Distributed acoustic sensing (DAS) deployed on existing telecommunication dark optical fibers can turn the existing infrastructure into a dense seismic array with meter-level channel spacing. This technology provides a novel manner to record seismic phenomena with unprecedented high-temporal and spatial resolution; especially, in volcanic and geothermically active areas. We present the application of subsurface seismic imaging within the Long Valley caldera region, located in the north portion of the eastern Sierra in California. This area presents significant volcanic activity (e.g., hot springs and fumaroles) since a large volume of magma is emplaced within the deep subsurface (below 10 km from the surface). Monitoring this region is essential to assess the volcanic hazard but also for the exploitation of geothermal energy. In fact, a 30 MW binary cycle geothermal power plant is currently generating energy in this region. In this work, we deploy two DAS OptaSense Plexus units to record more than 6,000 local and regional earthquakes and trained a machine-learning (ML) algorithm to accurately pick more than 12 million P- and S-wave arrivals times that are then employed within an efficient tomographic workflow. Compared to tables based on conventional, even dense, seismic arrays, DAS data provides arrival time tables with a total number of picks that is 2 to 3 orders of magnitude larger, which represents a computational challenge for existing tomographic approaches. Moreover, volcanic areas present subsurface structures with significant velocity contrasts that would result in complex raypaths. To properly take into account complex ray geometries and handle the large number of ML-measured traveltimes, we develop a double-difference (DD) Eikonal traveltime tomography workflow based on the adjoint-state method. The figure below shows the tomographic images of the Long Valley caldera, in a side-by-side comparison with the latest tomography S-wave speed model, which is also our initial model, based on full waveform inversion of surface waves between 6 and 20 s. With the improved data coverage from the two DAS arrays, we substantially improve the model resolution in the top 15 km to 20 km. The heterogeneous shallow structures within the caldera, which only appear as a smooth low-velocity anomaly in the initial model, become sharp in our new tomographic model and correlate well with surface geology. These shallow velocity reductions are potentially related to the filling material deposited in the depression and the extensive surface hydrothermal system that could be employed to further expand the geothermal operations in the area. The successful application of our workflow in the Long Valley caldera demonstrates the potential of using the existing telecommunication infrastructure for subsurface imaging purposes. In the continental US, there are multiple zones in which the seismicity level and existing fiber network could be employed for imaging goals. For instance, one of the areas of interest is the Salton Sea area in which our method could better characterize the subsurface structures related to this geothermal field. We are currently exploring various areas where to apply our methodology.

Press the Back button in your browser, or search again.

Copyright 2024, Stanford Geothermal Program: Readers who download papers from this site should honor the copyright of the original authors and may not copy or distribute the work further without the permission of the original publisher.