Study On The Antarctic Mass Variations With GRACE Time-variable Gravity Models

The earth's gravity field reflects the spatial distribution, movement and change of the earth's mass. High precision gravity field is one of the main tasks of geodesy, and it is plays also an important role in the researches of geodesy and earth sciences.In this paper, the earth's gravity models derived from GRACE are used to derive the mass variations in Antarctic.The development of satellite gravity is briefly introduced in the introduction.In chapter 2, the scientific objectives, the payloads and the data products of GRACE mission are described. Meanwhile the methods for deriving earth's gravity field model from GRACE data are also expatiated, such as dynamic method and the Energy Conservation etc.In chapter 3, the mass variation of the earth is derived from GRACE time-variable gravity model. Several filtering technique are investigated, especially focused on isotropic and non-isotropic Gaussian filtering, wiener filtering and filtering of correlated error.In chapter 4, the effect of Post Glacial Rebound (PGR) to the mass variations is studied. The variations of Ice sheet mass in Antarctic are inferred from GRACE models together with ICE-5G PGR model. Because C20 from GRACE is not very accurate, C20 from SLR (Satellite Laser Ranging) is adopted instead, and effect to the result is also given.The conclusions and outlooks are given in chapter 5.The main works are:1. The methods for deriving mass variations of the earth from GRACE model are studied.2. The isotropic and non-isotropic Gaussian filter are compared. The spectrum by using non-isotropic Gaussian filter depends not only on degree but also on order. The non-isotropic smoother yields enhanced GRACE signals with significantly higher spatial resolution in latitude and the same resolution in longitude without reducing the accuracy as compared to isotropic Gaussian filter.3. An alternate spatial averaging method based on Wiener optimal filtering is presented. Compared to the Gaussian method, the wiener filter is constructed using the degree-power spectra of the desired signal and the contaminating noise.4. Because there exists correlated error in GRACE coefficients, using the correlated error filter can eliminate it effectively.5. The mass variations of ice sheet in Antarctic are derived from GRACE models and ICE-5G PGR model. The rates of ice mass change in the west Amundsen Sea Embayment and the Enderby Land in east, are estimated as -58±23km~3/yr and 11±14km~3/yr respectively, which is comparable with other researches. The Ice sheet is melting fast in Amundsen Sea Embayment, whereas it is accumulating slowly in Enderby Land.