The gravity field and its variations in time are important observables for the understanding of the Earth's dynamic system. The twin satellites of the GRACE (Gravity Recovery And Climate Experiment) mission have been designed to measure such temporal variations as well as the long-wavelength part of Earth's gravity field with unprecedented accuracy on a global scale. Due to the sensitivity of GRACE to this time variable signal, mass redistributions which cause temporal gravity field variations have to be considered in the gravity field recovery process. This is typically done by the application of geophysical models to the satellite data. These models however do not perfectly resemble reality, resulting in residual time variable signal in the measurements which deteriorates derived gravity field solutions. To obtain reliable estimates, an appropriate modeling of the time variable gravity signal is therefore unavoidable. The incorporation of temporal variations into the least squares adjustment process is however accompanied by computational challenges. When modeling daily variations as spherical harmonic coefficients up to degree and order 40 within the adjustment process, an additional 1677 unknowns per day have to be considered. For the whole GRACE observation period starting from 2003 until today, this yields total of approximately 7.2 million unknown parameters. A least squares adjustment of this size is not solvable in a sensible time frame, therefore measures to reduce the problem size have to be taken. In this thesis an integrated approach for the combined estimation of the static gravity field and temporal variations of different time scales is presented. The developed approach is applied to GRACE-L1B data and the effect of different temporal representations is investigated. The capability of the method will be demonstrated on the basis of three computed GRACE-only gravity field models.
|Betreuer/-in / Berater/-in|
|Publikationsstatus||Veröffentlicht - 2014|