NRC Research and Fellowship Programs
Fellowships Office
Policy and Global Affairs

Participating Agencies

RAP opportunity at Air Force Science and Technology Fellowship Program     AF STFP

Geophysical Machine Learning Applications


Space Vehicles Directorate, RV/Battle Space Environment

opportunity location
13.40.12.C0247 Kirtland Air Force Base, NM 871175776


name email phone
Glenn Eli Baker 505.846.6070


With each magnitude unit decrease in threshold for nuclear explosion monitoring, the number of seismic events that must be detected and classified increases by a factor of ten. Similarly, as many more stations are needed to achieve lower thresholds, the number of signals detected at monitoring stations that must be detected, classified, and associated with a seismic source may increase by more than a factor of 100. Processing such data streams in near real time must be accomplished at very low error rates, as it’s not practical to significantly increase the number of human analysts who provide quality control. To achieve this will require significant improvements to automated approaches in all aspects of the problem, including signal detection, phase identification, association of signals with hypothesized events, the event locations and times, and event classification. Complicating this already challenging problem is that signals from smaller events will likely have lower signal-to-noise ratios, and will likely be more affected by scattering as their propagation paths will generally be through the more heterogeneous shallowest layers in the Earth than those of larger events recorded at greater distances experience.

Application of machine learning techniques to problems in seismic monitoring is a new and rapidly developing area of research (for a broad review see Kong et al., 2019). Applications include, but aren’t limited to, signal detection (e.g. Yoon, et al., 2015; Ross et al., 2018), phase identification, classification, and association. Methods that build on these early efforts and apply machine learning techniques to seismic monitoring problems are sought.

Kong, Q., D. T. Trugman, Z. E. Ross, M. J. Bianco, B. J. Meade, P. Gerstoft (2019) Machine Learning in Seismology: Turning Data into Insights. Seismological Research Letters ; 90 (1): 3–14. doi:

Ross, Z. E., M.-A. Meier, E. Hauksson, and T. H. Heaton (2018). Generalized seismic phase detection with deep learning, Bull Seismol. Soc. Am. doi: 10.1785/0120180080

Yoon, C. E., O. O'Reilly, K. J. Bergen, and G. C. Beroza (2015). Earthquake detection through computationally efficient similarity search, Sci. Adv. 1, no. 11, e1501057, doi: 10.1126/sciadv.1501057

key words
Machine Learning; Seismology; Monitoring


Citizenship:  Open to U.S. citizens
Level:  Open to Postdoctoral and Senior applicants


Base Stipend Travel Allotment Supplementation
$95,000.00 $5,000.00

Experience Supplement:
Postdoctoral and Senior Associates will receive an appropriately higher stipend based on the number of years of experience past their PhD.

Copyright © 2024. National Academy of Sciences. All rights reserved.Terms of Use and Privacy Policy