5.5.6 Total Budget of the Global Mean Sea Level Change

The various contributions to the budget of sea level change are summarised in Table 5.3 and Figure 5.21 for 1961 to 2003 and 1993 to 2003. Some terms known to be small have been omitted, including changes in atmospheric water vapour and climate-driven change in land water storage (Section 5.5.5), permafrost and sedimentation (see, e.g., Church et al., 2001), which very likely total less than 0.2 mm yr^–1. The poorly known anthropogenic contribution from terrestrial water storage (see Section 5.5.5.4) is also omitted.

Figure 5.21. Estimates of the various contributions to the budget of the global mean sea level change (upper four entries), the sum of these contributions and the observed rate of rise (middle two), and the observed rate minus the sum of contributions (lower), all for 1961 to 2003 (blue) and 1993 to 2003 (brown). The bars represent the 90% error range. For the sum, the error has been calculated as the square root of the sum of squared errors of the contributions. Likewise the errors of the sum and the observed rate have been combined to obtain the error for the difference.

For 1961 to 2003, thermal expansion accounts for only 23 ± 9% of the observed rate of sea level rise. Miller and Douglas (2004) reached a similar conclusion by computing steric sea level change over the past 50 years in three oceanic regions (northeast Pacific, northeast Atlantic and western Atlantic); they found it to be too small by about a factor of three to account for the observed sea level rise based on nine tide gauges in these regions. They concluded that sea level rise in the second half of the 20th century was mostly due to water mass added to the oceans. However, Table 5.3 shows that the sum of thermal expansion and contributions from land ice is smaller by 0.7 ± 0.7 mm yr^–1 than the observed global average sea level rise. This is likely to be a significant difference. The assessment of Church et al. (2001) could allow this difference to be explained by positive anthropogenic terms (especially groundwater mining) but these are expected to have been outweighed by negative terms (especially impoundment). We conclude that the budget has not yet been closed satisfactorily.

Given the large temporal variability in the rate of sea level rise evaluated from tide gauges (Section 5.5.2.4 and Figure 5.17), the budget is rather problematic on decadal time scales. The thermosteric contribution has smaller variability (though still substantial; Section 5.5.3) and there is only moderate temporal correlation between the thermosteric rate and the tide gauge rate. The difference between them has to be explained by ocean mass change. Because the thermosteric and climate-driven land water contributions are negatively correlated (Section 5.5.5.3.), the apparent difference implies contributions during some 10-year periods from land ice, the only remaining term, exceeding 2 mm yr^–1 (Figure 5.17). Since it is unlikely that the land ice contributions of 1993 to 2003 were exceeded in earlier decades (Figure 4.14 and Section 4.6.2.2), we conclude that the maximum 10-year rates of global sea level rise are likely overestimated from tide gauges, indicating that the estimated variability is excessive.

For 1993 to 2003, thermal expansion is much larger and land ice contributes 1.2 ± 0.4 mm yr^–1. These increases may partly reflect decadal variability rather than an acceleration (Section 5.5.3; attribution of changes in rates and comparison with model results are discussed in Section 9.5.2). The sum is still less than the observed trend but the discrepancy of 0.3 ± 1.0 mm yr^–1 is consistent with zero. It is interesting to note that the difference between the observed total and thermal expansion (assumed to be due to ocean mass change) is about the same in the two periods. The more satisfactory assessment for recent years, during which individual terms are better known and satellite altimetry is available, indicates progress since the TAR.