Retrieval Of Cloud Radiative Properties From FY-2C Satellite Data

It's well known that clouds have a strong effect on adjustment of energy budget of the earth and atmosphere system. The radiative properties and spatial-temporal varieties of clouds play a considerably important role for formation of the climate and synoptic system. The cloud optical thickness, cloud particle effective radius and cloud top temperature are important parameters to determine cloud radiative properties, and it's absolutely necessarily to make use of satellite data to get a worldwide information of cloud parameters and their spatial-temporal varieties. Since the past decades, people have increasingly concentrated on the research in retrieval of cloud radiative parameters by satellite data and respectable papers on research and experiment in retrieval of cloud parameters by polar-orbit satellite observations, while the relevant work by geostationary satellites has not been seen. As Geostationary satellites have an advantage of higher temporal resolution than polar-orbit satellite on cloud parameters exploring, it's valuable to make use of geostationary satellites to retrieve the radiative properties of clouds.This paper describes the theories of remote sensing and algorithms for retrieving cloud radiative properties such as cloud optical thickness, cloud particle effective radius and cloud top temperature, using multi-spectral image such as visible, middle-infrared and far infrared channel reflectance and brightness temperature by FY-2C sensor. Because infrared channel information correlates with surface temperature, which is required as accessorial data in the cloud retrieval. NCEP/NCAR analysis is widely utilized in climate diagnose and synoptic analysis etc., providing initial field for climate and meso-scale mode. In order to use NCEP/NCAR analysis surface temperature and 1.5m-above air temperature data in cloud retrieval, we firstly make a systematic comparison and analysis between the NCEP/NCAR Global Final Analysis (FNL) dataset and station observation over China. The results show that in spite of some difference between NCEP/NCAR analysed and station observed surface temperature data, it's feasible as accessorial data in the cloud retrieval. Then, the sensitivity-analysis is made between FY-2C visible, middle-infrared and far infrared channel radiance and radiative parameters such as cloud optical thickness, cloud particle effective radius, cloud top temperature, surface albedo and surface temperature etc. by radiative transfer calculations. According to the response of each channel, an iterative algorithm for retrieval of cloud optical thickness, cloud particle effective radius and cloud top temperature simultaneously by FY-2C visible, middle-infrared (3.75μm) and far-infrared (11μm) channel information is developed and applied for experimental study.The retrieval results are compared with MODIS cloud products. Finally, the retrieval errors are analyzed. Conclusions are below:1) Surface temperature of the NCEP/NCAR data is lower (about 2K) than the observed data in general, and 1.5m-above air temperature of NCEP/NCAR data is also consistent with the station observation in spatial distribution. Generally speaking, NCEP/NCAR analysed surface temperature and 1.5m-above air temperature as accessorial data for cloud retrieving are feasible in China.2) Based on the sensitivity-analysis and the iterative algorithm for retrieval of cloud optical thickness and cloud particle effective radius simultaneously by two channels (fengsheng zhao et al.), a three-channel iterative algorithm for retrieval of cloud optical thickness, cloud particle effective radius and cloud top temperature simultaneously by FY-2C visible, middle-infrared (3.75μm) and far-infrared (11μm) channel information is advanced.3) The iterative algorithm is applied for experimental study and compared with MODIS cloud optical thickness, cloud particle effective radius and cloud top temperature products, the results show a good consistence of cloud optical thickness and cloud particle effective radius between FY-2C and MODIS, while cloud top temperature of FY-2C retrieval results is lower than MODIS generally caused by lower-valued of 11μm channel radiance on FY-2C.