Detection of Fog Using Temporally Consistent Algorithm With INSAT-3D Imager Data Over India

A fog detection algorithm using INSAT-3D imager data has been developed and tested. The algorithm emphasizes continuous fog detection as the current fog detection algorithm which is used for the operational fog product generation over India is having temporal discontinuity during dawn and dusk. Moreover, currently, the INSAT-3D fog product over India is being generated with different algorithms for day and night. In this paper, a single algorithm using brightness temperature difference of 3.9 and 10.8 μm wavelength for both day and night has been used. The thermal infrared, water vapor, and visible channel spectral information (for daytime) has been used as additional criteria to reduce false alarm. The thresholds used are determined dynamically using an image-based method. The spatial homogeneity test, which is a special characteristic of fog, has also been carried out to minimize false detection. Validation was carried out against visibility data over five airports in the Indo-Gangetic (IG) plains of India. It shows that probability of detection and false detection to be 66% and 10%, respectively, for 2013-2014 and 68% and 10% for 2015-2016. The algorithm is able to capture the entire life cycle of fog from formation to dissipation, consistently for both the season with different weather conditions.