Land surface albedo from the synergistic use of polar (EPS) and geo-stationary (MSG) observing systems: An assessment of physical uncertainties

This investigation aims at quantifying the various sources of uncertainties in the derivation of albedo products from geo-stationary and polar orbiting optical systems with emphasis on the sensor and surface type (soil, snow, vegetation) spectral characteristics, atmospheric condition, and angular sampling issues. This research specifically takes into account the uncertainties in albedo that we can expect due to the synergistic use of the European Polar System (EPS)/Advanced Very High Resolution Radiometer (AVHRR) and Meteosat Second Generation (MSG)/Scanning Enhanced Visible and Infrared Imager (SEVIRI) instruments data. Satellite orbital models and a Scattering by Arbitrarily Inclined Leaves (SAIL)-hotspot canopy radiative transfer model were used to simulate synthetic bidirectional reflectance data sets for a broad range of vegetation canopies. The surface bidirectional reflectance distribution function (BRDF) database derived from the Polarization and Directionality of Earth Reflectance (POLDER) data was used in support of the simulations. Spectral atmospheric effects were generated using the 6S atmospheric radiative transfer code. Linear BRDF models, which are candidates for operational use, were inverted with the synthetic reflectance data sets of MSG, AVHRR and MSG-AVHRR combined. The BRDF parameters were then used to derive albedo by hemispherical angular integration. The retrieved model coefficients are discussed with regard to angular sampling and the implications these yield on spectral and broadband albedo determinations. It appears that the sampling problem is better conditioned thanks to the synergistic nature of MSG and AVHRR data. This synergy contributed to enhanced albedo retrievals, but can introduce larger uncertainties due to issues related to the differences in spectral bands and atmospheric state. Hence, the quality of the operational albedo products derived from one or more sources of satellite data will depend to a large extent on sensor characteristics (spectral, radiometric, and geometric), cloud detection, atmospheric correction, the angular distribution of the observations, and finally, the narrow-to-broadband albedo conversion.