The GLOBCARBON initiative aims to develop a service to provide the Global Carbon Project (GCP) and wider IGBP communities information on global land surface properties. The baseline inputs are 1 km satellite sensor data, in the first instance data from the ATSR-2, VEGETATION, AATSR and MERIS sensors. The service generates products at resolutions of 10 km, 0.25 degrees and 0.5 degrees covering the period 1998-2002 with VEGETATION and ATSR-2. From 2003 this changes to AATSR, VEGETATION and MERIS. The introduction of MERIS has required the addition of new processing modules. One challenge in particular is the generation of a cloud detection system for MERIS Level 1b data. The development of such a system is conditioned by the need to process efficiently approximately 5000 MERIS Reduced Resolution scenes per year for GLOBCARBON and the absence on MERIS of either the shortwave or thermal wavebands used by existing cloud detection systems such as APOLLO or CLAVR. The ultimate aim of the development is to provide an operational cloud detection method for MERIS Fine Resolution data for projects like GLOBCOVER. This paper reports on the pragmatic solution adopted for GLOBCARBON which exploits the data content in the so-called 'Bright Layer' using additional tests to separate cloud, snow and bright land.
展开▼
