Impacts of Model Mean Biases and Climate Sensitivity on Regional Dynamic Sea Level Projections in the CMIP5 and CMIP6 models

Abstract:

As an important component of regional sea level projections, the ocean dynamic sea level (DSL) change reflects ocean dynamical adjustment to the changing climate. In this study, we analyze mean states and future projections of the DSL from the global coupled climate models participating in phase 5 and phase 6 of the Coupled Model Intercomparison Project (CMIP5 and CMIP6, respectively).

Despite persistent biases relative to observations, both CMIP5 and CMIP6 simulate the mean sea level reasonably well. The equatorward bias of the Southern Hemisphere westerly wind stress is reduced from CMIP5 to CMIP6, which improves the simulated mean sea level in the Southern Ocean. The CMIP5 and CMIP6 DSL projections exhibit very similar features and intermodel uncertainties. With several models having a notably high climate sensitivity, CMIP6 projects larger DSL changes in the North Atlantic and Arctic associated with a larger weakening of the Atlantic meridional overturning circulation (AMOC).

We further identify linkages between model mean states and future projections by looking for their intermodel relationships. The common cold-tongue bias leads to an underestimation of DSL rise in the western tropical Pacific. Models with their simulated midlatitude westerly winds located more equatorward tend to project larger DSL changes in the Southern Ocean and North Pacific. In contrast, a more equatorward location of the North Atlantic westerly winds or a weaker AMOC under current climatology is associated with a smaller weakening of the AMOC and weaker DSL changes in the North Atlantic and coastal Arctic.

Our study provides useful emergent constraints for regional DSL projections and highlights the importance of reducing model mean-state biases and uncertainties in climate sensitivity for future projections.