IAFE   05512
INSTITUTO DE ASTRONOMIA Y FISICA DEL ESPACIO
Unidad Ejecutora - UE
congresos y reuniones científicas
Título:
Atmospheric correction of OLCI imagery over very turbid waters using the RED/NIR/SWIR bands
Autor/es:
GOSSN, J.I; RUDDICK, K.; DOGLIOTTI, A.I.; NECHAD, B.; GOSSN, J.I; RUDDICK, K.; DOGLIOTTI, A.I.; NECHAD, B.
Lugar:
Lisboa
Reunión:
Encuentro; International Ocean Color Science Meeting 2017; 2017
Institución organizadora:
International Ocean Colour Coordinating Group
Resumen:
Atmospheric correction using near infrared bands (700nm-900nm) may work for clear to moderately turbid waters, but generally fails for extreme turbidities because of flattening of the water reflectance spectrum. The use of SWIR bands at 1.6µm and/or 2.3µm can be effective in extremely turbid waters, but adds significantly to the mission cost. The new 1016nm band on OLCI may help improving atmospheric correction in turbid waters at a lower marginal cost.In this work, we endeavored to find "atmospheric invariant" quantities to construct an atmospheric correction scheme over turbid waters. We found that if we choose a series of spectrally-close band triplets (such as 620nm-709nm- 779nm), the Rayleigh-corrected reflectance of the ?middle? band after baseline subtraction (baseline residual, BLR) is essentially independent on the atmospheric conditions (including for very hazy scenes). In this work, we used three consecutive band triplets of the group of bands 620-709- 779-865- 1016nm to relate these three BLRs to water reflectance. To establish this relation, we used in-situ radiometric data from the La Plata Estuary (Argentina), Gironde River (France), Scheldt Estuary and Belgian Coast. We also performed a match-up analysis between OLCI and in situ-radiometric data from two measurement campaigns in a Fishermen´s Pier in La Plata Estuary on January 6 th -26 th , 2017.Comparison between satellite and in-situ measurements suggests that using three BLRs together effectively removes the aerosol signal. Although the BLR algorithm still needs to be improved for extremely high water turbidities, we obtained a plausible correlation between in-situ and OLCI Rayleigh-corrected BLRs from the match-up analysis.