[A-CG31] Climate Variability and Predictability on Subseasonal to Centennial Timescales

[ACG31-P01] Multi-year predictive skill of the wintertime heavy rainfall potentials in western Japan

*Takashi Mochizuki¹ (1.Department of Earth and Planetary Sciences, Kyushu University)

Keywords:decadal prediction, trans-basin variability, heavy rainfall potential, climate extreme, large ensemble simulation

Statistical changes in extreme weather and climate events have attracted much attention as potential sources of natural disasters. Numerical experiments to obtain huge ensembles of multiyear predictions with initialization using a high-resolution model represent a direct approach but a difficult solution for the computational resources available currently. Here, we validated the multiyear predictive skills of wintertime heavy rainfall potential in western Japan, by combining initialized decadal hindcasts of the global climate model and large ensemble simulations from a high-resolution global atmospheric model. By analyzing a set of initialized hindcasts, we identified one of the major predictive components of sea surface temperature (SST) variability beyond interannual timescales as the so-called trans-basin variability. A set of 100 ensemble simulations using a high-resolution atmospheric model show a significantly large signal-to-noise ratio for the wintertime heavy rainfall potential in western Japan, which is closely related to the trans-basin variability. When the SST around the maritime continent is higher, the anomalously low pressure in the northwestern Pacific enhances low-level cold air transport due to the winter monsoon. Consequently, the resultant weaker baroclinicity in the lower atmosphere reduces storm activity and wintertime heavy rainfall potential in western Japan.