International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences, Vol XXXV, Part B7. Istanbul 2004
meteorological events, winds and tides. It is one of the most
productive inland seas in the world, contributing significantly to
the economy of eastern Canada with its finfish and invertebrate
fisheries, apart from its important role as a seaway.
The estuary is fed by the St. Lawrence River input, with a
yearly mean of 12000 mis! (but highly variable) and with a
drainage area of 1 320 000 km? (Larouche, 2000). Estuarine
waters are thus characterized by a salinity gradient and a flux of
terrigenous particles and dissolved organic matter. The gulf is a
triangular shaped semi-enclosed sea with an approximate
surface area of 226 000 km°. It has only two connections with
the Atlantic Ocean: the Cabot Strait and the Strait of Belle-Isle
(Koutitonsky and Bugden, 1991) .
Considering the size of the area and the variability of the
system, remote sensing becomes a useful tool in order to
observe large scale distribution of oceanographic parameters
such as chlorophyll concentration in the surface waters.
1 .3 Remote sensing of Case II waters of the Estuary
In estuarine waters of the St. Lawrence system, at least three
relevant quantities (phytoplankton chlorophyll, SS and
gelbstoff) can vary independently of each other, a case which is
conventionally named as Case-Il (Morel and Prieur, 1977;
Gordon and Morel, 1983; cited in Sathyendranath, 2000).
Because of this inherent variability of the bio-optical
constituents, remote sensing algorithms for the retrieval of
chlorophyll must be site and season specific, and relatively
complex with respect to Case I algorithms where phytoplankton
are alone the principal agents responsible for variations in
optical properties of the water (Jacques et. al., 1998).
The complexity of Case-II waters arises from the fact that the
interactions between the optically active constituents are non-
linear, and have spectrally varying effects on the remotely
sensed signal. The change in optical signal may be minor with
respect to the change in the concentration of a variable
component, and moreover, two or more substances may have a
similar influence on the signal at some wavelengths. This signal
competition makes it difficult to decouple chlorophyll
information from the optical contributions of the other two
major constituents; Gelbstoff and suspended sediments.
(Sathyendranath, 2000).
1.4 Present study
In this work we will focus on ship-based spectral reflectance
data acquired using an Analytical Spectrometer Device (ASD-
FieldSpec®). This instrument has two principal advantages for
this type of works: first, it has a narrow bandwidth (1,4 nm)
relative to airborne and satellite sensors, and thus allows a
detailed investigation of the spectral signatures. Secondly, the
ship-based radiometer data do not impose atmospheric
correction, even though it will provide valuable data for
atmospheric correction needed for validation of the airborne
and satellite data.
The ultimate usage of the ship-based spectral reflectance data
consists on the validation of satellite remote sensing data.
2. MATERIAL AND METHODS
An Analytical Spectrometer Device (ASD-FieldSpec^) has
been used to collect field data in this study. The instrument is a
326
field-portable spectroradiometer, which operates in the visible-
near infrared portion of the spectrum between 350-1050 nm
wavelength domain. It has a 512 channel silicon photodiode
array overlaid with an order separation filter. The integration
time is manually adjustable through the controlling software.
Dark current can be measured and sampled manually at any
time, and it can be corrected either manually or automatically.
Data collecting procedure consists of taking uncalibrated
radiance measurements of the seasurface radiance, the sky and
the reflectance panel. In this approach, the uncalibrated sensor
is used to measure signals proportional to the seasurface
radiance, sky radiance and the radiance reflected from a
horizontal reference panel, having a known bi-directional
reflectance for the solar and viewing directions. (Mueller et. al.,
Chapter 10, 2000)
2.1 Seasurface radiance
In our cases, we accept the upwelling radiance above water
composed of the waterleaving radiance, which has the
information we seek for, and the sky radiance reflected from the
seasurface. The ideal is to omit sunglints, foam, ship motion
effects, ship’s shadow while minimizing the surface reflected
radiance. The viewing angles should be adjusted according to
the previous principles, and the last convention (Mueller et. al.,
Chapter 10, 2000) defines the optimal angles as 40? for the
zenith and 130? with the azimuth relative to the sun. In any
circumstance, the sensor should view the seasurface at a zenith
angle within the range 30-50? and at an azimuth angle in the
range 90? to 180?. The field of view of the sensor can be
adjusted to 2? or 18? (in this study it is 2?).
2.2 Reflectance panel
The information about the spectral downwelling radiance is
obtained by measuring the radiance L,(A) reflected from a
reflectance panel (Spectralon) which is held horizontally. The
reflectance of Spectralon is known and the relationship that
relates the measured radiance to the incident irradiance E4(A) is
given by:
E44) QU). Vp, - Lol.) (1)
Where Q(X) = angular and wavelength dependent factor
relating the radiance to the irradiance (Q(A) = m for a
Lambertian surface) Here, p, is the irradiance reflectance of
Spectralon (Mobley, 1999).
The surface of the panel must be cleaned regularly by gently
rubbing the surface with a fine-sand paper under steady stream
of distilled water, to get rid of the specularly reflecting film at
the panel surface, as well as other particles like the seasalt.
The reflected diffuse sky radiance signal is estimated by
measuring the reflected radiance signal from the shaded
reflectance panel. This provide an independent estimate of the
sky contribution, and it is also useful for estimating the direct
solar influence on the reflected signal. The shadowing on the
panel should be done with a sun shade which is as far as
possible from the panel, and the geometrical configuration
Internation
mere
(angles, the
the sensor)
2.3 Sky ra
The spectr:
ASD sensc
which pre:
viewing zel
whose spec
beam eque
measureme
interest sil
uniform,
dependency
sky.
24 Remo
The reflect
formula:
Where 6, «
wavelength
is the wate
the viewir
component
seasurface
radiance (
seasurface |
where Lt s
pointed to
azimuth an
the total ra.
by supposi
reflected b:
Here, p is |
reflected fr
radiance at
starting fr
geometry a
assuming
Lambertiar
3.
Within the
been perfo
