ility
onal
ause
lese
lites
ning
RS
tion
onal
ents
RS
RS
985;
th's
ites.
the
rs to
the
d by
tion,
lite
se is
ot of
tice,
| RS
ility
ems,
has
This
rture
face
wave
ered
ginal
e of
cted
ight.
, the
nder
rain,
SAT
1 28
ness,
any
| EO
jonal
rk of
ard
data recorders which enable data acquisition even from areas
outside the range of ground stations. It is expected that eight
ground stations will be receiving the RADARSAT SAR data at
the end of 1997. The RADARSAT SAR system works at 5.3
GHz frequency (C-band), corresponding to 5.6 cm wavelength,
with horizontal polarization (HH). Its unique feature is the
wide choice of imaging modes. There are 25 possible choices,
depending on the selection of incidence angle (SAR beam
position), ground resolution (SAR beam operational mode)
and size of RADARSAT SAR scene. The choice of incidence
angles, at which the ground scene can be viewed by the
RADARSAT’s SAR steerable antenna, ranges from 10 to 59
degrees. This enables frequent coverage of selected areas,
which is important for monitoring the impacts of natural
disasters, such as floods. RADARSAT SAR image data can be
recorded at 6 different ground resolutions, ranging from 10m
to 100m. Its actual ground resolution is better than the
specified nominal values. For example, the first tests indicate
that the actual fine resolution is between 8m and 8.5m, rather
than 10m (see below). The size of RADARSAT SAR scenes
ranges from 50km x 50km (recorded with 10m ground
resolution) to 500km x 500km (recorded with 100m ground
resolution). This provides an adequate flexibility to select the
imaging mode best suited to a particular application. (Engel et
al., 1995; Nazarenko et al., 1996).
Analysis of first RADARSAT images has already yielded
better than expected results. For example, a complex pattern
of forest clearcuts, which is usually difficult to delineate on
SAR images, was clearly delineated on RADARSAT SAR
image recorded at a shallow incidence angle (43-46 degrees)
with actual ground resolution between 8m and 8.5m. (Ahern &
Banner, 1996).
The strongest benefits of satellite SAR image data are
expected in monitoring applications, when data have to be
often recorded under adverse weather conditions, unsuitable
for optical RS systems. For example, monitoring of extensive
floods, oil spills or icebergs is needed day and night, during all
weather conditions. Furthermore, the combination of image
data from the SAR and optical RS systems will usually
increase the interpretability of land cover classes and thus
result in higher information contents and accuracies of natural
resources assessments and land cover maps.
The interpretation of SAR images requires different expertise
and skills than the interpretation of images from optical
sensors. In order to overcome this problem, the Canada Centre
for Remote Sensing (CCRS), with funding support from the
Canadian Space Agency (CSA) and International Development
Research Centre (IDRC), has initiated an innovative
GlobeSAR program for training prospective users of
RADARSAT SAR images in developing countries. The
GlobeSAR training, which is implemented by CCRS jointly
with private sector companies, is based on pilot projects and
focused on the effective application of SAR images to tasks
defined by participants. (Campbell, 1993, 1994 & 1995;
St-Pierre, 1995).
2.2 NOAA/TIROS Series of EO Satellites
The NOAA/TIROS series of polar orbiting meteorological
satellites operated by the U.S. National Oceanic and
Atmospheric Administration. (NOAA) is of particular interest
to global environmental monitoring programs, because of its
Advanced Very High Resolution Radiometer (AVHRR).
Actually, the ground resolution of AVHRR data is only 1.1 km
at best, which is rather coarse for land applications, but these
are very high resolution data for meteorological applications.
AVHRR image data are recorded at 4-5 spectral bands (red,
near-IR, mid-IR and 1-2 thermal IR). The size of AVHRR
image scenes is about 2500 km x 2500 km. Twice daily
global coverage by NOAA/TIROS satellites, coupled with
about 1 km ground resolution yielding a manageable size of
regional and global datasets, and with relatively low-cost
access to data, have contributed to their popularity in global
and regional vegetation cover monitoring programs. Processing
of AVHRR data into normalized difference vegetation index
products (NDVI) further enhanced their usefulness for
vegetation monitoring. The NDVI products, sometimes
combined with the AVHRR thermal imagery, have been
successfully used for the regional and global monitoring of
forest cover, forest and grassland fires, assessment of
agricultural drought risk, monitoring of desert locust recession
areas, etc. (Cihlar et al., 1996 a & b; Gutman, 1991; Gutman
415
& Ignatov, 1994; Townshend, 1994; Tucker et al, 1985).
2.3 Landsat Program
The Landsat program of EO satellites started in 1972, as a
civilian spin-off from the military satellite reconnaissance
technology. While the meteorological applications of satellite
RS started about a decade earlier, it was the success of
Landsat program which provided the basis for mapping and
monitoring of the Earth’s surface from space platforms. The
main sensor system of the first 3 satellites was a Multispectral
Scanner (MSS), with 4 spectral bands (green, red, and 2 near-
IR) and about 80m ground resolution. Although the MSS has
been retained in the RS payload of Landsats 4 & 5 in order to
provide a continuity of MSS coverage, the more advanced
Thematic Mapper (TM) has become their main sensor system.
It has 7 spectral bands (blue, green, red, near-IR, 2 mid-IR and
thermal-IR), with ground resolution of 120m for the thermal-
IR band and 30m for the remaining 6 bands. The size of MSS
and TM scenes is 185km x 185km, covering about 34000
sq.km of the Earth’s surface. Such a large area of Landsat
scenes covered with 30m ground resolution in 6 optical
spectral bands, and the availability of long-term Landsat
database, are the main comparative advantages of Landsat
program. Its weakness is the uncertainty about its future in the
near term (before the launch of Landsat 7, and long term (the
follow-on to Landsat 7). Landsat 6 failed to reach its
operational orbit. Landsat 7 is scheduled for launch in 1998
as part of the EOS program (Section 3.3).
2.4 SPOT Program
The first of three SPOT satellites was launched by the French
Space Agency (CNES) in 1986. SPOT satellites have two
identical RS sensor systems onboard, the High Resolution
Visible (HRV). They can operate in either panchromatic
mode, producing a single-band image with 10m ground
resolution, or in multispectral mode, in which 3 spectral bands
are recorded (green, red and near-IR) with 20m ground
resolution. The size of a SPOT scene is 60km x 60km,
covering 3600 sq.km of the Earth's surface. SPOT has
introduced a steerable sensor system, which enables recording
of ground scenes located up to 450km on each side of SPOT
International Archives of Photogrammetry and Remote Sensing. Vol. XXXI, Part B4. Vienna 1996
