| Sea fog is a kind of catastrophic weather affecting maritime transportation and military activities,which has the characteristics of high concentration,wide range and long duration.As a sea fog-prone area in China,Fujian coast is also a national strategic area,and it is crucial to conduct sea fog monitoring research in this area.Cloud-Aerosol Lidar with Orthogonal Polarization(CALIOP)is suitable for sea fog monitoring because of its vertical penetration capability and its ability to detect the vertical structural features of clouds and fog.In this paper,we first carry out remote sensing monitoring of sea fog in Fujian by CALIOP L1 level532 nm total attenuation backscattering and Vertical Feature Mask(VFM)data of the atmosphere to expand the sea fog detection range based on the physical characteristics of sea fog.Using CALIOP to monitor sea fog in Fujian from 2017 to 2020,a total of 30 sea fogs were monitored.The algorithm was verified using visibility data from coastal stations,and no misclassification points were found.The draping characteristics of Fujian sea fog were also analyzed,and it was found that fog events under clouds were frequent.Although the CALIOP sea fog monitoring algorithm is highly accurate,it still has disadvantages such as long revisit period,small coverage area,and inability to penetrate when thick clouds are encountered.Therefore,in this paper,two representative domestic and foreign geostationary satellites: FY-4A and Himawari-8 are selected to monitor sea fog in Fujian.The sea fog monitoring results of CALIOP and VFM data are used to establish a sample data set of sea fog and low clouds.The statistical analysis of the spectral features and image texture features of the sea fog and the determination of the threshold values are carried out using this sample dataset,and the statistical results show that for the Himawari-8 satellite,the bright temperature difference in the 11 th and 14 th bands,the bright temperature difference in the 14 th and 15 th bands,and the normalized difference snow index(NDSI)calculated using the visible and near-infrared bands Index(NDSI)and the Normalized Water Vapor Index(NWVI)calculated using the visible and near-infrared bands can effectively separate sea fog and low clouds.For FY-4A satellite,the 1st band reflectivity and the 8th,12 th and13th band bright temperature are more effective in separating sea fog and low clouds.The image texture features are selected from the gray co-generation matrix with 3×3 windows,and the image elements with image homogeneity less than or equal to 0.5 and standard deviation less than or equal to 1.5 are judged as sea fog.The threshold values and methods are combined to establish two daytime monitoring algorithms for Fujian sea fog.The results of the algorithms are compared with the visibility information of the measured sites,and the accuracy of the Himawari-8 algorithm is 79.8% and that of the FY-4A algorithm is 83.5%,and the FY-4A algorithm is more capable of monitoring Fujian sea fog.To explore the spectral characteristics of fog under clouds,this paper compares the spectral differences among clouds,pure fog and fog under clouds based on Himawari-8 data.The results show that there is no significant difference between the spectral features of pure fog and sub-cloud fog in all bands during daytime,and the reflectance in the 1-4 band is much lower than that of cloud image elements.At night,the bright temperature of pure fog in the3.9 μm band is generally lower than that of sub-cloud fog.The above spectral feature differences distinguish cloud,pure fog and sub-cloud fog,which is expected to improve the accuracy of Fujian sea fog monitoring in the subsequent work. |
