Crustal Strain Rate Tensors From Dense GPS Networks

Some of the most destructive seismic events such as the 1994 Northridge MW 6.7 and the 1995 Kobe MW 6.8 earthquakes showed in a dramatic way the shortcoming of traditional seismic hazard assessment, which is mainly based on the recorded seismic history. As recent satellite based geodetic measurements like GPS and InSAR provide insights of crustal motion and deformation at new scales, both in accuracy and spatial and temporal resolution, they are of increasing importance for seismic hazard assessment. In particular, with the deployment of dense networks like the one of SCEC in southern California and others in different parts of Europe, southern America and Asia. However, in order to characterize active faults in terms of slip rates, locking depths, and earthquake recurrence times the geodetic data have to be interpreted using models and can be biased by the implicit assumptions. Furthermore a preliminary knowledge of the location of the active faults is often required for the modeling, increasing the difficulty of identifying unknown active structures. To avoid these uncertainties we apply geostatistical interpolation schemes to interpolate geodetic data and to obtain a continuous velocity field for each component. These fields are the basis for continuous (time dependent) strain rate tensors, which provide information like the spatial distribution of strain rates, direction of maximum shear strain rate, dilatation, etc. In regions with dense geodetic networks (e.g. Southern California and Western China) we can easily identify areas of high strain often associated with active faults, and estimate the errors introduced by the interpolation. Our interpolation scheme, that can easily be applied wherever geodetic data are available, clearly identifies and characterizes the active fault systems. For example in the eastern margin of the Tibetan Plateau our results are compatible with strain rate fields by Liu and Bird (2008) obtained by finite element interpolation of earthquake focal mechanism. Our results can be utilized as a starting point for further numerical models and/or geological investigation to estimate current activities of faults.