NASA's satellite altimetry program was originally formulated at the 1969 Williamstown Conference (Kaula, 1970). An altimeter was actually used to scan the Moon by Apollo 14 (Kaula et al, 1974). The first altimeter used for experimental Earth measurements was operated on the Skylab manned space station (McGoogan et al., 1974). The first altimeter on an unmanned satellite was the Geodynamics Explorer Ocean Satellite 3 (GEOS-3), launched in 1975 (Stanley, 1979). Geos-3 provided useful data on eddy variability of the Gulf Stream region of the Atlantic ocean and demonstrated the potential of satellite altimetry for oceanography and marine geodesy (e.g., Douglas et al, 1983). In 1978, the Seasat altimeter satellite was launched. In addition to the altimeter, it carried a microwave radiometer for correcting the altimeter measurements for delays caused by tropospheric water vapor. While Seasat's lifetime was only 3 months, it provided the first data sets sufficiently accurate to be used for global oceanographic studies (Lame and Born, 1982; Cheney et al, 1983).
Geosat, launched in 1985, had performance similar to that of Seasat except it did not carry a microwave radiometer. It was operational for nearly 4 years and achieved many successes (see/. Geophys. Res. 95(C3) and 95(C10), 1990). In 1991, the European Space Agency launched the initial non-U.S. radar altimeter onboard ERS-1. It provided better performance than the Geosat altimeter, but still did not meet the requirements for regional or global sea level change studies as defined for the T/P mission.
T/P, launched in 1992, ushered in an entirely new era in satellite altimetry. Both the altimeter and orbit errors were only a few centimeters, giving sea level measurements accurate to 3-4 cm. These improvements were achieved by (1) making altimeter measurements at two different radio frequencies (Ku and C bands), which enabled the frequency-dependent ionosphere errors to be removed; (2) making simultaneous measurements of nadir columnar water vapor using a microwave radiometer for correcting the delay due to the wet troposphere; and (3) improving the orbit determination accuracy by using improved tracking techniques and geopotential models. Remarkably, T/P has been collecting measurements for 7 years as of this writing, well past its design life. The results shown in this chapter are entirely based on T/P, but a summary of the early sea level change results from Seasat, Geosat, and ERS-1 will be given to place the T/P results in a historical context.
A number of attempts have been made to measure global mean sea level variations from earlier satellite altimeter missions. Results using Seasat's 3-day repeat orbit showed 7-cm variations for estimates of global sea level over a month (Born et al., 1986). Tapley et al. (1992) used 2 years of Geosat altimeter data and found 17-day values of variations in mean sea level with an RMS error of 2 cm and a rate of 0 ± 5 mm/yr. The largest errors were attributed to the orbit determination, ionosphere, wet troposphere delay corrections, and unknown drift in the altimeter bias (not independently calibrated for Geosat). Wagner and Cheney (1992) used a collinear differencing scheme and 2.5 years of Geosat altimeter data to determine a rate of global sea level rise of -12 ± 3 mm/yr. When this data was compared to a 17-day Seasat data set, a value of +10 mm/yr was found (Wagner and Cheney, 1992). The RMS error of the Geosat variations was still a few centimeters, even after the application of several improved measurement corrections. The ionosphere path delay correction was identified as the single largest error source, but there were many other contributions including errors in the orbit, wet troposphere correction, ocean tide models, altimeter clock drift, and drift in the altimeter electronic calibration. Since the Geosat study of Wagner and Cheney (1992), several improvements have been made to the Geosat altimeter measurement corrections (ionosphere, tides) and the orbit determination. However, Nerem (1995b) and Guman (1997) still find that Geosat mean sea level measurements are not of sufficient quality to allow a determination of the rate of mean sea level change accurate to 1 mm/yr, although the latter study succeeded in developing a tie between Geosat and T/P that allowed a reasonable estimate of sea level change (+1.0 ± 2.1 mm/yr) over the decade spanning the two missions. Some investigators have examined ERS-1 altimeter data for long-term sea level change (Anzenhofer and Gruber, 1998; Cazenave et al., 1998; Guman, 1997), but in general these results are less accurate because the ERS-1 altimeter is single frequency, and the orbits are less precise than T/P. Due to instrument problems, ERS-2 will be of limited use for studies of mean sea level change (Moore et al., 1999). The Geosat Follow-On (GFO) mission has suffered from instrument problems since its launch in 1998, although the mission engineers are still optimistic that this satellite will one day produce useful altimeter measurements. Clearly T/P altimeter data represent the highest accuracy currently attainable from satellite altimetry.
Was this article helpful?