The most sensitive initial error modes modulating intensities of CP- and EP- El Niño events

Abstract The two types of El Nino events simulated by the Geophysical Fluid Dynamics Laboratory Climate Model version 2p1 (GFDL CM2p1) model and its “spring predictability barrier” (SPB) associations are examined. By conducting the ensemble hindcast experiments related to the sea temperature on the whole Pacific, both of the predictions for CP- and EP-El Nino show a significant SPB phenomenon, but the CP-El Nino features a much weaker SPB compared to the EP-El Nino. Further analyses revealed that, for CP-El Nino events, the initial sea temperature errors of the North Pacific with triple-like shape, referred to as negative Victoria Mode (VM) induces the largest prediction errors in Nino4 areas and modulates the intensities of CP-El Nino events. While for EP-El Nino events, the initial sea temperature errors in the subsurface layer of the western equatorial Pacific and the upper layer of the Southeast Pacific (15oS-30oS) with the meridional mode induce the largest prediction errors in Nino3 areas and modulates the intensities of EP-El Nino events. Obviously, results stress that, in order to reduce final prediction errors and obtain better predictions in terms of intensity on the two types of El Nino events, we should mainly focus on initial sea temperature accuracy in not only the subsurface layer of the west equatorial Pacific but also the surface layer of southeast Pacific and the region covered by the VM-like mode in the North Pacific.