ABSTRACT
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.
HIGHLIGHTS
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.
NOTATIONS AND ABBREVIATIONS
- 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
INTRODUCTION
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.
MATERIALS AND METHODS
Hydrology
Euphrates basin long-years monthly average evaporation (mm) (Aytac 2023).
Rainfall
Average total precipitation of the stations representing the basin (Aytac 2023).
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.
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.
A general network diagram of nodes and reservoirs used in the model.
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.
Dam data (Aytac et al. 2024)
Dam ID . | Installed 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 ID . | Installed 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 |
Long-term monthly average electricity prices (cent$/kWh) (Aytac 2023).
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 = Ereservoir − Etailwater.
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.
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).
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.
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 classification according to the SPI and the RDI method is given in Table 2.
SPI and RDI drought severity classes
SPI and RDI limit values . | Drought category . |
---|---|
≥ 2 | Heavy rain |
2 > SPI ≥ 1 | Rainy |
1 > SPI > −1 | Normal |
−1 ≥ SPI > −2 | Dry |
−2 ≥ | Heavy dry |
SPI and RDI limit values . | Drought category . |
---|---|
≥ 2 | Heavy rain |
2 > SPI ≥ 1 | Rainy |
1 > SPI > −1 | Normal |
−1 ≥ SPI > −2 | Dry |
−2 ≥ | Heavy dry |
RESULTS AND DISCUSSION
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.
SPI and RDI index classification
Water year . | SPI index . | RDI index . | Classification 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 year . | SPI index . | RDI index . | Classification 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 |
Hydropower
Year-based change in the SPI and annual total energy of nine dams in the study area.
Year-based change in the SPI and annual total energy of nine dams in the study area.
Energy-income variation with the SPI
Year type . | Index . | Total 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 type . | Index . | Total 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 |
Surface water storage
CONCLUSIONS
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.
AUTHORS CONTRİBUTİONS
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 AVAILABILITY STATEMENT
Data cannot be made publicly available; readers should contact the corresponding author for details.
CONFLICT OF INTEREST
The authors declare there is no conflict.