Using a fish entrainment model assistant in a reservoir operation in China Open Access

Reservoir construction and operation impose significant changes in the flow transfer, sediment transport, and fish entrainment risk (ER) on intakes and spillways (Langford et al. 2015; Deng et al. 2017; Amaral et al. 2018; Rentschler et al. 2018). Researchers working on reservoir operations have shifted their focus from maximizing hydropower energy generation to protecting reservoir ecosystems. Among the new areas of focus are fish ER, fish protection technology, and suitable eco-friendly reservoir operation schemes (Langford 2016; Yang et al. 2016; Xu et al. 2017; Silva et al. 2018). This has made it timely and important to investigate the fish ER in reservoirs, especially the forebay of reservoirs (Rodriguez et al. 2015; Sedighkia et al. 2021). Additionally, to ensure that hydropower operations are eco-friendly, it is necessary to combine both hydrodynamic factors (velocity, water depth, turbulence, kinetic energy, temperature, etc.) and biological factors (fish species types, life stage, etc.) to simulate the fish ER.

Fish entrainment is the term used to describe the loss of fish whenever people divert their streams, creeks, or rivers. When the fish follow the flow of diverted water, they become entrained. Fish entrainment results in the loss of fish and affects the fish community and riverine ecosystems (Mork & Council 2014; Olden 2016; Eslamian et al. 2018; Xue et al. 2019). Fish entrainment is caused by various factors including high velocity, turbulence density, turbine blades, and high sediment concentrations. Many studies have examined fish mortality caused by hydro turbines (Coutant & Whitney 2000; Langford 2016; Yin et al. 2020). Researchers have also focused on how to reduce the rate of fish entrainment in the intakes (Cornett et al. 2015; Huang et al. 2015; Rytwinski et al. 2017; Stoot et al. 2018). Once the fish are located inside the hydropower intake, there is a high probability of entrainment into the turbines, and the fish will likely be killed because the flow velocities and turbines are so powerful.

Field measurements are the most direct way to evaluate the forebay of reservoir fish ER. These can be traced to the 1960s (Marghzar et al. 2003; Khan et al. 2004; Bryant et al. 2008) and mainly focus on fish species behavior under various hydrodynamic velocities and turbulence conditions (Liao 2006; Pavlov et al. 2008; Trinci et al. 2017). Evaluating the risk of fish entrainment from a hydrodynamic perspective is relatively new. However, over the last decade, numerical modeling methods have become more prominent in evaluating the risk of fish entrainment. Reservoir hydrodynamic simulations are the foundation of fish ER modeling. Software and tools available to calculate reservoir hydrodynamics include ANSYS, Flow3D, Mike system, OpenFOAM, StarCD, etc. (ADAPCO 2004; Hirt 2011; DHI 2014; Chen et al. 2014a, 2014b; Stolarski et al. 2018). Hydrodynamics research provides the hydraulic characteristics near water intakes, spillways, and flood discharge tunnels. These hydrodynamic characteristics are the basis for fish entrainment and can quantify the fish ER.

It is well known that fish ERs are related to reservoir operation and water elevation. However, the ER is not been quantified. The main aim of this study is to propose a fish entrainment model framework to evaluate the Schizothorax and Garra fish (Garrapingi) ER zone at the Dawei Reservoir operation. The first part of the study involved evaluating the fish entrainment zone under various intake operations and spillway operation scenarios in wet, normal, and dry years. The second part involved determining the maximum and minimum fish ERs under various operation schemes and different water levels. The last part involved predicting the long-term reservoir operation effects on the fish entrainment zone.

The Dawei Reservoir (E100 °10′–102 °00′, N31 °42′–33 °37′) is located in southwest Sichuan Province in China. The forebay of the Dawei computational domain is approximately 7.77 × 10⁵ m². The mean width and length are 1.42 × 10³ and 5.38 × 10² m, respectively (Figure 1). The Dawei dam hydropower system consists of five individual turbines (3 × 310 and 2 × 306 MW) and three spillway gates. The intakes are situated on the right side and the spillway is situated on the left side of the river basin. The seasonal pattern of forebay surface elevation in the wet, normal, and dry years range from 2,638 to 2,680 m, with the highest reservoir surface elevation occurring in late summer and autumn and the lowest reservoir surface elevation occurring in late winter and early spring. The mean annual rainfall and the evaporation in the Dawei Reservoir are 896 and 875 mm, respectively. The study area of the Dawei Reservoir geological area has three types of geological materials: conglomerate, cataclasite, and other surficial deposits. The riverbed materials are composed of an active layer and a substrate layer according to their erosion, deposition, and geological structure (Li et al. 2012).

The Dawei Reservoir plays a significant role in creating and maintaining diverse hydrodynamic conditions for many fish species including endemic fish species and several endangered fish species. These include Schizothorax (Racoma), Euchiloglanis davidi (Sauvage), Schizothorax, and Leptobotia elongata (Bleeker). Among these fish species, the Schizothorax and Garra fish (Garrapingi) represent the dominant and endangered fish species, respectively. A full list of fish species known to occur in the Dawei Reservoir can be found in the Supplementary materials of Yao (2016). Each year, thousands of fish are entrained into the intakes and spillways and are injured by the high-speed flow and hydro turbines. The different intakes and spillway operation scenarios may cause different hydrodynamic conditions. These conditions are the main factors contributing to fish ER in the Dawei hydropower system (Li et al. 2018).

The modeling and analysis of fish entrainment in the forebay of a reservoir is often a complex task. The proposed model must capture the key variables and describe the specific problem in the reservoir ecosystem. The fish entrainment model in this study includes a reservoir hydrodynamic module, reservoir long-term operation module, fish entrainment rules setup, and fish entrainment analysis module. The flowchart of the module framework is shown in Figure 2. The model can predict the fish ER zone and define the fish-friendly dam operation in the Dawei Reservoir. It can also predict the fish ER zone in the Dawei Reservoir based on dam operation schemes.

The reservoir bathymetric data, intakes and spillway outlines, and reservoir operation schemes can be obtained from engineering drawings, a field survey and the scientific report of the Dawei Reservoir domain. Input data for the hydrodynamic module simulation came from the reservoir water elevation in all seasons, bathymetric data, substrates data, intakes, and spillway operation schemes. The two-dimensional hydrodynamic modules were used to predict the velocity distribution and turbulence kinetic energy. After the hydrodynamic variables were simulated, the fish species ER curves in the Dawei Reservoir were established. The fish ER curves, shown in Figure 3, are based on field survey work, a literature review, and professional judgment. After the fish species ER curve is established, the fish entrainment risk values (ERVs) under all reservoir operation schemes can be predicted. The evaluating fish ER interaction process is shown in Figure 4. The module theory mentioned here will be described in detail in the following sections.

After velocity and turbulent energy were obtained, the fish ER was calculated based on the Schizothorax and Garra fish (Garrapingi) risk curves in Figure 4. The value for ER ranged from 0 to 1, with 0 representing the lowest ER and 1 representing the highest ER. The fish ERV was defined based on the fish ER by applying equal weight to the hydrodynamic variables of velocity and turbulent energy. The ERV was used to evaluate two fish ERs under various reservoir operation scenarios.

The fish ER at the Dawei Reservoir for all intakes and spillway operation scenarios was predicted using the fish entrainment model. In this study, the typical seasonal intake operation schemes (four, five, four, and two turbines operating during the spring, summer, autumn, and winter, respectively) under wet, normal, and dry years were calculated. In order to determine the highest and lowest Dawei Reservoir operation schemes, a system-wide risk entrainment scenarios in the reservoir were predicted. The fish ER is based on the water elevation (from 2,630 to 2,670 m with 10 m interval) and all intake operation schemes (one, two, three, four, and five intake gates opening) (Figure 4). In addition, the fish ER under spillway operation schemes (Spillways 1 and 3 operated under water levels of 2,655 and 2,665 m) were also predicted. As part of this study, the effects of the long-term operations at the Dawei Reservoir, including the effects of erosion and deposition, were also investigated. Thus, a total of 31 scenarios in this study were simulated. The ERV, WERS, OFRV, DRZ, MRZ, and LRZ were calculated for each scenario.

The seasonal pattern of velocity and turbulent energy distribution for the Dawei Reservoir under wet, normal, and dry years are shown in Figure 5. The simulation results indicated that the areas affected by velocity and turbulent energy increased from spring to summer and decreased from autumn to winter. The velocity and turbulent energy affected areas under the wet year were bigger than during the normal and dry years. The maximum velocity near the intakes was 1.4, 1.0, and 0.8 m/s under the wet, normal, and dry years, respectively. The maximum turbulent energy was 0.16, 0.15, and 0.14 J/kg under the wet, normal, and dry years, respectively. Notably, the velocity and turbulent energy affected areas occupy a small proportion of the whole computational area.

The results of the fish ER zone distribution based on seasonal intake operation schemes at the Dawei Reservoir are depicted in Figure 6. For the Schizothorax, the HRZs for spring, summer, autumn, and winter were 9.75 × 10³, 1.13 × 10⁴, 9.75 × 10³, and 6.81 × 10³ m², respectively. The MRZs for the spring, summer, autumn, and winter were 1.49 × 10⁴, 1.32 × 10⁴, 1.49 × 10⁴, and 9.99 × 10³ m², respectively. For the same time periods, the HRZs were 1.26, 1.46, 1.26, and 0.88% and the MRZs were 1.92, 1.70, 1.92, and 1.29%, respectively.

For the Garra fish (Garrapingi), ER showed similar trends, while the HRZs for all four seasons were slightly bigger than those in the Schizothorax ER zone (Figure 7). The overall distribution trends for the HRZ were similar for all high fish ERVs located near the intakes. This means that the Schizothorax and Garra fish (Garrapingi) ER is low at the Dawei intakes during typical seasonal operations.

The simulated fish ERVs at each scenario based on different water elevation and intake operation schemes are depicted in Table 1. According to the simulation result, the maximum fish ER operation occurred with five intake operations at a water level of 2,670 m. The HRZs for the maximum fish ER operation were 1.87 and 3.56% for Schizothorax and Garra fish (Garrapingi), respectively. The maximum fish entrainment areas were 1.44 × 10⁴ and 2.78 × 10⁴ m² for Schizothorax (Schizothorax) and Garra fish (Garrapingi), respectively (Figure 8). In contrast, the minimum fish ER operation occurred with two intakes at a water level of 2,640 m. The HRZs for the maximum fish ER operation were 0.36 and 0.88%, and the corresponding areas were 2.77 × 10³ and 6.78 × 10³ m² for Schizothorax and Garra fish (Garrapingi), respectively (Figure 9). The fish ER analysis results indicated that the difference between these two scenarios is not significant compared to the entire forebay reservoir area.

The spillway operation scenarios had significant effects on both Schizothorax and Garra fish (Garrapingi) ER. The impacts were not significantly modified by the Dawei Reservoir's spillway operation pattern (Figures 10 and 11, Table 2). Specifically, increasing the spillway gate operation was associated with increasing the high fish ER zones. The spillway operation under normal reservoir water surface elevation and designed reservoir water elevation also had a strong effect on fish ER, but there was no significant difference in the fish ER pattern. In addition, when considering the hydrodynamic variables as a measure of fish ER at the forebay of the Dawai Reservoir, the velocity overrides other variables and contributes to the high ER to both Schizothorax and Garra fish (Garrapingi).

To exclusively analyze the relationship between reservoir operations and fish ERs, Schizothorax and Garra fish (Garrapingi) ER zones were simulated in the model. The simulation included 20 years of the Dawei Reservoir hydromorphology effects on velocity distribution, erosion, and deposition and fish entrainment. The results revealed that the long-term hydromorphology effects strongly influenced the reservoir bottom erosion and deposition, but had no significant effects on velocity distribution and Schizothorax and Garra fish (Garrapingi) ER (Figure 12, Table 3). The maximum erosion and deposition occurred in the areas near the dam and in the middle of the computational domain with values of 10 and 15 m, respectively. However, the velocity field on the forebay of the reservoir was not significantly influenced by the sedimentation and erosion processes or the deposition effects. Thus, in general, the long-term Dawei Reservoir regulation had no significant effects on Schizothorax and Garra fish (Garrapingi) ER.

Our simulation results demonstrate that hydropower operation-induced fish entrainment can have different impacts on Schizothorax and Garra fish (Garrapingi) depending on a reservoir's regulation pattern. When the Dawei Reservoir operation scenarios were applied, the fish ER showed different responses to the intakes and spillway operation patterns. These quantitative analyses of reservoir operation patterns have important implications for managing the effects of fish entrainment caused by hydropower operations in reservoirs.

The Schizothorax and Garra fish (Garrapingi) ER zone predicted in this study based on seasonal intake operation patterns matched well with the trends surveyed and recorded by Li et al. (2018). The sampled Garra fish (Garrapingi) also displayed a distinct seasonal pattern on the forebay of the Dawei Reservoir. These findings suggest that the ER for both fish species reaches the highest level during the summer, when the reservoir water elevation and the number of operating intakes are highest.

Due to the complexity of the Dawei Reservoir operation, it is difficult to validate the maximum and minimum ER based on intake operations and reservoir surface elevation. Because of the restrictions and limitations imposed by reservoir management regulations and life cycles, it is also difficult to validate the fish ER based on spillway and long-term hydromorphology effects. Nonetheless, the simulation results can provide valuable quantitative fish entrainment information for reservoir operation planning, operation optimization, and long-term predictions. This information is also strongly needed for fish-friendly hydraulic facility design, optimization, and operation (Thorncraft et al. 2013; Airody et al. 2017; Silva et al. 2018; Samma et al. 2020). Our findings indicated that the long-term hydromorphology had no significant impact on fish entrainment effects. However, the spillway operation pattern had the clearest entrainment effects on the forebay of the Dawei Reservoir. The simulated information could be valuable to hydraulic engineers, scientists, and hydropower managers (Hu et al. 2008; Holm 2017; Wild et al. 2019), trying to balance anthropogenic reservoir operation demands and fish conservation within a sustainable water resources management framework (Silvano et al. 2009; Yin & Yang 2011).

The fish entrainment model could be helpful in a fish-friendly Dawei Reservoir operation and be used to optimize the current operation schemes, decreasing fish ER while sustaining hydropower production. Based on model prediction and application, the following measures could be applied. First, because of the fish entrainment, the risk zone would be increased along with the operation of intakes and the increase in water level. Reservoir surface water elevation would have to be kept at a suitable level and intakes would have to be kept relatively stable. The reservoir management authority also recommends such measurements, along with using the reservoir capacity volume to increase the time period for low fish entrainment and decrease the period for high fish entrainment (Li et al. 2018). Because spillway operation schemes lead to significant entrainment impacts on Schizothorax and Garra fish (Garrapingi), spillway operations should be limited when a fish-friendly hydropower operation pattern has been considered or applied. Current reservoir operation guidelines also suggest reducing the operation of the spillway, especially during the spawning season. Moreover, because the long-term reservoir operation would not significantly affect the ER for these two species, it is suggested that the reservoir operation pattern should be adjusted when the appropriate pattern is determined.

The balance and optimization of a reservoir operation are not only important to the Dawei Reservoir, but also to all reservoirs that experience fish entrainment, especially large reservoirs. It is estimated that there may be 16.7 million reservoirs worldwide, more than 59,000 of which are behind large dams (Lehner et al. 2011; ICOLD 2018). Many studies have documented negative impacts on fish entrainment caused by intake and spillway operations and reservoir sediment discharge (Miranda 2001; Clarke et al. 2008; Dreyer 2018). The fish entrainment model provides a new way to determine the quantity and quality of risk, balance, and optimization between fish ecosystems and hydropower generation. The analysis between reservoir regulation and fish entrainment helps to understand the impacts of fish ER associated with reservoir regulation.

Although this fish entrainment model is useful for providing detailed quantitative information about fish-friendly reservoir operations, it requires further validation, especially of real fish data, since there is no surveyed fish number to calibrate and validate this model at the current stage in the study area. In addition, currently there is no comparable fish data on reservoirs, which may reduce confidence in the model.

Nevertheless, the model proved to be a very useful tool. It simulates the fish ERV based on fish risk curves, which is suitable to deal with vague and rough data or knowledge. In addition, work has begun on a fish entrainment monitoring program to address the fish ER problem by enabling us to address the model's weaknesses and to increase the predicting accuracy. Thus, in the next stage, when more data are gathered and more testing done, the model efficiency and accuracy could be improved.

This study presents the results of fish ER modeling in the Dawei Reservoir in China. Modeling was performed in order to better understand the fish ER in the forebay of the reservoir in response to reservoir regulation. The models used were able to predict seasonal and spatial distribution reasonably well for fish entrainment zones in the Dawei Reservoir. The modeling results can be used to help in management decision-making for the Dawei Reservoir.

The study shows that among all proposed reservoir operation scenarios, the intake operations and long-term reservoir operation patterns have no significant impact on the forebay of the reservoir, while the spillway operation pattern has significant entrainment impacts. The development of fish ER modeling in the Dawei Reservoir should focus on fish entrainment data testing and suitable fish-friendly reservoir management. These model results were specific to the Dawei Reservoir, but can be extended to other reservoir and water bodies with a high degree of robustness. As highlighted by the hydropower project development on eco-friendly management, the urgent need to focus attention on reservoir development and operations required the fish ER evaluation model. If unsuitable reservoir management is applied, the whole water body and fish community could become unstable and fragile. Fish ER assessment and evaluation models in the reservoir system are likely to become the foundation for efforts to prevent the reservoir from causing ecologically damage.

We thank W. Zhang and Y. Zhong for helping in the field. The authors thank the BC-Hydro and Hydro-China for valuable discussions.

All relevant data are included in the paper or its Supplementary Information.

The authors declare there is no conflict.