Phoebe M. R. DeVries’s research while affiliated with Harvard University and other places

What is this page?


This page lists works of an author who doesn't have a ResearchGate profile or hasn't added the works to their profile yet. It is automatically generated from public (personal) data to further our legitimate goal of comprehensive and accurate scientific recordkeeping. If you are this author and want this page removed, please let us know.

Publications (9)


Comparisons of spatial patterns of stress metrics
a–d, Analogous to Fig. 2e–h, but for an idealized thrust earthquake. The fault plane dips 45° to the north and the red line is the trace of the fault at the surface. Depth shown is 10 km.
Mainshock–aftershock examples
a–h, Analogous to Fig. 1a–h, using the same sign conventions for Coulomb failure stress change, but with results based on a training dataset (Supplementary Table 1) that excludes grid cells more than 5 km below the maximum depth of each slip distribution.
Comparisons of performance
a–h, Analogous to Fig. 2a–h, using training and test datasets (Supplementary Table 1) that exclude grid cells more than 5 km below the maximum depth of each slip distribution.
ROC curves associated with realization 6 of the datasets
a–d, Curves incorporate grid cells down to a depth of 50 km. e–h, Curves including grid cells down to 5 km beyond the maximum depth of each slip distribution. Thus, the neural network in d is trained and evaluated on a version of dataset realization 6 (Supplementary Table 2) that incorporates grid cells down to a depth of 50 km, whereas that in h is trained and evaluated on the same realizations of slip distributions, but incorporating only grid cells down to 5 km below each slip distribution.
Forward predictions of the neural networks from each realization of the training dataset, incorporating all grid cells down to 50 km
Each panel is analogous to Fig. 2h, but uses one of ten distinct neural networks trained on one of ten different realizations of the training dataset (Supplementary Table 2). See Methods for further discussion.

+3

Deep learning of aftershock patterns following large earthquakes
  • Article
  • Publisher preview available

August 2018

·

3,087 Reads

·

312 Citations

Nature

Phoebe M. R. DeVries

·

Fernanda Viégas

·

·

Aftershocks are a response to changes in stress generated by large earthquakes and represent the most common observations of the triggering of earthquakes. The maximum magnitude of aftershocks and their temporal decay are well described by empirical laws (such as Bath’s law¹ and Omori’s law²), but explaining and forecasting the spatial distribution of aftershocks is more difficult. Coulomb failure stress change³ is perhaps the most widely used criterion to explain the spatial distributions of aftershocks4–8, but its applicability has been disputed9–11. Here we use a deep-learning approach to identify a static-stress-based criterion that forecasts aftershock locations without prior assumptions about fault orientation. We show that a neural network trained on more than 131,000 mainshock–aftershock pairs can predict the locations of aftershocks in an independent test dataset of more than 30,000 mainshock–aftershock pairs more accurately (area under curve of 0.849) than can classic Coulomb failure stress change (area under curve of 0.583). We find that the learned aftershock pattern is physically interpretable: the maximum change in shear stress, the von Mises yield criterion (a scaled version of the second invariant of the deviatoric stress-change tensor) and the sum of the absolute values of the independent components of the stress-change tensor each explain more than 98 per cent of the variance in the neural-network prediction. This machine-learning-driven insight provides improved forecasts of aftershock locations and identifies physical quantities that may control earthquake triggering during the most active part of the seismic cycle.

View access options

Figure 1. Examples of two static stress fields considered in this analysis and corresponding ROC curves for one slip distribution (Cohee & Beroza, 1994) from the 1992 M w = 7.3 Landers earthquake in California. (a-c) Map view of ΔCFS(σ, 0.4) values within 100 km of the fault at 2.5 km, 7.5 km, and 22.5 km depth, respectively. Black squares represent grid cells in which one or more aftershocks occurred within 1 year of the mainshock. Thick yellow and black line represents extent of the mainshock rupture at each depth. Scale bar is shown in Figure 1a as a thick black line. (d) ROC curve for this particular slip distribution (Cohee & Beroza, 1994) and ΔCFS(σ, 0.4), including all grid cells and all aftershocks within a year after the mainshock. Black dotted line is a 1:1 line for reference. (e-h) Analogous to Figures 1a-1d for a different static stress field, Δτ max (σ).
Figure 2. ROC curves for all 38 stress metrics, in ranked order by merged AUC value (A m ). Category I metrics (1-23) have red titles; Category II metrics (24-38) have black titles. See Table 1 for detailed descriptions of the symbols in the titles. Vertically averaged AUC values A v , threshold averaged AUC values A t , and the fraction of statistically significant (α = 0.005) empirical p values Ψ are included for each metric. ROC curves for every slip distribution (213 in total) are shown with thin gray lines for each stress metric. Gray circles represent the locations on the ROC curves where Youden's index is maximized. Merged, vertically averaged, and threshold averaged ROC curves across all slip distributions are shown in thick blue, black, and red lines, respectively.
What Is Better Than Coulomb Failure Stress? A Ranking of Scalar Static Stress Triggering Mechanisms from 10 5 Mainshock-Aftershock Pairs: Static Stress Triggering Mechanisms

November 2017

·

265 Reads

·

42 Citations

·

Phoebe M. R. DeVries

·

Jeremy Faller

·

[...]

·

Aftershocks may be triggered by the stresses generated by preceding mainshocks. The temporal frequency and maximum size of aftershocks are well described by the empirical Omori and Bath laws, but spatial patterns are more difficult to forecast. Coulomb failure stress is perhaps the most common criterion invoked to explain spatial distributions of aftershocks. Here we consider the spatial relationship between patterns of aftershocks and a comprehensive list of 38 static elastic scalar metrics of stress (including stress tensor invariants, maximum shear stress, and Coulomb failure stress) from 213 coseismic slip distributions worldwide. The rates of true-positive and false-positive classification of regions with and without aftershocks are assessed with receiver operating characteristic (ROC) analysis. We infer that the stress metrics that are most consistent with observed aftershock locations are maximum shear stress and the magnitude of the second and third invariants of the stress tensor. These metrics are significantly better than random assignment at a significance level of 0.005 in over 80% of the slip distributions. In contrast, the widely-used Coulomb failure stress criterion is distinguishable from random assignment in only 51-64% of the slip distributions. These results suggest that a number of alternative scalar metrics are better predictors of aftershock locations than classic Coulomb failure stress change.


FIGURE 1. a) Structure and setup of an example artificial neural network with hidden layers, input layers, and output layers highlighted. Calculations move from left to right in the feedforward neural networks used here. b) Inputs to an example neuron from the previous layer. c) The calculation performed by the example neuron, weighting the inputs relative to one another, adding a bias, and applying an activation function, in order to calculation a value a, referred to as the activation of the neuron. d) The activation of the example neuron serves as one of the inputs to the next layer of neurons. Each neuron in the successive layers of the ANN are performing this same operation with different values of tunable parameters b and w i .  
FIGURE 2. Performance of the artificial neural network for a simple elastic case. a) Okada [1992] solution for a point source at 7.9 km depth. Black dots indicate 500 randomly distributed training points used to train an ANN; predictions from this ANN are shown in (b); c) Residuals between the ANN prediction and the true Okada [1992] solution; d) Mean absolute residual as a function of the number of training points for this elastic case; (e-h) Analogous to (a-d) for 13450 training points; (i-l) Analagous to (a-d) for 42200 training points.
Enabling large-scale viscoelastic calculations via neural network acceleration

January 2017

·

279 Reads

·

65 Citations

One of the most significant challenges involved in efforts to understand the effects of repeated earthquake cycle activity are the computational costs of large-scale viscoelastic earthquake cycle models. Computationally intensive viscoelastic codes must be evaluated 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 time-dependent deformation over short (<10 yr) 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 resolution. We demonstrate that this machine learning approach accelerates viscoelastic calculations by more than 50,000%. 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.


Enabling large-scale viscoelastic calculations via neural network acceleration

January 2017

One of the most significant challenges involved in efforts to understand the effects of repeated earthquake cycle activity are the computational costs of large-scale viscoelastic earthquake cycle models. Computationally intensive viscoelastic codes must be evaluated 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 time-dependent deformation over short (<10 yr) 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 resolution. We demonstrate that this machine learning approach accelerates viscoelastic calculations by more than 50,000%. 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.


Statistical tests of simple earthquake cycle models: Testing Earthquake Cycle Models

December 2016

·

20 Reads

·

1 Citation

A central goal of observing and modeling the earthquake cycle is to forecast when a particular fault may generate an earthquake: a fault late in its earthquake cycle may be more likely to generate an earthquake than a fault early in its earthquake cycle. Models that can explain geodetic observations throughout the entire earthquake cycle may be required to gain a more complete understanding of relevant physics and phenomenology. Previous efforts to develop unified earthquake models for strike-slip faults have largely focused on explaining both preseismic and postseismic geodetic observations available across a few faults in California, Turkey, and Tibet. An alternative approach leverages the global distribution of geodetic and geologic slip rate estimates on strike-slip faults worldwide. Here we use the Kolmogorov-Smirnov test for similarity of distributions to infer, in a statistically rigorous manner, viscoelastic earthquake cycle models that are inconsistent with 15 sets of observations across major strike-slip faults. We reject a large subset of two-layer models incorporating Burgers rheologies at a significance level of α = 0.05 (those with long-term Maxwell viscosities ηM <~ 4.0 × 10¹⁹ Pa s and ηM >~ 4.6 × 10²⁰ Pa s) but cannot reject models on the basis of transient Kelvin viscosity ηK. Finally, we examine the implications of these results for the predicted earthquake cycle timing of the 15 faults considered and compare these predictions to the geologic and historical record.


Figure 3. Model geometry cross section and rheologies used in this study.
Figure 5. Map of the locations and sizes of the historic earthquakes included in the viscoelastic block models. For the purposes of this
Figure 7. (a) Observed (blue) and modeled (red) interseismic GPS velocity fields for the best-fit viscosity structure highlighted in Figure 6b (Maxwell viscosity η M ˆ 10 19:0 Pa·s and a Kelvin viscosity η K ˆ 10 19:0 Pa·s). The observed velocity field shown here is the corrected interseismic velocity field v η M ; η K † (equation 8). (b) Residual velocities for the case shown in (a). Note the different scale from (a).
Detailed Parameters of the Earthquakes Taken into Account in the Viscoelastic Block Models
Viscoelastic Block Models of the North Anatolian Fault: A Unified Earthquake Cycle Representation of Pre‐ and Postseismic Geodetic Observations

November 2016

·

299 Reads

·

26 Citations

Bulletin of the Seismological Society of America

Along the North Anatolian fault (NAF), the surface deformation associated with tectonic block motions, elastic strain accumulation, and the viscoelastic response to past earthquakes has been geodetically observed over the last two decades. These observations include campaign-mode Global Positioning System (GPS) velocities from the decade prior to the 1999 Mw 7.4 İzmit earthquake and seven years of continuously recorded postseismic deformation following the seismic event. Here, we develop a 3D viscoelastic block model of the greater NAF region, including the last 2000 yrs of earthquake history across Anatolia, to simultaneously explain geodetic observations from both before and after the İzmit earthquake. With a phenomenologically motivated simple two-layer structure (schizosphere and plastosphere) and a Burgers rheology (with Maxwell viscosity log10 ηM ≈ 18:6–19:0 Pa·s and Kelvin viscosity log10 ηK ≈ 18:0–19:0 Pa·s), a block model that incorporates tectonic plate motions, interseismic elastic strain accumulation, transient viscoelastic perturbations, and internal strain can explain both the pre- and post-İzmit earthquake observations with a single unified model. Viscoelastic corrections to the interseismic GPS velocity field with the unified model reach magnitudes of ~2:9 mm=yr. Geodetically constrained slip-deficit rate estimates along the central NAF and northern strand of the NAF in the Sea of Marmara vary nonmonotonically with Maxwell viscosity and change by up to 23% (~4 mm=yr) for viscosities ranging from 10¹⁸ to 10²³ Pa·s. For the best-fit viscosity structures, central NAF slip-deficit rates reach 22 mm=yr, increasing to 28 mm=yr in the Sea of Marmara. Along the central NAF, these rates are similar to the fastest geologic slip-rate estimates. The fastest slip-deficit rate estimates along the entire fault system (~27–28 mm=yr) occur less than 50 km from Istanbul, along the northern strand of the NAF in the Sea of Marmara.


Figure 2: The geometry and rheologies used in this study.
Figure 3: The components of   (equation (2)) for a two-layer Maxwell model with a Maxwell viscosity of ηM = 1 × 1019 Pa s at three times (0, 25, and 50 years) after an idealized earthquake of MW = 7.3. The idealized fault is 50 km long, it ruptures to 15 km depth, and slip is assumed to be a uniform 5 m.
Figure 4: Maps of estimated   at 10 km depth on the day before the 1942, 1943, 1944, 1951, 1957, 1967, and 1999 earthquakes for end-member viscosity structure VS1 (ηM = 4 × 1018 Pa s, ηK = 1 × 1018 Pa s). Projected map of   in 2020 in the absence of additional earthquakes is also included. Because the NAF fault system consists of largely east-west striking faults,   changes are plotted with respect to vertical east-west striking receiver planes. Analogous figure for end-member viscosity structure VS2 (ηM = 1 × 1019 Pa s, ηK = 1 × 1019 Pa s) is included in supporting information Figure A1. Circles, colored by local   value resolved on the fault, highlight the location of the next earthquake in the sequence.
Figure 8: Evolution of   resolved along the western NAF in the Sea of Marmara at 10 km depth with viscosity structure VS1 (ηM = 4 × 1018 Pa s, ηK = 1 × 1018 Pa s). (bottom) Analogous figure for VS2 (ηM = 1 × 1019 Pa s, ηK = 1 × 1019 Pa s). Gray lines indicate the   values for the 1942, 1943, 1944, 1951, 1957, 1967, and 1999 earthquakes along the NAF. Inset shows the western NAF in the Sea of Marmara, with individual points along the fault colored with the same color scheme as in Figures 8a and 8b.
The Parameters of the Eight Earthquakes Included in the Models of Stress Transfer
Geodetically constrained models of viscoelastic stress transfer and earthquake triggering along the North Anatolian fault

June 2016

·

161 Reads

·

13 Citations

Over the past 80 years, 8 MW > 6.7 strike-slip earthquakes west of 40° longitude have ruptured the North Anatolian fault (NAF) from east to west. The series began with the 1939 Erzincan earthquake in eastern Turkey, and the most recent 1999 MW = 7.4 Izmit earthquake extended the pattern of ruptures into the Sea of Marmara in western Turkey. The mean time between seismic events in this westward progression is 8.5 ± 11 years (67% confidence interval), much greater than the timescale of seismic wave propagation (seconds to minutes). The delayed triggering of these earthquakes may be explained by the propagation of earthquake-generated diffusive viscoelastic fronts within the upper mantle that slowly increase the Coulomb failure stress change ( CFS) at adjacent hypocenters. Here we develop three-dimensional stress transfer models with an elastic upper crust coupled to a viscoelastic Burgers rheology mantle. Both the Maxwell (ηM = 4 × 1018−1 × 1019 Pa s) and Kelvin (ηK = 1 × 1018−1 × 1019 Pa s) viscosities are constrained by studies of geodetic observations before and after the 1999 Izmit earthquake. We combine this geodetically constrained rheological model with the observed sequence of large earthquakes since 1939 to calculate the time evolution of CFS changes along the North Anatolian fault due to viscoelastic stress transfer. Apparent threshold values of mean CFS at which the earthquakes in the eight decade sequence occur are between ∼0.02 to ∼3.15 MPa and may exceed the magnitude of static CFS values by as much as 177%. By 2023, we infer that the mean time-dependent stress change along the northern NAF strand in the Marmara Sea near Istanbul, which may have previously ruptured in 1766, may reach the mean apparent time-dependent stress thresholds of the previous NAF earthquakes.


Figure 1. The evolution of ΔCFS classic (Figure 2, row 1) and ΔCFS KC (Figure 2, row 2) at two specific locations over many earthquake cycles. In this example calculation, T = 100 years with a two-layer Burgers model (log 10 η M ≈ 19.0 Pa · s and log 10 η K = 21.0 Pa Á s). The locations of these specific points are shown as white dots in Figure 2. The inset is a schematic diagram of the model geometry and rheology.
Figure 4. Histograms of the individual times t max p (in percent of earthquake cycle), at which points in an area 250 × 250 km around the fault at 10 km depth reach a maximum value of ΔCFS KC > 0.0001 MPa. The thin black vertical lines represent the mean time across all points for which t max p is defined ( t max p D E , see text) and the median time, med t max p . Opacity encodes recurrence intervals (T): the most transparent histograms (background) correspond to T = 100 years and the least transparent histograms correspond to T = 500 years.
Kinematically Consistent Models of Viscoelastic Stress Evolution

April 2016

·

91 Reads

·

10 Citations

Following large earthquakes, coseismic stresses at the base of the seismogenic zone may induce rapid viscoelastic deformation in the lower crust and upper mantle. As stresses diffuse away from the primary slip surface in these lower layers, the magnitudes of stress at distant locations (>1 fault length away) may slowly increase. This stress relaxation process has been used to explain delayed earthquake triggering sequences like the 1992 Mw=7.3 Landers and 1999 Mw=7.1 Hector Mine earthquakes in California. However, a conceptual difficulty associated with these models is that the magnitudes of stresses asymptote to constant values over long time scales. This effect introduces persistent perturbations to the total stress field over many earthquake cycles. Here we present a kinematically consistent viscoelastic stress transfer model where the total perturbation to the stress field at the end of the earthquake cycle is zero everywhere. With kinematically consistent models, hypotheses about the potential likelihood of viscoelastically triggered earthquakes may be based on the timing of stress maxima, rather than on any arbitrary or empirically constrained stress thresholds. Based on these models, we infer that earthquakes triggered by viscoelastic earthquake cycle effects may be most likely to occur during the first 50% of the earthquake cycle regardless of the assumed long-term and transient viscosities. Key Points: A framework for kinematically consistent viscoelastic stress through the earthquake cycle is presented In the models, earthquake-generated stresses are zero at the end of the earthquake cycle and may reach an interseismic maximum Based on the timing of stress maxima, the mean time of viscoelastically triggered earthquakes may be ~5-35% through the earthquake cycle


Earthquake cycle deformation in the Tibetan plateau with a weak mid-crustal layer: EARTHQUAKE CYCLE WITH A WEAK LAYER

June 2013

·

140 Reads

·

35 Citations

[1] Geodetic observations of interseismic deformation across the Tibetan plateau contain information about both tectonic and earthquake cycle processes. Time-variations in surface velocities between large earthquakes are sensitive to the rheological structure of the subseismogenic crust, and, in particular, the viscosity of the middle and lower crust. Here we develop a semianalytic solution for time-dependent interseismic velocities resulting from viscoelastic stress relaxation in a localized midcrustal layer in response to forcing by a sequence of periodic earthquakes. Earthquake cycle models with a weak midcrustal layer exhibit substantially more near-fault preseismic strain localization than do classic two-layer models at short (<100 yr) Maxwell times. We apply both this three-layer model and the classic two-layer model to geodetic observations before and after the 1997 MW = 7.6 Manyi and 2001 MW = 7.8 Kokoxili strike-slip earthquakes in Tibet to estimate the viscosity of the crust below a 20 km thick seismogenic layer. For these events, interseismic stress relaxation in a weak (viscosity ≤1018.5 Pa⋅s) and thin (height ≤20 km) midcrustal layer explains observations of both preseismic near-fault strain localization and rapid (>50 mm/yr) postseismic velocities in the years following the coseismic ruptures. We suggest that earthquake cycle models with a localized midcrustal layer can simultaneously explain both preseismic and postseismic geodetic observations with a single Maxwell viscosity, while the classic two-layer model requires a rheology with multiple relaxation time scales.

Citations (7)


... Human intervention cannot halt natural disasters like earthquakes, but machine learning application expertise can be utilized to detect patterns in data and increase understanding and predictive power [11][12][13]. Most of the machine learning (ML) algorithms fall into one of two categories: Supervised Learning (SL) and Unsupervised Learning (USL) (Figure 1). ...

Reference:

Machine learning for predicting earthquake magnitudes in the Central Himalaya
Deep learning of aftershock patterns following large earthquakes

Nature

... Numerous studies have shown a strong correlation between aftershocks spatial pattern and changes in static Coulomb Failure Stress (CFS) resulting from mainshocks (e.g., Asayesh et al., 2018Asayesh et al., , 2019Asayesh, Zarei, & Zafarani, 2020;Harris & Simpson, 1992;Jamalreyhani et al., 2020;King et al., 1994;Toda et al., 1998). Recently, some studies using receiver operating characteristic (ROC) analysis have shown that several alternative scalar metrics are better predictors of aftershock locations than ΔCFS (DeVries et al., 2018;Meade et al., 2017;Mignan & Broccardo, 2019). Later, Sharma et al. (2020), by using available slip models from the SRCMOD database (Mai & Thingbaijam, 2014) and repeating the analysis with more appropriate stress calculations, showed that ΔCFS resolved on optimally oriented planes (OOP) or calculated for variable mechanism (VM) significantly improved the ΔCFS results for receiver mechanisms identical to the mainshock mechanism (on master fault orientation, MAS). ...

What Is Better Than Coulomb Failure Stress? A Ranking of Scalar Static Stress Triggering Mechanisms from 10 5 Mainshock-Aftershock Pairs: Static Stress Triggering Mechanisms

... ML has been applied in different scientific fields to replace simulations with the goal of reduced processing time (Kasim et al., 2021;Lam et al., 2023). Examples in Seismology include speeding-up viscoelastic calculations (DeVries et al., 2017), seismic wave simulation (Moseley et al., 2020), synthetic earthquake waveform generation (Florez et al., 2022), and so forth. In the field of earthquake forecasting, Dascher-Cousineau et al. (2023), Stockman et al. (2023), andZlydenko et al. (2023) have recently proposed ML models as alternatives to ETAS models, achieving similar or slightly superior forecasting results. ...

Enabling large-scale viscoelastic calculations via neural network acceleration

... Satellite geodetic data, such as Global Navigation Satellite System (GNSS) and Interferometric Synthetic Aperture Radar (InSAR), are regularly used to determine fault slip rates. A growing number of geodetic modeling approaches are being applied, including elastic block models (Evans, 2022;Hammond et al., 2024;McCaffrey, 2009;Meade & Loveless, 2009;Shen et al., 2015;Styron, 2022), deep dislocation models (Elston et al., 2024;Zeng, 2022;Zeng & Shen, 2014, NeoKinema-a kinematic finite element approach to estimate fault slip rates and off-fault strains (Bird, 2009;Shen & Bird, 2022), viscoelastic fault models (Pollitz, 2022;Pollitz et al., 2010;Pollitz & Evans, 2017), and viscoelastic earthquake cycle block models (Chuang & Johnson, 2011;DeVries et al., 2017;Johnson & Fukuda, 2010;Pollitz & Evans, 2017). Among these techniques, geodetically-derived surface velocity data are used as the primary constraints in models, alongside geologicallyderived slip rates typically applied as a prior constraint. ...

Viscoelastic Block Models of the North Anatolian Fault: A Unified Earthquake Cycle Representation of Pre‐ and Postseismic Geodetic Observations

Bulletin of the Seismological Society of America

... The physical mechanisms for earthquake interactions are reasonably well understood; among these are (1) coseismic static stress changes caused by the offset on the rupturing faults (e.g., Toda et al., 2005); (2) transient dynamic stress changes caused by the emitted seismic waves, which can trigger events even at thousands of kilometers distance (Gomberg et al., 2001;Felzer and Brodsky, 2006;Pollitz et al., 2012;Brodsky and van der Elst, 2014); (3) fluids propagating stress at a distance from the source region and causing delayed triggering (Prejean et al., 2004;Brodsky and Prejean, 2005;Miller, 2013;van der Elst et al., 2013); and (4) postseismic viscoelastic stress transfer caused by the gradual response of the upper-mantle to crustal deformation (Freed and Lin, 2001;DeVries et al., 2016), which perturbs the crustal stress at an extended time scales and over large distances. For example, Pollitz and Cattania (2017) were able to relate both the static and viscoelastic stress change to the spatial pattern of southern California seismicity. ...

Geodetically constrained models of viscoelastic stress transfer and earthquake triggering along the North Anatolian fault

... The difference in time-dependent loading between purely elastic fault models and those considering viscoelastic deformation suggests that viscoelastic interactions are an important ingredient for efforts aimed at modeling regional tectonics and multifault interactions, particularly given that the spatial footprint of this distributed deformation can be much larger than that of slip on individual faults (Figures 3 and 4). Viscoelastic stress interactions have been noted to be relevant to along-strike stress transfer and timing of a recent sequence of great earthquakes on the North Anatolian Fault (Devries & Meade, 2016;Devries et al., 2017), and Southern California (e.g., Freed & Lin, 2001). More generally, time-dependent loading alters the stress state on the fault preceding dynamic rupture. ...

Kinematically Consistent Models of Viscoelastic Stress Evolution

... These findings do not invalidate previous work on estimating the effective viscosity from postseismic, postglacial and lake rebound deformation observations assuming a linear Maxwell rheology (e.g., Devries & Meade, 2013;England et al., 2013;Johnson & Segall, 2004;Kaufmann & Amelung, 2000;Kenner & Segall, 2003;Larsen et al., 2005;Tamisiea et al., 2007). However, the important implication is that these estimates of the average viscosity, or viscosity structure, are tied to the observational window. ...

Earthquake cycle deformation in the Tibetan plateau with a weak mid-crustal layer: EARTHQUAKE CYCLE WITH A WEAK LAYER
  • Citing Article
  • June 2013