ENSO diversity impacts on CMIP6 projections of tropical cyclone activity

The El Niño-Southern Oscillation (ENSO) modulates tropical cyclone (TC) activity globally. Given that anthropogenic climate change can affect both ENSO and TCs, we explore if and how future projections of the ENSO-TC relationship could be affected by ENSO characteristics and biases in the CMIP6 models. In particular, we examine whether the ability of the CMIP6 models in reproducing ENSO diversity has an impact on their simulation of the ENSO-TC relationship. To this end, we analyze large-scale environmental fields that are typically associated with TC activity (e.g. vertical wind shear, potential intensity), as well as TC proxies, such as genesis indices and the ventilation index in the CMIP6 models. We consider 45 CMIP6 model historical simulations and 3 future scenarios (ssp245, ssp370, ssp585). We also include all ensembles available for each of the model simulations. ENSO diversity in the models is measured by the nonlinearity coefficient alpha described in Karamperidou et al. (2017) and is based on the models’ pre-industrial simulations. We classify the CMIP6 models into 3 groups, based on ENSO diversity: A (similar to observed), B (too weak diversity), and A+ (too strong diversity). ENSO composites of the large-scale environmental variables in the historical and future scenarios are constructed for the 3 groups. We show that models in the A+ group have very different future projections of the ENSO composites from the other two groups. While future projections of the ENSO modulation of TC proxies for models in groups A and B are very similar to those in historical simulations and observations, this is not the case for the A+ models, which show significant changes between future and historical patterns. Finally, we compare these results with those of CMIP5, to determine whether these relationships have changed since the previous generation of models.

Reference:

Karamperidou, et al, Clim. Dyn., doi: 10.1007/s00382-016-3475-y.