2.1.2. Monitoring surface motion with GPS and Galileo

In Europe, the number of high quality GPS stations, the time series they provide and their precision has steadily increased over the last years. Furthermore, the recently started GALILEO project of European satellite systems will in the future provide access to fundamental data on contemporary kinematics and deformation of the Earth’s surface (Fig. 5). These observations will have a major impact on qualitative and quantitative research on geodynamic processes and physical properties of the crust and upper mantle by delivering independent constraints on the boundary conditions and rheological properties of numerical models of plate tectonic forces (Fig. 6). Since GPS data first became available, numerical modelling research also commenced. Only recently however, has increased computer power permitted to construct high resolution numerical models that reflect the geometrical complexity of the crust and its heterogeneous material properties. GPS data will play a major role in the 4th generation of seismic hazard assessment. Here, 4D deterministic numerical models of stress evolution will permit to quantify stress concentrations in the seismogenic crust and their evolution on time scales of several seismic cycles.

Fig. 6. Comparison of GPS observed velocities in the Eastern Mediterranean relative to a fixed Eurasia plate (McClusky et al., 2000) with results of a 3D numerical model with non-linear, temperature controlled visco-elastic-plastic rheology (Heidbach and Drewes, 2003). The model is driven by slab pull forces, the indentation of the Arabian plate into the Eurasia plate and gravitationally induced body forces due to the changing Moho depth. Major results are that trench suction due to roll-back of the Hellenic arc and collisional forces are required to promote lateral extrusion of the Anatolia-Aegean complex, as observed by the GPS data.

To accurately use GPS signals as independent constraints for numerical stress evolution models, the causes of these signals need to be fully understood and quantified. In seismogenic regions, first order contributions to the GPS signal come from two groups of sources acting on different spatial and time scales, namely from (1) tectonic plate boundary forces, such as slab pull, ridge push, gravitational collapse of elevated regions and mantle drag, and (2) co-seismic displacements (Fig. 7). In seismogenically active intraplate regions, extensional and compressional intraplate deformations, including lithospheric folding may also contribute to the signal.

Fig. 7. Observations at the permanent GPS station Arequipa in Peru before and after the June 2001 Mw 8.4 earthquake in South-Peru. Interseismic (tectonic) movement, co-seismic displacement and post-seismic signal are clearly distinguishable. The results of a 2D finite element model can explain the unexpected post-seismic signal with the relaxation of tensional stresses in the lower crust and upper mantle (Hergert and Heidbach, 2006). GPS data were provided by the German Geodetic Research Institute (DGFI) in Munich (Wolfgang Seemüller, pers. comm., 2005).

Second order causes involve a wide range of other processes such as visco-elastic relaxation of co-seismically induced stress changes in the lower crust and upper mantle (e.g. Hergert and Heidbach, 2006), after slip and after creep (e.g. Melbourne et al., 2002), silent slip (e.g. Douglas et al., 2005), poroelastic rebound (e.g. Cocco and Rice, 2002), and mass redistribution due to deglaciation, sedimentation and erosion (e.g. Fischer et al., 2004). In relatively stable regions, characterized by low seismicity, first order mechanisms play a subordinate role whilst second order processes are probably the driving mechanism of displacements.

Incorporation of time series from dense networks of continuously observing GPS stations permits to set up 4D numerical stress evolution models for time-dependent seismic hazard assessment. Currently the European permanent GPS station network of EUREF consists of 189 stations. In addition, a large number of regional and local networks is available, which provide additional detailed kinematics data (McClusky et al., 2000; Roberts and Ganas, 2000; Bürgmann et al., 2002; Calais et al., 2002; Caporali et al., 2003; Fernandez et al., 2003; Hollenstein et al., 2003; Lenk et al., 2003; Van der Hoeven et al., 2004; Wdowinski et al., 2004; Tesauro et al., 2005).