The purpose of this handbook is to provide basic information about the 1997 release of the Geosat altimeter JGM-3 geophysical data records (GDRs).
From March 1985 until January 1990, the U.S. Navy satellite Geosat generated a new data set with unprecedented spatial and temporal coverage of the global oceans (Douglas and Cheney, 1990). Designed and built at the Johns Hopkins University Applied Physics Laboratory, Geosat carried a radar altimeter which produced profiles of sea level, wave height, and wind speed along the satellite ground track. Such records have applications in many areas of geodesy, ocean dynamics, and global climate research. Experience with GEOS-3 and Seasat in the 1970s had demonstrated the enormous potential of altimetry, but neither mission provided such complete long-term global coverage. Geosat paved the way for a series of highly-successful altimeter missions that followed it: ERS-1 (1991-96), TOPEX/POSEIDON (1992- ), and ERS-2 (1995- ).
Geosat was launched on March 12, 1985, and altimeter data collection began on March 30. The acronym "Geosat" was derived from geodetic satellite, because its primary mission was to obtain a high-resolution description of the marine geoid up to latitudes of 72 degrees. This goal was achieved during the first 18 months, known as the geodetic mission (GM). During this time the ground track had a near-repeat period of about 23 days (330 revolutions in 23.07 days; average orbital period of 6039.84 sec). The drifting orbit resulted in a dense, global network of sea level profiles separated by about 4 km at the equator. Because of the military significance of this unique set of observations, the GM data were initially classified but in 1995 were released to NOAA in their entirety for public distribution.
At the conclusion of the GM on September 30, 1986, the satellite orbit was changed, and the exact repeat mission (ERM) began on November 8, 1986. This produced sea level profiles along tracks that repeated themselves within 1-2 km at intervals of about 17 days (244 revolutions in 17.05 days; average orbital period of 6037.55 sec). The ERM covered 62 complete 17-day cycles before tape recorder failure in October 1989 terminated the global data set. A limited amount of data was subsequently collected by direct broadcast in the North Atlantic and Gulf of Mexico. By January 1990 continued degradation of the altimeter output power finally ended the Geosat mission.
The JGM-3 GDRs for Geosat represent the culmination of a decade of progress in refining the data set first distributed by NOAA in 1987. The previous release of the ERM data, referred to as the "T2" GDRs, is documented by Cheney et al. (1991). The initial full release of the GM data in 1995 was based on the original NSWC orbits. The 1997 release comprises the entirety of the GM + ERM data, with consistent JGM-3 orbits and enhanced geophysical corrections.
Most of the improvement has come in the area of orbit determination. For example, the initial satellite orbits for the ERM contained radial errors of 2-3 m. These were reduced to the 50-cm range in the T2 release in 1991. The present ephemeris, computed by the NASA Goddard Space Flight Center, is accurate at the level of 10 cm for most of the mission (the exception being the last year of the ERM mission, see Chapter 5 Precautions).
This remarkable achievement was made possible by (1) the 1993 release of the complete archive of Doppler tracking data and (2) development of the Joint Gravity Model-3 (JGM-3) (Tapley et al., 1994). Derivation of the JGM-3 orbit for Geosat, and the improvements gained from it, are discussed by Williamson and Nerem (1994). In the process of replacing the old NSWC (GM) and T2 (ERM) orbits, a timing bias correction of 5 msec was also applied to the GDRs. The time tags on the altimeter data were increased by a constant 5 msec over the duration of the mission, and the JGM-3 orbits were interpolated at the corrected GDR times. This timing bias correction greatly reduces the "twice-per-rev" orbit errors present in earlier releases of the Geosat data.
In addition to the improved orbits, nearly all of the geophysical corrections have been upgraded:
The enhanced GDRs yield a relatively consistent 13-cm rms global crossover difference (computed over periods of about 3 weeks) during the first 3.5 years of the mission, as shown in the figure below. By comparison, a value of about 10 cm is obtained for similar spans of data from Topex/Poseidon (T/P), the present state-of-the-art altimeter satellite.
The Geosat data can therefore be used, without performing separate orbit adjustments, to derive relatively accurate sea level variability. An example is provided by comparison with the tide gauge at Christmas Island in the central equatorial Pacific Ocean. For the 3.5-year period beginning in April 1985, monthly averages computed from Geosat crossover differences in an 8x1 (longitude x latitude) cell adjacent to the gauge yield an rms difference of 2.9 cm with 0.96 correlation, as shown below. Comparison of T/P data with the same tide gauge yields 1.4 cm and 0.98 correlation for collinear differences over a similar time span during 1992-96.
