Research On Application Of GNSS Occultation Detection Data In Atmospheric Science

Global Navigation Satellite System Radio Occultation Detection Technology(GNSS-RO)has the characteristics of all-weather,uniform global coverage,long-term stability,high precision and vertical resolution.The RO detection data has been widely used in Atmospheric Science research.Existing studies have shown that the presence of reflected signals in the occultation signal is an indicator of good occultation data quality.It is important to identify and separate the reflected signals in the occultation detection signal and help to assimilate the occultation data into the numerical weather prediction system.On the other hand,gravity waves are one of the most important dynamic processes in the middle and upper atmosphere,and gravity wave parameterization is also a key component of climate and weather models.Obtaining gravity wave parameters using RO data is of great significance for improving the forecast accuracy of weather and climate models.Therefore,this paper uses COSMIC-2,FY-3D occultation data and ERA5 reanalysis data to carry out the following two researches:(1)Using deep learning models to identify reflected signals from occultation detection data;(2)Using S-transform extracts and analyzes the characteristic parameters of the stratospheric gravity wave.In the first part,based on radio holography technology,this paper uses the near real-time con Phs data of COSMIC-2 RO from January 1 to 9,2020 to establish an improved GoogLeNet(Im-GNet)deep learning model that can automatically identify occultation reflection signals.The accuracy of the Im-GNet model has reached 96.4%,which is significantly better than the support vector machine(SVM)method,and achieves a good occultation reflection signal recognition effect.In addition,by analyzing the geographical distribution of occultation events,and introducing the NCEP forecast value to compare with COSMIS-2 RO data,it is concluded that the data quality of occultation events with reflected signals is better and contains more atmospheric information.In the second part,based on the S-transform method,this paper presents a detailed process for inversion of gravity wave pseudomomentum flux(PMF)by combining the RO temperature profile and Reanalysis of horizontal wind data,including the calculation methods of gravity wave vertical wave parameters and horizontal wave parameters.Using COSMIC-2 and FY-3D RO data and ERA5 Reanalysis data for a total of 363 days in 2020,various parameters of the stratospheric gravity wave were calculated.First,the error of the temperature disturbance of gravity waves obtained by these three kinds of data is analyzed,and it is found that COSMIC-2 is an ideal data source for the study of gravity waves in the stratosphere.Secondly,by comparing the horizontal wavelength,potential energy(Ep)and PMF space-time distribution of gravity waves obtained from the three kinds of data,we analyzed the gravity wave activity in different height ranges and seasons of the stratosphere,and the characteristics of the space-time distribution of gravity waves obtained are basically the same as previous results.Finally,the ERA5 average data from January 2015 to December 2020 was used to obtain the time-latitude distribution of global stratospheric gravity waves’ Ep and PMF.At the same time,the zonal wind was superimposed on the time-latitude distribution of gravity wave PMF.Above,the inter-annual variation of gravity waves was studied.Through analysis,we found that there are obvious seasonal changes in gravity waves at mid-latitudes in the northern hemisphere,which corresponds to the Quasi-Biennial Oscillation(QBO)phenomenon in tropical regions;the Stratospheric Sudden Warming(SSW)that occurs in the polar regions of the northern hemisphere affects gravity Wave activity;ERA5 Reanalysis data has led to a serious underestimation of PMF in the subtropical regions at mid-latitudes and on both sides of the equator in the southern hemisphere.