Interdecadal change of interannual variability and predictability of two types of ENSO using a MME method

Thursday, 18 December 2014
Hye-In Jeong, APEC Climate Center, Busan, South Korea, Joong-bae Ahn, Pusan National University, Busan, South Korea, June-yi Lee, Busan National University, Busan, South Korea, Andrea Alessandri, Agenzia Nazionale per le nuove Tecnologie, l'energia e lo sviluppo economico sostenibile, Rome, Italy and Harry Hendon, CAWCR, Melbourne, VIC, Australia
A significant interdecadal climate shift of interannual variability and predictability of two types of the El Nino-Southern Oscillation (ENSO), namely the canonical or eastern Pacific (EP)-type and Modoki or central Pacific (CP) type, are investigated. Using the retrospective forecasts of six-state-of-the-art coupled models and their multi-model ensemble (MME) for December–January–February during the period of 1972–2005 along with corresponding observed and reanalyzed data, we examine the climate regime shift that occurred in the winter of 1988/1989 and how the shift affected interannual variability and predictability of two types of ENSO for the two periods of 1972–1988 (hereafter PRE) and 1989–2005 (hereafter POST). The result first shows substantial interdecadal changes of observed sea surface temperature (SST) in mean state and variability over the western and central Pacific attributable to the significant warming trend in the POST period. In the POST period, the SST variability increased (decreased) significantly over the western (eastern) Pacific. The MME realistically reproduces the observed interdecadal changes with 1- and 4-month forecast lead time. It is found that the CP-type ENSO was more prominent and predictable during the POST than the PRE period while there was no apparent difference in the variability and predictability of the EP-type ENSO between two periods. Note that the second empirical orthogonal function mode of the Pacific SST during the POST period represents the CP-type ENSO but that during the PRE period captures the ENSO transition phase. The MME better predicts the former than the latter. We also investigate distinctive regional impacts associated with the two types of ENSO during the two periods.