Results From The Topexposeidon Mission Overview of the Achievements of Topexposeidon

There is a rich collection of journal publications describing the scientific results from the TOPEX/POSEIDON mission. The initial results are described in two special issues of the Journal of Geophysical Research: "TOPEX/POSEIDON: Geophysical Evaluation," 99(C12), 24,369-25,062, 1994, and "TOPEX/ POSEIDON: Science Results," 100(C12), 24,893-25,382, 1995. The former collection of papers contains overviews of the mission, gravity model development, orbit determination, and evaluations of data quality and accuracy, while the latter focuses more on scientific problems such as ocean tides, the mean sea surface, global sea level change, sea level variability, and ocean circulation. Another major scientific accomplishment not covered in these special issues is the monitoring of ENSO events using TOPEX/POSEIDON altimetry (Chel-

Sensitivity of TOPEX drift rate to varying land motion assumptions

For ocean rate of 1.5 (0.5), drift estimate is about -0.4 (0.4)

Sensitivity of TOPEX drift rate to varying land motion assumptions

For ocean rate of 1.5 (0.5), drift estimate is about -0.4 (0.4)

Different curves are for different geodetic length scales

Figure 6.7 A sensitivity test of the method used to assign land motion estimates to the tide gauge sea levels. There are two parameters in the land motion corrections: the true ocean rise rate and the length scale over which the variance in the geodetic estimates of the land motion rate increases. The T/P drift rate was estimated using a wide range of these parameters and these rates are plotted versus the assumed ocean rise rate. The different lines on the figure correspond to different values for the length scale for the variance increase of the geodetic rate. The technique is sensitive to the ocean rise rate chosen, but relatively insensitive to the geodetic length scale chosen.

Different curves are for different geodetic length scales

Figure 6.7 A sensitivity test of the method used to assign land motion estimates to the tide gauge sea levels. There are two parameters in the land motion corrections: the true ocean rise rate and the length scale over which the variance in the geodetic estimates of the land motion rate increases. The T/P drift rate was estimated using a wide range of these parameters and these rates are plotted versus the assumed ocean rise rate. The different lines on the figure correspond to different values for the length scale for the variance increase of the geodetic rate. The technique is sensitive to the ocean rise rate chosen, but relatively insensitive to the geodetic length scale chosen.

ton and Schlax, 1996). Here we provide an overview of only the results on estimating global mean sea level variations using TOPEX/POSEIDON altime-try. There have been numerous papers describing the global mean sea level variations observed by T/P (Cazenave et al, 1998; Minster et al, 1995, 1999; Nerem, 1995a,b; Nerem et al, 1997a, 1999) and their spatial variation (Hendricks et al, 1996; Nerem et al., 1997b). Altimeter measurements have also been used to monitor mean water level in semienclosed or enclosed seas (Cazenave et al., 1997; Larnicol et al., 1995; and Le Traon and Gauzelin, 1997 among others) and lakes (e.g., Morris, 1994; Birkett, 1995) and rivers (e.g., Birkett, 1998), although these applications are outside the scope of this chapter.

6.4.2 Measuring Changes in Global Mean Sea Level

For the purposes of completeness, a summary of our processing of the T/P data for the determination of global mean sea level variations will be presented. While the T/P data have been processed in a variety of different ways (Cazen-ave et al, 1998; Minster et al, 1995,1999; Nerem, 1995a,b; Nerem et al., 1997a, 1999), the results are all quite similar. The data processing employed for the results presented here is essentially identical to that used in Nerem (1995a,b, 1997a) and thus will not be reproduced in detail.

In our processing, mean sea level variations are computed every 10 days by using equi-area-weighted averages (Nerem, 1995b) of the deviation of sea level from a 6-year along-track mean (1993-1998) computed exclusively from the T/P data (as opposed to using a more general multimission mean sea surface). Although T/P cannot measure "global" mean sea level because it covers only ±66° latitude, tests have indicated the mean sea level estimates are very insensitive to this gap in latitudinal coverage (Minster et al, 1995; Nerem, 1995b).

All of the usual altimeter corrections discussed earlier in this chapter (ionosphere, wet/dry troposphere, ocean tides, sea state, etc.) have been applied to the data, with one exception. No inverted barometer (IB) correction was applied to these data (Nerem, 1995a,b), because we were concerned about errors in the IB correction (Fu and Pihos, 1994, Raofi, 1998). Although improvements in the IB correction have been made (Dorandeu and Le Traon, 1999; Raofi, 1998), we argue it is the total sea level change signal that is of interest, and not its IB-corrected equivalent. If there was a secular change in mean atmospheric pressure, resulting in a secular change in mean sea level, we do not want to remove this signal from the results. Nevertheless, the IB contribution to secular changes in mean sea level over the T/P mission is less than 1 mm/yr (Dorandeu and Le Traon, 1999). Since the CSR 3.0 ocean tide model was used (which was developed with IB-corrected T/P data), not applying the IB correction introduced a slight error with a 58.7-day period associated with the S2 atmospheric pressure tide normally modeled in the IB correction.

Modified geophysical data records (MGDRs) covering Cycles 10-233 from both the TOPEX and POSEIDON altimeters have been used in this study. Cycles 1-9 were omitted because they are suspect (Nerem, 1995b). The single-frequency POSEIDON altimeter does not provide an ionosphere correction directly, but since this altimeter is used in only about 10% of the mission, the somewhat lower accuracy of the DORIS-derived ionosphere correction (Minster et al, 1995) does not significantly affect the results presented here. The MGDRs differ from the GDRs (used in earlier studies) in that an improved EM bias algorithm is employed (Gaspar et al, 1994), and updated estimates of the cr0 calibration were used (Callahan et al, 1994). The on-board TOPEX altimeter internal calibration estimates have also been applied (Hayne et al., 1994) and are designed to measure changes in the instrument calibration using measurements from a calibration loop in the instrument electronics. In addition, a correction of 1.2 mm/yr over 1993-1996 (flat from 1997 to present) has been applied for an apparent drift in the TOPEX microwave radiometer (TMR), which provides the wet troposphere correction (Keihm et al., 1998). The latest improved orbits using the improved JGM-3 gravity model (Tapley et al., 1996) have been employed. While some data editing is performed (shallow water, outliers, high mesoscale variability, etc.), Nerem (1995b) and Minster et al. (1995) have shown the mean sea level estimates to be very insensitive to this editing.

Figure 6.8 shows the cycle-by-cycle (10 days) estimates of global mean sea level for Cycles 10-233 computed using the techniques described in Nerem (1995b). The RMS variation of the mean sea level is roughly 7 mm after removal of a trend. Spectral analysis reveals much of the variability lies near periods of 511, 365, 182, and 59 days, all of which have amplitudes of 2 mm and greater. The variability at a period of 59 days is near the fundamental period at which the T/P orbit samples semidiurnal varying phenomena, such

Year

Figure 6.8 A time series of 10-day estimates of global mean sea level computed from the TOPEX/POSEIDON mission altimeter data and the same results after smoothing using a 60-day boxcar filter.

Year

Figure 6.8 A time series of 10-day estimates of global mean sea level computed from the TOPEX/POSEIDON mission altimeter data and the same results after smoothing using a 60-day boxcar filter.

as atmospheric and ocean tides (such as the aforementioned S2 error) and the ionosphere. This variability can be significantly reduced by smoothing the mean sea level values using a 60-day boxcar filter. As shown in Fig. 6.8, this reduces the RMS variation of the time series to less than 4 mm after removal of a trend. Also note the large 15- to 20-mm increase in mean sea level at the end of the time series, which we will show later is related to the 1997-1998 ENSO event (Nerem et al, 1999). The power of satellite altimetry to spatially map the sea level change signal will also be reviewed later.

Was this article helpful?

0 0

Post a comment