Seasonal predictability of the ensemble seasonal prediction by tier-1 and tier-1.5 models

By using a hybrid model tier-1 that is coupled in the Indo-Pacific tropical ocean, we perform a set of 10 ensemble runs with different initial condition for two-season period that starts from November 1st and May 1st respectively in 1982 though 2004. The tier-1.5 that is only coupled from India Ocean to dateline and prescribes model monthly climatology plus observed SSTA in east tropical Pacific is also used to perform another suite of 10 ensemble runs with same initial conditions. The sensitivity of Asia monsoon and Australian monsoon to SST in central and eastern tropical Pacific has been investigated. The results of experiments suggest that, 1) the climatology precipitation bias has close relationship with climatology SST bias which possibly cause overabundant rainfall over Arab Sea and Bay of Bengal, northward shift rain belt over West Pacific in summer, deficient rainfall of ITCZ over tropical Pacific in summer, too long south branch of ITCZ over southern Pacific in winter. 2) To east of dateline, the precipitation external variance and internal variance in tier-1.5 are much larger than those in tier-1.5. 3) The similar signal to noise ratio or error to external ratio as well as anomaly correlation over Asia monsoon region between in tier-1 and tier-1.5 suggests that precipitation predictability over Asia monsoon region has little relationship with the improvement of central and east Pacific SST. 4) The improvement spatial correlation in winter by tier-1.5 over El Nino region during La Nina year, especially when the cold SST over Nino3.4 turn into warm SST suggest that the winter precipitation predictability over El Nino region can possible be improved by SST predictability.