Add a table with the project cost break-down
Model closure
Flexible exchange rate with fixed foreign savings
Savings driven
CPI numeraire
Flexible government savings
Impoundment and operations phase assessment
In this study, hydro-energy generation from Bui and Akosombo dams was estimated in two distinct periods using different approaches. For the period 2009 to 2016, actual energy generation data were used for the scenario of Akosombo with Bui online. For the same period, the initial impoundment volume of Bui Dam was passed to Akosombo Dam to assess energy generation with Bui offline. This was done using a water balance model driven by reservoir water level data and actual energy generation data. Water level data for Akosombo were obtained from the Copernicus Land Monitoring Service \cite{CGLS2019}, which uses radar altimetry to estimate the water levels of several water bodies worldwide. For Bui Dam, the water level data were derived from Landsat satellite images. We refer the reader to \citet{Basheer_2019} for a detailed description of the methodology used to derive reservoir water levels from satellite imagery.
For the period 2017 to 2067, a calibrated water balance model was used to estimate hydro-energy generation from the river system for the scenarios with and without the Bui Dam online. The model was driven by 50 stochastic streamflow conditions that were developed by bootstrapping from the historic flow record \cite{Efron_1992}. Figure 6(a) shows a schematic of the model developed herein, which includes three inflow nodes (Black Volta, White Volta, and Oti), two storage reservoirs (Akosombo and Bui), three channel loss nodes, and evaporation from Bui and Akosombo reservoirs. In the calibration and validation periods (1993 to 2013), the model was driven by the historic river flow data and the operating rules of dams. We are aware that several small-scale dams are located upstream from the model domain. We have not included them in the model due to unavailability of data and assuming that they would have little impact on Akosombo. This assumption is valid judging by the model performance, as shown in Figures 6(b-d) and Table 3. The figures and the table show a good correspondence between the simulated and satellite-observed water levels and suggest only a slight systematic error. Monthly net evaporation rates were obtained from XXX and used to estimate evaporation from Bui and Akosombo reservoirs. Furthermore, elevation-volume and elevation-area tables for each of Bui and Akosombo reservoirs were acquired from XXX and then populated to the model. The historic river flow data were obtained from XXX, which provides a flow record based on XXX. Throughout the model period, Akosombo was operated to target a constant power generation rate while keeping the reservoir water level higher than the Minimum Operating Level and lower than the Full Supply Level. Because the initial filling of Bui Reservoir was in 2011-2013, the observed reservoir water levels of Bui were used to model the operation of the dam in this period.