Multichannel singular spectrum analysis of thegravity field data from GRACE satellites

摘要

The NASA/DLR GRACE (Gravity Recovery and Climate Experiment) twin-satellite mission was launched on 17.03.2002. They provide data products including monthly gravitational field solutions, which reflect the mass redistribution within the Earth system. However, these data need filtering and reduction of the so-called stripes or correlated high-frequency errors due primarily to model errors and imperfect observability of the gravity field inversion solution. We applied the multichannel singular spectrum analysis (MSSA) method to the post-process monthly GRACE data in the spectral and spatial domains. This method allows an improved separation of signals represented by the principal components containing periodic and secular terms, which can be associated with seasonal or longer redistribution of water masses, ice melting, glacial isostatic adjustment, ocean mass variations, and transient events such as the Sumatra earthquake in 2004. Flexibility of MSSA, which allows to filter out stripes and high-frequency noise, makes the method useful for GRACE data post-processing.