Multi-century sea-level projections and human impacts under the indicative ScenarioMIP-CMIP7 forcings

 

Projections of future sea-level rise are critical for informing adaptation planning and risk assessments. However, physical modelling frameworks, due to significant computational requirements, lack the flexibility required for rapid analysis and exploration of the latest global emission scenarios. Data-driven and statistical sea-level emulators can fill this requirement and continue to calibrate themselves to the latest large physical modelling experiments, such as ISMIP, and literature evidence which is published in slower-time - while providing new insights derived from observational constraints and combinations of multiple lines of evidence. Here, we present multi-century sea-level projections using an enhanced version of the ProFSea sea-level emulator tool, and quantify human exposure under high and low-likelihood ice-sheet processes. Due to the flexibility, performance and probabilistic structure of the model, we can explore a large suite of scenarios as well as their observationally constrained counterparts, identify dominant sources of model and process uncertainty, and go further in our analysis to determine human-relevant impacts for vulnerable regions around the world. In addition, we push the emulator out-of-sample to explore its behaviour under idealised very-high emission and overshoot scenarios - a potentially critical limitation of non-physically based emulators.