Research On The Influence Of Assimilated Multi-Source Satellite Radiation Data On Precipitation Simulation In Shiyang River Basin

The accuracy of numerical weather forecast is limited by the accuracy of initial values.Data assimilation method is an effective method to improve the initial values,and has been gradually applied to the regional weather simulation in China.However,there are relatively few studies on the arid regions of Northwest China.To a certain extent,satellite data can make up for the shortage and uneven distribution of station observation data in northwest China,so it is of great significance to apply satellite data to improve the high-precision temporal and spatial distribution simulation of precipitation in the basin.Combining the characteristics of the Shiyang River Basin and the current research status of precipitation in this area,introducing three-dimensional variational assimilation method and multi-source observation data on the basis of exploring the optimal parameterization scheme combination,the high-precision simulation of the precipitation process in the Shiyang River Basin is achieved by improving the initial field of the model.First,based on the different physical processes,initial field data,and horizontal resolution of the WRF model,the optimal parameterization scheme combination suitable for precipitation simulation in the Shiyang River Basin is explored,and the spatial distribution of precipitation,the error and correlation coefficient of precipitation monitoring from ground stations Analysis and comparison were carried out in other aspects.The results show that: microphysical processes and cumulus convection respectively have a greater impact on the precipitation magnitude and spatial distribution of the precipitation simulation in the study area;in the land surface scheme,the Noah scheme simulates the best,while the RUC scheme is not suitable for Shiyang River Watershed precipitation simulation;the boundary layer has a significant impact on precipitation,and the MRF scheme is more advantageous than other boundary layer schemes;among different initial field data,ERA5 data performs best in the simulation of the precipitation center,and the precipitation magnitude simulated by the NCEP FNL data set It is more in line with the actual situation;comprehensive advantages in all aspects,the final determination of the microphysical process,land surface process,planetary boundary layer,cumulus convection parameterization scheme is the combination of WSM5,Noah,MRF,GD as the optimal simulation scheme.Secondly,the NCEP ADP data set and AMSU-A satellite data were assimilated using the WRF-3DVAR system.,in order to explore the impact of different data on the precipitation forecast effect when the different nesting layers of the model are assimilated,and according to the spatial distribution of precipitation,TS,ETS precipitation The score and MAE,RMSE and other indicators were comprehensively evaluated.The results show that NCEP ADP data has a better assimilation effect in the innermost layer of the model,while AMSU-A satellite data has the best assimilation effect in the outermost layer of the model.When both data are assimilated in the outermost layer,the effect of precipitation forecast is greatly improved.Finally,14 sets of satellite data assimilation schemes are designed for the typical precipitation process to further explore the improvement effect and influence mechanism of assimilating AMSU-A,IASI and MHS satellite data on precipitation forecast under different background error covariance matrices.In general,assimilated satellite data has significantly improved the accuracy of precipitation forecasts.IASI data performs better than AMSU-A and MHS data.AMSU-A data has the greatest impact on the humidity field.Assimilation of IASI and MHS data has significantly improved the model temperature field and wind field.