Seasonal climate predictability with Tier-one and Tier-two prediction systems

In this study seasonal predictability of Tier-one and Tier-two predictions are evaluated and compared. Through the comparison of these two predictions, it is demonstrated that the air-sea coupled process is an important factor not only for climatological simulation but also for seasonal predictability. In particular, the air-sea coupling plays a crucial role over the warm pool region, as the atmosphere tends to lead the ocean in anomalous variability. In this region, the Tier-one prediction has better climatology compared to the Tier-two prediction despite the presence of a climatological SST bias. Furthermore, the Tier-one has a relatively higher seasonal predictive skill than that of the Tier-two although its SST prediction skill is relatively poor. It is suggested that the air-sea coupled process plays a role to reduce both the climatological and anomalous biases in the uncoupled AGCM by means of the negative feedback of the SST-heat flux-precipitation loop. Using the CliPAS and DEMETER seasonal prediction data, the robustness of these results are demonstrated in the multi-model frame works.