arXiv ScienceSearch

arXiv subjects

John Rundle

Publications and source records attributed to John Rundle.

3 recordsLinked to original sources

Simulating Subterranean Fluid Injection through Iteration on the VirtualQuake Model

This work extends the VirtualQuake earthquake simulation framework to incorporate the effects of fluid injection on fault stability and induced seismicity. Reworking VirtualQuake into a system using stress point sources, instead of rectangular segments, the new model offers increased geometric flexibility, greater stability, and the re-addition of cross-fault interactions. This approach is paired with the addition of fluid injection modeling, through the distribution of inflationary stress sources, according to invasion percolation, simulating both the stress effect of the injection on nearby faults, and deformation from the injection itself. The model captures both immediate and long-term impacts of injection cycles, including hydraulic fracturing processes and post-injection pressure dissipation. Results show that while single injections produce limited stress changes, repeated injections generate persistent high-pressure regions that progressively destabilize nearby faults, increasing the likelihood of seismic events. This model evolution offers a tool for the evaluation and characterization of long term risks from commercial injection.

physics.geo-ph

Earthquake Nowcasting with Deep Learning

We review previous approaches to nowcasting earthquakes and introduce new approaches based on deep learning using three distinct models based on recurrent neural networks and transformers. We discuss different choices for observables and measures presenting promising initial results for a region of Southern California from 1950-2020. Earthquake activity is predicted as a function of 0.1-degree spatial bins for time periods varying from two weeks to four years. The overall quality is measured by the Nash Sutcliffe Efficiency comparing the deviation of nowcast and observation with the variance over time in each spatial region. The software is available as open-source together with the preprocessed data from the USGS.

physics.geo-ph

Simulated Dynamical Weakening and Abelian Avalanches in Mean-Field Driven Threshold Models

Mean-field coupled lattice maps are used to approximate the physics of driven threshold systems with long range interactions. However, they are incapable of modeling specific features of the dynamic instability responsible for generating avalanches. Here we present a method of simulating specific frictional weakening effects in a mean field slider block model. This provides a means of exploring dynamical effects previously inaccessible to discrete time simulations. This formulation also results in Abelian avalanches, where rupture propagation is independent of the failure sequence. The resulting event size distribution is shown to be generated by the boundary crossings of a stochastic process. This is applied to typical models to explain some commonly observed features.

nlin.AO