A Global Cloud Detection Approach for Geostationary Satellites

Abstract:

Geostationary satellites provide continuous cloud and meteorological observations over a fixed portion of the Earth’s surface, allowing them to monitor and track the development and movement of severe storm systems in real time. For climate studies, geostationary observations provide valuable insight of cloud formation and evolution and how they influence the Earth’s radiation budget. Five well-positioned geostationary satellites can cover most of the globe and their products can be merged into a uniform data set. Presently the constellation of these geostationary satellites consists of the NOAA GOES-West (135° W) and East (75° W), the EUMETSAT Meteosat-9 (9.5° E), the Chinese Feng-Yun (105° E), and the Japanese MTSAT-2 (140° E) platforms.

Stitching together observations from these imagers requires an understanding of their different spectral characteristics and calibrations. This paper presents the cloud detection algorithms developed using the imager data on the five geostationary satellites. The algorithms are used operationally in NASA’s Cloud and Earth’s Radiant Energy System (CERES) Time and Space Averaging (TISA) gridded cloud products and for near-real-time retrievals used for weather and nowcasting applications. Examples of cloud mask results for different surface background (ocean, land, desert) and their diurnal changes will be described. Additionally, merged global cloud fractions and zonal cloud distributions are compared with CERES-MODIS, CALIPSO, and NASA LaRC AVHRR cloud fractions. Potential future improvements are discussed.