Propagation Velocities of Mesoscale Features from Along-Track Sea Level Anomalies

Measurement and mapping of seal level anomaly (SLA) from satellite systems has revolutionized oceanography. However, the present standard gridded products may not properly factor propagation of mesoscale features into the optimal interpolation procedure used to produce the datasets. We present composites of SLA variance from SSALTO/DUACS to demonstrate that gridded-SLA variability is higher in proximity to high-repeat rate Jason satellite tracks. Correcting this bias could improve representation of mesoscale features away from Jason tracks, particularly at intermediate latitudes where the zonal separation of Jason tracks is large relative to the decorrelation scale. Properly resolving mesoscale features reduces errors in propagation velocities; eddy tracking and SLA variability with resulting decreased uncertainties in transport of heat and fresh water; less misidentification of striations as eddies (or vice versa) and lower uncertainties in eddy kinetic energy. To quantify some effects of this bias we use along-satellite-track SLA to generate empirical correlation functions for specific latitude and longitude bands. We then use singular value decomposition to remove seasonal signals and apply radon transforms to the residuals to obtain zonal propagation velocities. Preliminary analysis indicates qualitative agreement between our results, Fu (2009) and Chelton et al (2011) in the meridional structure of propagation velocities away from the equator, but suggests that use of the gridded dataset reduces the magnitude of zonal propagation velocities by 0.5-2 km/day