The seasonal march of the equatorial Pacific upper-ocean and its El Niño variability

Abstract

Based on two modern data sets, the climatological seasonal march of the upper-ocean is examined in the equatorial Pacific for the period 2004–2014, because of its large contribution to the total variance, its relationship to El Niño, and its unique equatorial wave phenomena. Argo provides a broadscale view of the equatorial Pacific upper-ocean based on subsurface temperature and salinity measurements for the period 2004–2015, and satellite altimetry provides synoptic observations of the sea surface height (SSH) for the period 1993–2015. Using either 11-year (1993–2003/2004–2014) time-series for averaging, the seasonal Rossby waves stands out clearly and eastward intraseasonal Kelvin wave propagation is strong enough in individual years to leave residuals in the 11-year averages, particularly but not exclusively, during El Niño onset years. The agreement of altimetric SSH minus Argo steric height (SH) residuals with GRACE ocean mass estimates confirms the scale-matching of in situ variability with that of satellite observations. Surface layer and subsurface thermohaline variations are both important in determining SH and SSH basin-wide patterns. The SH/SSH October-November maximum in the central-eastern Pacific is primarily due to a downward deflection of the thermocline (∼20m), causing a warm subsurface anomaly (>1°C), in response to the phasing of downwelling intraseasonal Kelvin and seasonal Rossby waves. Compared with the climatology, the stronger October-November maximum in the 2004–2014 El Niño composites is due to higher intraseasonal oscillations and interannual variability. Associated with these equatorial wave patterns along the thermocline, the western warm/fresh pool waters move zonally at interannual timescales through zonal wind stress and pressure gradient fluctuations, and cause substantial fresh (up to 0.6psu) and warm (∼1°C higher than the climatology) anomalies in the western-central Pacific surface-layer during the El Niño onset year, and of the opposite sign during the termination year. These El Niño-related patterns are then analyzed focusing on the case of the onset of the strong 2015/2016 episode, and are seen to be around two times larger than that in the 2004–2014 El Niño composites. The present work exploits the capabilities of Argo and altimetry to update and improve the description of the physical state of the equatorial Pacific upper-ocean, and provides a benchmark for assessing the accuracy of models in representing equatorial Pacific variability.