Runoff Prediction And Research In Middle Reaches Of Yarlung Zangbo River

The water resources is one of most basic essential factors which supports the human life and the economical growth and sustainable development, but the runoff essential factor change is leading the overall system change. Since runoff in one basin is influenced by the climate, the basin physical geography, the social development as well as the human beings activities etc, its characteristic is very complicated, showing multi-time-scale, randomicity, sudden-change, nonlinear, and so on. So, prediction of runoff, is very difficult, and always be a hotspot or a nodus in the prediction field.Based on the runoff data of the middle reaches of Yarlung Zangbo River, prediction of annual runoff, lunar runoff and daily runoff is developed:In the prediction and research of annual runoff part, the traditional model as auto-regression model is applied. Then the nearest neighbor bootstrapping regressive model, and artificial neural network model are also tried. Not only that, the multi-stations auto-regression model, multi-stations linear regression model and multi-stations artificial neural network model are attempted. Among these prediction models, the multi-stations auto-regression model can give a best prediction result.In the prediction and analysis of lunar runoff part, seven models are introduced: seasonal auto-regression model, threshold regression model, seasonal nearest neighbor bootstrapping regressive model, seasonal artificial neural network model, seasonal multi-stations auto-regression model, seasonal multi-stations linear regression model and seasonal multi-stations artificial neural network model. The result of analysis showed that the seasonal multi-stations auto-regression model is best.During the prediction of daily runoff study, also is the last part. The time-steady auto-regression model, nearest neighbor bootstrapping regressive model, artificial neural network model, multi-stations linear regression model and multi-stations artificial neural network model are suggested. As a result, the usability of all of these models are best in the middle reaches of Yarlung Zangbo River.Through Runoff Prediction and Research in Middle Reaches of Yarlung Zangbo River, both traditional models and recent new models applies wildly are...