| Precipitation is the most important influencing factor in the global water cycle,as an important forcing data of the hydrological simulation system,the accuracy of precipitation affects the accuracy of hydrological simulation and forecasting directly.Conventional ground rainfall station observations and ground rainfall radar observations have limited their application in distributed watershed hydrological simulation systems due to their respective defects.The satellite precipitation products have a wide range of coverage,and it has the advantages of quick and convenient access.In the area where there has no or less precipitation,satellite precipitation products can make up for the deficiency of traditional precipitation data,which provides a strong support for the application and research of meteorological and hydrological.The ground rainfall station network can provide accurate point precipitation information,but its spatial estimation ability is insufficient due to its uneven distribution;the satellite precipitation product can realize the full coverage of the space in the form of continuous grid information,and has the unique advantage of accurately reflecting the spatial distribution of precipitation.This thesis selected the Ganjiang River Basin as the study area,and carried out the precipitation fusion of satellite remote sensing precipitation and ground rainfall station and then the hydrological simulation application based on integrated precipitation was carried out,aiming to obtain a more accurate precipitation data source to further improve the hydrological simulation accuracy.The data accuracy and variation characteristics of TRMM 3B42V7 satellite precipitation products is carried out at the day and month scale at on the grid scale of 0.25 degrees,based on the observational precipitation of ground rainfall from 2009 to 2014.The conclusions had been drawn as follows.First,the 3B42V7 satellite precipitation has good correlation with the ground observation precipitation on both day and month scale,and further strengthened with the increase of time scale.Second,the 3B42V7 satellite precipitation has obvious seasonal variation characteristics in the Ganjiang River Basin,and the accuracy of the flood season is higher than that of the non-flood season.Thirdly,the 3B42V7 satellite precipitation has strong ability of daily precipitation event detection and can detect the occurrence of most precipitation events.The linear regression model,the mean bias correction and the Bayesian combination method were used to construct the satellite precipitation and ground observation precipitation data fusion scheme,and the accuracy of the merged precipitation was quantitatively evaluated by the cross validation.The results showed that the linear regression model fusion method can not obtain more accurate fusion precipitation;the mean bias correction and the Bayesian fusion method can improve the accuracy of the original 3B42V7 satellite precipitation data.The accuracy of the Bayesian fusion method is higher than that of the average deviation.On the basis of this,the distributed hydrological model of the Ganjiang River Basin based on the integrated precipitation is constructed,and the hydrological simulation of the daily runoff in the Ganjiang River Basin is carried out by using the precipitation of the 3B42V7 satellite precipitation,the ground rainfall station network observation and the Bayesian fusion precipitation.The results show that the three kinds of precipitation can simulate the daily runoff process in the Ganjiang River Basin,and the time of the flood peak can be better reproduced,but the flood peak flow rate is different degree of underestimation;Bayesian fusion precipitation simulation accuracy Ground observation precipitation and 3B42V7 satellite precipitation are higher,indicating that the Bayesian method after the correction of 3B42V7 satellite precipitation accuracy has been improved for the basin hydrological simulation to provide a more accurate precipitation data sources. |
PDF Full Text Request