Climate Change 2001:
Working Group I: The Scientific Basis
Other reports in this collection

11.3.2.3 Mean sea level change from satellite altimeter observations

In contrast to the sparse network of coastal and mid-ocean island tide gauges, measurements of sea level from space by satellite radar altimetry provide near global and homogenous coverage of the world's oceans, thereby allowing the determination of regional sea level change. Satellite altimeters also measure sea level with respect to the centre of the earth. While the results must be corrected for isostatic adjustment (Peltier, 1998), satellite altimetry avoids other vertical land movements (tectonic motions, subsidence) that affect local determinations of sea level trends measured by tide gauges. However, achieving the required sub-millimetre accuracy is demanding and requires satellite orbit information, geophysical and environmental corrections and altimeter range measurements of the highest accuracy. It also requires continuous satellite operations over many years and careful control of biases.

To date, the TOPEX/POSEIDON satellite-altimeter mission, with its (near) global coverage from 66°N to 66°S (almost all of the ice-free oceans) from late 1992 to the present, has proved to be of most value to direct estimates of sea level change. The current accuracy of TOPEX/POSEIDON data allows global average sea level to be estimated to a precision of several millimetres every 10 days, with the absolute accuracy limited by systematic errors.

Careful comparison of TOPEX/POSEIDON data with tide gauge data reveals a difference in the rate of change of local sea level of -2.3 ± 1.2 mm/yr (Mitchum, 1998) or -2 ± 1.5 mm/yr (Cazenave et al., 1999). This discrepancy is caused by a combination of instrumental drift, especially in the TOPEX Microwave Radiometer (TMR) (Haines and Bar-Sever, 1998), and vertical land motions which have not been allowed for in the tide gauge data. The most recent estimates of global average sea level rise from the six years of TOPEX/POSEIDON data (using corrections from tide gauge comparisons) are 2.1 ± 1.2 mm/yr (Nerem et al., 1997), 1.4 ± 0.2 mm/yr (Cazenave et al., 1998; Figure 11.8), 3.1 ± 1.3 mm/yr (Nerem, 1999) and 2.5 ± 1.3 mm/yr (Nerem, 1999), of which the last assumes that all instrumental drift can be attributed to the TMR. When Cazenave et al. allow for the TMR drift, they compute a sea level rise of 2.6 mm/yr. Their uncertainty of ± 0.2 mm/yr does not include allowance for uncertainty in instrumental drift, but only reflects the variations in measured global sea level. Such variations correlate with global average sea surface temperature, perhaps indicating the importance of steric effects through ocean heat storage. Cazenave et al. (1998) and Nerem et al. (1999) argue that ENSO events cause a rise and a subsequent fall in global averaged sea level of about 20 mm (Figure 11.8). These findings indicate that the major 1997/98 El Niño-Southern Oscillation (ENSO) event could bias the above estimates of sea level rise and also indicate the difficulty of separating long-term trends from climatic variability.

Figure 11.10: Estimated sea level rise from 1910 to 1990. (a) The thermal expansion, glacier and ice cap, Greenland and Antarctic contributions resulting from climate change in the 20th century calculated from a range of AOGCMs. Note that uncertainties in land ice calculations have not been included. (b) The mid-range and upper and lower bounds for the computed response of sea level to climate change (the sum of the terms in (a) plus the contribution from permafrost). These curves represent our estimate of the impact of anthropogenic climate change on sea level during the 20th century. (c) The mid-range and upper and lower bounds for the computed sea level change (the sum of all terms in (a) with the addition of changes in permafrost, the effect of sediment deposition, the long-term adjustment of the ice-sheets to past climate change and the terrestrial storage terms).

After upgrading many of the geophysical corrections on the original European Remote Sensing (ERS) data stream, Cazenave et al. (1998) find little evidence of sea level rise over the period April 1992 to May 1996. However, over the time span of overlap between the ERS-1 and TOPEX/POSEIDON data, similar rates of sea level change (about 0.5 mm/yr) are calculated. For the period April 1992 to April 1995, Anzenhofer and Gruber (1998) find a sea level rise of 2.2 ± 1.6 mm/yr.

In summary, analysis of TOPEX/POSEIDON data suggest a rate of sea level rise during the 1990s greater than the mean rate of rise for much of the 20th century. It is not yet clear whether this is the result of a recent acceleration, of systematic differences between the two measurement techniques, or of the shortness of the record (6 years).



Other reports in this collection