Ground Motion Prediction Equations (GMPEs)

This is the 1^st subproject of the “Seismological research concerning Swiss nuclear installations” project.

This subproject aims to improve source-scaling and attenuation models and to develop methods for the prediction of strong ground motion in Switzerland both at the surface as well as at depth. Two main approaches are investigated: ground motion prediction equations (GMPEs) and stochastic simulation models. Both approaches require adaptions to the local seismicity and careful consideration of their calibration to Swiss conditions. The Fourier spectral and stochastic models correspond to the current state of research and have some advantages over the empirical attenuation relationships, as it is possible to adjust the model to specific local site conditions. The complete understanding in terms of physical parameterization of such models is crucial in order to decouple different effects, which allow building robust predictive models that scale appropriately to large magnitude events. In this regard, variability in source parameter such as stress drop and corner frequency is crucial. Similarly, variability in site-related attenuation parameter kappa (local intrinsic and scattering attenuation) is also need to be well understood. Developed stochastic and duration models from Japanese data allow the review of the Swiss model for large magnitudes in the different distance ranges and at various rock sites which have, as yet, not been instrumentally recorded in Switzerland. The local weak-to-moderate seismicity is used to calibrate the predictive models. The long-term goal is to develop an improved stochastic simulation model for Switzerland allowing existing uncertainties to be rigorously evaluated and reduced. In future, such models will also allow an assessment of ground motion caused by induced seismicity due to the activation of existing fractures and/or the generation of new fractures. Moreover, we developed, validated, and applied a physics-based stochastic model to characterize high-frequency ground motion at depth in the Fourier domain. The goal of it is to be able to predict future ground motions at deep geological disposals.