[A-OS17] Climate variability and predictability on subseasonal to decadal timescales

[AOS17-P01] Predictability of two flavors of El Nino and statistical downscaling by SVD analysis using the MIROC5 seasonal prediction system

★Invited Papers

*Yukiko Imada¹, Hiroaki Tatebe², Masayoshi Ishii¹, Yoshimitsu Chikamoto⁶, Masato Mori⁴, Miki Arai², Shinjiro Kanae⁵, Masahiro Watanabe³, Masahide Kimoto³ (1.Meteorological Research Institute, Japan Meteorological Agency, 2.Japan Agency for Marine-Earth Science and Technology, 3.Atmosphere and Ocean Research Institute, the University of Tokyo, 4.Research Center for Advanced Science and Technology, the University of Tokyo, 5.Tokyo Institute of Technology, 6.University of Utah)

Keywords:Seasonal prediction, ENSO, CGCM

It is known that there are two flavors of ENSO events in the tropical Pacific Ocean: traditional eastern Pacific (EP) events and central Pacific (CP) events. Several studies indicated that the CP event is more difficult to predict than the EP event. Here, we assessed the difference of the seasonal predictability for two prominent types of El Nino using the coupled atmosphere-ocean general circulation model (AOGCM) MIROC5 co-developed by Atmosphere and Ocean Research Institute (AORI), National Institute for Environmental Studies (NIES), and Japan Agency for Marine-Earth Science and Technology (JAMSTEC). The spatial resolution is a horizontal triangular spectral truncation at total wave number 85 (T85) with 40 vertical layers, and eight ensemble forecast members are generated according to the protocol of the WCRP Climate-system Historical Forecast Project (CHFP). Hindcast products showed high predictability of tropical climate signals with the significant anomaly correlation coefficient skill scores, even though the ocean anomaly data assimilation is applied to the initialization process.
Overall, the predictable months of CP events are shorter than EP events because CP events are sensitive to atmospheric noises. Characteristics of each error growing process were also investigated.
We also developed a statistical downscaling technique using SVD analysis which is effective for rainfall prediction over the Indochina Peninsula and Australian continent.