Enabling largescale viscoelastic calculations via neural network acceleration
Abstract
One of the most significant challenges involved in efforts to understand the effects of repeated earthquake cycle activity is the computational costs of largescale viscoelastic earthquake cycle models. Computationally intensive viscoelastic codes must be evaluated at thousands of times and locations, and as a result, studies tend to adopt a few fixed rheological structures and model geometries and examine the predicted timedependent deformation over short (<10 years) time periods at a given depth after a large earthquake. Training a deep neural network to learn a computationally efficient representation of viscoelastic solutions, at any time, location, and for a large range of rheological structures, allows these calculations to be done quickly and reliably, with high spatial and temporal resolutions. We demonstrate that this machine learning approach accelerates viscoelastic calculations by more than 50,000%. Finally, this magnitude of acceleration will enable the modeling of geometrically complex faults over thousands of earthquake cycles across wider ranges of model parameters and at larger spatial and temporal scales than have been previously possible.
 Authors:

 Harvard Univ., Cambridge, MA (United States). Dept. of Earth and Planetary Sciences
 Publication Date:
 Research Org.:
 Krell Inst., Ames, IA (United States)
 Sponsoring Org.:
 USDOE National Nuclear Security Administration (NNSA); National Science Foundation (NSF); Southern California Earthquake Center; USGS
 OSTI Identifier:
 1465348
 Alternate Identifier(s):
 OSTI ID: 1402388
 Grant/Contract Number:
 FG0297ER25308; EAR‐1033462; 6239; G12AC20038
 Resource Type:
 Accepted Manuscript
 Journal Name:
 Geophysical Research Letters
 Additional Journal Information:
 Journal Volume: 44; Journal Issue: 6; Journal ID: ISSN 00948276
 Publisher:
 American Geophysical Union
 Country of Publication:
 United States
 Language:
 English
 Subject:
 58 GEOSCIENCES
Citation Formats
DeVries, Phoebe M. R., Thompson, T. Ben, and Meade, Brendan J. Enabling largescale viscoelastic calculations via neural network acceleration. United States: N. p., 2017.
Web. doi:10.1002/2017GL072716.
DeVries, Phoebe M. R., Thompson, T. Ben, & Meade, Brendan J. Enabling largescale viscoelastic calculations via neural network acceleration. United States. doi:10.1002/2017GL072716.
DeVries, Phoebe M. R., Thompson, T. Ben, and Meade, Brendan J. Tue .
"Enabling largescale viscoelastic calculations via neural network acceleration". United States. doi:10.1002/2017GL072716. https://www.osti.gov/servlets/purl/1465348.
@article{osti_1465348,
title = {Enabling largescale viscoelastic calculations via neural network acceleration},
author = {DeVries, Phoebe M. R. and Thompson, T. Ben and Meade, Brendan J.},
abstractNote = {One of the most significant challenges involved in efforts to understand the effects of repeated earthquake cycle activity is the computational costs of largescale viscoelastic earthquake cycle models. Computationally intensive viscoelastic codes must be evaluated at thousands of times and locations, and as a result, studies tend to adopt a few fixed rheological structures and model geometries and examine the predicted timedependent deformation over short (<10 years) time periods at a given depth after a large earthquake. Training a deep neural network to learn a computationally efficient representation of viscoelastic solutions, at any time, location, and for a large range of rheological structures, allows these calculations to be done quickly and reliably, with high spatial and temporal resolutions. We demonstrate that this machine learning approach accelerates viscoelastic calculations by more than 50,000%. Finally, this magnitude of acceleration will enable the modeling of geometrically complex faults over thousands of earthquake cycles across wider ranges of model parameters and at larger spatial and temporal scales than have been previously possible.},
doi = {10.1002/2017GL072716},
journal = {Geophysical Research Letters},
number = 6,
volume = 44,
place = {United States},
year = {2017},
month = {3}
}
Works referenced in this record:
Viscosity of Oceanic Asthenosphere Inferred from Remote Triggering of Earthquakes
journal, May 1998
 Pollitz, F. F.
 Science, Vol. 280, Issue 5367
TimeDependent Distributed Afterslip on and Deep below the Izmit Earthquake Rupture
journal, February 2002
 Burgmann, R.
 Bulletin of the Seismological Society of America, Vol. 92, Issue 1
How do “ghost transients” from past earthquakes affect GPS slip rate estimates on southern California faults?: “GHOST TRANSIENTS” AND FAULT SLIP RATES
journal, April 2013
 Hearn, E. H.; Pollitz, F. F.; Thatcher, W. R.
 Geochemistry, Geophysics, Geosystems, Vol. 14, Issue 4
Viscoelastic Block Models of the North Anatolian Fault: A Unified Earthquake Cycle Representation of Pre‐ and Postseismic Geodetic Observations
journal, November 2016
 DeVries, Phoebe M. R.; Krastev, Plamen G.; Dolan, James F.
 Bulletin of the Seismological Society of America, Vol. 107, Issue 1
Modelling of stress patterns along the western part of the north anatolian fault zone
journal, September 1988
 Roth, Frank
 Tectonophysics, Vol. 152, Issue 34
Dynamic models of interseismic deformation and stress transfer from plate motion to continental transform faults: MODELS OF INTERSEISMIC DEFORMATION
journal, May 2012
 Takeuchi, Christopher S.; Fialko, Yuri
 Journal of Geophysical Research: Solid Earth, Vol. 117, Issue B5
Multilayer feedforward networks are universal approximators
journal, January 1989
 Hornik, Kurt; Stinchcombe, Maxwell; White, Halbert
 Neural Networks, Vol. 2, Issue 5
Elastic and inelastic triggering of earthquakes in the North Anatolian Fault zone
journal, October 2006
 LorenzoMartín, Francisco; Roth, Frank; Wang, Rongjiang
 Tectonophysics, Vol. 424, Issue 34
Stress Triggering of the 1999 Hector Mine Earthquake by Transient Deformation Following the 1992 Landers Earthquake
journal, May 2002
 Pollitz, F. F.
 Bulletin of the Seismological Society of America, Vol. 92, Issue 4
Viscoelastic stresstriggering of the 1999 Hector Mine Earthquake by the 1992 Landers Earthquake
journal, August 2001
 Zeng, Yuehua
 Geophysical Research Letters, Vol. 28, Issue 15
Gradientbased learning applied to document recognition
journal, January 1998
 Lecun, Y.; Bottou, L.; Bengio, Y.
 Proceedings of the IEEE, Vol. 86, Issue 11
Fault interaction and stress triggering of twentieth century earthquakes in Mongolia: STRESS TRIGGERING OF EARTHQUAKES IN MONGOLIA
journal, October 2003
 Pollitz, Fred; Vergnolle, Mathilde; Calais, Eric
 Journal of Geophysical Research: Solid Earth, Vol. 108, Issue B10
Stress transfer and nonlinear stress accumulation at the North Anatolian fault, Turkey
journal, February 1990
 Tselentis, G. Akis; Drakopoulos, J.
 Pure and Applied Geophysics PAGEOPH, Vol. 132, Issue 4
Delayed triggering of the 1999 Hector Mine earthquake by viscoelastic stress transfer
journal, May 2001
 Freed, Andrew M.; Lin, Jian
 Nature, Vol. 411, Issue 6834
Quasistatic internal deformation due to a dislocation source in a multilayered elastic/viscoelastic halfspace and an equivalence theorem
journal, July 2006
 Fukahata, Yukitoshi; Matsu'ura, Mitsuhiro
 Geophysical Journal International, Vol. 166, Issue 1
Postearthquake ground movements correlated to porepressure transients
journal, July 2003
 Jónsson, Sigurjón; Segall, Paul; Pedersen, Rikke
 Nature, Vol. 424, Issue 6945
General expressions for internal deformation fields due to a dislocation source in a multilayered elastic halfspace
journal, May 2005
 Fukahata, Yukitoshi; Matsu'ura, Mitsuhiro
 Geophysical Journal International, Vol. 161, Issue 2
Izmit earthquake postseismic deformation and dynamics of the North Anatolian Fault Zone
journal, January 2009
 Hearn, E. H.; McClusky, S.; Ergintav, S.
 Journal of Geophysical Research, Vol. 114, Issue B8
Approximation by superpositions of a sigmoidal function
journal, December 1989
 Cybenko, G.
 Mathematics of Control, Signals, and Systems, Vol. 2, Issue 4
Geodetically constrained models of viscoelastic stress transfer and earthquake triggering along the North Anatolian fault: VISCOELASTIC STRESS TRANSFER MODELS
journal, July 2016
 DeVries, Phoebe M. R.; Krastev, Plamen G.; Meade, Brendan J.
 Geochemistry, Geophysics, Geosystems, Vol. 17, Issue 7