Explaining the Spread in Global Mean Thermosteric Sea Level Rise in CMIP5 Climate Models

The ocean stores more than 90% of the energy excess associated with anthropogenic climate change. The resulting ocean warming and thermal expansion are leading contributors to global mean sea level (GMSL) rise. Confidence in projections of GMSL rise therefore depends on the ability of climate models to reproduce global mean thermosteric sea level (GMTSL) rise over the twentieth century. This study first compares the GMTSL of the climate models from phase 5 of the Coupled Model Intercomparison Project (CMIP5) to observations over 1961-2005. Although the model ensemble mean is within the uncertainty of observations, the model ensemble exhibits a large spread. The authors then aim to explain the spread in CMIP5 climate model GMTSL over the twentieth and twenty-first centuries. It is shown that the climate models' GMTSL rise depends linearly on the time-integrated radiative forcing F (under continuously increasing radiative forcing). The constant of proportionality expresses the transient thermosteric sea level response of the climate system, and it depends on the fraction of excess heat stored in the ocean, the expansion efficiency of heat, the climate feedback parameter, and the ocean heat uptake efficiency. The across-model spread in explains most (>70%) of the across-model spread in GMTSL rise over the twentieth and twenty-first centuries, while the across-model spread in time-integrated F explains the rest. The time-integrated F explains less variance in the across-model GMTSL rise in twenty-first-century than in twentieth-century simulations, as the spread in F is reduced over the twenty-first century because the anthropogenic aerosol forcing, which is a large source of uncertainty in F, becomes relatively smaller.