We evaluate retrieval of particle backscattering at 555 nm (bbp (555)) in the Mediterranean Sea with the quasi-analytical algorithm (QAA), using new in-situ concurrent data of remote-sensing reflec- tance (Rrs) and inherent optical properties (IOP). When Rrs is cor- rected for Raman scattering, retrieved bbp(555) are reduced between ~7% in open waters and ~3% in coastal waters. After Raman effect is accounted for, partitioned statistics within the data set for cruise and water type are rather homogeneous. QAA- retrieved bbp(555) slightly underestimates our in-situ bbp(555), thus contradicting the few previous studies, in which high over- estimation was either reported or suggested. Retrieval errors are mostly caused by the Rrs-IOP model. The 'new' Rrs-IOP model that separates the influences of water and particle phase functions produces mean bias ~-3% for open waters and ~-7% for coastal waters. In contrast, the 'old' model, ambiguous with respect to the phase function, produces a bias ~-5% for open waters and ~- 9% for coastal waters. It is also shown that regional calibration can virtually suppress bias. In all cases, RMS error remains ~19% and accounts for all errors involved.
Use of the quasi-analytical algorithm to retrieve backscattering from in-situ data in the Mediterranean Sea
Jaime Pitarch;Marco Bellacicco;Gianluca Volpe;Simone Colella;Rosalia Santoleri
2016
Abstract
We evaluate retrieval of particle backscattering at 555 nm (bbp (555)) in the Mediterranean Sea with the quasi-analytical algorithm (QAA), using new in-situ concurrent data of remote-sensing reflec- tance (Rrs) and inherent optical properties (IOP). When Rrs is cor- rected for Raman scattering, retrieved bbp(555) are reduced between ~7% in open waters and ~3% in coastal waters. After Raman effect is accounted for, partitioned statistics within the data set for cruise and water type are rather homogeneous. QAA- retrieved bbp(555) slightly underestimates our in-situ bbp(555), thus contradicting the few previous studies, in which high over- estimation was either reported or suggested. Retrieval errors are mostly caused by the Rrs-IOP model. The 'new' Rrs-IOP model that separates the influences of water and particle phase functions produces mean bias ~-3% for open waters and ~-7% for coastal waters. In contrast, the 'old' model, ambiguous with respect to the phase function, produces a bias ~-5% for open waters and ~- 9% for coastal waters. It is also shown that regional calibration can virtually suppress bias. In all cases, RMS error remains ~19% and accounts for all errors involved.I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.