← Back to Gallery

Improving Event Depth Constraint of a Local-scale Surface Seismic Network With Downhole DAS

In recent years, distributed acoustic sensing (DAS) has garnered significant interest from the induced seismic monitoring community due to the potential for simultaneous improvements in earthquake monitoring, subsurface imaging, and reservoir characterization. Despite this interest, the practicality of using DAS as a means of real-time induced seismic monitoring remains unproven due to the large data volumes generated by DAS and the cost of providing adequate azimuthal coverage of the study area with downhole DAS alone. We propose an alternate approach wherein a real-time surface network of traditional wideband seismometers is enhanced by downhole DAS observations, providing dependable and efficient microseismic monitoring with improved depth constraint. First, we demonstrate that a targeted array of 9 surface stations with approximately 2 km spacing and pertinent noise-reduction techniques can sufficiently capture induced microseismicity to the extent that is required by governmental regulations. Next, we preprocess snippets of DAS data obtained from the lateral section of an injection well at 2 km depth for 114 microseismic events detected by the surface network during a one-week period of hydraulic fracturing. Leveraging the neural network model PhaseNet DAS (Zhu et al., 2023), we then generate P and S arrivals on all channels. To prevent these DAS observations from dominating relocations, we calculate average arrival times for different sections of the cable before relocating the events with NonLinLoc (Lomax et al., 2000) using both DAS and surface network observations as inputs. Comparing the resulting relocations to the original catalog, we find significant reductions in vertical uncertainty. Depth constraint is critical in the analysis of fracture networks and often seen as a weakness of surface networks. As a result, these results demonstrate the utility of a single downhole DAS cable when used in tandem with a robust surface array. Future developments of this workflow will explore potential improvements to cable orientation and the pre-trained model used in PhaseNet DAS.

Session: Seismology for the Energy Transition [Poster]

Type: Poster

Room: Exhibit Hall

Date: 4/16/2025

Presentation Time: 08:00 AM (local time)

Presenting Author: Alex Dzubay

Student Presenter: No

Invited Presentation: 

Poster Number: 128

Additional Authors