Théodora Cohen Liechti
Worldwide, almost 50,000 dams over 15 m height have been built during the last six decades with an aggregated storage capacity of 6,000 km3. The fact that large dams, by increasing irrigation and hydroelectricity production, can sustain development and reduce poverty has led developing countries to undertake major investment in dam construction. However, scientific progress is required in order to design and manage the future complex hydrologic/hydraulic systems in sustainable ways. The African Dams project (ADAPT) aimed at enhancing the scientific basis of integrated water resource management. New models for the real-time control and multi- objective optimization of large hydraulic structures were created and data resources enhanced through field survey. The case study considered to apply this new knowledge is the Zambezi River basin located in southern Africa. It contains many of southern Africa’s largest and most intact freshwater and estuarine wetlands, e. g. the Kafue flats, the Mana Pools and the Zambezi delta as well as several free-flowing yet unprotected river reaches. Three of Africa’s largest dams (Kariba, Cahora Bassa and Kafue) inundate hundreds of square kilometers of river habitat and modify the natural flow patterns that sustain floodplains. Increasing human activity by cities and industry is causing a regional energy shortage, leading governments and investors to plan yet more dams in the basin. In the framework of the ADAPT project, the major contribution of the present research is to set-up a hydrologic-hydraulic model of the whole catchment area which includes all relevant elements as hydraulic structures and schemes as well as floodplains. The multi- objective modelling and simulation define how dam operation can be adapted to get the highest environmental results under highest energy production. Three main steps structure the project: (1) the evaluation of the quality of available input data, (2) the definition of the specific hydrological processes needed for hydraulic-hydrological modeling of the Zambezi Basin along with the establishment of a calibration outline, (3) the assessment of the impacts of the planned new hydraulic structures and the refurbishment of the existing hydropower plants on the flow regime at critical points of the basin. At first, three operational and acknowledged satellite derived precipitation products (the Tropical Rainfall Measuring Mission product 3B42 -TRMM 3B42-, the Famine Early Warning System product 2.0 -FEWS RFE2.0- and the National Oceanic and Atmospheric Administration/Climate Prediction Centre (NOAA/CPC) morphing technique -CMORPH-) are analyzed in terms of spatial and temporal distribution of the precipitation. They are compared to ground data at daily, 10-daily and monthly time steps. Based on the results, TRMM 3B42 was chosen as input data for the hydrological modeling. Secondly, the Soil and Water Assessment Tool (SWAT 2009), a semi-distributed physically based continuous time model, was selected to simulate the hydrology of the basin. Due to the specificities of the Zambezi River basin, two main additional functions were developed. (1) A floodplain sub-model based on a reservoir approach was implemented. The model separates the outflow of the reservoir simulating the floodplain into main channel flow and flow over the floodplain area. (2) The hydropower plant operations are simulated based on the rule curve and the technical characteristics of the dams. The pertinence of the implemented approach was verified by modeling the existing hydropower plants. Given the complexity and the size of the basin, an automated calibration procedure based on A Multi- ALgorithm Genetically Adaptive Multi-objective method (AMALGAM) was applied to optimize the relative error and the volume ratio at multiple discharge stations. The observed volume at the artificial reservoir derived from the measured water level was included in the calibration. Thirdly, scenarios combining different levels of environmental requirements as well as multiple hydropower development schemes were simulated at a daily time step with the hydraulic-hydrological model. The hydropower operation rules are simulated in detail. The mean annual energy produced, the firm power and the spilled volume during flood season are computed for each scenario. The impact on flow regime is characterized by Pardé coefficients, a set of indicators based on the Range of Variability Approach (RVA) and duration curves. In a global perspective, the analysis shows that it is possible to reach a compromise between energy production and environmental sustainability. Finally, the data of two Global Circulation Models (GFDL-CM2.0 and CCCma- CGCM3) for the emission scenario SRES A2 of the IPCC report were used to simulate the hydrological input during 2045-2065 and 2080-2100 periods. The prediction of the climate models diverge in terms of precipitation as GFDL-CM2.0 forecasts an increase while CCCma-CGCM3 a diminution but agree on the increase of temperature. The impacts of climate change were assessed both on energy production and flow pattern changes. The development and application of a hydraulic-hydrological model to a large river basin in a context of data scarcity and particular hydrologic units is particularly important for water resource management. The use of open source data and adapted calibration tools allows operators, authorities and researchers to assess together the impacts of hydropower development.
EPFL2013
