Retrieval bias analysis of ice cloud optical thickness based on the FY-2 satellite

Ice cloud microphysical retrieval is a research hot in cloud parameter retrieval. Until now, the FengYun-2 satellite has not provided related operational products. Based on the visible and middle infrared channels, the cloud optical thickness retrieval algorithm is developed. The cloud phase and multilayer cloud detection is the base of the cloud optical thickness retrieval. The reflectivity at the visible, middle infrared channel and the brightness temperature at the water vapor and infrared channel are used to detect cloud thermodynamic phase. For the ice cloud optical thickness retrieval, the form of hexagonal solid column is used. The EOS/ Terra and EOS/ Aqua MODIS cloud optical thickness data are selected as validation data to evaluate the FY-2 retrieval cloud optical thickness. The retrieval test was done based on the August 2013 FY-2 data. 34 matched data were gotten to evaluate. The analysis results show that the mean bias between the FY-2 and EOS/MODIS cloud optical thickness is 6.41. The mean correlation coefficient between FY-2 and EOS/MODIS is 0.92. The slope of fitting linear is 0.74. It shows that the FY-2 retrieval data has the same pattern as the EOS MODIS cloud optical thickness, but FY-2's cloud optical thickness is smaller than that of the EOS MODIS data, especially for thicker cloud. The reason why the bias between FY-2 and EOS is caused comes from some factors. Except the different retrieval algorithm, different satellite data is a main factor. The reflectivity of the FY-2 visible channel is smaller than EOS's. The more similar the physical value of the retrieval channel data is, the smaller the bias between the two kinds of satellite retrieval result is.