Evolving CO2 rather than SST accelerates GCM convergence by a factor of ten

Submitter

Romps, David — Lawrence Berkeley National Laboratory

Area of research

General Circulation and Single Column Models/Parameterizations

Journal Reference

Zhang Y, J Bloch‐Johnson, D Romps, and D Abbot. 2021. "Evolving CO2 rather than SST leads to a factor of ten decrease in GCM convergence time." Journal of Advances in Modeling Earth Systems, 13(11), e2021MS002505, 10.1029/2021MS002505.

Science

An earlier paper (Romps 2020) applied an inverse climate-modeling method to rapidly equilibrate SST and CO2 in limited-area cloud-resolving simulations and proposed a way to extend that inverse method to global climate models. Here, that approach is tested in a global aquaplanet simulation of perpetual equinox and is found to accelerate the equilibration by a factor of ten.

Impact

This speedup enables the equilibration of 51 climate simulations to measure, in fine detail, the bump in equilibrium climate sensitivity that has been reported around a global mean temperature of 310-320 K. Other potential applications include the ability to probe unstable climate states and climate states close to bifurcations, such as the runaway greenhouse transition.

Summary

Testing a proposal for rapid equilibration of global climate simulations finds a successful factor-of-ten speedup of climate simulations, enabling the probing of climate states not readily accessed via the standard slab-ocean methods.