Study On The Application Of Xin'anjiang Model For Flood Forecasting In Medium-sized Reservoir

Flood forecasting plays an important role in all non-engineering flood control measures. Timely and accurate flood forecasting can assist decision-making on flood control effectively, so as to minimize or avoid flood damage. In some cases it even may turn the harmful into the beneficial. Therefore, flood forecasting has always been attracting the governments and various research institutions'attentions. It has been making important contributions to assure people's lives and properties as well as the sustained and healthy development of the national economy.Xin'anjiang model that used widely in the humid and sub-humid regions of China is representative in various models. People think that the model is appropriate to flood forecasting especially in the southern part of China. In this paper, a typical medium-sized reservoir in Jiangxi Province-Panxi reservoir is selected to validate the model. Based on the Xin'anjiang model, Panxi reservoir flood forecasting system is set up.Luoxi hydrologic station is located next to the Panxi reservoir. Based on the Luoxi watershed (Controlled by Luoxi hydrologic station), Day Model and Case Model are studied. As for the Day model, the author selected series data of 1972-1984 for fixing parameters and 1985-1987 for testing the precision. Results show that both runoff depth and deterministic coefficient hit the qualified rate of 100% in the above two periods. As for the Case model, the author set 50 floods for the fixing and 12 for the testing. Four methods (1 and 3 hour(s) delay algorithms, 1 and 3 hour(s) Unit-graph algorithms) were used to compare the outputs with each other. Only 1 hour delay algorithm has high precision, the qualified rate meeting the requirements. By comparing the computational efficiency, delay algorithms are superior to Unit hydrograph algorithms as well. Finally, the flood season of 1986 and 1987 are simulated, the results show that the model is also suitable for long-term hydrological forecasting. In conclusion, the 1 hour delay algorithm was recommended as the prediction method for flood forecasting for the Panxi reservoir.Through fixing parameters and testing the precision of the Xin'anjiang model apply to Panxi reservoir, the author gained some useful experience about flood orecasting in a typical medium-sized reservoir. In fact, after building a dam the proportion of water surface area will rise in a reservoir basin. Taking this fact into account, the author attempted to divide the Runoff yield module of Xin'anjiang model into two parts - land surface runoff and water surface runoff. The water surface in thereservoir should be calculated by direct runoff, while land surface must consider conversion coefficient. This division is beneficial in practical applications to improve flood forecasting accuracy. Research shows that this method is more conducive to irrigation reservoirs located in Jiangxi Province. Panxi reservoir flood forecasting system is based on it.Research suggests that when considering the characteristics of medium-sized reservoir basin based on its own, using Xin'anjiang model for flood forecasting in medium-sized reservoirs is feasible. Its reliability and accuracy can meet the requirements. This study promoted the application of Xin'anjiang model in medium-sized reservoirs and it's still worthy of further extension.