The Study Of Surface Observation Data Assimilation Scheme In GRAPES

With the advances of numerical weather simulation and forecasting, the spatial and temporal resolution of numerical model is much higher and the data assimilation updating cycle is reduced. Surface Observation Data Assimilation (SODA) becomes more and more important in data assimilation system. The density of surface observation data is very great in China, but it is in a low utilization ratio, and almost uses as qualitative analysis. This may proved to be a largely waste of resources. Currently, the research of SODA just begins in our country, so it has important science worth and realistic meaning to the research of SODA.The Surface Observation Data is under the great influence of topography , meanwhile, the topography distribution of our country is very complex and there exists different elevation between surface observation station measurement and numerical model surface, so how to resolve availably the different elevation to become a basic problem of SODA. Considering on the basis of the restriction of thermodynamic process in PBL, to design the Surface Observation Operator (SOO) is an effective way to resolve the problem spoken of earlier, and is the way of general international practice in Data Assimilation.The paper describes the foundation of the new SOO, which contains thermodynamic process in PBL, and considers the different elevation between surface observation station measurement and numerical model surface. In order to research the ability of assimilation and forecast with the new SOO, we carried on some experiments using the GRAPES model. The major results and conclusions are summarized as follows:(1) It is based on the GRAPES 3D_VAR framework, using surface similarity Theory, considering the influence of thermodynamic process in atmospheric boundary layer, to build the new SOO. Theoretically, the new SOO can better describe the function of surface observation in Data Assimilation.(2) It carried on the strict examination and test to the SOO, including the accuracy test of tangent linear and adjoint codes. The result was correct.(3) Different elevation difference, which between surface observation station measurement and numerical model surface, can result in different forecast effect, so it is necessary to confirm which elevation difference is best and critical. To study the forecast effect in the condition of the different elevation difference, we made 6 tests. And the results indicate that 750 meters is the better elevation difference, because the root mean square(rms) of Re_fcst for wind field is smaller than others at this elevation.(4) The GRAPES₃DVAR system using the new SOO can correctly reflect the interaction relationship between the wind field and other variables. Through an ideal test, we can conclude that the background structure still remain using the new SOO, and it can describe the multivariable analysis character exactly(5) According to compare the statistic character of innovation in new scheme with the old. It indicates that we can get the more perfect statistic character of innovation using the new SOO. It is may concern the earth's surface character, such as roughness, and static stability of atmosphere.(6) Compared the forecast effect in new scheme with which in old scheme, we have found that the rms of Re_fcst in new scheme is slightly bigger than it in old scheme. Reasons as follow: for the u component wind, the upper elevation difference( -50 meters) may be smaller; for the v component wind, it ignore the different pressure between u and v in different horizontal grid-point, replacing the v pressure with u pressure.(7) The change of wind field can result in the change of atmospheric-circulation situation, and have an impact on water vapour transportation. Designed two serial precipitation forecast tests, and the results indicate that the new scheme of surface wind observation data assimilation has influence on precipitation , which is positive influence in general.