| Forecasting the Nile River flow is of vital interest for African nations such as Sudan and Egypt. These nations use the Nile's water for agriculture and hydropower. An accurate forecast of water availability would be beneficial for efficient management of water resources facilities such as reservoirs and diversions. Any improvement in the forecast accuracy or increase in prediction horizon will have a significant influence on improving the water management in these nations.; This research makes intensive use of sea surface temperatures as predictors. Linear correlation analysis was used to establish the connection between the Nile River flows and the leading climatic indicators. Multiple linear regression, principal component regression, canonical correlation analysis and artificial neural network were used to develop the forecast models. The statistics for equations based on principal component analysis showed improvement in the forecasting accuracy over the equations developed using original variables. Multivariate models improved the performance measures over univariate models.; A crucial step in developing empirical formulae for long-range forecasting requires the selection of appropriate predictors. This dissertation has been guided by an objective search among a large number of predictors. The sea surface temperature gradient between two locations (referred to as oscillation, dipole, or seesaw) emerged as stable and consistent predictors.; In comparison with the Atlantic and Indian Oceans, the Pacific Ocean shows the highest contribution to the variability of the Blue Nile River flows. Its contribution is as high as the contribution from the three oceans combined. The Nile River flows at Aswan, as expected, show similar patterns with lower adjusted R2 magnitudes. The Atbara River shows higher response the large-scale variables, followed by the Blue Nile River, the Nile River at Aswan and the Sobat River.; The major contribution of this research is the development of further knowledge in the field of hydroclimatology, and in doing so, the development of long-range streamflow forecasting models. Future work includes the incorporation of the real time operation and investigation of the stable predictors in relation with large-scale phenomenon and other hydroclimatic variables. Further, tests should be on large rivers flows, such as the Mississippi and the Amazon. |
