Effects Of Different Initial Fields On GRAPES Numerical Prediction

In this paper, by using the new generation numerical weather prediction modelsystem GRAPES, case experiments are conducted with the initial fields and lateralboundaries provided with T213 L31 and NCEP fnl, respectively. The sensitivity ofGRAPES numerical prediction products to different initial fields and effects ofthree-dimensional variational assimilation on simulation are discussed. The resultsshow that the ability of GRAPES numerical prediction depends on different initialfields and lateral boundaries to some extend, and the different initial fields will greatlyaffect the GRAPES simulated results.In this paper, a case of flood-causing torrential rain in Huaihe R. basin during9th～10th July 2005 is selected. Firstly large-scale background analyses and synopticand dynamic diagnosis are made. The vertical structure and the dynamic conditions ofphysical elements as vorticity, divergence are analyzed. And then 4 groups experiments byT213,NCEP initial fields are contrasted. After analyzing the differences between the twoinitial fields and the different simulated results, the mnemonic ability to initial fieldsof the model and the influence of different initial fields on precipitation forecast aremainly analyzed, it shows that:1,The obvious differences are existed in sub-synoptic scale characteristics ofT213 and NCEP initial fields. And also the different GRAPES simulated results havecontrast in circulation situation,precipitation and station elements forecast. Namelydifferent initial fields cause the different simulated results.2,For T213 and NCEP initial fields, the GRAPES simulated results haveobvious differences and the differences don't disappear with the integration running.But for the same initial field whether it has assimilation or not, it only markedlyinfluence the GRAPES model results in initial 0～12hours. And the differences willminish gradually.3,It shows that both the location and intensity of heavy rain forecasted by thenumeric model is very nearly the fact. But the forecasting area of strong torrentialrain has some differences with the fact.