Application Of HY - 2A SCAT Data In 3DVAR And Hybrid Assimilation Methods

Hybrid data assimilation(DA) method has been becoming the hotspot in assimilation research field with its remarkable advantages.This method incorporates the flow-dependent estimate of the background error covariance based on the three-dimensional variational(3DVAR) data assimilation(DA) method. We all know that the real background error covariance of the atmospheric system varies with the weather flow-pattern,and the background error covariance of 3DVAR is static,and Hybrid method make up for the shortage just right,so Hybrid method is better than 3DVAR theoretically.Satellite scatter data becomes the important source of sea surface wind data with its advantages such as high temporal and spatial resolution,wide coverage, continuous observation and so on. HY-2A satellite scatter data is provided by HY-2A satellite scatter which developed by our country. The data enriched the rare wind data of sea surface. Select the data as the observation data in this paper in order to make the best use of limited resources to improve the initial field of numerical prediction.This paper aims to study whether the Hybrid and 3DVAR method assimilating HY-2A SCAT wind data is able to improve the initial field and model forecast skill or not. Meanwhile,we also expect to get a conclusion that which method is better than the other by comparing the results of two kinds of methods. We conducted corresponding work and got the preliminary conclusion based on the above purposes,as follows:(1) we use the NCEP FNL global analysis data and Buoy data to check the HY-2A SCAT wind data respectively before making the data assimilation expirement.The result shows that the accuracy of the HY-2A SCAT wind data meets the need of data assimilation.Besides, we conducted valid quality control for HY-2A SCAT wind data.(2) Taking the Typhoon Soulik(201307) and Fitow(201323) for example,three experiments are designed:The expirement that taking the ensemble-mean as the initial field and making 72 hours forecast is named as Ctrl; The expirement that ingesting the HY-2A scatter wind data into the initial field with 3DVAR and Hybrid DA methods are named as 3DVAR and Hybrid respectively.The result is as follows:(a) The result of adjustment of initial field shows that the performance of two kinds of adjustment between using two different data assimilation methods is different for different variable in different cases. Overall,assimilating HY-2A SCAT wind data can improve the initial field for the case of "Soulik" to some extent, and have little improvement for the case of "Fitow".Besides,the performance of Hybrid DA method is better than 3DVAR’s.(b)The result of typhoon forecast shows that the adjustment of two kinds of DA methods for initial field both get positive effects for track forecast and intensity forecast.In these two cases,the improvement of track is obvious and the performance of Hybrid DA method whose improvement rate could reach up to 40% is better than 3DVAR’s.The improvement of intensity is weak comparing with track’s whose improvement is below 10%,and the performance of 3DVAR DA method is better than Hybrid’s. So ingesting the HY-2A scatter wind data into the initial field can improve both the track forecast result and intensity forecast result obviously,meanwhile,the use of the flow-dependent co variance can improve the track forecast result only.(c)We adopt three kinds of random perturbation generation methods to generate ensemble-perturbation and then get three kinds of ensemble-mean.The results show that the performance is different somewhat among three perturbation schemes but the difference is little.