English
Various Earth science disciplines strongly benefit from precise knowledge of the global gravity field of the Earth. High-quality gravity field models are needed, in particular, for: (1) precise estimations of the marine geoid, which is required for the quantitative determination of absolute ocean currents and the associated transport of heat, salt and other substances; this information forms a part of the input for building a global climate change picture; (2) defining a better global height reference system for datum connection, which serves as a reference surface for the study of various topographic processes, including the evolution of ice-sheets and land-surface topography; (3) better understanding the physics of the Earths interiors including geodynamics associated with the lithosphere, mantle composition and rheology, uplifting and subduction processes; (4) estimating the thickness of the polar ice sheets through the combination of bedrock topography and ice-sheet surface topography; and (5) accurate orbit determination for low-orbiting satellites. A series of dedicated low-orbiting satellite missions have been recently launched or will be launched soon in order to study the Earths gravity field: CHAMP in 2000, GRACE in 2002, and GOCE in 2007. The latter two missions are of a particular importance since GRACE is especially accurate at low spatial frequencies (above 200 km in terms of half-wavelengths), whereas GOCE will be beyond comparison in the range of half-wavelengths from 70 to 200 km. Furthermore, there are evidences that the GRACE and GOCE missions will be highly complementary: the accuracy of a joint gravity field model may be several times higher than that of a stand-alone model produced from either GRACE or GOCE data separately. The primary objective of the project is to compile a state-of-the-art model of the Earths gravity field up to high spherical harmonic degree (300 or even larger) from GRACE and GOCE satellite data together. Our ambition is to produce the model of the highest quality by fully exploiting the information contained in level-1b GRACE and GOCE data. The computed model will be supplied with a full covariance matrix.
