Abstract:The inversion calculation of the atmospheric visibility data of the FY-3A meteorological satellite was performed in this paper. The LAPS (Local Analysis and Prediction System) of the United States was used to calculate the inversion of the visibility data of the night fog and the multi-element inversion of the FNL reanalysis data. Data was analyzed by multi-source feature fusion. The low-visibility distribution of fog resulting from the inversion of FY-3A satellite was compared with the weather map of the Micaps (Meteorological Information Comprehensive Analysis and Process System) system, the infrared cloud map of the FY-2E geosynchronous satellite, and the statistical yearbook of the elemental statistics, showing that the range and intensity of the invisible low visibility area were reasonable. In particular, the absence of conventional observational data networks in the sea areas provided information on the distribution of visibility at night. Through the LAPS system comparison of satellite data, reanalysis data, and satellite and reanalysis data fusion results, the fusion effect of the satellite monitoring data and the FNL reanalysis data was shown, bringing a better improvement on the fog low visibility distribution retrieved from single-source data. After the fusion, the satellite orbital blind area had been bridged for satellite data. Secondly, information on the distribution of low-visibility areas at sea had been obtained, and the visibility of the sea and coastal fog areas had been reasonably improved. For FNL reanalysis data, the strong gradients of the original elements were reasonably smoothed. The scope of low visibility had also been improved. The distribution of low-visibility areas in the sea fog was vital, and the mutual verification and information integration of numerical simulations and FY-3A satellite information were obtained and the credibility was enhanced. Therefore, LAPS data fusion effect is worthy of recognition, and it has good reference value for forecasting and early warning.