Research On Retrieving Atmospheric Parameters From FY-3D Satellite Microwave Brightness Temperature Data

Geophysical parameters such as temperature and humidity profile,surface temperature,cloud liquid water,total precipitable amount and rainfall rate can provide important information for numerical weather forecasting and research on earth climate change.Satellite-borne passive microwave remote sensing has good temporal and spatial continuity,high global coverage,and all-day and all-weather observation capabilities.It is currently the main means of achieving atmospheric sounding in the field of meteorological research.FY-3D satellite is the latest polar-orbiting meteorological satellite in my country.Although its hardware level has reached the international advanced level,there is no standard atmospheric remote sensing data product release,so high-precision inversion research based on FY-3D observation brightness temperature data is of great value.Microwaves Temperature Sounder,Microwave Humidity and Temperature Sounder,and Microwave Radiation Imager are equipped with three microwave radiometers with operating frequencies distributed in50-60 GHz,89-183 GHz and 10-89 GHz,respectively.They can detect temperature,humidity,surface temperature,TPW,CLW,rainfall rate and other geophysical parameters information.Based on the above research background,this thesis carries out the inversion research of the atmospheric temperature and humidity profile,surface temperature,total liquid water,total precipitation and rainfall rate parameters based on the MWTS,MWHTS and MWRI observed brightness temperature data of FY-3D satellite,Using neural network and one-dimensional variational algorithm to invert the above parameters.In the inversion test,the inversion area is divided,and the brightness temperature data of different frequency bands are used for comparison test.Finally,a variety of data sources,including the microwave radiometer load ATMS secondary remote sensing data product on the American meteorological satellite Suomi NPP,are used to verify and evaluate the inversion accuracy,analyze the factors that affect the inversion accuracy,and evaluate the inversion effect.The test results show that the temperature inversion effect in the 50-60 GHz frequency band is better than that in the 118 GHz frequency band;in the humidity inversion based on neural network,adding temperature detection channels will improve the inversion effect;MWRI’s 10-89 GHz rainfall rate inversion accuracy is slightly higher than MWHTS’s 118+183GHz.Comparing the one-dimensional variational and neural network inversion results,the results show that for the temperature and humidity profile,the neural network and the one-dimensional variational algorithm have the same inversion effect;for the ocean clear sky TPW,the neural network inversion accuracy is higher;for rainfall rate,weak rainfall areas,the one-dimensional variational inversion accuracy is higher,and for heavy rainfall areas,the neural network inversion accuracy is higher;the inversion accuracy tested in this thesis is compared with the accuracy requirements of ATMS inversion products.The comparison results show that the inversion accuracy of the inversion parameters in this thesis has reached the accuracy level of ATMS products.