A kinematic rupture generator for ground-motion simulations: Validation and scenarios in South Iceland

Victor Moises Hernández Aguirre , Rajesh Rupakhety, Roberto Paolucci, Chiara Smerzini, Manuela Vanini, Bjarni Bessason, Sigurður Erlingsson

Physics-based ground-motion simulation can reduce epistemic uncertainty in regions with sparse strong-motion data, but hazard applications require rupture ensembles that are physically plausible, statistically controlled, and computationally efficient. We present a modular kinematic rupture generator for physics-based simulations (PBS) in which final slip, rupture speed ratio (VR/VS), and peak slip velocity (Vmax) are modelled as heterogeneous, mutually correlated spatial fields governed by prescribed one-point statistics and covariance-based two-point structure. The generator supports both event-constrained ruptures (for validation) and fully stochastic scenario ruptures, while enabling systematic propagation of source uncertainty through an effective stress-parameter scaling with Vmax. We demonstrate the approach using a regional 3D numerical model of South Iceland (accurate up to 1.9 Hz) solved with the spectral-element code SPEED. A broadband extension is obtained via the ANN2BB method. Validation against the June 2000 Mw 6.5 and Mw 6.4 earthquakes shows good agreement between recorded and simulated low-frequency waveforms and response spectra, supported by goodness-of-fit metrics across duration, peak measures, and long-period spectral ordinates. We then generate Mw 6.5 and Mw 7 scenario ensembles and compare their spectral-acceleration attenuation and variability with local and global GMMs. Mw 6.5 scenarios track the Icelandic model within its calibration range, whereas Mw 7 scenarios diverge from extrapolated local predictions, indicating substantial epistemic uncertainty in large-magnitude, short-distance scaling where observations are unavailable. The simulated variability is consistent with empirical expectations, supporting the use of the proposed rupture-generator–PBS framework to produce region-specific, non-ergodic ground-motion ensembles for scenario-based risk analysis and PBS-informed hazard assessment.