Detailed Characterization of aerosol properties from satellite Observations using GRASP algorithm

Abstract:

GRASP (Generalized Retrieval of Aerosol and Surface Properties) is rather sophisticated algorithm was developed recently by Dubovik et al. (2011, 2014) with objective of achieving more complete and accurate aerosols and surface retrieval. Specifically, GPASP searches in continuous space of solutions and doesn't utilize look-up-tables. It based on highly elaborated statistically optimized fitting. For example, it uses multi-pixel retrieval when statistically optimized inversion is implemented simultaneously for a group of satellite pixels. This allows using additional a priori information about limited variability of aerosol of surface properties in time and/or space. As a result, GRASP doesn't use any specific information about aerosol or surface type in the each observed pixel, and the results are essentially driven by observations. However GRASP retrieval takes longer computational time compare to most conventional algorithms that is the main practical challenge of employing GRASP for massive data processing. Nonetheless, in last two years, GRASP has been significantly optimized and adapted to operational needs. As a result of this optimization, GRASP has been accelerated to the level acceptable for processing large volumes of satellite observations. Recently GRASP has been applied to multi-years archives of PARASO/POLDER and ENVISAT/MERIS. Based, on the preliminary analysis GRASP results are very promising for comprehensive characterization of aerosol even for observations over bright surfaces and for monitoring very high aerosol loading events (with AOD 2 or 3). In addition, it was made the attempts to estimate such aerosol characteristics as aerosol height, air mass, radiative forcing, aerosol type, etc. The results and illustrations will be presented.