黄河中游吴堡-龙门段河道洪水概化预报研究.

Combining flood hydrograph generalization and Support Vector machine for Regression ( SVR) , a new gradual and continuous channel flood forecasting technique was presented in this paper. A research about channel flood forecast was performed on the Wubu-Longmen reach of the middle Yellow River. Twenty eight flood events during the period between 1980 and 2000 were used to generate the flood hydrograph model and calibrate the SVR-GA models applied to forecast flood peak and duration. With the amplifier of forecasted flood peak and duration ,the flood hydrograph model was amplified to obtain the forecasted flood hydrographs. Five flood events during the period between 2001 and 2006 were forecasted to test forecast accuracy of the new technique. The results indicate that the forecast accuracy of the new technique increases continuously by using the updating relevant data. In addition, forecasted flood hydrograph performes well during the whole flood process. [ABSTRACT FROM AUTHOR]