Comparison of Fog Detection between Maximum Entropy Method and Dual Channel Difference Method near Korean Peninsula

Currently, most widely used operational Fog detection method based on satellite remote sensing is Dual Channel Difference(DCD) method between IR and near IR channels. However, DCD methods has a lot of uncertainty including cannot distinguish fog from the low level clouds. Maximum entropy(MAXENT) is the method to find the most probable state that maximizing the entropy of information theory. From the given environmental variables as a constraint, MAXENT could produce probability distributions of occurrences and this method is widely used for species distribution model in the field of ecology and the natural language process in the field of computer science, etc. Using the fact that the difference of sea surface temperature and cloud top temperature of the fog is small, Cloud top temperature of CALIPSO with sea surface temperature of OSTIA is used as training data set to determine fog pixels from geostatonary satellite COMS over north-east Asia from May to July in the 2011 to 2013. For comparison,  the DCD method applied to COMS as well to retrieved foggy pixel. We compared fog distribution of the DCD with the MAXENT methods. In this presetation, we will present the pros and cons in fog retrieval between two methods.