The Impact Study Of Soil Moisture On Direct Assimilation Of Satellite Radiance Data

With the rapid development of satellite remote sensing technology, there has been increasing number of valuable kings of unconventional observations, thus direct assimilation of satellite radiance data in numerical prediction model has become a mainstream research directions. However, in the current technical conditions, only about5%of the satellite data has been used in the data analysis system, and due to the impact of surface radiation, cloud and precipitation, a lot of information has been eliminated during the pretreatment process thus cannot be assimilated in the system. Taking into account of the negative impact of suface information in satellite data assimilation, many of the low-level channels which are sensitive to land surface may be eliminated in previous studies, this greatly reduces the utilization of satellite data. How to use the low-level channels which are sensitive to land surface by changing the surface information and thus increase the utilization rate of resources?The experiments of direct assimilating microwave remote sensing data AMSU-A and AMSU-B were researched with the mesosacale numerical weather prediction model WRFV3.3.1and its3DVAR system. Land surface sensitivity test, numerical simulation and assimilation experiment research were conducted with a heavy rainfall process occurred at Jiangxi Province during19th June2010. On this basis, we change the soil moisture in the initial field of the model, and analysis the impact of improved soil moisture to model simulation and emissivity data directly assimilation. What’s more, we focus on the adjustments of the deviation of the observation and the background under different soil moisture conditions. The results show that:(1) During the simulation of WRF model coupled with several land surface schemes, there are some differences between the rainbands and precipitation center locations with the real observation, Noah scheme is the best one, which can roughly simulate the distribution and centers of the rain.(2) Soil moisture improved test shows that, precipitation is sensitive to the changes of the initial soil moisture, the change of initial field of soil moisture can cause significant differences in changes in the short-term rainfall and surface heat flux. The soil moisture from the output of the CLSMDAS can better performance the precipitation center and regional fluctuation characteristics of the24h accumulated precipitation simulation, he statistical test also shows that the TS score are higher, the miss rate, the false alarm and the bias are lower than the results from NCEP initial soil moisture. (3) The direct assimilation of satellite radiance data in numerical prediction model shows that, AMSU-A data assimilation under the new initial field of soil moisture adjust the temperature background more distinctly, and AMSU-B data assimilation under the new initial field of soil moisture adjust the wind background more distinctly, but the adjustment of relative humidity is not obvious. The simulation test shows that, AMSU-A data assimilation under the new initial field of soil moisture is the best to simulate the heavy rainfall process occurred at Jiangxi Province, while AMSU-B data can better simulate the6h accumulated precipitation during12h and18h in June19th, so the simulation resuts differ at different times.(4) The analysis of the deviation between observation and background in the assimilation of AMSU-A shows that, the soil moisture from the output of the CLSMDAS simulates the brightness temperature values much closer to actual observations, indicating us that the improved soil moisutre initial condition can get a positive adjustment, especially with band3, while for the common channels the results are similar and the adjustment are not obvious.