Uncertainties in the projection of sterodynamic sea level in CMIP6 models

Sterodynamic sea level (SdynSL) is an essential component of sea level changes that climate models can simulate directly. Here we untangle the impact of intermodel uncertainty, internal variability, and scenario uncertainty on SdynSL projections from the Coupled Model Intercomparison Project Phase 6 (CMIP6) models. At global scale, intermodel (scenario) uncertainty reigns before (after) ~2070, but internal variability is negligible. At regional scale, intermodel uncertainty is the largest contributor (50~100%), internal variability is secondary (20~50%) in the Indian Ocean and tropical Pacific in the near term and midterm. Scenario uncertainty is negligible until emerging in the long term. The anthropogenic signal of global  mean SdynSL emerges from the beginning of this century relative to the 1971-2000 climatology. However, only 70% of the ocean can detect anthropogenic SdynSL signals until the long term. Assuming that model differences are eliminated, anthropogenic SdynSL signals will emerge 38 almost globally after the midterm.