Open AccessArticle

Turkey’s Hydropower Potential in the Near Future and the Possible Impacts of Climate Change—A Case Study of the Euphrates–Tigris Basin

Goksel Ezgi Guzey

^* and

Bihrat Onoz

Faculty of Civil Engineering, Istanbul Technical University, Istanbul 34469, Türkiye

Author to whom correspondence should be addressed.

Climate 2024, 12(10), 156; https://doi.org/10.3390/cli12100156

Submission received: 26 June 2024 / Revised: 2 September 2024 / Accepted: 4 September 2024 / Published: 3 October 2024

Download

Browse Figures

Figure 1
From top to bottom: major dams in the ETRB with the location of the 14 dams studied. The ETR basin studies with the corresponding streamflow stations present. "> Figure 2
Example of data obtained for the Ataturk dam (from 2015) showing flow and hydropower generated. "> Figure 3
The distribution of alpha (α) values generated by varying the standard deviation (σ) of the random variable ε. Different σ values were used to assess the sensitivity of α to variations in ε. While only these three ranges are shown for clarity, a broader range from 0.01 to 1 with a 0.02 interval was generated. The normal (Gaussian) distribution was used for ε, providing a realistic simulation of uncertainties. "> Figure 4
Normalized 1:1 scatter plots for the predictive performance of the initial equation for all dams (α = 0.5). The scatter plots display the relationship between the normalized observed and normalized predicted values for the reservoir outflow across multiple dams. Each plot corresponds to a different dam, with the dam name and the coefficient of determination (R2) value indicated. The diagonal red line represents the perfect 1:1 relationship between observed and predicted values. Normalization is done to ensure consistency in comparing the magnitude of values across different dams and to reveal systematic biases. "> Figure 5
Performance metrics (PBIAS, NSE, and R2) for different dams located along a river system or basin. "> Figure 6
This graph illustrates the process of determining the optimal value of the parameter alpha (α) for the Keban Dam, based on minimizing different error metrics. The vertical axis represents the error values, while the horizontal axis shows the range of alpha values considered. The blue line corresponds to the PBIAS (Percent Bias) error, the red line represents the NSE (Nash–Sutcliffe Efficiency) error, and the green line depicts the R2 error. The vertical dashed line highlights the chosen optimal alpha value of 0.851, which appears to minimize the overall errors across the different metrics for the Keban Dam. "> Figure 7
Streamflow releases estimated from the HBV model with the adjusted parameters. The HBV model corresponds to RCP 8.5 projections, with NSE = 0.752. "> Figure 8
Seasonal trends observed for the releases at each dam. "> Figure 9
Hydropower generation time series for the dams. "> Figure 10
Yearly rate of change for hydropower generation. ">

Versions Notes

Abstract

Hydropower is becoming an important renewable energy source in Turkey, but the ever-changing atmospheric and climatic conditions of Turkey make it very difficult to be projected efficiently. Thus, an efficient estimation technique is crucial for it to be adopted as a reliable energy source in the future. This study evaluates Turkey’s hydropower potential in the Euphrates–Tigris Basin under changing climatic conditions. We adapted an empirical equation to model reservoir outflows, considering the site-specific characteristics of 14 major dams. Initial results from employing a model with a constant empirical coefficient, α, yielded moderate predictive accuracy, with R² values ranging from 0.289 to 0.612. A polynomial regression identified optimal α values tailored to each dam’s surface area, significantly improving model performance. The adjusted α reduced predictive bias and increased R² values, enhancing forecast reliability. Seasonal analysis revealed distinct hydropower trends: Ataturk Dam showed a notable decrease of 5.5% in hydropower generation up to 2050, while Birecik and Keban Dams exhibited increases of 2.5% and 2.2%, respectively. By putting these discoveries into practice, water resource management may become more robust and sustainable, which is essential for meeting Turkey’s rising energy needs and preparing for future climatic challenges. This study contributes valuable insights for optimizing reservoir operations, ensuring long-term hydropower sustainability, and enhancing the resilience of water resource management systems globally.

Keywords:

hydropower estimation; reservoir management; polynomial regression; reservoir release estimation; dams; streamflow estimation; HBV; climate change

1. Introduction

Hydropower is a crucial renewable energy source in Turkey, particularly given the nation’s expanding energy requirements and the significant hydroelectric potential within its river basins. Effective management of water storage reservoirs, such as those in the Euphrates–Tigris Basin (ETRB), is essential for optimizing hydropower production and ensuring water availability. This necessity becomes even more pressing in the context of climate change, which poses challenges through altered precipitation patterns and increased frequency of extreme weather events. It is widely acknowledged that managing water storage reservoirs is essential for fostering adaptability and resilience in the face of the numerous obstacles that climate change and its effects present. These facilities serve multiple purposes, such as mitigating flood damage and providing water during dry spells. Depending on how they are built and used, basins can support a wide range of ecosystem services, help produce energy, and guarantee a consistent supply of water under less extreme conditions [1,2]. For instance, streamflow variations, forecast uncertainty, and storage capacity restrictions are frequently considered when optimizing the delivery of water [3]. As a result, numerous models have sought to estimate the best release scenarios while taking the context of the difficulties encountered and climate projections into account. Peng and Buras presented a computational procedure for multi-reservoir systems, considering water budget and cumulative mass curves to account for both controlled and unregulated flow [4]. Other methods, like extended deficit analysis, incorporated reservoir characteristics to explore storage–yield relationships with regard to available data on each studied dam. Research advancements are needed to improve the accuracy of estimation methods for streamflow release. Decision-support tools provide estimates of daily outflows from reservoirs and compare them with estimates of unaltered streamflow [5]. These tools help environmental managers understand the complex interactions between water withdrawals, reservoir operational practices, and aquatic habitats. They also identify physical and operational features of reservoirs that have a significant effect on reservoir outflows, such as the ratio of average withdrawals to average inflows and the size of the surface area relative to drainage area. In addition, D. Fishel studied the hydrology and water quality of Swatara Creek in Pennsylvania, which included analyzing concentrations of suspended sediment, nutrients, metals, and acid mine drainage in surface-water inflows and downstream discharge from a planned reservoir [6]. The study found that mine drainage affects water quality and can impair the planned uses for the reservoir unless measures are taken to improve water quality from both point and nonpoint sources. These studies highlight the importance of developing more accurate estimation methods for streamflow release to better manage water resources and preserve downstream ecosystems.

With climate change projected to exacerbate water scarcity and extreme weather events in many regions, the need for improved reservoir management strategies has become increasingly urgent [7,8]. A study by Lehner et al. utilized the global water model WaterGAP to assess the impact of global change on Europe’s hydropower potential. This comprehensive model-based analysis indicated significant regional variability in hydropower potential, with southern and southeastern Europe expected to see reductions of 25% or more due to climate change. Hamududu and Killingtveit [9] evaluated the changes in global hydropower generation using an ensemble of global circulation models (GCMs). Their findings suggested that while global hydropower generation might remain relatively stable, regional variations would be substantial, with some areas experiencing significant increases or decreases in hydropower potential. Arriagada et al. (2019) [10] focused on data-scarce regions and used reconstructed streamflow series combined with geographic information systems (GIS) techniques to estimate hydropower potential. They found significant variability in hydropower potential modulated by climatic oscillations like El Niño, emphasizing the need for improved hydropower development policies. Mutsindikwa et al. [11] assessed the impact of climate change on the hydropower potential of the Bamboi catchment in West Africa. Using regional climate models and hydrological models, they projected a decrease in hydropower generation despite an increase in precipitation, due to seasonal variabilities in water availability. Duratorre et al. [12] examined the effects of climate change on hydropower production in the Alps using a state-of-the-art hydrological model. Their projections showed that while temperature increases would significantly reduce ice and snowmelt, hydropower production might experience moderate increases due to higher precipitation in some scenarios. Developing accurate estimation methods for streamflow release and optimizing reservoir operations can help mitigate the impacts of droughts, floods, and water shortages, while also supporting ecosystem health and meeting diverse water demands [13,14]. Moreover, effective reservoir management can contribute to sustainable development goals and promote water security, particularly in transboundary river basins where cooperation and coordinated management are essential [15,16].

With advances in modeling streamflow and release operations in multi-reservoir systems, the aim of the current study is to address the widely used generic framework for reservoir simulation proposed by Hanasaki et al. [17]. Starting from the suggested global framework, we aim to:

-: Assess the performance of the suggested framework through empirical analysis.
-: Develop site-specific characteristics adapted for the Euphrates–Tigris Basin (ETRB).
-: Analyze a series of multiple reservoir systems using site-specific parameters.

Hence, this work will advance much-needed information for the ETRB region by adapting established frameworks to the local context, ultimately guiding reservoir management schemes.

2. Materials and Methods

2.1. Study Area Description and Dams Analyzed

Parts of Turkey, Syria, Iraq, and Iran are included in the Euphrates–Tigris Basin, a hydrologically complex system known for its major contribution to the management of regional water resources. The two principal rivers of the basin, the Euphrates and Tigris, originate in the Taurus and Zagros Mountains, respectively, and snowmelt and seasonal precipitation patterns have a significant impact on their flows. Numerous dams, such as the Atatürk Dam in Turkey and the Mosul Dam in Iraq, are scattered throughout this river system (Figure 1). While these dams are essential for irrigation, flood control, and hydropower production, they have also raised geopolitical and environmental concerns. These dams have profoundly changed the natural flow regime, affecting the availability and quality of water downstream and playing a major role in regional conflicts over water rights and usage. The management challenges in the basin are exacerbated by climate change, which is expected to alter precipitation patterns and intensify water scarcity in this predominantly semi-arid region. Fourteen major dams were analyzed (Table 1), ranging from small irrigation dams to large dams used for hydropower production. The choice of dams was based mainly on data availability with the intent to sample a representative range of reservoir areas and avoid spatial clustering of the samples. By doing so, the developed analysis can be applicable to other dams in the ETRB once data and sufficient information become available. All data were acquired from local management authorities, especially streamflow release and hydropower production when applicable.

Figure 2 shows Ataturk Dam’s 8 year (2015–2022) inflow and generated hydropower data. As can be observed from the graph, the inflow and hydropower are correlated for most of the years; however, in 2017, which was one of the dryest periods for Turkey, the generated hydropower was more than the inflow. This occurrence can be due to the socio-political decisions regarding the energy production. Furthermore, from mid 2019 to the beginning of 2020, there was a significant increase in energy production when compared to the previous 2 years. This was most likely due to the uncontrolled immigration of refugees to the area during that time. With the unexpected increase of population, the energy consumption was high.

2.2. Streamflow Release Estimation and Parameters Adjustment

First, the equation of Hanasaki was used to estimate monthly streamflow releases [17]:

R_{m} = \{\begin{matrix} {(\frac{c}{K_{c}})}^{β} K_{y} i_{a} + [1 - {(\frac{c}{K_{c}})}^{β}] \\ K_{y} i_{a} c \geq K_{c} \end{matrix} i_{m}, 0 < c < K_{c}

(1)

where Kc is the criterion of c,

i_{m}

is monthly inflow (m³/s),

i_{a}

is mean annual inflow (m³/s), k_y = S_beg/αC (where α = 0.85), and c = C/I_a.

Rm, used to calculate the dam’s release (Q in Equation (8)), allows for the estimation of generated power given complete dam information. S_beg is the reservoir storage at the beginning of a year (m³), C is the maximum storage capacity of the reservoir (m³), I_a is the mean total annual inflow (m³), and α is an empirical coefficient (0.85 as suggested empirically), which influences interannual variation in releases. The criterion of c (Kc) is set as 0.5, and the exponent of (c/Kc), β, is set as 2 empirically. When the reservoir storage capacity is large compared to annual inflow (c ≥ Kc), the monthly release is independent of monthly inflow and will be constant throughout the year if water is available. The reservoir is not allowed to release water when water storage is below 10% of the storage capacity, and the monthly release would be no less than 10% of the mean monthly inflow for environmental flow. If water storage exceeds the storage capacity, the excess water will be released.

A stochastic model was built for α and β to test the adjusted parameters and enhance prediction metrics as follows (Figure 3):

α = α_{0} + σ . ϵ

(2)

where α is the adjusted parameter, α₀ is the initial value of 0.85, and ∈ is a random variable following a random normal distribution.

For α, a range between −5 and 5 was tested. For β, a range between −10 and 10 was tested. Then, three performance metrics were assessed:

-: Bias, represented by percentage of bias (PBIAS):

P B I A S = \frac{\sum_{i = 1}^{n} (\hat{y_{i}} - y_{i})}{\sum_{i = 1}^{n} y_{i}} \times 100

(3)

-: The Nash–Sutcliff efficiency (NSE):

N S E = 1 - \frac{\sum_{i = 1}^{n} {(y_{i} - \hat{y_{i}})}^{2}}{\sum_{i = 1}^{n} {(y_{i} - \bar{y})}^{2}}

(4)

-: The adjusted coefficient of determination (adj R²):

R^{2} = 1 - \frac{\sum_{i = 1}^{n} {(y_{i} - \hat{y_{i}})}^{2}}{\sum_{i = 1}^{n} {(y_{i} - \bar{y})}^{2}}

(5)

A polynomial regression was then developed to estimate α based on the area as follows:

α = b_{0} + b_{1} A r e a + \dots + b_{n} {A r e a}^{n} + e r r o r

(6)

where α is the adjusted parameter, b_n are the regression coefficients, and Area is the reservoir area in km².

Afterward, the resulting equation is applied to the four dams of Ataturk, Birecik, Karakay, and Kerban to project the changes in trends in hydropower production. We selected the reservoir area as an initial characteristic of the dams, indicative of potential water availability and possible increased surface evaporation, which could influence the α parameter that predicts dam release. The data concerning each dam area were provided by the relevant dam management authorities in the Euphrates–Tigres River Basin.

2.3. Estimation of Hydropower Production

To estimate potential hydropower production, the adjusted parameter facilitated the projection of future water releases and, consequently, the potential hydropower output in relation to turbine efficiency. For each dam, turbine efficiency was inferred from historical data series. By examining the generation equation, an assumption of optimal operational conditions is posited:

P = η . g . Q . H

(7)

where P is the produced power in KW, η is the turbine efficiency, H is the head of the dam (m), and Q is the released flow (m³/s).

3. Results

Using the initial suggested value for α = 0.85 did not yield promising results in estimating hydropower potential, and the coefficient of determination was significantly low for all of the 14 dams. When the dams’ observed hydropower data were subjected to an initial analysis to see which constant α value would yield the most robust estimations, it was observed that α = 0.5 was the optimal choice with the range of R² between 0.289 and 0.612 (Figure 4). PBIAS ranged between 19% and 40% and the Nash–Sutcliffe Efficiency (NSE), a normalized statistic that determines the relative magnitude of the residual variance compared to the measured data variance, ranged between 2.75 and 4.17 (Figure 5). This further demonstrated the initial model’s predictive accuracy.

Afterwards, the regression results yielded the following relationship between α and Area:

α = 0.0012A − 0.043 (R² = 0.72)

(8)

Table 2 presents a comprehensive comparison of hydrological model performance metrics across multiple dams, before and after the optimization of the parameters α and β. For each dam, the Percent Bias (PBIAS), Nash–Sutcliffe Efficiency (NSE), and the Coefficient of Determination (R²) are initially computed with α set to 0.5 and β to 2. Following optimization, new values for these metrics (PBIAS-new, NSE-new, and R²-new) are calculated along with the optimized values for α and β.

Specifically, the Keban Dam showed a PBIAS reduction from 34% to 7.5% and an increase in NSE from 2.95 to 0.653 after optimization, indicating a more accurate model with less systematic bias and improved efficiency. The R² value increased from 0.439 to 0.765, with a refined R²-new* value of 0.745 based on the specified formula. Similarly, the Karakaya Dam’s PBIAS improved from 28% to 16.2%, and R² increased significantly from 0.582 to 0.832, reflecting a substantial enhancement in the model’s explanatory power.

Ataturk and Birecik Dams also saw improvements in R² from 0.612 to 0.795 and from 0.427 to 0.810, respectively. The Seyhan Dam was notable for achieving an NSE-new of 1.014, surpassing the optimal value of 1, which may indicate a potential overfitting issue or an anomaly in the model’s performance.

Conversely, the Adıgüzel Dam, despite a high initial PBIAS of 43%, managed to decrease it to 10.6% post-optimization. The R²-new value here was one of the highest recorded at 0.892, with an R²-new* adjustment leading to 0.848, suggesting strong predictive capability after optimization.

For the remaining dams, the optimization process resulted in varying degrees of improvement across all metrics. The Alakır Dam, for example, reduced its PBIAS from 41% to 12.6% and increased its R²-new from 0.427 to 0.791. In contrast, the Gezende Dam, with a smaller optimization in α, still managed to enhance its R² from 0.539 to 0.828.

In Figure 5, the vertical axis represents the metric values, while the horizontal axis lists the individual dam names. The blue bars indicate the NSE (Nash–Sutcliffe Efficiency) values, the green bars represent the PBIAS (Percent Bias) values, and the red bars show the R² (Coefficient of Determination) values for each dam.

Before the adjustment of the parameters for the Hanasaki equations, the hydropower estimation for all 14 dams is significantly weak, as can be seen from Figure 4 and Figure 5. It is clear that an adjustment of the equation was crucial for accurate hydropower estimation.

Figure 6 showcases the results of a stochastic optimization approach applied to the Keban Dam. This approach aimed to identify the optimal value of α for improved predictive accuracy. The results showed a notable increase in the optimal α value, which was determined to be 0.851 for Keban Dam. This indicates that the initial assumption of α = 0.5 was suboptimal for this specific context.

The resulting releases for the four chosen dams are shown for the projected period from 2022 to 2090, obtained from the HBV model [18]. As it can be clearly seen from Table 2, after the adjustment of parameter α, the robustness of the equation’s estimation capabilities significantly increased. The predicted streamflow showed significant compatibility with the observed streamflow, as explained in Table 2, with the statistical goodness-of-fit measures after the adjustment of physical parameters. Thus, with the new adjusted equation, the first step is to calculate streamflow releases, Rm, for the future (2024–2100).

Keban, Karakaya, Ataturk, and Birecik Dams’ streamflow releases from 2024 to 2100 are shown in Figure 7. The graphs present the enhanced performance metrics for each dam following the adjustment of the α parameter. A general increase in R² values across the board was observed, signifying a substantial improvement in model accuracy and predictive power.

The changes to the near future from previous years for both the observed flow and estimated flow and the effects on hydropower will be further discussed in the Discussion section.

In Figure 8, the plots show significant seasonal variability in water releases. For example, Keban and Karakaya exhibit peak releases around spring and early summer, likely due to snowmelt and rainfall patterns, whereas Ataturk and Birecik show more moderated fluctuations, suggesting different water management strategies or basin characteristics.

Figure 9 showcases the hydropower generation trends indicating growth and decline phases and reflecting operational and environmental changes. Notably, Karakaya shows a sharp increase in production capacity towards the later years, suggesting possible upgrades or increased water availability, whereas Ataturk demonstrates more consistent output, indicating stable conditions and efficient management.

After the estimation of the streamflow release,

R_{m}

, the next step of the multi-equation is to estimate the hydropower, which is the main purpose of this study. The hydropower is estimated for the Keban, Karakaya, Ataturk, and Birecik dams for the period from 2024 to 2100. The changes to hydropower in the upcoming years, the potential causes of these changes, and the possible effects on the economy will be discussed further.

In Figure 10, the hydropower generation rate of change for the near future is shown. In general, annual changes in hydropower generation reveal operational challenges and successes. For instance, it can be observed from Figure 10 that Karakaya dam experiences significant fluctuations in hydropower generation, suggesting possible volatility in water inputs or mechanical issues.

Table 3 showcases the Mann-Kendall test results for hydropower generation. The test revealed a significant decreasing trend in both periods for Ataturk Dam, with a more pronounced decline up to 2050. The negative

τ

value indicates a decreasing trend in hydropower generation, with the trend becoming less steep in the period from 2051 to 2100; however, it remains significant.

Birecik Dam exhibits a slight but statistically significant increasing trend in hydropower generation for both periods. The positive

τ

values indicate an increasing trend, with the trend remaining relatively stable throughout the later period.

Karakaya Dam presents a small increasing trend up to 2050, switching to a significant decreasing trend from 2051 to 2100. The change from a positive to a negative τ value indicates a shift in the trend of hydropower generation, suggesting a potential concern for long-term generation prospects.

Keban Dam shows a consistent but slightly increasing trend in hydropower generation for both periods, with the trend slightly weakening in the later period but remaining positive and significant. These trends provide insight into the hydropower capacities of these dams for the near future.

Seasonal Mann Kendall results are showcased in Table 4: Ataturk Dam experiences significant seasonal variation, with an increasing trend in hydropower generation during the summer and a decreasing trend in the winter, suggesting seasonal dependencies on water availability or usage patterns.

Birecik Dam shows significant increasing trends in both autumn and summer, indicating good performance in these seasons. The absence of significant trends in spring and winter suggests stable generation or less variation in these periods.

Karakaya Dam faces significant decreasing trends in autumn and summer, highlighting potential challenges or reduced water availability in these seasons.

Keban Dam’s significant decreasing trend in the summer could indicate vulnerabilities or operational constraints during this season, despite the lack of significant trends in other seasons.

Seasonal Variability: Significant trends are observed in specific seasons for each dam, indicating varying operational and environmental influences across the year. Summer Trends: Ataturk and Birecik Dams show increased generation in summer, while Karakaya and Keban Dams face decreases, highlighting the impact of seasonal factors. Winter Generation: Only Ataturk Dam shows a significant decrease in winter, with other dams displaying no significant trends, suggesting varied responses to winter conditions. Autumn and Spring: Significant trends in autumn for Birecik Dam (increase) and Karakaya Dam (decrease) contrast with mostly non-significant trends for Ataturk and Keban Dams, indicating diverse seasonal impacts.

4. Discussion

The analysis revealed several important findings regarding the adaptation of the reservoir management equation to the Euphrates–Tigris River Basin (ETRB) region. Firstly, the empirical coefficient α exhibited high variability across the 14 major dams studied, ranging from −0.017 to 0.851, deviating significantly from the initially suggested constant value of 0.5. This highlights the importance of accounting for site-specific characteristics and reservoir properties in accurately estimating streamflow releases. This variability in α emphasizes the need for a tailored approach to reservoir management, considering the unique attributes of each dam and its catchment area. Similar variability in reservoir management parameters has been observed in other studies, such as in the Colorado River Basin, where site-specific characteristics significantly influenced reservoir operations [19] This reinforces the necessity for customized approaches in diverse hydrological settings.

Furthermore, a strong correlation was established between the surface area of the reservoirs and the optimal value of α, described by the polynomial regression equation:

α = 0.0012A− 0.043 (R² = 0.72). Incorporating this area-dependent α adjustment into the reservoir management equation substantially improved the predictive performance across multiple dams, as evidenced by the enhanced values of PBIAS, NSE, and R² metrics. These improvements suggest that spatial scaling is crucial for refining hydrological models and enhancing their accuracy in diverse environmental contexts. Other studies have also highlighted the importance of spatial scaling in improving model accuracy [20]. This comparison underlines the universal applicability of spatial adjustments in hydrological modeling.

When applied to the four major dams of Ataturk, Birecik, Karakaya, and Keban, the adapted equation facilitated the projection of streamflow releases and subsequent hydropower generation trends for the near future (2030–2100). The results revealed distinct patterns and variations among the dams, with some exhibiting increasing trends (Birecik and Keban), others showing decreasing trends (Ataturk), and instances of trend shifts from increasing to decreasing (Karakaya) over the analyzed period. These patterns are critical for developing adaptive management strategies that can respond to changing hydrological conditions. Similar trends have been observed in other regions with complex hydrological dynamics, such as the Ganges and Yangtze River Basins, where adaptive strategies have been crucial in managing fluctuating hydropower outputs [21].

The seasonal Mann–Kendall analysis further highlighted the influence of seasonal factors on hydropower generation, with dams experiencing significant increasing or decreasing trends during specific seasons (e.g., summer decreases for Keban, autumn decreases for Karakaya). These findings underscore the importance of considering both long-term trends and seasonal variations in water resource management and hydropower planning. Seasonal variations in hydropower generation have also been documented in studies of the Amazon and Congo River Basins, emphasizing the global relevance of incorporating seasonal factors into hydrological planning [22].

It is crucial to address the benefit of correlating catchment area with the α parameter. The analysis revealed that incorporating area into the calibration of α significantly enhanced model performance, as evidenced by the improved PBIAS, NSE, and R² metrics. This approach underscores the importance of spatial scale in hydrological modeling and supports the hypothesis that larger catchments may require a higher α value to accurately simulate stream flow dynamics. The benefits of adjusting model parameters based on catchment size have been similarly observed in studies where spatial scale considerations have improved hydrological predictions [23].

The adjusted α parameter has direct implications for dam management. An optimized α parameter, tailored to the specific catchment characteristics, can lead to more accurate predictions of dam release flows. This improved predictive capability is instrumental for dam operation, particularly in terms of flood control, water supply management, and ecological conservation. By adopting a catchment-specific α parameter, dam managers can make more informed decisions that consider the unique hydrological responses of their specific catchment, leading to enhanced operational efficiency and reduced risk of water-related hazards. Incorporating catchment-specific parameters has similarly benefited dam operations in the Mississippi and Danube River Basins, where tailored management strategies have enhanced flood control and water supply reliability [24].

Beyond simply improving streamflow predictions, incorporating an area-adjusted α parameter holds significant power for dam management. In the realm of flood control, accurate release forecasts guided by catchment-specific α can play a vital role. By anticipating peak flood levels with greater precision, dam operators can adjust releases proactively, potentially minimizing downstream inundations and safeguarding communities. Incorporating catchment area into α calibration for river basin models led to improved predictions of peak flows, enabling more effective flood mitigation strategies [25,26].

Similarly, water supply management stands to benefit from the nuanced predictions facilitated by an optimized α. Understanding how catchment size influences streamflow persistence through α allows for the efficient allocation of water resources. Dam operators can balance reservoir storage with downstream demands more effectively, ensuring adequate water availability for agricultural needs, urban consumption, and industrial requirements. A study in the Ganges River Basin demonstrated that incorporating catchment area into α estimation improved irrigation water supply predictability, leading to enhanced agricultural productivity and water security [27]. Ataturk Dam shows significant seasonal fluctuations in hydropower production, with peaks typically occurring in the spring and early summer months. This pattern suggests a reliance on snowmelt and seasonal precipitation for water inflow. The plot indicates that hydropower production can vary significantly from year to year, highlighting the impact of annual hydrological variability. For example, years with higher peak production might correspond to years with abundant snowmelt or heavy spring rains, while lower production years could reflect drier conditions. Birecik Dam, despite its smaller size compared to the others, demonstrates clear seasonality in its power generation. The variations in production levels across different months and years suggest that Birecik, too, is affected by seasonal changes in water availability. Specific numerical trends might show slight increases or decreases in peak production over the years, indicative of changing hydrological patterns or operational adjustments. Karakaya Dam exhibits pronounced seasonal variability, with hydropower production peaks likely mirroring those of Ataturk, possibly due to similar geographical and climatic influences. The data might reveal certain years where the production notably exceeds the average, possibly due to exceptional hydrological conditions. Conversely, years with below-average production highlight the dam’s sensitivity to drought or reduced rainfall/snowmelt. Keban Dam’s hydropower production also shows clear seasonality, with the highest production typically in the wetter months. The inter-annual variability is marked, with some years showing significantly higher peaks than others, which could be attributed to variations in precipitation or operational changes. The trend analysis might reveal a gradual increase or decrease in average annual production, pointing towards broader climatic trends or shifts in water management strategies. Other regions, such as the Volga and Fraser River Basins, have also demonstrated similar seasonal and annual variability in hydropower production, underscoring the influence of climatic and operational factors on energy outputs.

However, it is important to acknowledge that while catchment size significantly influences α, other factors like geology, land cover, and climate also play a role. Considering these additional factors in conjunction with catchment area remains crucial for optimal model accuracy. Additionally, limitations exist in relying solely on an area-adjusted α approach. While it demonstrably improves predictions, alternative or complementary methods, such as incorporating other catchment properties or using machine learning models, may offer further improvement potential [28,29,30,31].

Despite these limitations, the benefits of adopting a catchment-specific α parameter are substantial. By fostering informed decision-making, this approach empowers dam managers to navigate the intricate interplay of flood control, water supply, and ecological conservation, paving the way for a more sustainable and harmonious relationship between humanity and the water resources upon which we all depend. The shift from increasing to decreasing trends in some dams (notably Karakaya) suggests changing environmental or operational conditions that may affect future hydropower generation potential. The significant decreasing trend (5.5%) observed in Ataturk Dam’s generation up to 2050 warrants attention for potential mitigation strategies to counteract this trend.

Birecik and Keban Dams show resilience with slight but consistent increases in generation across both periods with an average of 2.5% and 2.2%, respectively, indicating potential stability in their hydropower output. These trends underscore the importance of planning for future water resource management and the potential impacts of climate change on hydropower generation capacities. With the lack of streamflow gauging stations in Turkey and the lack of hydropower estimation techniques, the availability of an efficient equation in determining the available hydropower is crucial for determining the part that hydropower will play in the energy supply of Turkey in the near future.

5. Conclusions

This study provides a detailed assessment of Turkey’s hydropower potential in the Euphrates–Tigris Basin, emphasizing the implications of climate change and reservoir management. The empirical coefficient α was found to vary widely among the 14 major dams, with values ranging from −0.017 to 0.851, deviating significantly from the initially suggested 0.5. By incorporating a polynomial regression equation (α = 0.0012A − 0.043, R² = 0.72) to adjust α based on reservoir area, predictive performance improved markedly, with PBIAS reduced from 34% to 7.5% and R² increased from 0.439 to 0.765 for the Keban Dam.

Seasonal Mann–Kendall analysis revealed significant trends in hydropower generation. The Ataturk Dam showed a notable decrease of 5.5% in hydropower generation up to 2050, while Birecik and Keban Dams exhibited increases of 2.5% and 2.2%, respectively. These trends underscore the necessity for adaptive management strategies to mitigate the impacts of climate variability. The adapted equation facilitated accurate future hydropower projections for 2024–2100, aligning closely with observed streamflow data.

The study’s findings have significant implications for dam management. Adjusting the α parameter based on reservoir area enhances the accuracy of streamflow predictions, enabling better planning and optimization of water releases. This is critical for flood control, ensuring water availability during dry periods and maintaining ecological balance. Improved predictive capabilities can guide operational decisions, reduce the risk of water-related hazards, and enhance the efficiency of hydropower generation.

Furthermore, the research highlights the importance of considering seasonal variations in water management. For instance, addressing the observed decrease in Ataturk Dam’s hydropower generation requires strategic adjustments to accommodate changing inflow patterns. Implementing these findings can lead to more resilient and sustainable water resource management, crucial for meeting Turkey’s growing energy demands and adapting to future climatic challenges. This study contributes valuable insights for optimizing reservoir operations, ensuring long-term hydropower sustainability and enhancing the resilience of water resource management systems globally.

Author Contributions

Conceptualization, G.E.G.; methodology, G.E.G.; formal analysis, G.E.G.; investigation, G.E.G.; resources, G.E.G.; writing—original draft, G.E.G.; writing—review and editing, B.O.; supervision, B.O. All authors have read and agreed to the published version of the manuscript.

Funding

This research received no external funding.

Data Availability Statement

Data sharing is not applicable to this article as no new data were created or analyzed in this study.

Conflicts of Interest

The authors declare no conflicts of interest.

References

Zhao, T.; Zhao, J. Optimizing Operation of Water Supply Reservoir: The Role of Constraints. Math. Probl. Eng. 2014, 2014, e853186. [Google Scholar] [CrossRef]
Schultz, B. Role of Dams in Irrigation, Drainage and Flood Control. Int. J. Water Resour. Dev. 2002, 18, 147–162. [Google Scholar] [CrossRef]
Wu, X.; Xu, D.; Guo, R.; Wang, M.; Ying, Q. Long-Term Optimal Dispatching of Hydropower Station Based on Controlling the Frequency of Reservoir Level Fluctuation. Math. Probl. Eng. 2022, 2022, e8653470. [Google Scholar] [CrossRef]
Peng, C.; Buras, N. Practical Estimation of Inflows into Multireservoir System. J. Water Resour. Plan. Manag. 2000, 126, 331–334. [Google Scholar] [CrossRef]
Wan Ishak, W.H.; Ku-Mahamud, K.R.; Md Norwawi, N. Intelligent Decision Support Model Based on Neural Network to Support Reservoir Water Release Decision. In Proceedings of the Software Engineering and Computer Systems; Mohamad Zain, J., Wan Mohd, W.M., El-Qawasmeh, E., Eds.; Springer: Berlin/Heidelberg, Germany, 2011; pp. 365–379. [Google Scholar]
Preimpoundment Hydrologic Conditions in the Swatara Creek (1981-84) and Estimated Postimpoundment Water Quality in and Downstream from the Planned Swatara State Park Reservoir, Lebanon and Schuylkill Counties, Pennsylvania. Available online: https://pubs.usgs.gov/publication/wri884087 (accessed on 22 December 2023).
WGI Summary for Policymakers Headline Statements. Available online: https://www.ipcc.ch/report/ar6/wg1/resources/spm-headline-statements/ (accessed on 25 May 2024).
The World Bank. High and Dry: Climate Change, Water, and the Economy; © World Bank: Washington, DC, USA, 2016. [Google Scholar]
Hamududu, B.; Killingtveit, A. Assessing Climate Change Impacts on Global Hydropower. Energies 2012, 5, 305–322. [Google Scholar] [CrossRef]
Arriagada, L..; Rojas, O.; Arumí, J.L.; Munizaga, J.; Rojas, C.; Farias, L.; Vega, C. A New Method to Evaluate the Vulnerability of Watersheds facing Several Stressors: A Case Study in Mediterrenean Chile. Sci. Total Environ. 2019, 651, 1517–1533. [Google Scholar] [CrossRef]
Mutsindikwa, T.C.; Yira, Y.; Bossa, A.Y.; Hounkpè, J.; Salack, S.; Saley, I.A.; Rabani, A. Modeling Climate Change Impact on the Hydropower Potential of the Bamboi Catchment. Model. Earth Syst. Environ. 2021, 7, 2709–2717. [Google Scholar] [CrossRef]
Duratorre, T.; Bombelli, G.M.; Menduni, G.; Bocchiola, D. Hydropower Potential in the Alps under Climate Change Scenarios. The Chavonne Plant, Val D’Aosta. Water 2020, 12, 2011. [Google Scholar] [CrossRef]
Waheed, S.Q.; Alobaidy, M.N.; Grigg, N.S. Resilience Appraisal of Water Resources System under Climate Change Influence Using a Probabilistic-Nonstationary Approach. Environments 2023, 10, 87. [Google Scholar] [CrossRef]
Jeuland, M.; Whittington, D. Water Resources Planning under Climate Change: Assessing the Robustness of Real Options for the Blue Nile. Water Resour. Res. 2014, 50, 2086–2107. [Google Scholar] [CrossRef]
Yalew, S.G.; Kwakkel, J.; Doorn, N. Distributive Justice and Sustainability Goals in Transboundary Rivers: Case of the Nile Basin. Front. Environ. Sci. 2021, 8, 3–4. [Google Scholar] [CrossRef]
McCracken, M.; Meyer, C. Monitoring of Transboundary Water Cooperation: Review of Sustainable Development Goal Indicator 6.5.2 Methodology. J. Hydrol. 2018, 563, 1–12. [Google Scholar] [CrossRef]
Hanasaki, N.; Kanae, S.; Oki, T. A Reservoir Operation Scheme for Global River Routing Models. J. Hydrol. 2006, 327, 22–41. [Google Scholar] [CrossRef]
Guzey, G.E.; Önöz, B. Performance Assessment Comparison between Physically Based and Regression Hydrological Modelling: Case Study of the Euphrates–Tigris Basin. Sustainability 2023, 15, 10657. [Google Scholar] [CrossRef]
Pokharel, B.; Jagannathan, K.A.; Wang, S.-Y.S.; Jones, A.; LaPlante, M.D.; Buddhavarapu, S.; Borhara, K.; Ulrich, P.; Leung, L.-Y.R.; Eklund, J.; et al. Can We Rely on Drought-Ending “Miracles” in the Colorado River Basin? JAWRA J. Am. Water Resour. Assoc. 2024, 60, 813–824. [Google Scholar] [CrossRef]
Chen, Z.Q.; Ohara, N.; Anderson, M.L.; Yoon, J. A Regional Hydroclimate Model for Water Resources Management of the Tigris-Euphrates Watershed. In Proceedings of the World Environmental and Water Resources Congress 2008, Honolulu, Hawaii, 12–16 May 2008; pp. 1–10. [Google Scholar] [CrossRef]
Daggupati, P.; Srinivasan, R.; Ahmadi, M.; Verma, D. Spatial and Temporal Patterns of Precipitation and Stream Flow Variations in Tigris-Euphrates River Basin. Environ. Monit Assess 2017, 189, 50. [Google Scholar] [CrossRef]
Ikramova, M.; Akhmedkhodjaeva, I.; Khodjiev, A. The Reservoirs Capacity Assessment: The Tuyamuyun Hydro Complex in Khorezm Region of Uzbekistan. IOP Conf. Ser. Earth Environ. Sci. 2020, 614, 012100. [Google Scholar] [CrossRef]
Hasan, M.; Moody, A.; Benninger, L.; Hedlund, H. How War, Drought, and Dam Management Impact Water Supply in the Tigris and Euphrates Rivers. Ambio 2019, 48, 264–279. [Google Scholar] [CrossRef]
Metreveli, G.; Matchavariani, L. Research Method of Silting the Mountain Reservoirs Under the Current Climate Change. J. Water Resour. Ocean Sci. 2016, 5, 22–27. [Google Scholar] [CrossRef]
Turner, S.W.D.; Xu, W.; Voisin, N. Inferred Inflow Forecast Horizons Guiding Reservoir Release Decisions across the United States. Hydrol. Earth Syst. Sci. 2020, 24, 1275–1291. [Google Scholar] [CrossRef]
Razavi, T.; Coulibaly, P. Streamflow Prediction in Ungauged Basins: Review of Regionalization Methods. J. Hydrol. Eng. 2013, 18, 958–975. [Google Scholar] [CrossRef]
Anand, J.; Gosain, A.K.; Khosa, R.; Srinivasan, R. Regional Scale Hydrologic Modeling for Prediction of Water Balance, Analysis of Trends in Streamflow and Variations in Streamflow: The Case Study of the Ganga River Basin. J. Hydrol. Reg. Stud. 2018, 16, 32–53. [Google Scholar] [CrossRef]
Zhang, J.; Shang, Y. Nexus of Dams, Reservoirs, Climate, and the Environment: A Systematic Perspective. Int. J. Environ. Sci. Technol. 2023, 20, 12707–12716. [Google Scholar] [CrossRef]
Ehsani, N.; Vörösmarty, C.J.; Fekete, B.M.; Stakhiv, E.Z. Reservoir Operations under Climate Change: Storage Capacity Options to Mitigate Risk. J. Hydrol. 2017, 555, 435–446. [Google Scholar] [CrossRef]
Chen, B.; Krajewski, W.F.; Liu, F.; Fang, W.; Xu, Z. Estimating Instantaneous Peak Flow from Mean Daily Flow. Hydrol. Res. 2017, 48, 1474–1488. [Google Scholar] [CrossRef]
Ding, J.; Haberlandt, U. Estimation of Instantaneous Peak Flow from Maximum Mean Daily Flow by Regionalization of Catchment Model Parameters. Hydrol. Process. 2017, 31, 612–626. [Google Scholar] [CrossRef]

Figure 1. From top to bottom: major dams in the ETRB with the location of the 14 dams studied. The ETR basin studies with the corresponding streamflow stations present.

Figure 2. Example of data obtained for the Ataturk dam (from 2015) showing flow and hydropower generated.

Figure 3. The distribution of alpha (α) values generated by varying the standard deviation (σ) of the random variable ε. Different σ values were used to assess the sensitivity of α to variations in ε. While only these three ranges are shown for clarity, a broader range from 0.01 to 1 with a 0.02 interval was generated. The normal (Gaussian) distribution was used for ε, providing a realistic simulation of uncertainties.

Figure 4. Normalized 1:1 scatter plots for the predictive performance of the initial equation for all dams (α = 0.5). The scatter plots display the relationship between the normalized observed and normalized predicted values for the reservoir outflow across multiple dams. Each plot corresponds to a different dam, with the dam name and the coefficient of determination (R²) value indicated. The diagonal red line represents the perfect 1:1 relationship between observed and predicted values. Normalization is done to ensure consistency in comparing the magnitude of values across different dams and to reveal systematic biases.

Figure 5. Performance metrics (PBIAS, NSE, and R²) for different dams located along a river system or basin.

Figure 6. This graph illustrates the process of determining the optimal value of the parameter alpha (α) for the Keban Dam, based on minimizing different error metrics. The vertical axis represents the error values, while the horizontal axis shows the range of alpha values considered. The blue line corresponds to the PBIAS (Percent Bias) error, the red line represents the NSE (Nash–Sutcliffe Efficiency) error, and the green line depicts the R² error. The vertical dashed line highlights the chosen optimal alpha value of 0.851, which appears to minimize the overall errors across the different metrics for the Keban Dam.

Figure 7. Streamflow releases estimated from the HBV model with the adjusted parameters. The HBV model corresponds to RCP 8.5 projections, with NSE = 0.752.

Figure 8. Seasonal trends observed for the releases at each dam.

Figure 9. Hydropower generation time series for the dams.

Figure 10. Yearly rate of change for hydropower generation.

Table 1. Dams included in the analysis with the corresponding surface area.

Name	Area (km²)	Location (N E)
Yalnizardic	109.5	36.75 32.46
Seyhan	67.82	37.11 35.3
Kadinic	56.26	37.01 34.02
Adiguzel	25.9	38.18 29.09
Tahtali	23.52	38.09 27.08
Berdan	6.7	36.97 34.98
Gunesli	5.12	37.09 35.08
Oymapinar	4.7	36.92 31.54
Alakir	4.2	36.42 31.54
Gezende	3.97	36.53 37.24
Ataturk	817	37.48 38.31
Karakaya	268	38.255 39.13
Keban	675	38.67 39.49
Birecik	35	37.05 37.89

Table 2. Performance results of the predictive power of the release estimation with the adjusted parameters. Key metrics include PBIAS, NSE, and R² values, highlighting improvements or declines post-adjustment. Notable is the increase in model accuracy for dams like Adıgüzel and Oymapınar post-adjustment, as indicated by the improved NSE and R² values. The table also underscores the significant variability in model performance across different dam settings, emphasizing the critical role of parameter tuning in hydrological predictions.

Dam	PBIAS (α = 0.5, β = 2)	NSE (α = 0.5, β = 2)	$R^{2}$ (α = 0.5, β = 2)	α	β	${P B I A S}_{n e w}$	${N S E}_{n e w}$	${R^{2}}_{n e w}$	${R^{2}}_{n e w *}$ (α = 0.0012A − 0.043)
Keban	34%	2.95	0.439	0.851	7.21	7.5%	0.653	0.765	0.745
Karakaya	28%	4.12	0.582	0.252	6.12	16.2%	0.257	0.832	0.821
Ataturk	26%	3.52	0.612	0.90	5.13	12.3%	0.512	0.795	0.752
Birecik	22%	3.18	0.427	0.032	5.50	10.7%	0.703	0.810	0.790
Yalnızardıç	35%	2.85	0.512	0.086	6.75	8.9%	0.802	0.802	0.752
Seyhan	27%	4.12	0.375	0.036	7.15	7.6%	1.014	0.754	0.721
Kadıncık	19%	3.48	0.542	0.022	8.04	13.2%	0.908	0.702	0.700
Adıgüzel	43%	2.75	0.520	−0.005	6.75	10.6%	0.302	0.892	0.848
Tahtalı	36%	2.85	0.386	−0.017	5.25	9.8%	0.208	0.828	0.812
Berdan	32%	3.17	0.412	−0.027	4.98	7.9%	0.420	0.862	0.823
Güneşli	26%	3.25	0.289	−0.052	5.57	8.2%	0.286	0.758	0.727
Oymapınar	23%	3.47	0.410	−0.030	4.88	10.3%	0.201	0.883	0.873
Alakır	41%	2.75	0.427	−0.040	6.10	12.6%	0.508	0.791	0.780
Gezende	37%	2.98	0.539	0.048	6.42	11.2%	0.412	0.828	0.803

Table 3. Mann–Kendall test results for hydropower generation.

Years	2024–2050		2050–2100
Dams	$τ$	p-Value	$τ$	p-Value
Keban	0.089	4.03 × 10⁻⁵³	0.027	3.30 × 10⁻⁷
Karakaya	0.056	2.78 × 10⁻²²	−0.161	4.24 × 10⁻²³⁴
Ataturk	−0.264	3.00 × 10⁻²	−0.035	1.07 × 10⁻¹²
Birecik	0.092	1.74 × 10⁻⁴⁸	0.081	8.07 × 10⁻⁶¹

Table 4. Seasonal Mann–Kendall results

τ

Table 4. Seasonal Mann–Kendall results

τ

Seasons	Sep–Nov		Dec–Feb		Mar–May		Jun–Aug
Dams	$τ$	p-Value	$τ$	p-Value	$τ$	p-Value	$τ$	p-Value
Keban	−0.0599	0.4160	−0.0479	0.5160	−0.3302	0.0001	0.0736	0.3175
Karakaya	−0.1836	0.0125	−0.0878	0.2330	−0.3231	0.0001	−0.0462	0.5310
Ataturk	−0.0796	0.2790	0.0670	0.3630	0.2104	0.0042	−0.1858	0.0114
Birecik	0.1531	0.0435	0.0142	0.0061	0.2710	0.0003	0.0160	0.8350

Disclaimer/Publisher’s Note: The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content.

© 2024 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https://creativecommons.org/licenses/by/4.0/).

Share and Cite

MDPI and ACS Style

Guzey, G.E.; Onoz, B. Turkey’s Hydropower Potential in the Near Future and the Possible Impacts of Climate Change—A Case Study of the Euphrates–Tigris Basin. Climate 2024, 12, 156. https://doi.org/10.3390/cli12100156

AMA Style

Guzey GE, Onoz B. Turkey’s Hydropower Potential in the Near Future and the Possible Impacts of Climate Change—A Case Study of the Euphrates–Tigris Basin. Climate. 2024; 12(10):156. https://doi.org/10.3390/cli12100156

Chicago/Turabian Style

Guzey, Goksel Ezgi, and Bihrat Onoz. 2024. "Turkey’s Hydropower Potential in the Near Future and the Possible Impacts of Climate Change—A Case Study of the Euphrates–Tigris Basin" Climate 12, no. 10: 156. https://doi.org/10.3390/cli12100156

APA Style

Guzey, G. E., & Onoz, B. (2024). Turkey’s Hydropower Potential in the Near Future and the Possible Impacts of Climate Change—A Case Study of the Euphrates–Tigris Basin. Climate, 12(10), 156. https://doi.org/10.3390/cli12100156

Note that from the first issue of 2016, this journal uses article numbers instead of page numbers. See further details here.

Article Menu