A Study On Real Time Retrieving Low-Level Wind Field In Quasi-operational Application With 4D-VAR Assimilation Technique

It is effective to improve new generation weather radar application that reliable 3D wind fields are retrieved from real time radar radial velocities. The retrieved wind field can help forecaster to identify the meso-scale structures. The paper predigest the 4D-VAR assimilation system seted up by Sun etc. according to each different impact of the model parameter to the result of wind field. The feasible study is done on real-time retrieving low-level wind field in quasi-operational application with single Doppler-radar data's 4D-VAR assimilation technique. The main conclusions are summarized:(1) The wind fiels is almost consistent between the wet model and double-based radar data retrieved,sounding data etc. The main structure of meso-scale weather system can be retrieved. When the wind field considered only, the difference of dry model and wet is small. The wind field retrieved by dry model can get the similar effect.(2)Different initial and background field can impact the effect of retrieved wind. The wind field structure is better if using the forward result as background, especially when iteration number is less. When the forward result can't be getted, and the variation of wind is not great, the VAD analysis is used as initial wind field. Beacause of VAD analysis denotating the average wind profile in the field, the variation of wind is greater, the error is bigger. Zero wind field as initial shouldn't be utilized because of greater error. It is little impact to the result when changed the resolving and area of retrieved. Considering of the charater of meso- and micro-scale weather system, we should select the high-resolving and biger area of retrieved to compute.(3)Transplanting the model to Pentium4 Processor 2.0G/RAM 1G/PC, the dry model with 15-20 iteration can gain the result within 6 minute. Given the VAD analysis as initial wind field and the forword result as background, the basic characteristics in low-level wind fields can be presented from dry model with 15-20 iterations. Selected the familiar strong weather cases, it is indicated that the retrieved wind character accords with the representative weather style. In a word, we can use dry model with 15-20 iteration, the forward result as background to quasi-operational application.(4) It is compared that Computing divergence and shear using retrieved wind data and using radar radial velocities. And the result is good at the shape and location. In addition, divergence and shear computed by retrieved wind can be identified which wind vector is perpendiculary with radar radial velocity. It is indirectly indicated that the wind field retrieved by the simple 4D-VAR assimilation system.(5) Then the simple 4D-VAR assimilation system is merged into software RPG of Beijing METSTAR Radar CO.Ltd. The retrieved wind field is displayed as a product of PUP. The system is stabiliy, and the result is logical by and large. It is tested that the simple 4D-VAR assimilation is competent for the real-time operational application.(6) When the distance of some radar is close but not enough closer as dual or multiple Doppler radar, we can combine the result of each as the products of dualor multiple Doppler radar.