Towards overshoot-proof multi-century sea level rise projections

Temporarily crossing and subsequently returning below 1.5°C, a so-called temperature ‘overshoot’, is a scenario of increasing relevance and interest. Potential impacts and risks of such an overshoot, including triggering irreversible ice loss and a large multi-century sea-level rise (SLR) commitment, need to be better understood to support well-informed policy and decision making.

Here, we use a set of eight overshoot scenarios from the PROVIDEv1.2 ensemble, covering a wide range of peak temperatures and emission reduction rates, to force an updated MAGICC-SLR emulator to explore the multi-century responses of the main sea level components. The emulator updates include a new calibration for the Greenland solid ice discharge component and different land water storage representations following population assumptions as represented in the Shared Socioeconomic Pathway (SSP) framework. These are the first steps of a comprehensive MAGICC-SLR update to provide overshoot-proof state-of-the-art probabilistic SLR projections.

Under a scenario that extrapolates mitigation efforts resulting from current climate policies out to 2100 and thereafter decreases global average temperatures back to 1.5°C, we project a global mean SLR of 1.4 m (median, 0.7-3.2 m very likely range) by 2300, relative to 1995-2014 levels. By extending the 2300 radiative forcing levels further into the future, we explore SLR projections until 2500, with greatly increasing uncertainties and decreasing robustness of the sea level response. In case of a temperature overshoot below 2.0°C, our results suggest that global mean SLR is reduced by following a SSP1 rather than SSP2 population pathway through dam impoundment and groundwater extraction management. For the updated MAGICC-SLR emulator, we find that the relative contribution of the Greenland solid ice discharge component steadily increases over time and becomes the dominant SLR driver across all scenarios beyond 2300. Our results suggest that by 2500, the committed global SLR from overshooting 1.5°C cannot be returned to levels of a 1.5°C stabilization scenario.

We highlight and discuss the limitations and caveats when projecting SLR under overshoot with simplified modeling approaches and outline next steps to continue overshoot-proofing MAGICC-SLR. We emphasize the need for a careful evaluation of the parameterizations for each SLR component to ensure a physically robust representation of the (ir)reversibile multi-century SLR response under overshoot.