Shubham Sharma
Research Subjects
Research Interests
Shubham Sharma has a total of 29 co-authors. Here are the top co-authors by number of shared publications:
- Supriyo Mitra
Department of Earth Sciences Indian Institute of Science Education and Research Kolkata Mohanpur India
9 shared publications - Sebastian Hainzl
GFZ Helmholtz Centre for Geosciences, Potsdam, Germany
4 shared publications Gert Zöller
Institute of Mathematics, University of Potsdam , 14469 Potsdam , Germany
3 shared publicationsMonumoy Ghosh
3 shared publications- Siddharth Dey
Department of Earth Sciences, Indian Institute of Science Education and Research Kolkata , Mohanpur - 741246, West Bengal
3 shared publications Debarchan Powali
2 shared publications
Latest Publications
- Open Access
- 20 January 2025
- Published by Copernicus GmbH
Deformation of the Indian Lithosphere from radial anisotropy: Signatures of laterally varying plate geometry beneath Tibet and hotspot volcanism beneath the Deccan Plateau.
The Indian Lithosphere has been shaped by multiple tectonic processes, which include break-up from the Gondwana Supercontinent, traversing over the Reunion and Kerguelen hotspots, collision with Eurasia, and underthrusting beneath the Himalaya and Tibetan Plateau. Seismic velocity structure and radial anisotropy of the lithosphere preserves imprints of these tectonic processes and related deformation. We perform joint-modeling of fundamental-mode Rayleigh (LR) and Love (LQ) wave group-velocity dispersion, for periods between 10 and 120s, to obtain radially anisotropic shear-wave velocity structure across India, Himalaya and Tibet. 1D path-average dispersion curves, computed for ~14700 regional earthquake-receiver raypaths, has been passed through systematic quality control of signal-to-noise ratio (>3), elimination of multipathed energy using polarization analysis, and removal of overtone interference, by synthetic tests. These 1D dispersion data are combined through a tomographic formulation to obtain 2D maps. The tomographic parametrization is done using 4906 nodes as apex of triangular elements of side 1°. LR and LQ fundamental-mode group-velocity dispersion data at these nodes are the observation input to the joint inversion. The inversion is done in 2-steps, first by parameterizing the model as isotropic layers and using an isotropic inversion scheme to obtain the best fitting Vs model; second using this output Vs model into an anisotropic inversion scheme, implemented using Genetic Algorithms (GA). GA exhaustively searches the model-space composed of Vsh, Vph and Xi[Vsh^2/Vsv^2] as free parameters. The fit to both LR and LQ datasets significantly improve in the anisotropic inversion. Results are presented as 2D depth-slice maps and cross-sections constructed using bilinear interpolation. The main findings from our models are lateral variation in the voigt-average Vs beneath the Tibetan Plateau at depth between 80-140 km. Western Tibet has high Vs and positive Xi, while Central-Eastern TIbet has Low Vs and negative Xi. From cross-sections across both regions, we infer that the dip and underthrusting of the Indian Plate beneath Tibet has lateral variation. The high Vs and positive anisotropy in Western Tibet indicates a shallow underthrusting of the Indian lithosphere up to the Tarim Basin, with simple-shear deformation. Where as, the lower Vs and negative anisotropy in Central-Eastern Tibet is a result of partial-underthrusting of India at a steeper-angle up to the Bangong-Nujiang Suture, and pure-shear deformation of thickened Tibet Lithosphere beneath North-Central Tibet. A negative anisotropy signature along the Reunion volcanic track is observed between 100 and 160 km depth. We infer this to be the signature of Reunion hotspot volcanism in the Indian lithosphere caused by the vertical ascent of a huge volume of melt arising from the plume-head. Similar observations are also made beneath the track of the Kergulean hotspot.
- 20 January 2025
- Published by Copernicus GmbH
Open AccessDeformation of the Indian Lithosphere from radial anisotropy: Signatures of laterally varying plate geometry beneath Tibet and hotspot volcanism beneath the Deccan Plateau.
The Indian Lithosphere has been shaped by multiple tectonic processes, which include break-up from the Gondwana Supercontinent, traversing over the Reunion and Kerguelen hotspots, collision with Eurasia, and underthrusting beneath the Himalaya and Tibetan Plateau. Seismic velocity structure and radial anisotropy of the lithosphere preserves imprints of these tectonic processes and related deformation. We perform joint-modeling of fundamental-mode Rayleigh (LR) and Love (LQ) wave group-velocity dispersion, for periods between 10 and 120s, to obtain radially anisotropic shear-wave velocity structure across India, Himalaya and Tibet. 1D path-average dispersion curves, computed for ~14700 regional earthquake-receiver raypaths, has been passed through systematic quality control of signal-to-noise ratio (>3), elimination of multipathed energy using polarization analysis, and removal of overtone interference, by synthetic tests. These 1D dispersion data are combined through a tomographic formulation to obtain 2D maps. The tomographic parametrization is done using 4906 nodes as apex of triangular elements of side 1°. LR and LQ fundamental-mode group-velocity dispersion data at these nodes are the observation input to the joint inversion. The inversion is done in 2-steps, first by parameterizing the model as isotropic layers and using an isotropic inversion scheme to obtain the best fitting Vs model; second using this output Vs model into an anisotropic inversion scheme, implemented using Genetic Algorithms (GA). GA exhaustively searches the model-space composed of Vsh, Vph and Xi[Vsh^2/Vsv^2] as free parameters. The fit to both LR and LQ datasets significantly improve in the anisotropic inversion. Results are presented as 2D depth-slice maps and cross-sections constructed using bilinear interpolation. The main findings from our models are lateral variation in the voigt-average Vs beneath the Tibetan Plateau at depth between 80-140 km. Western Tibet has high Vs and positive Xi, while Central-Eastern TIbet has Low Vs and negative Xi. From cross-sections across both regions, we infer that the dip and underthrusting of the Indian Plate beneath Tibet has lateral variation. The high Vs and positive anisotropy in Western Tibet indicates a shallow underthrusting of the Indian lithosphere up to the Tarim Basin, with simple-shear deformation. Where as, the lower Vs and negative anisotropy in Central-Eastern Tibet is a result of partial-underthrusting of India at a steeper-angle up to the Bangong-Nujiang Suture, and pure-shear deformation of thickened Tibet Lithosphere beneath North-Central Tibet. A negative anisotropy signature along the Reunion volcanic track is observed between 100 and 160 km depth. We infer this to be the signature of Reunion hotspot volcanism in the Indian lithosphere caused by the vertical ascent of a huge volume of melt arising from the plume-head. Similar observations are also made beneath the track of the Kergulean hotspot.
- Open Access
- 26 December 2024
- Published by GeoScienceWorld in Lithosphere
- Vol. 2024 (Special 15)
Role of Intraplate Strike-Slip Earthquakes in Accommodating Convergence Across the Eastern Himalayan Plate Boundary System
North-East India, at the eastern extremity of the Himalaya, is one of the most rapidly deforming intraplate regions. The tectonics of this region is shaped by oblique convergence between two nearly perpendicular plate boundaries of the Eastern Himalaya and the Indo-Burman convergence zone. This region of distributed deformation is associated with intraplate strike-slip and oblique-slip earthquakes. We model the source mechanisms of six recent moderate-to-strong intraplate earthquakes (5.0≤ Mw ≤6.7) using teleseismic P- and SH-waveform inversion and use source directivity and rupture back-projection, for the largest event, to isolate the fault plane. We combine these mechanisms with previous earthquake source studies, GPS-geodetic-velocity vectors, and GPS-derived strain-rate field, to build a kinematic model. Majority of the earthquakes have strike-slip to oblique-slip (thrust) motion and originate in the middle-to-lower crust. These reveal that the entire NE-Indian crust is seismogenic. The oblique-thrust earthquakes occur due to high in-plane compressive stresses in the flexed Indian Plate. The region north of the Dawki Fault, in the vicinity of the Kopili and Dhubri-Chungthang Fault Zones, deforms through dextral strike-slip faulting and anticlockwise rotation of blocks along NW-SE trending transverse structures. The transitional crust of the Bengal Basin has several NE-SW trending paleo-rifts which are reactivated as sinistral strike-slip faults and the intervening blocks undergo clockwise rotation. The oblique convergence between the Indian and Eurasian Plates is partitioned into dextral and sinistral strike-slip motions across NE-India. The GPS velocity vectors and the strain-rate field indicate that the region north of the Dawki Fault has strong coupling between the surface deformation and the earthquake faulting. However, in the region south of the Dawki Fault, the coupling is weaker. The strike-slip earthquakes beneath Indo-Burma probably occur due to a complex interplay between the trench-normal slab-pull forces and lateral-shear forces set up by the strike-parallel components of the interplate-coupling resistance and the mantle-drag forces.
- 26 December 2024
- Published by GeoScienceWorld in Lithosphere
- Vol. 2024 (Special 15)
Open AccessRole of Intraplate Strike-Slip Earthquakes in Accommodating Convergence Across the Eastern Himalayan Plate Boundary System
North-East India, at the eastern extremity of the Himalaya, is one of the most rapidly deforming intraplate regions. The tectonics of this region is shaped by oblique convergence between two nearly perpendicular plate boundaries of the Eastern Himalaya and the Indo-Burman convergence zone. This region of distributed deformation is associated with intraplate strike-slip and oblique-slip earthquakes. We model the source mechanisms of six recent moderate-to-strong intraplate earthquakes (5.0≤ Mw ≤6.7) using teleseismic P- and SH-waveform inversion and use source directivity and rupture back-projection, for the largest event, to isolate the fault plane. We combine these mechanisms with previous earthquake source studies, GPS-geodetic-velocity vectors, and GPS-derived strain-rate field, to build a kinematic model. Majority of the earthquakes have strike-slip to oblique-slip (thrust) motion and originate in the middle-to-lower crust. These reveal that the entire NE-Indian crust is seismogenic. The oblique-thrust earthquakes occur due to high in-plane compressive stresses in the flexed Indian Plate. The region north of the Dawki Fault, in the vicinity of the Kopili and Dhubri-Chungthang Fault Zones, deforms through dextral strike-slip faulting and anticlockwise rotation of blocks along NW-SE trending transverse structures. The transitional crust of the Bengal Basin has several NE-SW trending paleo-rifts which are reactivated as sinistral strike-slip faults and the intervening blocks undergo clockwise rotation. The oblique convergence between the Indian and Eurasian Plates is partitioned into dextral and sinistral strike-slip motions across NE-India. The GPS velocity vectors and the strain-rate field indicate that the region north of the Dawki Fault has strong coupling between the surface deformation and the earthquake faulting. However, in the region south of the Dawki Fault, the coupling is weaker. The strike-slip earthquakes beneath Indo-Burma probably occur due to a complex interplay between the trench-normal slab-pull forces and lateral-shear forces set up by the strike-parallel components of the interplate-coupling resistance and the mantle-drag forces.
- Open Access
- 17 May 2023
- Published by Oxford University Press (OUP) in Geophysical Journal International
- Vol. 235 (1) , 509-517
Seismicity parameters dependence on main shock-induced co-seismic stress
The Gutenberg-Richter (GR) and the Omori-Utsu (OU) law describe the earthquakes’ energy release and temporal clustering and are thus of great importance for seismic hazard assessment. Motivated by experimental results, which indicate stress-dependent parameters, we consider a combined global dataset of 127 mainshock-aftershock sequences and perform a systematic study of the relationship between mainshock-induced stress changes and associated seismicity patterns. For this purpose, we calculate space-dependent Coulomb Stress (ΔCFS) and alternative receiver-independent stress metrics in the surrounding of the mainshocks. Our results indicate a clear positive correlation between the GR b-value and the induced stress, contrasting expectations from laboratory experiments and suggesting a crucial role of structural heterogeneity and strength variations. Furthermore, we demonstrate that the aftershock productivity increases nonlinearly with stress, while the OU parameters c and p systematically decrease for increasing stress changes. Our partly unexpected findings can have an important impact on future estimations of the aftershock hazard.
- 17 May 2023
- Published by Oxford University Press (OUP) in Geophysical Journal International
- Vol. 235 (1) , 509-517
Open AccessSeismicity parameters dependence on main shock-induced co-seismic stress
The Gutenberg-Richter (GR) and the Omori-Utsu (OU) law describe the earthquakes’ energy release and temporal clustering and are thus of great importance for seismic hazard assessment. Motivated by experimental results, which indicate stress-dependent parameters, we consider a combined global dataset of 127 mainshock-aftershock sequences and perform a systematic study of the relationship between mainshock-induced stress changes and associated seismicity patterns. For this purpose, we calculate space-dependent Coulomb Stress (ΔCFS) and alternative receiver-independent stress metrics in the surrounding of the mainshocks. Our results indicate a clear positive correlation between the GR b-value and the induced stress, contrasting expectations from laboratory experiments and suggesting a crucial role of structural heterogeneity and strength variations. Furthermore, we demonstrate that the aftershock productivity increases nonlinearly with stress, while the OU parameters c and p systematically decrease for increasing stress changes. Our partly unexpected findings can have an important impact on future estimations of the aftershock hazard.
- Open Access
- 15 May 2023
- Published by Copernicus GmbH
Radially anisotropic shear-wave velocity structure of northern India, Himalaya and Tibet
3-component regional waveforms for ~14700 raypaths sampling India, Himalaya and Tibet, have been used for multi-taper polarization analysis of surface waves between periods of 10 and 120 s. Rayleigh (LR) and Love (LQ) wave energy arriving at the station within +/- 10 degree of the theoretical back azimuth have been used to compute 1D path average fundamental mode group velocity dispersion. Theoretical dispersion of fundamental and first two higher modes have been computed using each 1D path average velocity structure constructed from CRUST1.0 over IASP91 mantle model. These are compared with the observed dispersion dataset to identify and remove those periods with higher mode overlap with the observed fundamental mode picks. The shortlisted dispersion datasets consist of ~90% of the original dataset. To ascertain the lateral variation in the group velocity, the observed dispersion has been used to compute 2D tomography maps of LR and LQ group velocities at discrete periods between 10 and 120 s. From these maps, seven 2D profiles across northern India, Himalaya and Tibet have been extracted for modeling the radially anisotropic shear-wave velocity structure of the lithosphere. Haskell-Thompson (H-T) matrix method is used to calculate synthetic LQ dispersion. For the LR dispersion, the H-T method with reduced delta matrix has been used, considering Vph≠Vpv and Eta≠1. The inversion scheme uses genetic algorithms (GA) to search the model space parameterized using Vsh, Vph, Xi and thickness for 3 crustal layers and 2 mantle layers underlain by a mantle half-space. Synthetic tests have been performed using theoretical LR and LQ dispersion curves, computed from global models with 5% and 10% anisotropy, introduced in the mantle layers. The fit to the synthetic LR and LQ dispersion data and the model recovery using GA inversion is satisfactory for such tests. This inversion scheme is being applied to the observed LR and LQ dispersion data from the seven profiles and the results will be presented.
- 15 May 2023
- Published by Copernicus GmbH
Open AccessRadially anisotropic shear-wave velocity structure of northern India, Himalaya and Tibet
3-component regional waveforms for ~14700 raypaths sampling India, Himalaya and Tibet, have been used for multi-taper polarization analysis of surface waves between periods of 10 and 120 s. Rayleigh (LR) and Love (LQ) wave energy arriving at the station within +/- 10 degree of the theoretical back azimuth have been used to compute 1D path average fundamental mode group velocity dispersion. Theoretical dispersion of fundamental and first two higher modes have been computed using each 1D path average velocity structure constructed from CRUST1.0 over IASP91 mantle model. These are compared with the observed dispersion dataset to identify and remove those periods with higher mode overlap with the observed fundamental mode picks. The shortlisted dispersion datasets consist of ~90% of the original dataset. To ascertain the lateral variation in the group velocity, the observed dispersion has been used to compute 2D tomography maps of LR and LQ group velocities at discrete periods between 10 and 120 s. From these maps, seven 2D profiles across northern India, Himalaya and Tibet have been extracted for modeling the radially anisotropic shear-wave velocity structure of the lithosphere. Haskell-Thompson (H-T) matrix method is used to calculate synthetic LQ dispersion. For the LR dispersion, the H-T method with reduced delta matrix has been used, considering Vph≠Vpv and Eta≠1. The inversion scheme uses genetic algorithms (GA) to search the model space parameterized using Vsh, Vph, Xi and thickness for 3 crustal layers and 2 mantle layers underlain by a mantle half-space. Synthetic tests have been performed using theoretical LR and LQ dispersion curves, computed from global models with 5% and 10% anisotropy, introduced in the mantle layers. The fit to the synthetic LR and LQ dispersion data and the model recovery using GA inversion is satisfactory for such tests. This inversion scheme is being applied to the observed LR and LQ dispersion data from the seven profiles and the results will be presented.
- Open Access
- 27 March 2022
- Published by Copernicus GmbH
Seismic Attenuation of India, Himalaya and Tibet using Lg-coda waves
We present maps to show the lateral variation of Lg coda attenuation at 1-Hz across India, Himalaya and Tibet. We use vertical component waveforms from regional earthquakes (epicentral distance5) recorded by the IISER-K seismological network, ones operated by the Indian Meteorological Department, and data acquired from the IRIS-DMC. Lg-coda waves are modeled as single back-scattered energy, sampling an ellipsoidal volume. The attenuation of Lg-coda is quantified using the quality factor (Q), which is frequency dependent. We use the stacked spectral ratio (SSR) method to calculate the single-trace Lg-coda Q at 1 Hz (Qo) and its frequency dependence (η). A moving-window stack of scaled-logarithmic ratios of spectral amplitudes, for window length of 25.6 s and different central lapse time, is computed for each frequency. Through a linear regression of log (stacked spectral ratio) and log (frequency), using least-squares fitting, we obtain (1-η) and log(Qo), respectively. Lg-coda is selected in a frequency range of 0.2-5 Hz, with coda window starting at 3.15 km/s. Our total coda window lengths vary between 140 s to 780 s. Our preliminary results show low Q values (~200-400) in the Eastern and Western Himalaya - possibly because of scattering of seismic energy from structural heterogeneities. Most of the Indian Shield and the intraplate regions of Shillong Plateau and Brahmaputra valley are characterized by intermediate to high Q values (~600-800), indicating fairly efficient propagation of seismic energy. Intermediate values of Q (~400-500) occur in the Indo-Burman Ranges which may be due to the cold elastic subducting oceanic lithosphere. Patches of low Q in the Tibetan Plateau (~200) are possibly the result of high temperatures and partial melts present in the crust. Our results show how the nature of the Indian Plate changes as we go from an active continent-continent collision zone in the north to eastward subduction of transitional material at the Indo-Burma ranges. Our plots of Qo and η as a function of epicentral distance, coda length and magnitude show no systematic variations.
- 27 March 2022
- Published by Copernicus GmbH
Open AccessSeismic Attenuation of India, Himalaya and Tibet using Lg-coda waves
We present maps to show the lateral variation of Lg coda attenuation at 1-Hz across India, Himalaya and Tibet. We use vertical component waveforms from regional earthquakes (epicentral distance5) recorded by the IISER-K seismological network, ones operated by the Indian Meteorological Department, and data acquired from the IRIS-DMC. Lg-coda waves are modeled as single back-scattered energy, sampling an ellipsoidal volume. The attenuation of Lg-coda is quantified using the quality factor (Q), which is frequency dependent. We use the stacked spectral ratio (SSR) method to calculate the single-trace Lg-coda Q at 1 Hz (Qo) and its frequency dependence (η). A moving-window stack of scaled-logarithmic ratios of spectral amplitudes, for window length of 25.6 s and different central lapse time, is computed for each frequency. Through a linear regression of log (stacked spectral ratio) and log (frequency), using least-squares fitting, we obtain (1-η) and log(Qo), respectively. Lg-coda is selected in a frequency range of 0.2-5 Hz, with coda window starting at 3.15 km/s. Our total coda window lengths vary between 140 s to 780 s. Our preliminary results show low Q values (~200-400) in the Eastern and Western Himalaya - possibly because of scattering of seismic energy from structural heterogeneities. Most of the Indian Shield and the intraplate regions of Shillong Plateau and Brahmaputra valley are characterized by intermediate to high Q values (~600-800), indicating fairly efficient propagation of seismic energy. Intermediate values of Q (~400-500) occur in the Indo-Burman Ranges which may be due to the cold elastic subducting oceanic lithosphere. Patches of low Q in the Tibetan Plateau (~200) are possibly the result of high temperatures and partial melts present in the crust. Our results show how the nature of the Indian Plate changes as we go from an active continent-continent collision zone in the north to eastward subduction of transitional material at the Indo-Burma ranges. Our plots of Qo and η as a function of epicentral distance, coda length and magnitude show no systematic variations.
- Open Access
- 27 March 2022
- Published by Copernicus GmbH
3D transversely isotropic shear-wave velocity structure of India and Tibet from joint modeling of Rayleigh and Love waves group velocity dispersion.
We use regional Rayleigh and Love wave data, from 4750 earthquakes (M >= 4.0) recorded at 726 stations across India and Tibet, to compute fundamental mode group velocity dispersion between 10 s and 120 s, using the Multiple Filter Technique (MFA). These result in the dense coverage of 14,706 and 14,898 ray-paths for Rayleigh and Love waves, respectively. The dispersion data at discrete periods have been combined through a ray-theory based tomographic formulation to obtain 2D maps of lateral variation in group velocities, where the best resolution is upto 2.5° and 4° for Rayleigh and Love waves tomographic maps, respectively. The Peninsular Shield, the Himalayan foreland basin, the Himalayan collision-zone and the Tibetan Plateau, have been sampled at unprecedented detail. Rayleigh and Love wave dispersion curves, at each node point of the tomography, have been inverted for 1D isotropic shear-wave velocity structure of Vsv and Vsh, respectively, which are combined to obtain 3D Vsv and Vsh structures across India and Tibet. We jointly invert the two datasets at each node to obtain an isotropic 1D velocity structure. The isotropic inversion fits the two datasets reasonably well, however, the misfit in the dispersion dataset both at high and low periods is high. For this, we incorporate radial anisotropy in the velocity structure and parameterize the crust with three layers and upper mantle with two layers. Assuming this radially anisotropic earth structure, we use Genetic Algorithms (GA) to explore the model space extensively. The synthetic dispersion curves are computed using Thomson-Haskell method with reduced delta matrix. The free parameters used in the inversion are VPH and VSH, layer thickness (h) and Vs anisotropy represented by Xi (ξ=VSH/VSV)2. The non-linear inversion technique converges to a best-fitting model by iteratively minimising the misfit between the observed and the data. The 2D group velocity dispersion heterogeneities, the 3D structures of Vsv and Vsh (both isotropic and transversely isotropic) will be presented with a focus to characterize a) the structure of the Indian plate and it’s extent of underthrusting beneath Tibet, and b) to quantify the low-velocity zone at the base of the Himalayan wedge, across the basal decollement, which ruptures in megathrust earthquakes.
- 27 March 2022
- Published by Copernicus GmbH
Open Access3D transversely isotropic shear-wave velocity structure of India and Tibet from joint modeling of Rayleigh and Love waves group velocity dispersion.
We use regional Rayleigh and Love wave data, from 4750 earthquakes (M >= 4.0) recorded at 726 stations across India and Tibet, to compute fundamental mode group velocity dispersion between 10 s and 120 s, using the Multiple Filter Technique (MFA). These result in the dense coverage of 14,706 and 14,898 ray-paths for Rayleigh and Love waves, respectively. The dispersion data at discrete periods have been combined through a ray-theory based tomographic formulation to obtain 2D maps of lateral variation in group velocities, where the best resolution is upto 2.5° and 4° for Rayleigh and Love waves tomographic maps, respectively. The Peninsular Shield, the Himalayan foreland basin, the Himalayan collision-zone and the Tibetan Plateau, have been sampled at unprecedented detail. Rayleigh and Love wave dispersion curves, at each node point of the tomography, have been inverted for 1D isotropic shear-wave velocity structure of Vsv and Vsh, respectively, which are combined to obtain 3D Vsv and Vsh structures across India and Tibet. We jointly invert the two datasets at each node to obtain an isotropic 1D velocity structure. The isotropic inversion fits the two datasets reasonably well, however, the misfit in the dispersion dataset both at high and low periods is high. For this, we incorporate radial anisotropy in the velocity structure and parameterize the crust with three layers and upper mantle with two layers. Assuming this radially anisotropic earth structure, we use Genetic Algorithms (GA) to explore the model space extensively. The synthetic dispersion curves are computed using Thomson-Haskell method with reduced delta matrix. The free parameters used in the inversion are VPH and VSH, layer thickness (h) and Vs anisotropy represented by Xi (ξ=VSH/VSV)2. The non-linear inversion technique converges to a best-fitting model by iteratively minimising the misfit between the observed and the data. The 2D group velocity dispersion heterogeneities, the 3D structures of Vsv and Vsh (both isotropic and transversely isotropic) will be presented with a focus to characterize a) the structure of the Indian plate and it’s extent of underthrusting beneath Tibet, and b) to quantify the low-velocity zone at the base of the Himalayan wedge, across the basal decollement, which ruptures in megathrust earthquakes.
- Open Access
- 4 March 2021
- Published by Copernicus GmbH
Is Coulomb Stress the Best Choice for Aftershock Forecasting?
The Coulomb failure stress (CFS) criterion is the most commonly used method for predicting spatial distributions of aftershocks following large earthquakes. However, large uncertainties are always associated with the calculation of Coulomb stress change. The uncertainties mainly arise due to nonunique slip inversions and unknown receiver faults; especially for the latter, results are highly dependent on the choice of the assumed receiver mechanism. Based on binary tests (aftershocks yes/no), recent studies suggest that alternative stress quantities, a distance‐slip probabilistic model as well as deep neural network (DNN) approaches, all are superior to CFS with predefined receiver mechanism. To challenge this conclusion, which might have large implications, we use 289 slip inversions from SRCMOD database to calculate more realistic CFS values for a layered half‐space and variable receiver mechanisms. We also analyze the effect of the magnitude cutoff, grid size variation, and aftershock duration to verify the use of receiver operating characteristic (ROC) analysis for the ranking of stress metrics. The observations suggest that introducing a layered half‐space does not improve the stress maps and ROC curves. However, results significantly improve for larger aftershocks and shorter time periods but without changing the ranking. We also go beyond binary testing and apply alternative statistics to test the ability to estimate aftershock numbers, which confirm that simple stress metrics perform better than the classic Coulomb failure stress calculations and are also better than the distance‐slip probabilistic model.
- 4 March 2021
- Published by Copernicus GmbH
Open AccessIs Coulomb Stress the Best Choice for Aftershock Forecasting?
The Coulomb failure stress (CFS) criterion is the most commonly used method for predicting spatial distributions of aftershocks following large earthquakes. However, large uncertainties are always associated with the calculation of Coulomb stress change. The uncertainties mainly arise due to nonunique slip inversions and unknown receiver faults; especially for the latter, results are highly dependent on the choice of the assumed receiver mechanism. Based on binary tests (aftershocks yes/no), recent studies suggest that alternative stress quantities, a distance‐slip probabilistic model as well as deep neural network (DNN) approaches, all are superior to CFS with predefined receiver mechanism. To challenge this conclusion, which might have large implications, we use 289 slip inversions from SRCMOD database to calculate more realistic CFS values for a layered half‐space and variable receiver mechanisms. We also analyze the effect of the magnitude cutoff, grid size variation, and aftershock duration to verify the use of receiver operating characteristic (ROC) analysis for the ranking of stress metrics. The observations suggest that introducing a layered half‐space does not improve the stress maps and ROC curves. However, results significantly improve for larger aftershocks and shorter time periods but without changing the ranking. We also go beyond binary testing and apply alternative statistics to test the ability to estimate aftershock numbers, which confirm that simple stress metrics perform better than the classic Coulomb failure stress calculations and are also better than the distance‐slip probabilistic model.
- Open Access
- 4 March 2021
- Published by Copernicus GmbH
Testing physics and statics based hybrid ETAS models
Currently, the Epidemic Type Aftershock Sequence (ETAS) model is state-of-the-art for forecasting aftershocks. However, the under-performance of ETAS in forecasting the spatial distribution of aftershocks following a large earthquake make us adopt alternative approaches for the modelling of the spatial ETAS-kernel. Here we develop a hybrid physics and statics based forecasting model. The model uses stress changes, calculated from inverted slip models of large earthquakes, as the basis of the spatial kernel in the ETAS model in order to get more reliable estimates of spatiotemporal distribution of aftershocks. We evaluate six alternative approaches of stress-based ETAS-kernels and rank their performance against the base ETAS model. In all cases, an expectation maximization (EM) algorithm is used to estimate the ETAS parameters. The model approach has been tested on synthetic data to check if the known parameters can be inverted successfully. We apply the proposed method to forecast aftershocks of mainshocks available in SRCMOD database, which includes 192 mainshocks with magnitudes in the range between 4.1 and 9.2 occurred from 1906 to 2020. The probabilistic earthquake forecasts generated by the hybrid model have been tested using established CSEP test metrics and procedures. We show that the additional stress information, provided to estimate the spatial probability distribution, leads to more reliable spatiotemporal ETAS-forecasts of aftershocks as compared to the base ETAS model.
- 4 March 2021
- Published by Copernicus GmbH
Open AccessTesting physics and statics based hybrid ETAS models
Currently, the Epidemic Type Aftershock Sequence (ETAS) model is state-of-the-art for forecasting aftershocks. However, the under-performance of ETAS in forecasting the spatial distribution of aftershocks following a large earthquake make us adopt alternative approaches for the modelling of the spatial ETAS-kernel. Here we develop a hybrid physics and statics based forecasting model. The model uses stress changes, calculated from inverted slip models of large earthquakes, as the basis of the spatial kernel in the ETAS model in order to get more reliable estimates of spatiotemporal distribution of aftershocks. We evaluate six alternative approaches of stress-based ETAS-kernels and rank their performance against the base ETAS model. In all cases, an expectation maximization (EM) algorithm is used to estimate the ETAS parameters. The model approach has been tested on synthetic data to check if the known parameters can be inverted successfully. We apply the proposed method to forecast aftershocks of mainshocks available in SRCMOD database, which includes 192 mainshocks with magnitudes in the range between 4.1 and 9.2 occurred from 1906 to 2020. The probabilistic earthquake forecasts generated by the hybrid model have been tested using established CSEP test metrics and procedures. We show that the additional stress information, provided to estimate the spatial probability distribution, leads to more reliable spatiotemporal ETAS-forecasts of aftershocks as compared to the base ETAS model.
- Open Access
- 22 September 2020
- Published by American Geophysical Union (AGU) in Journal of Geophysical Research: Solid Earth
- Vol. 125 (9)
Is Coulomb Stress the Best Choice for Aftershock Forecasting?
The Coulomb failure stress (CFS) criterion is the most commonly used method for predicting spatial distributions of aftershocks following large earthquakes. However, large uncertainties are always associated with the calculation of Coulomb stress change. The uncertainties mainly arise due to nonunique slip inversions and unknown receiver faults; especially for the latter, results are highly dependent on the choice of the assumed receiver mechanism. Based on binary tests (aftershocks yes/no), recent studies suggest that alternative stress quantities, a distance‐slip probabilistic model as well as deep neural network (DNN) approaches, all are superior to CFS with predefined receiver mechanism. To challenge this conclusion, which might have large implications, we use 289 slip inversions from SRCMOD database to calculate more realistic CFS values for a layered half‐space and variable receiver mechanisms. We also analyze the effect of the magnitude cutoff, grid size variation, and aftershock duration to verify the use of receiver operating characteristic (ROC) analysis for the ranking of stress metrics. The observations suggest that introducing a layered half‐space does not improve the stress maps and ROC curves. However, results significantly improve for larger aftershocks and shorter time periods but without changing the ranking. We also go beyond binary testing and apply alternative statistics to test the ability to estimate aftershock numbers, which confirm that simple stress metrics perform better than the classic Coulomb failure stress calculations and are also better than the distance‐slip probabilistic model.
- 22 September 2020
- Published by American Geophysical Union (AGU) in Journal of Geophysical Research: Solid Earth
- Vol. 125 (9)
Open AccessIs Coulomb Stress the Best Choice for Aftershock Forecasting?
The Coulomb failure stress (CFS) criterion is the most commonly used method for predicting spatial distributions of aftershocks following large earthquakes. However, large uncertainties are always associated with the calculation of Coulomb stress change. The uncertainties mainly arise due to nonunique slip inversions and unknown receiver faults; especially for the latter, results are highly dependent on the choice of the assumed receiver mechanism. Based on binary tests (aftershocks yes/no), recent studies suggest that alternative stress quantities, a distance‐slip probabilistic model as well as deep neural network (DNN) approaches, all are superior to CFS with predefined receiver mechanism. To challenge this conclusion, which might have large implications, we use 289 slip inversions from SRCMOD database to calculate more realistic CFS values for a layered half‐space and variable receiver mechanisms. We also analyze the effect of the magnitude cutoff, grid size variation, and aftershock duration to verify the use of receiver operating characteristic (ROC) analysis for the ranking of stress metrics. The observations suggest that introducing a layered half‐space does not improve the stress maps and ROC curves. However, results significantly improve for larger aftershocks and shorter time periods but without changing the ranking. We also go beyond binary testing and apply alternative statistics to test the ability to estimate aftershock numbers, which confirm that simple stress metrics perform better than the classic Coulomb failure stress calculations and are also better than the distance‐slip probabilistic model.
- 1 August 2020
- Published by Elsevier in Tectonophysics
- Vol. 789
A reappraisal of the 2005 Kashmir (M 7.6) earthquake and its aftershocks: Seismotectonics of NW Himalaya
- 1 August 2020
- Published by Elsevier in Tectonophysics
- Vol. 789
A reappraisal of the 2005 Kashmir (M 7.6) earthquake and its aftershocks: Seismotectonics of NW Himalaya