Nowadays, due to the climate change,
urban development and other human interventions, flood is still a major
challenge to humanity.
In recent years, some infrastructure plans to prevent the downstream areas from flooding and to increase the flood control are proposed and implemented by the government like constructing long dikes along the river. Fully implementation of the plan, constructing these measures, is delayed due to doubt on real cause of the problem of flooding in addition to a need of huge investments. The concern of the present research is to identify causes of flooding and to develop flood management strategies to reduce disaster.
For this reason, a MIKE 11 1D hydrodynamic model is built downstream of the reservoir to identify the causes of flooding by providing different boundary conditions and compare and identify critical sections on the reach using five years of measured stream flow data. The results showed that both the discharge from Koka and tributaries cause the flooding. Therefore, operation rule scenarios of the reservoir and downstream flood forecasting models were developed.
Result shows that without endangering the power supply, peak discharge
reductions up to 300 m3/s can be achieved
by advance release
Flood forecasting on the main tributary with a lead-time of 3 days for warning purpose using ANN and M5 model tree was developed. It shows that discharge peaks can be forecasted several days in advance. However, but with the present models, some peaks are underestimated or lagged. Further research can improve the accuracy of the models.
The overall experimental result showed that non-structural measures like reservoir management and flood forecasting could reduce flooding damage of farmlands and industries in Awash basin. A joint application of structural measures and non-structural measures could solve the flooding problem.
Key words: Awash basin, ANN, Koka reservoir, M5, MIKE 11, Flood control, flood forecasting and flooding