A mean dynamic topography computed over the world ocean from altimetry, in situ measurements, and a geoid model

The lack of an accurate geoid still prevents precise computation of the ocean absolute dynamic topography from satellite altimetry and only sea level anomalies (SLA) can be accurately deduced. In the new context of Global Ocean Data Assimilation Experiment (GODAE) where models are assimilating satellite altimetry, the estimation of a realistic mean dynamic topography (MDT) consistent with SLA is a crucial issue. In a first direct'' approach, a MDT is computed by subtracting the geoid model EIGEN-2 from the Mean Sea Surface Height CLS01, determined from 7 years of altimetric data (TOPEX and ERS1,2) at spherical harmonic degree 30. To provide the scales shorter than 660 km, the Levitus climatology is merged with the resulting MDT, both weighted by their respective errors. This solution provides a first guess'' for the computation of a global and higher resolution MDT. Then, a synthetic'' technique is used to combine in situ measurements and altimetric data: TOPEX and ERS1,2 altimetric anomalies are subtracted from in situ measurements of the full dynamical signal ( based on buoy velocities from the WOCE-TOGA program and XBT, CTD casts). The resulting values provide local estimates of the mean field, in terms of currents or dynamic topography, which are used to improve the first guess using an inverse technique. The MDT obtained is compared to other mean dynamic fields, and a verification using independent in situ data shows improvements in most areas. It exhibits a more energetic representation of the subtropical and subpolar gyres; sea level gradients associated with the main currents are strongly enhanced. Differences with independent velocity observations are globally lower than 13 cm/s rms.