and improving the GEOSAT coverage. This altimetry ftle provides the necessary gravity
infonnation for computing the EGM in the ocean areas between 80° and -72° latitude.
ALTIMETRY SOURCES
GEOSAT produced 49 million data points in its 18 month Geodetic Mission(GM). These
are distributed over the latitude range of +72°. Intertrack spacing at the equator is 4
kilometers. The precision of the altimeter was 3.5 centimeters for a 2 meter significant
wave height(MacArthur, Marth, and Wall, 1987). The largest error source was radial orbit
(ephemeris) error. The radial orbit error before modeling or adjustment, based on crossing
point statistics, is approximately 80 centimeters but can be over 2 meters in local areas.
ERS-l generated 16 million points during its one year GM. These are distributed over the
latitude range of +81.5°. The intertrack spacing at the equator is 8 kilometers. KMS used
the D-P AF prepared Quick Look OPR ERS-l data in the computation of their June 1995
anomaly set(Anderson, Knudsen, and Tscherning, 1995). The initial radial orbit error of
ERS-l based on crossing points is approximately 22 centimeters. The improvement in
initial radial orbit error of ERS-l compared to GEOSAT is a result of more accurate
tracking, the denser DORIS tracking network, and better quality gravity models for more
precise ephemeris. Mter adjustment the ERS-l radial orbit accuracy based on crossing
points is 10 centimeters(Gruber, Massmann, and Reigber, 1993). The NOAA ERS-l
values were derived from 35 days of the full-echo wavefonn data using special techniques
for reducing altimetry over ice(Laxon and McAdoo, 1994).
DMA GEOSA T PROCESSING AND DERIVED PRODUCT
DMA's crossing point adjustment and editing. A least squares adjustment using crossing
points is a common method of modeling radial orbit error. However, DMA's procedure is unique
in several ways. It uses a control network. A control network is a set of arcs that are more
accurate but wider spaced than the geodetic mission data being adjusted to it DMA's procedure
uses crossings of both the GM arcs with the Network and the GM arcs with other GM arcs. The
Network arcs are held fIxed during the adjustment. Initially, the GEOSAT GM data were received
and adjusted in weekly sets to a SEASAT network. Later DMA readjusted the entire GEOSAT
GM data to a network developed by William E. Rankin at the U. S. Naval Oceanographic OffIce
(NA VOCEANO), Stennis Space Station, Bay St Louis, MS from the fIrst year of the
GEOSAT Exact Repeat Mission(ERM) data reduced with precise orbits computed by the
Naval Surface Warfare Center, Dahlgren, VA. NAVOCEANO average the up to 24
collinear revolutions for each ground track. The resulting 244 revolutions, representing
each distinct ground track, were then adjusted using the method ofCloutier(1983). The
DMA adjustment was by diamond shaped areas bounded by network af(;S. Automated
crossing point cell editing algorithms, developed by DMA, were used to identify bad or
suspicious data. Only edited GEOSAT data were used in the computation of the 5'x5' mean
geoid heights which in turn were used to compute the 30'x30' mean gravity anomalies
One degree cell statistics of adjusted crossing points provide a measure of the regional
variation and consistency of the data The readjustment improved the RMS of the majority
of cells from about 13 centimeters to about 7 centimeters(see Figure 1). Before the
readjustment there is evidence of remaining orbital error(Figure 2). No such evidence is
seen after the readjustment Major currents cause a valid geographically correlated
variability that is seen in the crossing point statistics(Figure 3). Mter the readjustment, the
overall variability in the current areas has been reduced but there is an even stronger contrast
between the current area and the surrounding cells(Figure 3). The remaining variability after
72
infonnation for computing the EGM in the ocean areas between 80° and -72° latitude.
ALTIMETRY SOURCES
GEOSAT produced 49 million data points in its 18 month Geodetic Mission(GM). These
are distributed over the latitude range of +72°. Intertrack spacing at the equator is 4
kilometers. The precision of the altimeter was 3.5 centimeters for a 2 meter significant
wave height(MacArthur, Marth, and Wall, 1987). The largest error source was radial orbit
(ephemeris) error. The radial orbit error before modeling or adjustment, based on crossing
point statistics, is approximately 80 centimeters but can be over 2 meters in local areas.
ERS-l generated 16 million points during its one year GM. These are distributed over the
latitude range of +81.5°. The intertrack spacing at the equator is 8 kilometers. KMS used
the D-P AF prepared Quick Look OPR ERS-l data in the computation of their June 1995
anomaly set(Anderson, Knudsen, and Tscherning, 1995). The initial radial orbit error of
ERS-l based on crossing points is approximately 22 centimeters. The improvement in
initial radial orbit error of ERS-l compared to GEOSAT is a result of more accurate
tracking, the denser DORIS tracking network, and better quality gravity models for more
precise ephemeris. Mter adjustment the ERS-l radial orbit accuracy based on crossing
points is 10 centimeters(Gruber, Massmann, and Reigber, 1993). The NOAA ERS-l
values were derived from 35 days of the full-echo wavefonn data using special techniques
for reducing altimetry over ice(Laxon and McAdoo, 1994).
DMA GEOSA T PROCESSING AND DERIVED PRODUCT
DMA's crossing point adjustment and editing. A least squares adjustment using crossing
points is a common method of modeling radial orbit error. However, DMA's procedure is unique
in several ways. It uses a control network. A control network is a set of arcs that are more
accurate but wider spaced than the geodetic mission data being adjusted to it DMA's procedure
uses crossings of both the GM arcs with the Network and the GM arcs with other GM arcs. The
Network arcs are held fIxed during the adjustment. Initially, the GEOSAT GM data were received
and adjusted in weekly sets to a SEASAT network. Later DMA readjusted the entire GEOSAT
GM data to a network developed by William E. Rankin at the U. S. Naval Oceanographic OffIce
(NA VOCEANO), Stennis Space Station, Bay St Louis, MS from the fIrst year of the
GEOSAT Exact Repeat Mission(ERM) data reduced with precise orbits computed by the
Naval Surface Warfare Center, Dahlgren, VA. NAVOCEANO average the up to 24
collinear revolutions for each ground track. The resulting 244 revolutions, representing
each distinct ground track, were then adjusted using the method ofCloutier(1983). The
DMA adjustment was by diamond shaped areas bounded by network af(;S. Automated
crossing point cell editing algorithms, developed by DMA, were used to identify bad or
suspicious data. Only edited GEOSAT data were used in the computation of the 5'x5' mean
geoid heights which in turn were used to compute the 30'x30' mean gravity anomalies
One degree cell statistics of adjusted crossing points provide a measure of the regional
variation and consistency of the data The readjustment improved the RMS of the majority
of cells from about 13 centimeters to about 7 centimeters(see Figure 1). Before the
readjustment there is evidence of remaining orbital error(Figure 2). No such evidence is
seen after the readjustment Major currents cause a valid geographically correlated
variability that is seen in the crossing point statistics(Figure 3). Mter the readjustment, the
overall variability in the current areas has been reduced but there is an even stronger contrast
between the current area and the surrounding cells(Figure 3). The remaining variability after
72