Tropical tropospheric warming pattern explained by shifts in convective heating in the Matsuno-Gill model

Abstract

Horizontal temperature gradients in the tropical free troposphere are fairly weak, and tropical tropospheric warming is usually treated as uniform. However, here we show that projected tropospheric warming is spatially inhomogeneous in CMIP6 models, as well as in a storm-resolving climate model. We relate the upper tropospheric warming pattern to sea surface temperature changes that reorganise convection and thereby cause spatial shifts in convective heating. Using the classical Gill model for tropical circulation and forcing it with precipitation changes that arise due to greenhouse gas warming we can understand and reproduce the different warming patterns simulated by a range of global climate models. Forcing the Gill model with precipitation changes from a certain region demonstrates how local tropospheric temperature changes depend on local changes in convective heating. Close to the equator anomalous geopotential gradients are balanced by the dissipation term in the Gill model. The optimal dissipation timescale to reproduce the warming pattern varies depending on the CMIP6 model, and is between 1 and 10 days. We demonstrate that horizontal advection and eddy momentum fluxes have large enough equivalent dissipation timescales to balance the gradients in geopotential and thereby shape the warming pattern. While climate models show a large spread in projections of tropical sea surface temperature and precipitation changes, our results imply that once these predictions improve, our confidence in the predicted upper tropospheric warming pattern should also increase.