A global comparison of Argo and satellite altimetry observations

Abstract

Abstract. Differences, similarities and complementarities between Sea Level Anomalies (SLA) deduced from altimeter measurements and dynamic height anomalies (DHA) calculated from Argo in situ temperature (T) and salinity (S) profiles are globally analyzed. SLA and DHA agree remarkably well and, compared to previous studies, Argo dataset allows an improvement in the coherence between SLA and DHA. Indeed, Argo data provides a much better spatial coverage of all oceans and particularly the Southern Ocean, the use of an Argo mean dynamic height, the use of measured salinity profiles (versus climatological salinity), and the use of a deeper reference level (1000 m versus 700 m). The large influence of Argo salinity observations on the consistency between altimetry and hydrographic observations is particularly demonstrated with an improvement of 35% (relative to the SLA minus DHA signal) by using measured salinity profiles instead of climatological data. The availability of observations along the Argo float trajectories also provides a means to describe the sea level variability of the global ocean both for the low frequency and the mesoscale part of the circulation. Results indicate that sea level variability is dominated by baroclinic signal at seasonal to inter-annual periods for all latitudes. In the tropics, sea level variability is baroclinic for meso-scale to interannual periods and at high latitudes, sea level variability is barotropic with also deep baroclinic signals (i.e. influence of deep temperature and salinity signals) for intra seasonal and mesoscale periods. These results emphasize the need to separate the different time and space scales in order to improve the merging of the two data sets. The qualitative study of seasonal to interannual SLA minus DHA signals finally reveals signals related to deep ocean circulation variations and basin-scale barotropic signals. Future work is, however, needed to understand the observed differences and relate them to different forcing mechanisms.