The Study On Distributed Precipitation Estimation Model And Method

Precipitation is not only an important factor to affect resource, environment and disaster, but also is an important parameter in hydrology and in the research of water resources。But since a period of time,the precipitation data obtain is only for some disperse point,and that can not getting more material rainfall of area。Therefore,In this paper, a estimation model is built which utilizes the special interpolation technology and the MODIS datum, and get the distribute rainfall.Study on the special interpolation method of hydrological model have a great practical values, and the Artificial neural network has very capability in dealing with nonlinear process. in this paper,the modeling method of artificial neural networks studying on valley distributed rainfall model is adopted; And its performance is compared with other traditional algorithms such as Thiessen polygon, Inverse square distance, etc. In this paper, the parameters of cloud and atmosphere which are involved with the physical mechanics of rainfall are analyzed, and obtained with the cloud product data of MODIS sensor, which, together with data obtained by surface measurement of rainfall, compose the original sample, and build the rainfall estimation model which utilizes the BP network.Aiming at some weakness of the BP Neural Network in the actual applications, such as the astringency slow, easily falling into local solutions concerning bigger search space , another the Neural Network connection weights having no way in definition, Optimized the connection weights by means of genetic algorithm which has the best global search ability, built the optimized BP Neural Network estimation rainfall model, and tested with the original sample. the result indicates that the network model optimized by the genetic algorithm is improved both in the precision and stability, reflects the actual rainfall instance basically, at the same time, it reveals that it bears practical significance to apply the high spectral resolution and rich data products of MODIS to satellite estimation of precipitation.The precipitation data in this paper is obtained from observation data of rainfallin the automatic observation stations, Hubei Province, in 11,24,2006. The MODIS satellite data product, adopted in this thesis, is received and processed by the MODIS ground station of Digital Valley Engineering Center of Huazhong University of Science and& Technology.