Study On The Mid And Long Term Hydrological Forecast Of Mid-Short River In Nongjiushi Of Tacheng Prefecture Of Xinjiang

Unique topography in Tacheng Prefecture makes the west steam come easy, which formsnumerous medium-short river. For example, there are many rivers in the Ninth AgriculturalDivision of Tacheng Prefecture, bue 42 proportion is medium-short river. These riversmaintain local agricultural irrigation, industry and life. And Tacheng Prefecture mountainsystem mostly mountains low-to-middle in height, when seasonal snow-melt , the riverappears spring flood, which makes the runoff assign within the year unevenly. It is extremelyunfavorable in water resource distribution of Tacheng Prefecture. It is a prerequisitean ofirrigated area that strengthens water-planning and water-saving, It is significant to thewater resource rational utilization and rational distribution of Tacheng Prefecture.This paper directs against the actual existing problem, according to the hydrology andgeographical feature of the river of Tacheng Prefecture, take Wulasitai river as an example,according to changing the characteristic for a long time series (1966- 1995), adopt themultivariate linear regression,the artificial neural network and time series analysis predictmethods,predict the Wulasitai river annual runoff.The goal is to offer the valuable referencebasis for preventing flood of the reservoir and irrigation. The main research results are asfollows:(1) According to change characteristic for a long time of Wulasitai river annual runoff,propose the time series model. According to the result of the model, it has better adaptabilityand predicting precision, and the fitting is better, there is certain practicability to explain thiskind of prediction method.(2) Using physical cause analysis,mathematics and physics statistical analysis confirm thatthe precipitation,evaporate and relative humidity are main forecasters to Wulasitai river.(3) Utilizing the main forecasters confirmed, we set up the multivariate linear regression andBP neural network model separately. The result indicates that the prediction results to theWulasitai river are better, It can totally require in order to reach practically .(4) These different models for forecasting runoff forecast results of the comprehensivecomparative analysis of performance, in Tacheng region showed the long-term hydrological...