A data-driven artificial neural network model for the prediction of ground motion from induced seismicity: The case of The Geysers geothermal field

Ground-motion models have gained foremost attention during recent years for being capable of predicting ground-motion intensity levels future seismic scenarios. They are a key element estimating hazard and always demand timely refinement in order to improve the reliability maps. In present study, we propose ground motion prediction model induced earthquakes recorded The Geysers geothermal area. We use fully connected data-driven artificial neural network (ANN) fit parameters. Especially, used data from 212 at 29 stations Berkeley–Geysers between September 2009 November 2010. magnitude range is 1.3 3.3 moment (Mw), whereas hypocentral distance 0.5 20 km. motions predicted terms peak acceleration (PGA), velocity (PGV), 5% damped spectral (SA) T=0.2, 0.5, 1 s. values our deep learning compared with observed predictions made by empirical equations developed Sharma et al. (2013) same set using nonlinear mixed-effect (NLME) regression technique. For validation approach, on separate 25 region, magnitudes ranging 1.0 3.1 distances 1.2 15.5 km, ANN providing 3% improvement baseline GMM model. results obtained study show moderate unravel modeling features that were not taken into account comparison measured both R 2 statistic total standard deviation, together inter-event intra-event components.