Fig 1. Spectral range of the channels of the
sensors used during the "Adria 84"
water, this making soffisticated methods a must to
finish up with an acceptable precision.
The processing of the CZCS data consists essentially
of the following phases:
- Sensor sensitivity and callibration. The CZCS has
an internal calibration system, which has never
indicated an appreciable deterioration of the
sensors (except for CH 6, the thermal channel).
Nevertheless there were clear indications of a
deterioration, obviously an optical surface outside
the calibration circuit. Deterioration was
negligible for CH 4, but increased substantially
with decreasing wavelength for the other channels, a
phenomenon indicating (cosmic) dust deposit as the
most probable cause. Sensitivity loss coefficients
have been established for the channels 1,2 and 3
relative to CH 4, using the upwelling radiation of
the clear-water pixels of all CZCS scenes ever
elaborated at the JRC .
- Atmospheric correction. The fact that the sea
reflects so few radiation makes the atmospheric
correction a must. Although the basis of the
correction is still the one proposed by Gordon,
quite some refinements have been added eversince.
Three effects are'of importance :
- absorption on the illumination source welling
down into the sea and on the reflected radiation
welling up to the sensor.
- Rayleigh scattering or molecular scattering
- aerosol scattering
The first 2 effects can be calculated straight
forward, the 3d effect is an everchanging quantity
depending on place and time.
- Chlorophyll and Total Suspended Substance (TSS)
determination. An emperical relationship has been
developed, based on in-situ measurements and the RS
signature, to establish CHL and TSS concentrations.
The CZCS spectral resolution does not allow to
distinguish between CHL and Yellow Substance. In
coastal regions with no fixed relationship between
CHL and YS, this might lead to substantial errors.
Part of this work has been done by a visiting
scientist, M.Viollier.
- Geometric correction. For comparison of different
images, they must be brought on a common
geometrical footing, in our case the Mercator
projection (projection cylinder going thru the
equator) . The classical geometrical correction
method fails when not sufficient well distributed
reference point are available, as is regularly the
case over sea area's. NASA provides for each image
so-callled anchor points, synthetic reference
points of which the location is given, derived from
sensor relative data.
Experience however has learned that the anchor
points are not sufficiently precise and can give
rise to errors of some tens of kilometres.
A procedure has been developed, essentially based
on sensor relative data, only corrected with a
minimum of ground control points, which need not to
be equally distributed over the scene. The program
has been developed under contract (M.Langemann,
A.Popella) under the leader ship of prof. Ph. Hartl,
of the Technical University of Berlin.
- Final product. The elaborated images are hard
copied for subsequent visual interpretation, or
once again geometrically corrected to fit the grid
used for the hydrodynamic model of the Adriatic Sea.
-NOAA AVHRR
The sea has a much more uniform temperature compared
to land, reason why sea temperature, when to be
significant, must be established with much greater
precision.
The thermal channels of the AVHRR operate with an
equivalent precision of around .1 K (at 293 K), but
due to the atmospheric presence, and in particular to
the water vapor, errors of several degrees are
possible, the measured value always being lower than
the real one.
The water vapor quantity can be established in two
ways :
- with the split window technique, e.i. derived from
the difference between CH 4 and CH 5. Only NOAA 9
is equipped with a split window, NOAA 6 has only 4
channels.
- by means of the HIRS (High resolution IR Sounder),
installed on both NOAA satellites, but for this
scope considered less precise. A procedure has been
developed, based on the split window technique by a
visiting scientist G.Dalu (Istituto di Fisica dell'
Atmosfera, CNR, Roma).
The procedure consists eesentially of the following
phases:
- calibration.
- determination of brightness temperature for CH 4
and CH 5
- determination of the quantity of water vapor
- determination of the SST (Sea Surface Temperature)
from CH 4 and the water vapor quantity.
The objective was to arrive at a precision of .3 K.
Up to now published SST maps refer essentially to
scenes with relative large temperature differences,
e.i. the gulf stream eddies, or the strait of
Messina, but the Adriatic Sea, especially during the
summer has very small and slow temperature
differences, which makes high precision necessary.
The SST derived from surface radiation refers only
to the utmost top layer of the water body (20
microns) and may be more than 1 K below the water
bulk temperature (at 30 cm depth), depending on water
evaporating conditions and water turbulence. Actually
no procedure for geometric correction of AVHRR images
is available. The processing program is implemented
on the Amdahl mainframe.
-Landsat TM
With the probable shutdown of the CZCS in near
future, the marine community will be deprived from a
specific Ocean Color Sensor for an unknown stretch of
time. The TM might constitute at least a partial
substitute for the CZCS.
The major problems with the TM are :
- less spectral resolution, broader channels, thus
more difficulties to distinguish CHL and TSS.
- small swath, thus low daily coverage
- early morning pass, thus low illumination,
specially in winter,
- very expensive on a covered surface basis.
The JRC will participate in the North Sea experiment
in May 1986 during which the TM capabilities as an
ocean color sensor will be investigated.
-New sensors
When no risk exist of over stretching the activity,
new sensors will be investigated, as substitute for
actual opei
marine par:
Specific i
microwave s
which promj
-Image Pro<
For complet
processing
Main 1
computer
VAX 785
PDP 11/24
PDP 11/24
PDP 11/23
Satellite
computer
VAX 750
The satelld
and connect
The appli<
in-house, e
really a ti
Image Proc«
Communitar d
arrive at i
software e>
at least a
IN-SITU ME:
While the .
and does nc
airplanes,
during camj
year and 1:
The campaij
measurement
- biologic:
- in-situ c
- physical
- chemical
- in-situ i
- airborn s
Often the <
sometimes t
by others.
The most m:
now was in
The partied
The JRC org
a concertec
bearing on!
the rent ol
and where t
ships or fd
institutior
its own me:
elaboratior
all to brir
Essentially
the partied
more instit
only succee
institutior
The JRC or
the North t
participate
site, like:
- 1980 (29.
- 1983 (10.
- 1986 (5.5
700