hide
Free keywords:
-
Abstract:
Recently released global gravity field models generated solely from CHAMP and GRACE satellite observations allow with an unprecedented accuracy and resolution the recovery of the mean sea surface topography from the difference between an altimetry-based mean sea surface height model and the gravity model's derived geoid. Here the CHAMP EIGEN-2 gravity field model, and the first GFZ GRACE gravity model, EIGEN-GRACE01S, are used. The mean sea surface height model has been compiled from four years'; worth of TOPEX altimeter data. To evaluate the accuracy and resolution limits of the CHAMP and GRACE geoids for the envisaged application, a low pass filter in the spatial domain with different cut-off wavelengths has been applied to the geoid and sea surface data before subtraction. The minimum wavelength, where noisy and erroneous features in the recovered sea surface topography are minimised, can be interpreted as an indicator for the best suited common spatial resolution. The EIGEN-2 model's geoid has been tested to have a resolution of 1800 km, which corresponds to a truncation degree of l = 22 in terms of spherical harmonics. Using the EIGEN-GRACE01S model, the resolution could be extended to 1000 km (l = 40). These boundaries can be attributed to the geoid's error, exceeding 2 cm in case of the CHAMP model, and in case of the GRACE model to spurious systematic signals increasing with increasing spherical harmonic degree. The calculated sea surface topography models have been used to derive absolute geostrophic sea surface velocities. An error propagation shows that the requirement of 1 cm/s for geoid induced velocity errors is fulfilled at the given resolutions for all latitudes excluding a narrow equatorial band. Maximum geostrophic velocities are derived in the 1000 km-resolution model for the Kuroshio region with 40 cm/s, and for the Gulf Stream east off Cape Hatteras with 25 cm/s.
