4.6 Global Mean Sea Level (GMSL)
We construct an ensemble of GMSL projections by combining the ice emulator projections with the steric contribution from a corresponding CMIP6 model using the selection probability weights described in eqn. 1. The resulting ensemble of 500 GMSL projections is plotted in Figure 9. Both observations and projections show a near linear relationship between temperature and the rate of sea level rise. The TSLS slope for 2016-2050 is 5.3±1.0 mm/yr/K, which is greater than the 3.0±0.4 mm/yr/K estimated for 2051-2100. Observations indicate a sensitivity of 3.4±0.4 mm/yr/K. The central estimate of the sensitivity derived from the AR6 projections during the latter half of the 21st century is also 3.4 mm/yr/K. This is significantly larger than that obtained in the AR5 (2.7 mm/yr/K) (Grinsted and Christensen, 2021), which is likely due to the increased climate sensitivity of CMIP6 models (Forster et al, 2020). The observational estimates of TSLS determined here and in (Grinsted and Christensen, 2021) are not identical because different data were used. In particular, the AR6 assessed sea level rate for 1901-1990 (1.35±0.35 mm/yr) is greater than the rate in (Grinsted and Christensen, 2021) (1.1±0.3 mm/yr; (Dangendorf et al., 2017)). This greater rate has the effect that the acceleration into the satellite altimetry era appears smaller.