Energetic signatures of tropical rainband biases & shifts in CMIP6

Tropical and subtropical precipitation impact millions of people via agriculture and rainfall-driven disasters, driving interest in their potential future change. CMIP6 simulations broadly predict an increase in global monsoon precipitation. However, regional projections from individual models vary in magnitude and sign, and projected changes by the end of the century are often small compared with model biases. This motivates an interest in understanding model biases, and how to interpret the future shifts in rainfall.

The Hadley and Walker circulations transport Moist Static Energy in the direction of their upper branches, so that the change in sign of MSE transport acts as a proxy for mass convergence in the tropical rainband. MSE transport can then be interpreted in terms of top-of-atmosphere and surface energy fluxes using the column energy budget.

Recent work attributes contributions to annual- and zonal-mean divergent MSE transport to radiative fluxes, evaporative fluxes and sensible heat, and suggests that evaporative fluxes are key in setting the spatial structure of MSE transport. Here we extend this approach to regional and seasonal scales, and explore inter-model differences in CMIP6 historical simulations, and projected changes under SSP585.

MSE transport attributed to evaporative and radiative fluxes dominate the regional JJA & DJF transport. Empirical Orthogonal Functions (EOFs) are used to express historical intermodel differences in MSE transport in a 2-dimensional space of leading EOFs linked to land-sea thermal contrast and interhemispheric thermal contrast. Changes in MSE transport components under SSP585 are then projected into this space and decomposed into terms attributed to the top of atmosphere and surface fluxes. This reveals energetic signatures of model bias and future change, that illustrate how different processes contribute to the overall differences in energy transport.

Shared energetic signatures of bias are predominantly seen within model families. In contrast, shared signatures of future change emerge across (and differ within) model families, with a shared bias signature not implying a shared change signature. This suggests a set coherent but differing pathways through which climate change affects the energy budget and associated tropical rainfall in particular groups of models.