GSA Connects 2023 Meeting in Pittsburgh, Pennsylvania

Paper No. 88-5
Presentation Time: 9:05 AM

CHARACTERIZING SUBSURFACE STRUCTURES AT ROCK VALLEY, NEVADA WITH 2D SEISMIC DATA


LI, David1, GAO, Kai2, CHEN, Ting3, HUANG, Lianjie4, SWANSON, Erika5, SNELSON, Catherine6, HARDING, Jennifer7, PRESTON, Leiph7, BODMER, Miles7, ZEILER, Cleat8 and TURLEY, Reagan9, (1)Los Alamos National Laboratory, Mail Stop D452, Los Alamos, NM 87545; Los Alamos National Laboratory, Los Alamos, NM 87545, (2)Earth and Environmental Science, Los Alamos National Laboratory, Bikini Atoll Rd, PO Box 1663, Los Alamos, NM 87545; Los Alamos National Laboratory, Earth and Environmental Sciences Division, Los Alamos National Laboratory, Los Alamos, NM 87545, (3)Energy and Natural Resources Security, Los Alamos National Laboratory, Los Alamos, NM 87545, (4)Los Alamos National Laboratory, Los Alamos, NM 87545, (5)Los Alamos National Laboratory EES-17, P.O. Box 1663 D452, Los Alamos, NM 87545-0001, (6)Los Alamos National Laboratory, Earth and Environment Sciences Division, PO Box 1663, MS F665, Los Alamos, NM 87545, (7)Sandia National Laboratories, Albuquerque, NM 87120, (8)Nevada National Security Site, 232 Energy Way, North Las Vegas, NV 89030, (9)Nevada National Security Site, San Francisco, CA 94016

The Source Physics Experiment (SPE) aims to improve monitoring techniques by studying the source characteristics of underground chemical explosions. The main goal of SPE Phase III (Rock Valley Direct Comparison) is to execute two chemical explosions at hypocentral depths of the 1993 shallow earthquake sequence to compare the source characteristics of an underground explosion with those of a naturally occurring earthquake. To characterize the site before the main experiment, a 2D land seismic survey was conducted along two walkaway lines using accelerated weight drop sources (AWD). We processed the AWD seismic data and obtained 52,178 first-arrival P-wave times. Then built initial gradient velocity models along the two walkaway lines and applied a adjoint-state first-arrival traveltime tomography to update the P-wave velocity models down to 1.5 km depth. Applying reverse-time migration (RTM) we were able to obtain high-resolution subsurface image beneath the two lines. At this point, we then applied a machine-learning-based fault detection method to automatically pick faults on the RTM images. In comparing our RTM and automatically picked faults within a geologic framework model (GFM) of Rock Valley, we find that our results show more fidelity in the subsurface such as additional fault splays than the GFM model. Several of the automatically picked faults are consistent with the faults within the GFM. Our results will be incorporated in an update to the Rock Valley GFM, these data will be used in the output numerical meshes for the modeling & simulation to supporting underground monitoring.