International audienceIn this study, we analysed decadal and long-term steric sea level variations over 1966-2007 period in the Indo-Pacific sector, using an ocean general circulation model forced by reanalysis winds. The simulated steric sea level compares favourably with sea level from satellite altimetry and tide gauges at interannual and decadal timescales. The amplitude of decadal sea level variability (up to ~5 cm standard deviation) is typically nearly half of the interannual variations (up to ~10 cm) and two to three times larger than long-term sea level variations (up to 2 cm). Zonal wind stress varies at decadal timescales in the western Pacific and in the southern Indian Ocean, with coherent signals in ERA-40 (from which the model forcing is derived), NCEP, twentieth century and WASWind products. Contrary to the variability at interannual timescale, for which there is a tendency of El Niño and Indian Ocean Dipole events to co-occur, decadal wind stress variations are relatively independent in the two basins. In the Pacific, those wind stress variations drive Ekman pumping on either side of the equator, and induce low frequency sea level variations in the western Pacific through planetary wave propagation. The equatorial signal from the western Pacific travels southward to the west Australian coast through equatorial and coastal wave guides. In the Indian Ocean, decadal zonal wind stress variations induce sea level fluctuations in the eastern equatorial Indian Ocean and the Bay of Bengal, through equatorial and coastal wave-guides. Wind stress curl in the southern Indian Ocean drives decadal variability in the south-western Indian Ocean through planetary waves. Decadal sea level variations in the south-western Indian Ocean, in the eastern equatorial Indian Ocean and in the Bay of Bengal are weakly correlated to variability in the Pacific Ocean. Even though the wind variability is coherent among various wind products at decadal timescales, they show a large contrast in long-term wind stress changes, suggesting that long-term sea level changes from forced ocean models need to be interpreted with caution

International audienceEmerging decadal climate predictions call for an assessment of decadal climate variability in the Coupled Model Intercomparison Project (CMIP) database. In this paper, we evaluate the influence of El Niño Southern Oscillation (ENSO) on Pacific Decadal Oscillation (PDO) in 10 control simulations from the CMIP3 and 22 from the CMIP5 database. All models overestimate the time lag between ENSO forcing and the PDO response. While half of the models exhibit ENSO-PDO correlation which is close to that in observation (>0.5) when the time lag is accounted for, the rest of the models underestimate this relationship. Models with stronger ENSO-PDO correlation tend to exhibit larger PDO-related signals in the equatorial and south Pacific, highlighting the key role of ENSO teleconnection in setting the inter-hemispheric Pacific pattern of the PDO. The strength of the ENSO-PDO relationship is related to both ENSO amplitude and strength of ENSO teleconnection to the North Pacific sea-level pressure variability in the Aleutian Low region. The shape of the PDO spectrum is consistent with that predicted from a combination of direct ENSO forcing, atmospheric stochastic forcing over the North Pacific and the re-emergence process in 27 models out of 32. Given the essential role of ENSO in shaping the Pacific decadal variability, models displaying realistic ENSO amplitude and teleconnections should be preferentially used to perform decadal prediction experiments

We examine the consistency of Indo‐Pacific decadal sea level variability in 10 gridded, observation‐based sea level products for the 1960–2010 period. Decadal sea level variations are robust in the Pacific, with more than 50% of variance explained by decadal modulation of two flavors of El Niño–Southern Oscillation (classical ENSO and Modoki). Amplitude of decadal sea level variability is weaker in the Indian Ocean than in the Pacific. All data sets indicate a transmission of decadal sea level signals from the western Pacific to the northwest Australian coast through the Indonesian throughflow. The southern tropical Indian Ocean sea level variability is associated with decadal modulations of ENSO in reconstructions but not in reanalyses or in situ data set. The Pacific‐independent Indian Ocean decadal sea level variability is not robust but tends to be maximum in the southwestern tropical Indian Ocean. The inconsistency of Indian Ocean decadal variability across the sea level products calls for caution in making definitive conclusions on decadal sea level variability in this basin.