Toward the Development of Earthquake Recurrence Models from 3D GNSS Velocity Field in Europe

Probabilistic Seismic Hazard Assessment demands the development of reliable earthquake recurrence models, which are usually based on time and spatial distribution of the past seismicity contained on earthquake catalogs. This usually generate models rather well constrained on seismically active regions where large historical catalogs are available. But in low to moderate seismicity regions, where data is scarce, establishing earthquake recurrence from past events is a major challenge. On those regions, geodetic measurements can provide useful information for deriving alternative recurrence models based on strain rate.

The impact of the crust deformation on the processes that control the seismic activity is still controversial. The seismic activity is usually thought to be associated to the active tectonic deformation as estimated from the horizontal displacements field. But in regions with low horizontal deformation, getting the horizontal strain rates is difficult since the displacements field can be dominated by the noise of the geodetic data. Additionally, non-tectonic processes such as the Glacial Isostatic Adjustment (GIA) can exist, and may impact the seismicity rate of those regions. Then seismicity rates derived from the horizontal velocity fields might not adjust the observed seismicity rates on such regions.

We propose a methodology to build a combined GNSS velocity field dataset for Europe, that could be used for the development of earthquake recurrence models. For this, 5 different GNSS velocity field solution for Europe are considered. Using the velocity solutions of common stations, the different datasets are converted to a common reference frame. Based on the comparison of the velocity values, a methodology is established to generate a combined velocity field, considering the uncertainty of each independent solution. A criterion for automatic identification and outliers removal is implemented, as well an adaptive smoothing scheme that depends on the station density, the noise and the local tectonic deformation rate.

We propose a methodology to obtain strain rate maps from GNSS data based on the VISR software [Shen et al., 2015], not only considering the horizontal velocity field, but including also the vertical velocity field for Europe, considering the effects of flexure of the crust on regions where important GIA signals are observed.

Finally, earthquake recurrence models are derived and compared with catalog-based models in Europe to evaluate their mutual agreement, comparing also the results obtained on regions with significant tectonic deformation versus regions where important GIA signals are observed.