Improvement Of Precipitable Water Vapor Inversion Algorithm For Feng Yun Meteorological Satellites

As a member of the Global Integrated Earth Observation System,the Feng Yun(FY)Meteorological Satellite,with its mature technology and stable operation,is capable of continuously observing the global land,atmospheric,oceanic and surface environments around the clock,and thus has great potential for application services.The thermal infrared precipitable water vapor(PWV)products of the FY meteorologic-al satellite have become an important source of water vapor information in China's numerical forecasting models and data assimilation systems.FY-3D was successfully launched in 2017,with a second generation Medium Resolution Spectral Imager(MERSI-?)on board with significantly improved instrument performance compared to the previous sequence of satellites.The current operational algorithm generates the FY-3D MERSI ? PWV product with data accuracy that tends to overestimate water vapor in dry areas,and the algorithm has more room for improvement.In this paper,a water vapor sensitivity analysis of the four infrared channels of MERSI ? was carried out using the Community Radiative Transfer Model(CRTM).Based on the improved physical split-window algorithm,a combination of different channels was used to carry out water vapor inversion simulations,and the results show that the combination of adding 7.2?m channels and split-window channels can effectively improve the operational algorithm.Satellite observations from 1 April 2020 to 8 April 2020 were inverted using this scheme.When the surface emissivity was unknown,the correlation coefficient,root mean square error and mean deviation of the inverse water vapor compared with GPS water vapor data were 0.915,0.465 cm and 0.284 cm,respectively.When the surface emissivity was known,the correlation coefficient,root mean square error and mean deviation of the inverse water vapor compared with GPS water vapor data were 0.915,0.465 cm and0.283 cm.The analysis of the average relative error of the matching of each station shows that the error is controlled below 0.4 in the eastern part of China,and the stations with an average relative error greater than 1 in the dry area are significantly reduced.In summary,the improved physical split window algorithm has higher inversion accuracy,it can improve the current operational algorithm,and has a wider range of application capabilities.In addition,the successful and stable launch of FY-2H the following year was accompanied by the Stretched Visible and Infrared Spin Scan Radiometer(VISSR-2),which provided PWV products that were also generated using an improved physical split window algorithm inversion.This paper evaluates the quality of the water vapor product generated by the 2019 radiosonde in terms of both product accuracy and stability.Compared with the global sounding data for the first week of January,April,July and October 2019,the FY-2H PWV root mean squared error was 0.49 cm,the correlation coefficient reached 0.96,the mean deviation was 0.10 cm,and the data accuracy was high for both daytime and nighttime periods.The standard deviation of the monthly root mean squared error of the relative sounding data for the FY-2H PWV product was calculated to be 0.07 cm for all of 2019,indicating that the product was relatively stable over the test period.Analysing the synthetic monthly mean PWV product,the satellite's product can correctly reflect the spatial distribution of water vapor in the Belt and Road region.The study proves that the improved physical split window algorithm is more applicable.