – Energy Markets Operation Incorporated Company), public utility distribution firms (TEDAS – Turkiye Electricity Distribution Company), retail companies, and customers. The EPIAS energy exchange establishes prices for each day and hour via automatic auction, while distribution and retail companies supply electricity to the interconnected system at the most economical cost possible to meet demand. Monthly average wholesale energy prices are sourced from both EPIAS and TEDAS (Figure 7).
Figure 7

Long-term monthly average electricity prices (cent$/kWh) (Aytac 2023).

Figure 7

Long-term monthly average electricity prices (cent$/kWh) (Aytac 2023).

Close modal
Hydropower generation necessitates the inclusion of two fundamental inputs in addition to plant characteristics: reservoir flows and energy prices. Equation (5) illustrates the total power (Watts) that can be produced as a function of water density, under the conditions mentioned earlier. Density (ρ) is measured in kg/m3, gravitational acceleration (g) in m/s2, plant efficiency (η) is a constant, water head (H) is measured in meters, and flow through turbines (Q) is measured in m3/s. The installed power in Watts is the result of the product of these values:
(5)
Multiplying power by time, t (hours), results in energy production, measured in kWh, during a given period, ΔT (Equation (6)):
(6)
The calculation of hydroelectric revenue ($), as represented by Equation (7), requires the multiplication of energy produced by the price per kWh, p(t) ($/kWh):
(7)
To determine the total hydropower revenue across all power plants and time steps in a network, Equation (8) should be used:
(8)

Storage–head relationship

For a plant with a large storage capacity, the water head varies depending on the reservoir levels. As storage increases, the height of the drop increases, and as storage decreases, the height of the drop decreases. Depending on the topography of a reservoir site, there is a nonlinear relationship between water storage, elevation, and energy storage. The gross head is the difference between the reservoir height and the tailwater: H = EreservoirEtailwater.

In the FEHEM model of power plants, a polynomial curve is plotted between storage volume and elevation to obtain a storage–head relationship. Storage and elevation data are obtained from the volume-slope curves of the dams.

The coefficients in Equation (9) (θ, α, β, γ, and c) represent the polynomial parameters used by the model. Observed storage and elevation data are obtained from DSI (State Hydraulic Works). The index i represents the power plant, and H and S are head and storage, respectively:
(9)

Penalties

In hydro-economic modeling, water demands can be represented by functions of gross economic benefits over a given time period. In time periods with lower water deliveries than target demand, water scarcity costs are represented by penalty functions (Harou et al. 2009).

In the FEHEM model, water scarcity costs are represented by piecewise linear penalty functions for hydroelectric water demand. Penalties represent the loss of benefits from not producing energy. The x-axis of a penalty curve represents the amount of water supplied for hydroelectric, and the y-axis represents the total cost (Figure 8). A linearly reversed graph would represent the hydroelectric benefit curve.
Figure 8

Economic value of water (Dogan 2015).

Figure 8

Economic value of water (Dogan 2015).

Close modal

Hydroelectric generation is modeled with penalty curves in FEHEM, and power capacity, energy production, and revenue are calculated with a separate post-processor. A hydroelectric processor takes reservoir storage and turbine release data from the FEHEM output file and provides a monthly time series of power capacity, total monthly and annual energy production and revenue, amount of water spilled, and total turbine capacity used.

Drought indices

Standardized precipitation index

The SPI method, widely employed by the World Meteorological Organization, is one of the most dependable and frequently used drought indices for assessing droughts.

The SPI measures the deviation of precipitation from the long-term average over a specified period and relies solely on precipitation data. Its calculation is based on the statistical distribution of precipitation values. The SPI can be applied across different time scales, allowing for the assessment of both short- and long-term drought conditions.

This technique was initially formulated by Mckee et al. (1993). It is a potent drought modeling index that entails only precipitation as an input parameter (Svoboda et al. 2012). It is dimensionless, standardized, and appropriate for various regions.

The SPI is calculated by dividing precipitation deviation over a specified time frame from the average by the standard deviation (as shown in Equations (10) and (11)):
(10)
(11)
where H(Xk) is the cumulative probability and t is the gamma distribution with c0 = 2.515517, c1 = 0.802853, c2 = 0.010328, d1 = 1.432788, d2 = 0.189269, and d3 = 0.001308.

In a drought evaluation utilizing SPI values, a dry spell is defined as the duration during which the index remains persistently negative. The commencement of the drought is determined by when the index drops below zero for the first instance, and the conclusion is determined by the month in which it returns to a positive value. SPI values above zero indicate precipitation levels exceeding the median, while negative values indicate precipitation below the median. As the SPI is normalized, it can represent both wetter and drier climates equally, and the SPI can also be used to track wet periods.

Reconnaissance drought index

The RDI is an index that evaluates drought by using both precipitation and potential evapotranspiration data. It is calculated based on the water balance (precipitation minus evapotranspiration), taking into account the effects of evaporation. By relying on both precipitation and potential evapotranspiration data, the RDI provides a more comprehensive assessment of drought.

Both indices (SPI and RDI) can be used at different time scales, which can be adjusted depending on the type of drought being analyzed (e.g., short-term agricultural droughts or long-term hydrological droughts). Using SPI and RDI together provides a more comprehensive perspective on drought assessment.

The RDI is expressed as the initial value (ak) and normalized RDI, presented in an aggregate form with a monthly time step and calculated monthly, seasonally, or annually. Equation (12) is used to calculate k (monthly) on a time basis. Monthly calculations are based on the following equation:
(12)
where Pij and PETij represent the precipitation and potential evapotranspiration (PET) of month j in year i, respectively, and N denotes the total number of years of available data. The normalised RDI (RDIn) is calculated using the following equation:
(13)

The classification according to the SPI and the RDI method is given in Table 2.

Table 2

SPI and RDI drought severity classes

SPI and RDI limit valuesDrought category
≥ 2 Heavy rain 
2 > SPI ≥ 1 Rainy 
1 > SPI > −1 Normal 
−1 ≥ SPI > −2 Dry 
−2 ≥ Heavy dry 
SPI and RDI limit valuesDrought category
≥ 2 Heavy rain 
2 > SPI ≥ 1 Rainy 
1 > SPI > −1 Normal 
−1 ≥ SPI > −2 Dry 
−2 ≥ Heavy dry 

For 45 years, this study has presented the time series of the SPI results and the distribution of drought classes in the Upper Euphrates Basin by year. Different meteorological stations were used to represent varying elevation bands within the basin.

Table 3 and Figure 9 illustrate the drought index values from 1971 to 2016, where regions with an SPI and RDI of −1 or less are considered drought-prone areas. According to Table 3, areas with an SPI and RDI equal to or less than −1 are classified as moderately dry (−1 to −1.49), severely dry (−1.5 to −1.99), or extremely dry (−2 and below). According to Table 3, areas with an SPI and RDI equal to or less than −1 are classified as moderately dry (−1 to −1.49), severely dry (−1.5 to −1.99), or extremely dry (−2 and below). Based on the SPI value, the years 1989 and 2013 were the driest documented. However, the RDI results in the study identified 1989 as the driest year within the study area over the last half-century. The wettest year, based on the determination of the SPI and the RDI, was 1988. The basin encountered drought once in 10 years, according to both the SPI and the RDI. Between 1971 and 2016, RDI values remained higher than the corresponding SPI values.
Table 3

SPI and RDI index classification

Water yearSPI indexRDI indexClassification type
1971 −0.83 −1.21 Normal 
1972 −1.1 −1.34 Dry 
1973 −0.89 −0.97 Dry 
1974 −1.32 −1.46 Dry 
1988 2.12 2.44 Heavy rain heavy 
1989 −2.09 −2.01 Dry 
1990 −1.74 −1.23 Dry 
1991 0.28 0,78 Normal 
1998 1.03 1.26 Rainy 
1999 −1.00 −1.34 Dry 
2000 −1.10 −1.00 Dry 
2001 −0.01 0.35 Normal 
2002 −0.40 −0.55 Normal 
2006 −0.44 −0.25 Normal 
2007 0.56 0.77 Normal 
2008 1.00 1.23 Normal 
2009 −1.01 −1.30 Dry 
2013 −1.79 −1.97 Dry 
2014 −0.71 −0.90 Normal 
2015 0.13 0.44 Normal 
2016 0.07 0.39 Normal 
Water yearSPI indexRDI indexClassification type
1971 −0.83 −1.21 Normal 
1972 −1.1 −1.34 Dry 
1973 −0.89 −0.97 Dry 
1974 −1.32 −1.46 Dry 
1988 2.12 2.44 Heavy rain heavy 
1989 −2.09 −2.01 Dry 
1990 −1.74 −1.23 Dry 
1991 0.28 0,78 Normal 
1998 1.03 1.26 Rainy 
1999 −1.00 −1.34 Dry 
2000 −1.10 −1.00 Dry 
2001 −0.01 0.35 Normal 
2002 −0.40 −0.55 Normal 
2006 −0.44 −0.25 Normal 
2007 0.56 0.77 Normal 
2008 1.00 1.23 Normal 
2009 −1.01 −1.30 Dry 
2013 −1.79 −1.97 Dry 
2014 −0.71 −0.90 Normal 
2015 0.13 0.44 Normal 
2016 0.07 0.39 Normal 
Figure 9

Changes in the SPI and RDI indexes over the years.

Figure 9

Changes in the SPI and RDI indexes over the years.

Close modal

Hydropower

An energy generation analysis has been conducted for the Upper Euphrates Basin based on the SPI index of water year types. The analysis compares the anticipated electricity generation values in various water year types, including wet, dry, normal, severe wet, and severe dry. Figure 10 illustrates the yearly fluctuations in the SPI and annual total energy of the other nine major dams within the study area, except for the Keban Dam. As precipitation levels rise, hydroelectricity generation increases accordingly. There is no abrupt decline in hydroelectric power production when precipitation levels decrease. This occurrence can be clarified by the existence of dams that can stockpile substantial amounts of water. Generally, the overall energy output in reservoirs with great storage capacity does not diminish significantly on an annual basis. Energy production increases in direct proportion in years with high rainfall. Furthermore, an increase in SPI value has been observed to have a positive impact on energy production during the period between 1984 and 1989. The SPI index had its peak value in 1988, at 2.1, and the total energy generated reached its highest value at 5,871.2 GWh within the same year. In contrast, the SPI value hit its nadir in 1989 at −2.1, leading to a minimum annual total energy value of 2,999.1 GWh.
Figure 10

Year-based change in the SPI and annual total energy of nine dams in the study area.

Figure 10

Year-based change in the SPI and annual total energy of nine dams in the study area.

Close modal
Figure 11 displays the annual hydroelectric generation of the Keban Dam, alongside the SPI index alterations. This figure presents an increase in hydroelectric production as the SPI value grows annually. During 1988–1989, energy production fell with the rapid decline in the SPI values. In 1989, Keban Dam experienced its minimum value with an SPI index of −2.1. Notably, a total energy of 3,888.4 GWh was recorded as the minimum annual amount in 1974. In 1988, the SPI index attained its peak value of 2.1, whereas the total energy reached its maximum at 6,759.7.2 GWh during the same year. The abrupt drop in 1989 can be attributed to the difference in rainfall between 1988 and 1989. Notably, the bar depicting the energy value did not reduce at the same pace as the SPI index, given the Keban Dam's sizable storage capacity.
Figure 11

Year-based change of Keban (KBN) Dam SPI and annual total energy.

Figure 11

Year-based change of Keban (KBN) Dam SPI and annual total energy.

Close modal
Table 4 and Figure 12 present the total energy and revenue values of various year types, determined by the SPI index. During years when the SPI index is −2 or less, total energy and revenue are at their lowest. However, during years with high rainfall, a total energy of 12,630.9 GWh was produced annually. Furthermore, the total revenue reached its peak at $843.4 million per annum. The rise in precipitation results in an increment of hydroelectric energy yielded, and correspondingly, revenue totals escalate.
Table 4

Energy-income variation with the SPI

Year typeIndexTotal energy (GWh/year)Total revenue (M$/year)
Heavy rain SPI and RDI ≥ 2 12,630.9 843.4 
Rainy 2 > SPI–RDI ≥ 1 9,529.5 622.9 
Normal 1 > SPI–RDI > −1 9,578.1 625.9 
Dry −1 ≥ SPI–RDI > −2 9,005.9 586.5 
Heavy dry SPI–RDI ≤ −2 7,017.4 460.3 
Year typeIndexTotal energy (GWh/year)Total revenue (M$/year)
Heavy rain SPI and RDI ≥ 2 12,630.9 843.4 
Rainy 2 > SPI–RDI ≥ 1 9,529.5 622.9 
Normal 1 > SPI–RDI > −1 9,578.1 625.9 
Dry −1 ≥ SPI–RDI > −2 9,005.9 586.5 
Heavy dry SPI–RDI ≤ −2 7,017.4 460.3 
Figure 12

Total FEHEM revenue with SPI-RDI indexes.

Figure 12

Total FEHEM revenue with SPI-RDI indexes.

Close modal

Surface water storage

Average surface storages over a 46-year period are compared with SPI and RDI indices (Figure 13). FEHEM stores more water during rainy years. Dam reservoirs generally fill in the spring with precipitation and snowmelt runoff and drawdown in the summer to meet irrigation demands and generate hydropower. Both indexing and storage peaked in 1988. The lowest storage occurs in 1988, when rainfall and index values are the lowest. Following the years of maximization and minimization of the index values, energy production and hence energy revenue show a directly proportional change. Since there are dams with large storage volumes in the FEHEM region, energy production and index values do not change in the same year but usually in the following year. This is due to the use of stored water in the reservoir during dry years. Several factors affect reservoir storage levels. The location of reservoirs, whether they are in parallel or series, is important in hydropower management. Electricity prices and agricultural and urban demands determine the timing and magnitude of allocations. Downstream flow regulations also control releases from reservoirs. Considering all these effects, FEHEM decides on release amounts and timing.
Figure 13

Total FEHEM surface storages with SPI-RDI indexes.

Figure 13

Total FEHEM surface storages with SPI-RDI indexes.

Close modal

Drought forecasting studies are crucial for precise and timely water resource planning and management. Population growth and climate change have amplified the consequences of recurrent droughts. Hence, it is imperative to comprehend and characterize the frequency of drought and its impact on agricultural output, hydrological cycle, and hydropower generation. It is evident that forecasting drought hazards is challenging and intricate, but monitoring droughts through indices often forms a crucial foundation. Especially in dry years, when hydropower becomes more important in reservoir operations, generation discrepancies increase. To this end, the study conducted two distinct drought analyses. The research examined the drought conditions in the Upper Euphrates Basin during the past 50 years and explored its impact on the region, which serves as Turkey's primary hydroelectricity production base.

The findings indicate that RDI predicts a higher tendency toward drought than SPI, in general. In addition, the agreement between RDI and SPI is stronger during dry years than in wet years. In semi-arid, semi-humid, and humid regions, where the drought trends are established differently between the two indices, RDI shows a greater tendency for drought than SPI.

There is a direct correlation between index values and hydroelectric power generation. As the index values increase, these values increase and decrease with decreasing values. When analyzing the temporal fluctuations of SPI–RDI index and energy production graphs, a direct correlation is observed between the index curves and energy production values. Specifically, in years with high rainfall, the SPI–RDI index exhibits high values and the amount of energy production increases accordingly. In the years when the index values are above 2, the average hydroelectric energy across the basin is 12.6 billion kWh and the income is 843 million USD, while in the years when the index values are below −2, the average decreases to 7 billion kWh and 460 million USD. This pattern is consistent for dams with limited storage capacity. However, in high-capacity storage dams like the Keban Dam, fluctuations in index value do not have a direct impact on energy production in that given year. As a result, dams with high storage capacity are known to have more adaptable operating conditions.

In years when the index values are −1.5 and below, the storage amounts and production reliability of the reservoirs in the study region are significantly reduced. Decreases in river flow due to a lack of precipitation cause a decrease in the total energy of the dams as well as a decrease in their firm energy.

Based on the information provided, future studies should focus on identifying adaptation strategies to address demand deficits that may arise in the existing water supply system due to drought, climate change, and other factors. In addition, basin-based water operations should be approached with an integrated logic throughout the Euphrates Basin, which is the largest and most important source of hydroelectricity generation in the country.

A. A. and M. C. T. analyzed the data and wrote the paper. M.S.D. research conceptualized the whole article, developed the methodology, supervised the work, and reviewed and edited the article. All authors read and approved the final manuscript.

Data cannot be made publicly available; readers should contact the corresponding author for details.

The authors declare there is no conflict.

Ajjur
S. B.
&
Al-Ghamdi
S. G.
2021
Evapotranspiration and water availability response to climate change in the Middle East and North Africa
.
Clim. Change
166
,
28
.
doi:10.1007/s10584-021-03122-z
.
Aytac
A.
2023
Development of Upper Euphrates Basin Hydroeconomic Model and Determination of Climate Change Adaptation Strategies
.
Doktora Tezi
,
Fırat Üniversitesi, İnşaat Mühendisliği
,
Elazığ
.
Aytac
A.
,
Tuna
M. C.
&
Dogan
M. S.
2023
Development of Upper Euphrates Basin hydro-economic model and hydropower generation optimization
.
J. Water Clim. Change
14
(
9
),
3385
3397
.
https://doi.org/10.2166/wcc.2023.377
.
Aytac
A.
,
Dogan
M. S.
&
Tuna
M. C.
2024
Energy-based hydro-economic modeling of climate change effects on the Upper Euphrates Basin
.
J. Water Clim. Change
15
(
2
),
733
746
.
doi: https://doi.org/10.2166/wcc.2024.550
.
Dogan
M. S.
2015
Integrated Water Operations in California: Hydropower, Overdraft, and Climate Change.
MSc thesis. University of California Davis, Office of Graduate Studies
,
Davis
,
CA, USA
.
Dogan
M. S.
,
Fefer
M. A.
,
Herman
J. D.
,
Hart
Q. J.
,
Merz
J. R.
,
Medell__n-Azuara
J.
&
Lund
J. R.
2018
An open-source python implementation of California's hydroeconomic optimization model
.
Environ. Modell. Software
108
,
8
13
.
doi:10.1016/j.envsoft.2018.07.002
.
Draper
A. J.
,
Jenkins
M. W.
,
Kirby
K. W.
,
Lund
J. R.
&
Howitt
R. E.
2003
Economic- engineering optimization for California water management
.
J. Water Resour. Plann. Manage.
129
(
3
),
155
164
.
http://doi.org/10.1061/(ASCE)0733-9496(2003)129:3(155)
.
Hao
Z.
&
AghaKouchak
A.
2013
A nonparametric multivariate multi-index drought monitoring framework
.
J. Hydrometeorol.
doi:10.1175/JHM-D-12-0160.1
.
Harou
J. J.
,
Pulido-Velázquez
M.
,
Rosenberg
D. E.
,
Medellín-Azuara
J.
,
Lund
J. R.
&
Howitt
R. E.
2009
Hydroeconomic models: Concepts, design, applications, and future prospects
.
J. Hydrol.
375
,
627
643
.
Hart
W. E.
,
Laird
C. D.
,
Watson
J.-P.
,
Woodru_
D. L.
,
Hackebeil
G. A.
,
Nicholson
B. L.
&
Siirola
J. D.
2017
Pyomo Optimization Modeling in Python
, Vol.
67
.
Springer International Publishing
,
Cham
.
doi: 10.1007/978-3-319-58821-6
.
IPCC
2013
Summary for Policymakers
. In:
Climate Change 2013: The Physical Science Basis. Contribution of Working Group I to the Fifth Assessment Report of the Intergovernmental Panel on Climate Change
(
Stocker
T. F.
,
Qin
D.
,
Plattner
G.-K.
,
Tignor
M.
,
Allen
S. K.
,
Boschung
J.
,
Nauels
A.
,
Xia
Y.
,
Bex
V.
&
Midgley
P. M.
, eds.).
Cambridge University Press
,
Cambridge, UK
.
Kao
S.-C.
&
Govindaraju
R. S.
2010
A copula-based joint deficit index for droughts
.
J. Hydrol.
380
(
1–2
),
121
134
.
Lelieveld
J.
,
Proestos
Y.
,
Hadjinicolaou
P.
,
Tanarhte
M.
,
Tyrlis
E.
&
Zittis
G.
2016
Strongly increasing heat extremes in the Middle East and North Africa (MENA) in the 21st century
.
Clim. Change
137
,
245
260
.
doi:10.1007/s10584-016-1665-6
.
Li
B.
,
Liang
Z.
,
Zhang
J.
&
Wang
G.
2017
A revised drought index based on precipitation and pan evaporation
.
Int. J. Climatol.
37
(
2
),
793
801
.
McKee
T. B.
,
Doesken
N. J.
&
Kleist
J.
1993
The relationship of drought frequency and duration to time scales
. In:
8th Conference on Applied Climatology
,
17–22 January
.
American Meteorological Society
,
Anaheim, California
, pp.
179
184
.
Ozturk
T.
,
Ceber
Z. P.
,
Türkeş
M.
&
Kurnaz
M. L.
2015
Projections of climate change in the Mediterranean basin by using downscaled global climate model outputs
.
Int. J. Climatol.
35
,
4276
4292
.
doi:10.1002/joc.4285
.
Paltán
H. A.
,
Pant
R.
,
Plummer Braeckman
J.
&
Dadson
S. J.
2021
Increased water risks to global hydropower in 1.5 °C and 2.0 °C warmer worlds
.
J. Hydrol.
599
,
126503
.
doi:10.1016/j.jhydrol.2021.126503
.
Svoboda
M.
,
Hayes
M.
,
Wood
D.
&
Pappenberger
F.
2012
Standardized Precipitation Index User Guide
.
World Meteorological Organization
,
Geneva
.
Tallaksen
L. M.
&
van Lanen
H. A. J.
2004
Hydrological Drought: Processes and Estimation Methods for Streamflow and Groundwater
.
Developments in Water Science
, Vol.
48
.
Elsevier
,
Amsterdam
, pp.
579
.
Tam
B. Y.
,
Szeto
K.
,
Bonsal
B.
,
Flato
G.
,
Cannon
A. J.
&
Rong
R.
2019
CMIP5 drought projections in Canada based on the standardized precipitation evapotranspiration index
.
Can. Water Resour. J.
44
(
1
),
90
107
.
Tsakiris
G.
&
Vangelis
H.
2005
Establishing a drought index incorporating evapotranspiration
.
Eur. Water
9
(
10
),
3
11
.
Vicente-Serrano
S. M.
,
Miralles
D. G.
,
Dominguez-Castro
F.
,
Azorin-Molina
C.
,
El Kenawy
A.
,
McVicar
T. R.
,
Tomás-Burguera
M.
,
Beguería
S.
,
Maneta
M.
&
Peña-Gallardo
M.
2018
Global assessment of the standardized evapotranspiration deficit index (SEDI) for drought analysis and monitoring
.
J. Clim.
31
(
14
),
5371
5393
.
Waha
K.
,
Krummenauer
L.
,
Adams
S.
,
Aich
V.
,
Baarsch
F.
,
Coumou
D.
,
Fader
M.
,
Hoff
H.
,
Jobbins
G.
,
Marcus
R.
,
Mengel
M.
,
Otto
I. M.
,
Perrette
M.
,
Rocha
M.
,
Robinson
A.
&
Schleussner
C.-F
.
2017
Climate change impacts in the Middle East and Northern Africa (MENA) region and their implications for vulnerable population groups
.
Reg. Environ. Change
17
(
6
),
1623
1638
.
doi:10.1007/s10113-017-1144-2
.
Wu
J. F
.,
Chen
X. W.
,
Yao
H. X.
,
Gao
L.
,
Chen
Y.
&
Liu
M.
2017
Non-linear relationship of hydrological drought responding to meteorological drought and impact of a large reservoir
.
J. Hydrol.
551
,
495
507
.
http://dx.doi.org/10.1016/j.jhydrol.2017.06.029
.
Xiao
F.
,
Liu
Q.
,
Qin
Y.
,
Huang
D.
&
Liao
Y.
2024
Agricultural drought research knowledge graph reasoning by using VOSviewer
.
Heliyon
10
(
6
),
e27696
.
https://doi.org/10.1016/j.heliyon.2024.e27696
.
Zhang
R.
,
Zhao
C.
,
Ma
X.
,
Brindha
K.
,
Han
Q.
,
Li
C.
&
Zhao
X.
2019
Projected spatiotemporal dynamics of drought under global warming in Central Asia
.
Sustainability
11
(
16
),
4421
4440
.
Zhao
S.
,
Cong
D.
,
He
K.
,
Yang
H.
&
Quin
Z.
2017
Spatial-temporal variation of drought in China from 1982 to 2010 based on a modified Temperature Vegetation Drought Index (mTVDI).
Sci. Rep.
7
,
17473
.
https://doi.org/10.1038/s41598-017-17810-3
.
This is an Open Access article distributed under the terms of the Creative Commons Attribution Licence (CC BY-NC-ND 4.0), which permits copying and redistribution for non-commercial purposes with no derivatives, provided the original work is properly cited (http://creativecommons.org/licenses/by-nc-nd/4.0/).