11 The Marine Geoid and Satellite Altimetry
191
2008). EGM2008 geoid heights were compared against land geodetic surveys, tide
gauges, and long-term means of Topex/Poseidon/Jason altimeter data. RMS errors
in geoid height are around 5 cm, and RMS errors in deflections of the vertical are
around 0.3 a.s. (1.5 μrad). The geoid height at a point is sensitive to gravity anomalies over a wide area (Heiskanen and Moritz, 1967), and so the coastal tide gauge
and land survey comparisons are sensitive to the accuracy of the deep ocean data.
It should be noted that resolution of the marine gravity field by altimetry requires
spatially dense sampling. The presently available data come from the Geosat
Geodetic Mission (GM) of 1985–1986 and the ERS-1 Geodetic Phases E and F of
1994–1995. All other precise altimeter data have been acquired along exact repeat
mission ground tracks, the cross-track spacing of which is too large to resolve the
gravity anomalies in Fig. 11.3.
11.8 Implications for Oceanographic Satellite Altimetry
Satellite altimetry developed as a tool for observing ocean dynamics at a time
when the gravity field was not known well enough to furnish a geoid of sufficient accuracy to permit direct calculation of the dynamic topography. “Exact repeat
mission” (ERM) orbits were designed so that the satellite’s ground track would
repeat within ±1 km after a fixed number of orbital revolutions and rotations of
the Earth with respect to the satellite’s orbital plane, called synodic “days”. By
making repeated observations on the same ground tracks, altimeters furnished temporal changes in sea surface height that could be interpreted as changes in tidal
and dynamical signals, without requiring knowledge of the geoid height along the
tracks.
Different satellite series have used different ERM orbits with different space
and time sampling characteristics, with observations spanning different years and
decades. The long-term average sea surface height from each series has different
tidal aliasing and error characteristics, and different sampling of decadal and longer
temporal variations, complicating the blending of observations from different ERM
orbits. Even so, oceanographers are now in the habit of assuming that they must use
ERM orbits to observe their signals, that geoids are not sufficient to allow direct
observation of dynamic height signals, and that height anomalies must be referred
to a long-term average of sea surface height, rather than a geoid model.
In fact, however, errors in current geoid models are small, around 5 cm in
height and 1.5 μrad in slope (these are RMS values), and appear to be confined
to full-wavelengths shorter than 20 km, that is, spatial scales much shorter than the
correlation scales of some dynamical signals in the open ocean (Jacobs et al., 2001).
These virtues permit the observation of dynamic ocean signals from orbits other than
traditional oceanographic ERMs. For example, Scharroo and Smith (2009) find that
mesoscale eddies can be observed equally well from ERM and non-ERM orbits.
Scharroo and Smith’s result has prompted others to examine whether non-ERM
orbit data can be used to improve tide models (W. Bosch, personal communication,
at the 2nd Coastal Altimetry Workshop, Pisa, Italy, 2008).
191
2008). EGM2008 geoid heights were compared against land geodetic surveys, tide
gauges, and long-term means of Topex/Poseidon/Jason altimeter data. RMS errors
in geoid height are around 5 cm, and RMS errors in deflections of the vertical are
around 0.3 a.s. (1.5 μrad). The geoid height at a point is sensitive to gravity anomalies over a wide area (Heiskanen and Moritz, 1967), and so the coastal tide gauge
and land survey comparisons are sensitive to the accuracy of the deep ocean data.
It should be noted that resolution of the marine gravity field by altimetry requires
spatially dense sampling. The presently available data come from the Geosat
Geodetic Mission (GM) of 1985–1986 and the ERS-1 Geodetic Phases E and F of
1994–1995. All other precise altimeter data have been acquired along exact repeat
mission ground tracks, the cross-track spacing of which is too large to resolve the
gravity anomalies in Fig. 11.3.
11.8 Implications for Oceanographic Satellite Altimetry
Satellite altimetry developed as a tool for observing ocean dynamics at a time
when the gravity field was not known well enough to furnish a geoid of sufficient accuracy to permit direct calculation of the dynamic topography. “Exact repeat
mission” (ERM) orbits were designed so that the satellite’s ground track would
repeat within ±1 km after a fixed number of orbital revolutions and rotations of
the Earth with respect to the satellite’s orbital plane, called synodic “days”. By
making repeated observations on the same ground tracks, altimeters furnished temporal changes in sea surface height that could be interpreted as changes in tidal
and dynamical signals, without requiring knowledge of the geoid height along the
tracks.
Different satellite series have used different ERM orbits with different space
and time sampling characteristics, with observations spanning different years and
decades. The long-term average sea surface height from each series has different
tidal aliasing and error characteristics, and different sampling of decadal and longer
temporal variations, complicating the blending of observations from different ERM
orbits. Even so, oceanographers are now in the habit of assuming that they must use
ERM orbits to observe their signals, that geoids are not sufficient to allow direct
observation of dynamic height signals, and that height anomalies must be referred
to a long-term average of sea surface height, rather than a geoid model.
In fact, however, errors in current geoid models are small, around 5 cm in
height and 1.5 μrad in slope (these are RMS values), and appear to be confined
to full-wavelengths shorter than 20 km, that is, spatial scales much shorter than the
correlation scales of some dynamical signals in the open ocean (Jacobs et al., 2001).
These virtues permit the observation of dynamic ocean signals from orbits other than
traditional oceanographic ERMs. For example, Scharroo and Smith (2009) find that
mesoscale eddies can be observed equally well from ERM and non-ERM orbits.
Scharroo and Smith’s result has prompted others to examine whether non-ERM
orbit data can be used to improve tide models (W. Bosch, personal communication,
at the 2nd Coastal Altimetry Workshop, Pisa, Italy, 2008).