Gravity variations of geodynamic origin: Recognition and study on the basis of modern satellite gravity data

摘要

The possibility to study geodynamic processes from data on time variations in the Earth's gravitational field recorded by the GRACE (Gravity Recovery and Climate Experiment) and GOCE (Gravity Field and Steady-State Ocean Circulation Explorer) satellite systems is analyzed. As an example, processes in subduction zones are considered. The following problems are investigated by statistical methods: (1) the recognition of gravity variations caused by an earthquake provided that fault plane parameters are known but the faulting amplitude is unknown; (2) discrimination between two models of an earthquake source from satellite data; and (3) the search for a linear temporal trend in the gravity variations related to the strain accumulation in a locked fragment of a subduction zone in terms of a set of gravitational field models constructed for several time moments. Presently, only the first model of the Earth's gravitational field, constructed from GRACE data, is available. This model is much more accurate than previous models, although its expected instrumental accuracy has not been achieved as et. There are grounds to suggest that the accuracy of data will be significantly improved and, therefore, the problem of signal recognition is studied with various accuracy levels of satellite data (we mean the current and the basically attainable (expected) accuracy of the GRACE and GOCE systems). We show that deformations caused by strong earthquakes (such as the 1964 Alaska earthquake) produce a gravity effect that can be recovered from GRACE data even with their present-day accuracy. If the accuracy approaches its expected value, gravitational effects related to earthquakes of smaller magnitudes (such as the 2003 Hokkaido earthquake, which occurred after the GRACE launch) will be identified. Several fault plane models are available for many earthquakes, particularly for earthquakes whose epicentral zones are below sea level. In this work, we show that even the present-day accuracy of GRACE data is sufficient for discrimination between models of large earthquakes such as the 1960 Chile earthquake. An accuracy of the GRACE and GOCE data close to its expected value will enable discrimination between models of earthquakes of significantly smaller magnitudes (such as the 2003 Hokkaido earthquake). Numerical calculations demonstrate that gravity effects related to deformations of the Earth's surface in locked areas of subduction zones can be recognized if the accuracy of future models of the Earth's gravitational field derived from GRACE data is increased by an order of magnitude (still remaining an order of magnitude lower than the expected value).