ENSO-forced And ENSO-independent Mechanisms Of The Occurrence And Development Of The Indian Ocean Dipole

The Indian Ocean dipole(IOD)is an east-west dipole mode of interannual variability over the tropical Indian Ocean,which is generated by El Ni(?)o-Southern Oscillation(ENSO)remote forcing and local air-sea interaction.In this study,the mechanisms responsible for the growth and initiation of the ENSO-forced and ENSO-independent IOD modes are investigated.Based on the observational analysis and numerical simulation,some main conclusions associated with“Initiation and Growth Mechanisms for the ENSO-forced and ENSO-independent IOD modes”are summarized as follows:(1)The ENSO-forced and ENSO-independent IOD modes from observational data during 1900-2012 are extracted through a new approach of two extended EOF(Empirical Orthogonal Function)analyses.The growth of both the IOD modes depends on two positive feedbacks,“combined offshore-advection and thermocline” and “wind-evaporation-SST”feedbacks.The former is a variant of the classical thermocline feedback.Subsurface anomalous cold water is brought to the surface by climatological upwelling along the coast of SumatraJava and then transported by mean offshore currents into the interior southeastern Indian Ocean,where climatological mean convections are active.The cold water suppresses the overlying convections and drives equatorial easterly anomalies to the west.The equatorial zonal wind anomalies further strengthen subsurface anomalous cold water off the coast of Sumatra-Java through driving upwelling oceanic Kelvin waves.The initiation of the two types of IOD,however,is distinct.The ENSO-forced IOD mode is triggered by ENSO remote forcing through two distinctive atmospheric bridges involving an equatorially symmetric anomalous Walker circulation and cross-equatorial flows linking to the western North Pacific and Indian summer monsoons.The initiation of the ENSO-independent IOD mode is associated with the tropospheric biennial oscillation(TBO)involving the monsoon-warm ocean interactions and the delayed oceanic wave dynamics in the tropical Indian Ocean.The ENSO-forced and ENSO-independent IOD modes are also extracted based on the tropical Pacific pacemaker experiments(PM-TP)of the Community Earth System Model(CESM),and the growth and initiation mechanisms of the two types of IOD are analyzed.The model results indicate that the growth and initiation processes of the two types of IOD are mostly same with that of observed results,which confirms our conclusions.(2)The impacts of ENSO on IOD are examined based on the two sets of numerical experiments of the CESM.The first is a globally coupled control experiment(Picontrol run),the second is a no ENSO experiment(no ENSO run).In the latter,the ENSO variability is suppressed by restoring the SST to model climatological mean state in the tropical central eastern Pacific Ocean.The IOD mode is successfully simulated by the no ENSO run including the spatial distributions and seasonal evolutions,suggesting that the IOD can be triggered by the local air-sea interactions over the tropical Indian Ocean.In contrast of the two sets of model results,the power spectrum of IOD is highly consistent with that of ENSO in the Picontrol run,while in the absence of ENSO variability the IOD is dominant by the quasi-biennial component.In addition,the ENSO can strengthen the amplitude of IOD.The ENSO remote forcing,on the one hand,can induce the stronger antisymmetric heating about the equator in the Indo-Pacific warm pool region,and thus generate the stronger cross-equatorial wind anomalies off the coast of Sumatra-Java.On the other hand,the Indian summer monsoon precipitation tends to be weakened in El Ni(?)o developing summer,corresponding to weakened cross-equatorial monsoon circulation in the western tropical Indian Ocean.(3)We evaluate the IOD simulated by the FGOALS-f3 and FGOALS-g3 models.The FGOALS-f3 and FGOALS-g3 models are capable of capturing the IOD modes in their spatial patterns,periods and seasonal evolutions.Due to the different atmosphere components adopted in the FGOALS-f3 and FGOALS-g3 models,the amplitude of IOD is stronger in the FGOALSf3 results.Two factors contribute to this difference.First,the negative “cloud-radiation-SST”feedback is stronger in the FGOALS-g3 results.Second,the FGOALS-f3 model can reproduce the negative correlation between the ENSO and Indian summer monsoon.