The Tests And Application Of WRF-EnSRF Land Data Assimilation System Using Automatic Weather Station Data

Data assimilation refers to make full use of observational data, model prediction and error information, as far as possible to get optimal estimation model variables. Its basic meaning is using constraints of the physical and temporal continuity to integrate various space-time scattered distribution of data into model based on laws of physics. The core idea of land data assimilation is in the dynamic framework of land process model, combined with the different sources and different resolution of direct and indirect observation, the land process model and various observation operator sets become continue to automaticly adjust the trajectory of model rely on the observation and reduce the error forecasting system.Wang Shizhang initially completed the construction of WRF-EnSRF in the WRF model. Based on this system, Guo Yakai preliminary has constructed the WRF-EnSRF land surface data assimilation system, and the system for different materials, such as radar, satellite and automatic station data etc. The research work test the WRF-EnSRF land data assimilation system,and using automatic weather station data complete WRF-EnSRF land data assimilation baesd on AWS data.Finaly the validity and the feasibility of this system is verified.The research work results as follows:(1) Through the sensitivity test I analysis response of the model for the selection of coefficient of inflation and the localization distance,and finaly have determined the optimal resolution grid: coefficient of inflation has been chosen as0.8, meanwhile the localization distance has been selected as horizontal distance being10km, vertical distance being10cm. Through the sensitivity test of the initial disturbance intensity, the initial disturbance intensity have a certain effect on the assimilation. Disturbance intensity too small will cause the dispersion problems,but too big will cause the root mean square error being too large.So to determine the precise initial disturbance is very important to assimilation system forecast accuracy.(2) For data selection test, comparising root mean square error of the the anaylis field and the true field, we can get the result:when assimilating of surface temperature data and initial disturbance intensity being1k, assimilation effect is as expected.(3) After Data assimilation using the optimal grid, I analysis some elements.I find that the assimilation experiment effect is more close to the real field than the control experiment. It has improved small for precipitation and surface temperature, but for soil moisture, temperature and near surface wind field analysis field improved more. So it demonstrates the accuracy of analysis field. The final test shows that the prediction results is effective.