Validating Forecasting Models for Induced Seismicity Related to Deep Geothermal Energy Projects

Abstract

Deep geothermal systems have big potential for future green energy resources. Induced seismicity occurring during early stimulation periods in deep geothermal projects of past years, however, clearly document our limited understanding of the processes at depth that lead to significant seismic hazard and that may influence public acceptance of future projects. Managing induced seismicity related to deep geothermal projects with advanced traffic light systems require models that are forward looking, dynamically updated on the fly as new data arrive and probabilistic in the sense that the inherent uncertainties in our understanding of the processes and in the model parameters are included. Based on modeling of the 2006 Basel induced sequence, Goertz-Allmann and Wiemer (2013) and Gischig and Wiemer (2013) suggest a so-called hybrid geomechanical forecast model that loosely couples the evolving pressure perturbation at depth with a stochastic seed model of potential faults. We currently develop a fully coupled non-linear hydraulic-seismic 3D model joint with a hazard assessment procedure. The goal is to improve the forecasting ability owing to validated physical constraints. As one of the first steps, a forecast testing center is being built up and comparison of observed co- and post-stimulation seismicity (maximum magnitude, productivity, b-value, seismogenic index, spatial parameters) is carried out for various deep geothermal stimulation projects. We also report on the ongoing efforts to analyze the 2013 St. Gallen earthquake sequence where more than 700 earthquakes related to this project have been recorded to date, the largest one having a magnitude mw = 3.3. This project also pointed out our current limited understanding on complex processes of reservoir behavior, for instance the potential role of over-pressured methane gas in inducing earthquakes.