Cloud Properties Analysis and Its Application in FY-2 Cloud Detection

One pixel is cloud or not dependent on cloud detection scheme. Cloud detection approach includes histogram analysis, threshold detection, deviation analysis and so on. Cloud threshold is an important factor for cloud detection scheme. Now, cloud threshold usually can be gotten by one time and multi channel data in operational cloud detection method. But if surface is covered by snow or ice, or cloud covers one area for a long time, dynamic cloud threshold method always fails. So it should depend on clear temperature background. On the other hand, climate mean surface temperature can help to distinguish the dynamic cloud threshold' s validity. ISCCP data are proved by many researches that it is one of the best satellite cloud climate data in the world now. ISCCP data are used to analyze properties of cloud and clear temperature over China and neighboured areas. Cloud can be divided into three kinds, such as low, middle and high cloud. Different kind of cloud has different distribution property of cloud top temperature on a seasonly or daily basis along the latitude. Different kinds of cloud top temperature basically show strip distribution along longitude. The lower latitude, the higher cloud top temperature is. Along the same latitude, different area has different cloud properties during one day or one year.The analysis shows that the low cloud top temperature increases progressively in the fall and winter seasons from north to south, and has obvious linear increase tendency with reduced latitude. The middle level cloud top temperature distribution has distinct properties. Except July, August and September, the distribution of middle level cloud top temperature presents the meridional strip basically.The middle level cloud top temperature gradient is large in fall and winter, which is small in spring. The high cloud top temperature presents the meridional strip distribution in fall, winter and spring. The high cloud top temperature is lower than 240 Kand it' s latitudinal gradient is small to the north of 35°N. The high cloud top temperature meridional gradient increases and it' s diurnal variation is small to the south of 35°N. The temperature difference between clear surface and the warmest cloud top has stabile daily and seasonly change in the southeast part of China. At the same time, in the north part of China, the temperature difference between clear surface and the warmest cloud top has distinct daily and seasonly change. Above analyzed properties and mean clear temperature can be used as background information to distinguish the dynamic cloud detection threshold value' s validation and offer cloud threshold for areas that are covered by low cloud for a long time. An example is showed to testify mean clear temperature and temperature difference between clear and the warmest cloud can offer great information for cloud detection scheme.