June 15, 1999

Science Rationale for Seasat SAR Processing - Draft

Ben Holt

Summary.  Seasat SAR obtained extensive ocean and land coverage of North
America and portions of Europe plus surrounding seas during July-October
1978.  This L-band frequency,  HH-polarization data set represents an
historical archive of the land surface, that can be of considerable
scientific value for comparison with more recent and upcoming SAR sensors
to examine climate and natural hazard change detection.  For nearly a
decade, the data have been largely unavailable for processing and use.
With the recent transfer of the Seasat SAR raw data to Sony D1 media
through efforts at the Jet Propulsion Laboratory, we recommend that the
entire Seasat SAR data catalog be reprocessed and made available for
scientific use.

Background.  The Seasat SAR provided extensive of coverage of North America
(Figures 1-3) and portions of eastern Canada (Figure 4) and the United
Kingdom (Figure 5) from July -October 1978.  Within the three US ground
stations (Goldstone, Merritt Island, Fairbanks), nearly complete coverage
of the land areas was obtained during the mapping phase of the mission as
well as considerable acquisitions during the exact 3-day repeat orbital
periods of Seasat.  Also covered were surrounding oceans and sea ice.
Being the first spaceborne SAR imagery and of high quality, the data
provided extraordinary use for ocean and sea ice studies in particular plus
geology and other land investigations including interferometry. In fact,
because of the gain problems with both SIR-B (1984) and JERS-1 (1992-1998),
Seasat remained the best quality L-band data set until the flights of SIR-C
in 1994.  In the future, LightSAR and ALOS will carry L-band sensors.

	The Seasat SAR acquisitions totaled nearly 48 hours of
single-channel L-band HH-polarization data with 100 km wide swaths at a
fixed range of incidence angles between 20-26 degrees.  Over 95% of this
data was processed in a survey mode using optical laser techniques.  Using
the first digital SAR processor ever developed, approximately 25% of the
total data set were processed.  In 1988, the original raw data was
transferred from Honeywell high density recorder media to the more modern
Thorne EMI high density recorder media.  Processing continued on the
original SEL computer with FPS array processing until the early 1990's.
Due to high costs for maintaining aging hardware, the Seasat SAR processor
was officially decommissioned in 1995 after several years of inoperative
hardware.  In late 1997 JPL was able to identify sufficient funds to
transfer the raw data from Thorne to Sony D1 cassettes. This task was
completed in late 1998 and the tapes were subsequently shipped to ASF in
early 1999.

Scientific Value of Seasat SAR.  This data set proved to be extremely
valuable for a wide spectrum of scientific disciplines as well as for
exploring new techniques including interferometry and for geophysical
processing.  The sea ice data was used for ice motion, the fundamental
product of the ASF sea ice geophysical processor for ERS-1 and RADARSAT.
The ocean data provided the first comprehensive time-series of ocean waves,
internal waves, current features, and atmospheric interactions.  The
principal land uses were for mapping of geologic structures, glacial
extent, wetlands extent, and agriculture.  Additionally, the repeat-pass
data was used successfully in interferometry studies of topography.  On the
down-side, however, the data were not calibrated or at least very difficult
to calibrate, approaching absolute levels of 2-3 dB.  Also, the 3-day
repeat phase resulted in decorrelation over some types of surfaces for
interferometry.  However, the decorrelation problems were not extensively
evaluated over all conditions and key areas of interest, due to limited
processing throughput when this technique became widely publicized.  To
some extent, the comparatively limited processing capabilities available in
those early periods of digital processing (low throughput rates, lack of
standard geocoding product, lack of calibration) reduced interest.  With
current technology and processing capabilities, reprocessing Seasat SAR
would enable fuller and more complete investigation of the dataset
especially in terms of time series, mosaicing, and strip interferometry.

Current value of Seasat SAR data.  The most significant use of Seasat SAR
currently would be for examining climate and natural hazard land surface
change detection.  Seasat SAR data provides the following key qualities:
The nearly complete coverage of the continental US and Alaska could be used
for examining changes in land use, wetland extent, fire extent, and
deforestation; Glacial extent and conditions could be examined in Alaska,
Canada, Iceland, and limited portion of Greenland;  Nearly complete mosaics
within the 3 US station masks could be generated;  Interferograms would be
useful for determining change detection over land and glaciers;  The sea
ice and ocean data retain considerable interest, although by and large
C-band data is of higher value particularly for sea ice.  If available,
this data could be extremely useful for comparison with LightSAR, an L-band
interferometry-based mission with change detection as its highest science
priority.

Recommendations.  Specifically, the Seasat SAR data should be made
available in the following three formats: 1) Survey strip images processed
to a reduced resolution of 100m (with at least 16 looks) and available
on-line for browsing and distribution; 2) Full resolution images (25m with
4-looks)  either as 100 km by 100 km frames or as strip images, with
processing on demand; 3) Level 0 format suitable for interferometric
processing.  Furthermore, we recommend that the Seasat SAR data should be
4) Made available either through a stand-alone web-based ordering system
and/or through EOSDIS;  5) Radiometrically calibrated to 3 dB absolute and
1.5 relative radiometric values or better; and 6) Geometrically calibrated
to 500 m location accuracy or better.  Currently, the Sony tapes reside at
the Alaska SAR Facility.  We recommend that NASA supports the
implementation of the processing, catalog and distribution, and calibration
capabilities for this uniquely valuable active microwave fine-resolution
data set.


Draft -  1) Needs references added 2) need to scan coverage maps from
Pravdo et al-JPL pub
3) examine format of Seasat refined ephemeris available through JPL/U.Texas
for use in processing 4) Expand base of interested users