Introduction
Sedimentation processes in Sakuma reservoir in the Tenryuu River in Japan are gradually reducing the active storage of this reservoir. Plans for sediment management in Sakuma Reservoir include the excavation and downstream displacement of sediment. The removal of sediment from Sakuma reservoir reach will increase the storage capacity, thus contributing to the restoration of the flood-control function of the dam. The amount of sediment to be displaced (bypassed) are significant, and the impact of these measures are relevant for the overall behaviour of the full river system. To assess the impact of such measures, a 1D-morphological model was developed by (Sloff & Mosselman, 2009) using the SOBEK-RE modelling system. Of particular interest is a comparison of the morphological response of the river to a large sediment bypass with a low frequency (e.g. once per year), to the response due to continuous sediment bypassing with a small magnitude.
The presence of the hydropower dams as well as the large flux of suspended sediment are important characteristics, which define the bed level behaviour in the Tenryuu River. The new Delft3D-Flexible Mesh software includes a morphology module, with bed and suspended sediment loads. It is also possible to couple with the Real-Time Control module (D-RTC) in order to control the structures in the model. Therefore, it was decided to make the transition from the SOBEK-RE modelling system to the newer software package. In 2018, a start was made to convert the existing SOBEK-RE model, resulting in a one-dimensional hydrodynamic model using Delft3D-Flexible Mesh. In 2019, the work is continued to set up a morphodynamic model. The main advantages of Delft3D-Flexible Mesh over SOBEK-RE, are the possibility to include suspended sediment, and to connect the flow-model to other modules such as D-Real Time Control (D-RTC), which enables the user to control hydraulic structures in several ways. In a later stage, the connection can be made to the water quality module.
Aim
J-Power, aims to evaluate different sediment management options on a system scale. For that, we need a model that is able to reproduce the long-term bed development and to predict the morphological response to different types of sediment management strategies and gate operation protocols. Coupling the morphodynamic model to D-RTC allows for the simulation of different sediment management strategies. This study aims to set up such a model, covering the Tenryuu River between Hiraoka Dam and the river mouth. Scenario analyses are not in the scope of this project.
System description
The modelling reach considered within this project covers the Tenryuu River from Hiraoka Dam until the river mouth at Hamamatsu City. This reach has a length of 103 km and a bed level elevation ranging from -6 m+MSL at the mouth to 263 m+MSL at Hiraoka Dam. The reach also includes three hydropower dams:
- Sakuma Dam
- Akiba Dam
- Funagira Dam.
The upstream boundary of the model area is located at Hiraoka Dam (see Figure 3.1), which is about 34 km upstream of Sakuma dam. Due to sedimentation of the reservoir upstream of Hiraoka Dam, only a minor storage pool is left. Therefore, it can be assumed that all sediments from upstream pass Hiraoka Dam. The power station of this dam is still in operation, but during high flows the gates are opened, and sediments can freely pass through the dam.
The model also includes multiple tributaries, of which the most important are (from upstream to downstream): Ochise River, Misakubo River and Keta River. Ochise and Misakubo River enter the Tenryuu River between Sakuma Dam and Akiba Dam. The sediment supply coming from Ochise, Misakubo, and Keta rivers is relatively large. At the confluences, bars of gravel and sand mixture have been observed. During high flows the amount of fine sediments coming from these rivers is rather significant.
Model set up
Multiple bed level measurement campaigns have been carried out in the past years. However, we decided to set up the model with bed levels that are equal to those in the SOBEK-RE model. These bed levels are based on measurements from 2007. This has the advantage that we can compare the new model results with the old SOBEK-RE model. Also, we can validate the modelled long-term morphological behaviour based on the bed levels measurements of subsequent years.
The Tenryuu River downstream of Hiraoka Dam, includes three main tributaries: the Ochise, Misakubo, and Keta rivers. These rivers are included as branches, each having a length of 300 m, which is short, but sufficient to mimic their effects on the Tenryuu River.
The model includes three hydropower dams, which can be controlled by specifying a time series (similar to the SOBEK-RE system), or by feedback-control, via the Real-Time Control module.
A Manning roughness type is chosen for all branches; the initial roughness values were set to match the Sobek-RE model. In order to match the observed water levels at different discharge conditions, the manning roughness is made discharge dependent.
Morphodynamic calculations were carried out with graded sediment. The initial bed composition was updated to match the observed grain size distribution
As an added functionality with respect to the SOBEK-RE model, we included the transport of suspended sediment as silt.
Results
The SOBEK-RE model was successfully migrated to a Delft3D FM model, including hydrodynamics, morphodynamics, and structure-control. In general, it is concluded that the current Delft3D FM model is suitable for hydrodynamic and morphodynamic calculations on a system-scale.
Furthermore, Delft3D FM allows for coupling between flow calculations in 1D, 2D and 3D, and for coupling between flow, morphology, real-time control, and water quality simulations. For future applications, this implies that a local high-resolution model can be coupled to the 1D model. This would allow for an assessment of gate operation and local sediment dynamics, within a system-scale model.