Study On The Retrieval And Error Analysis Of Radio Occultation Data On FY Series

Since the successful launch of FY-3C polar orbiting meteorological satellite in 2013,China's Fengyun series of satellites have continued to operate and provide RO data and products.As a new branch,GNOS RO data has a series of scientific and technological problems that need to be solved in data preprocessing,product inversion,data quality analysis and so on.In this paper,the characteristics of GNOS instruments and data processing methods are comprehensively reviewed,and the error characteristics of GPS and Beidou occultation data are analyzed in detail.Aiming at the large error profiles caused by the low-frequency abnormal signal(L2),the research on the correction of low-frequency abnormal signal in the environment of low SNR is carried out.Besides,a new quality control for GNOS is introduced based on the noise factor derived from the correction method and the excess phase at the sensitive height.At last,an assimilation experiment is carried out to show the imapct of GNOS data.The main highlights of this paper are as follows:1.The error characteristics of Fengyun occultation data are fully analyzed.It is found that there is a large error in the low-level atmosphere result from the low-frequency signal(L2 signal)of FY3C/3D GNOS under the low SNR.Excluding the large error profiles,the average deviation of GNOS GPS occultation data is close to zero in the range of 0-45 km,which proves the unbiased feature of RO.The accuracy is the highest in the range of 5-25 km,and the standard deviation of refractivity is less than 1%.The stability and inheritance between FY3 D and FY3 C are good.However,the local multipath effect of occultation antenna,especially under the unfavorable installation environment on the satellite,may cause systematic error,which makes a new experience for the subsequent installation of satellite instruments and ground data processing.2.Combined with the mechanism of ionospheric effect,Chapman layer is studied based on the research of Culverwell and Healy.The relationship between the high and low frequency signals is established based on the good samples,and the correction method for the abnormal signals of the low frequency is derived subsequently in the low atmosphere.As a result,the accuracy of the profiles can be significantly improved.The large error profiles of FY3 C GPS RO accounted for about 18% before the correction,and it reduced to 2.5% after correction,which made the number of good samples increased significantly,and the overall accuracy was maintained,especially below 10 km.The result indicates that there are more high-quality samples in the low-level atmospheric after the correction.3.Based on the characteristics of GNOS,a new quality control method is developed.A new parameter named noise factor is introduced according to the correction method.In addition,in the case of large uncertainty in the signal-to-noise ratio of GNOS,the characteristics of excess phase between 60 to 80 km SLTA height,where reflect the observing ability of the instrument,are analyzed.The accuracy of the new quality control method for FY3 C GPS RO profile is 95.4%,while the error rate is 1.8%,which can identify most of the abnormal profiles.4.The best quality of Beidou occultation is 10-30 km,and the accuracy of Beidou occultation data is consistent with that of other GPS occultation data,which shows that as the first non GPS occultation profile in the world,the Beidou occultation at the core height is reliable.However,the open-loop signal tracking of FY3 D Beidou has not been brought into play.The detection depth rate and the accuracy in troposphere are still the bottleneck of Beidou RO sounding.Among the three kinds of Beidou orbit,the accuracy of MEO RO data is the highest.Because the precision orbit determination of GEO and IGSO is more difficult than that of MEO,the clock error accuracy is lower,which affects the detection accuracy.The proportion of large error profile of Beidou occultation is about 7.2%,which mainly appears in GEO and IGSO orbit.The reason is different from GPS RO.Further researches are still needed.Based on the researches of this paper,we have improved the understanding of our own RO sounder,and provided a new solution for the detection of low-frequency signals in low-level atmosphere in the environment of low signal-to-noise ratio,providing experience and solutions for the future instrument design.The method has been successfully implemented in the ground processing system of Fengyun satellite,which makes the GNOS data are widely used in the domestic and foreign numerical prediction centers,and play a role in the practical application of Fengyun satellite data.