Hyperspectral versus multispectral remote sensing approach to detect phytoplankton blooms in coastal waters : application to a Phaeocystis bloom.
Résumé
Identification of phytoplankton groups from space is essential to map and monitor
algal blooms in coastal waters, but remains a challenge due to the presence of suspended
sediments and dissolved organic matter which interfere with phytoplankton signal. On
the basis of field measurements of remote sensing reflectance (Rrs (l)), bio-optical
parameters, and phytoplankton cells enumerations, we assess the feasibility of using
multispectral and hyperspectral approaches for detecting spring blooms of Phaeocystis
globosa (P. globosa). The two reflectance ratios (Rrs (490)/Rrs (510) and Rrs (442.5)/
Rrs (490)), used in the multispectral inversion, suggest that detection of P. globosa blooms
are possible from current ocean color sensors. The effects of chlorophyll concentration,
colored dissolved organic matter (CDOM), and particulate matter composition on the
performance of this multispectral approach are investigated via sensitivity analysis. This
analysis indicates that the development of a remote sensing algorithm, based on the values
of these two ratios, should include information about CDOM concentration. The
hyperspectral inversion is based on the analysis of the second derivative of Rrs (l) (dl2Rrs ).
Two criteria, based on the position of the maxima and minima of dl2Rrs , are established
to discriminate the P. globosa blooms from diatoms blooms. We show that the position of
these extremes is related to the specific absorption spectrum of P. globosa and is
significantly correlated with the relative biomass of P. globosa. This result confirms the
advantage of a hyperspectral over multispectral inversion for species identification and
enumeration from satellite observations of ocean color
Domaines
Planète et Univers [physics]Origine | Fichiers éditeurs autorisés sur une archive ouverte |
---|