Author/Editor     Doxaran, David; Froidefond, Jean-Marie
Title     Retrieving suspended particulate matter concentrations in the gironde estuary from ocean-colour satellite data
Type     članek
Source     Image Anal Stereol
Vol. and No.     Letnik 20, št. Suppl 1
Publication year     2001
Volume     str. 475-80
Language     eng
Abstract     High-resolution ocean-colour satellite data (SPOT, Landsat-7) is used to retrieve suspended particulate matter (SPM) concentrations in highly turbid waters: the Gironde estuary. The exploitation of satellite data includes two stages. First, data is corrected for atmospheric effects (gaseous absorption, molecular absorption and scattering, aerosol scattering). A radiative transfer code is used to estimate ahnospheric characteristics; the aerosols effects are determined in near infrared wavelengths with the black target method and then extrapolated to visible wavelengths. The water-leaving signal (water reflectance) is finally extracted from satellite data. The second stage is based on in situ radiometric measurements, regularly carried out in the Gironde estuary since 1996, and which pennit to establish empirical relationships between water reflectance and SPM concentrations, in usual satellite sensors spectral bands. Concerning SPOT-HRV and Landsat-ETM+, the best relationships are obtained when considering the near infrared to yellowgreen spectral bands ratio. These relationships are logarithmic and allow to retrieve SPM concentrations from 10 to 1000 mg.L-1 with a percentage error lower than 35%. SPM concentration can be attributed to each pixel in the estuary and the horizontal distribution of the turbidity is retrieved in two opposite conditions: low and high river flows. Finally, satellite data is integrated in a hydrodynamic transport model. Satellite images are re-sampled in order to correspond to the limits of the estuary as defined in the model; fifteen easily identifiable points are used to realise the geographical corrections.
Descriptors     SPACECRAFT
SEAWATER
COLOR
PARTICLE SIZE
IMAGE PROCESSING, COMPUTER-ASSISTED