USGS Science⌕ Search

SEARCH · USGS Science

Results for “Journal of Geophysical Research”

Search indexed USGS publications on groundwater, aquifers, geologic maps, mineral resources and earthquakes. Explore source records by subject and place.

Quote a phrase for an exact phrase match. Source license links do not imply unrestricted reuse.

At least 289 records · Page 16Linked to original sources

Spatial heterogeneity of within-stream methane concentrations

Streams, rivers, and other freshwater features may be significant sources of CH 4 to the atmosphere. However, high spatial and temporal variabilities hinder our ability to understand the underlying processes of CH 4 production and delivery to streams and also challenge the use of scaling approaches across large areas. We studied a stream having high geomorphic variability to assess the underlying scale of CH 4 spatial variability and to examine whether the physical structure of a stream can explain the variation in surface CH 4 . A combination of high-resolution CH 4 mapping, a survey of groundwater CH 4 concentrations, quantitative analysis of methanogen DNA, and sediment CH 4 production potentials illustrates the spatial and geomorphic controls on CH 4 emissions to the atmosphere. We observed significant spatial clustering with high CH 4 concentrations in organic-rich stream reaches and lake transitions. These sites were also enriched in the methane-producing mcrA gene and had highest CH 4 production rates in the laboratory. In contrast, mineral-rich reaches had significantly lower concentrations and had lesser abundances of mcrA . Strong relationships between CH 4 and the physical structure of this aquatic system, along with high spatial variability, suggest that future investigations will benefit from viewing streams as landscapes, as opposed to ecosystems simply embedded in larger terrestrial mosaics. In light of such high spatial variability, we recommend that future workers evaluate stream networks first by using similar spatial tools in order to build effective sampling programs.

Journal of Geophysical Research G: Biogeosciences↗

Deposition of mercury in forests across a montane elevation gradient: Elevational and seasonal patterns in methylmercury inputs and production

Global mercury contamination largely results from direct primary atmospheric and secondary legacy emissions, which can be deposited to ecosystems, converted to methylmercury, and bioaccumulated along food chains. We examined organic horizon soil samples collected across an elevational gradient on Whiteface Mountain in the Adirondack region of New York State, USA to determine spatial patterns in methylmercury concentrations across a forested montane landscape. We found that soil methylmercury concentrations were highest in the midelevation coniferous zone (0.39 ± 0.07 ng/g) compared to the higher elevation alpine zone (0.28 ± 0.04 ng/g) and particularly the lower elevation deciduous zone (0.17 ± 0.02 ng/g), while the percent of total mercury as methylmercury in soils decreased with elevation. We also found a seasonal pattern in soil methylmercury concentrations, with peak methylmercury values occurring in July. Given elevational patterns in temperature and bioavailable total mercury (derived from mineralization of soil organic matter), soil methylmercury concentrations appear to be driven by soil processing of ionic Hg, as opposed to atmospheric deposition of methylmercury. These methylmercury results are consistent with spatial patterns of mercury concentrations in songbird species observed from other studies, suggesting that future declines in mercury emissions could be important for reducing exposure of mercury to montane avian species.

New York↗

Remote measurement of high preeruptive water vapor emissions at Sabancaya volcano by passive differential optical absorption spectroscopy

Water (H 2 O) is by far the most abundant volcanic volatile species and plays a predominant role in driving volcanic eruptions. However, numerous difficulties associated with making accurate measurements of water vapor in volcanic plumes have limited their use as a diagnostic tool. Here we present the first detection of water vapor in a volcanic plume using passive visible-light differential optical absorption spectroscopy (DOAS). Ultraviolet and visible-light DOAS measurements were made on 21 May 2016 at Sabancaya Volcano, Peru. We find that Sabancaya's plume contained an exceptionally high relative water vapor abundance 6 months prior to its November 2016 eruption. Our measurements yielded average sulfur dioxide (SO 2 ) emission rates of 800–900 t/d, H 2 O emission rates of around 250,000 t/d, and an H 2 O/SO 2 molecular ratio of 1000 which is about an order of magnitude larger than typically found in high-temperature volcanic gases. We attribute the high water vapor emissions to a boiling-off of Sabancaya's hydrothermal system caused by intrusion of magma to shallow depths. This hypothesis is supported by a significant increase in the thermal output of the volcanic edifice detected in infrared satellite imagery leading up to and after our measurements. Though the measurement conditions encountered at Sabancaya were very favorable for our experiment, we show that visible-light DOAS systems could be used to measure water vapor emissions at numerous other high-elevation volcanoes. Such measurements would provide observatories with additional information particularly useful for forecasting eruptions at volcanoes harboring significant hydrothermal systems.

Journal of Geophysical Research B: Solid Earth↗

Partitioning of pyroclasts between ballistic transport and a convective plume: Kīlauea volcano, 19 March 2008

We describe the discrete ballistic and wind-advected products of a small, but exceptionally well-characterized, explosive eruption of wall-rock-derived pyroclasts from Kīlauea volcano on 19 March 2008 and, for the first time, integrate the size distribution of the two subpopulations to reconstruct the true size distribution of a population of pyroclasts as it exited from the vent. Based on thinning and fining relationships, the wind-advected fraction had a mass of 6.1 × 10 5 kg and a thickness half distance of 110 m, placing it at the bottom end of the magnitude and intensity spectra of pyroclastic falls. The ballistic population was mapped, in the field and by using structure-from-motion techniques, to a diameter of > 10–20 cm over an area of ~0.1 km 2 , with an estimated mass of 1 × 10 5 kg. Initial ejection velocities of 50–80 m/s were estimated from inversion of isopleths. The total grain size distribution was estimated by using a mass partitioning of 98% of wind-advected material and 2% of ballistics, resulting in median and sorting values of −1.7 ϕ and 3.1 ϕ . It is markedly broader than those calculated for the products of magmatic explosive eruptions, because the grain size of 19 March 2008 clast population is unrelated to a volcanic fragmentation event and instead was “inherited” from a population of talus clasts that temporary blocked the vent prior to the eruption. Despite a conspicuous near-field presence, the ballistic subpopulation has only a minor influence on the grain size distribution because of its rapid thinning and fining away from source.

Hawaii↗

A 15-year catalog of more than 1 million low-frequency earthquakes: Tracking tremor and slip along the deep San Andreas Fault

Low-frequency earthquakes (LFEs) are small, rapidly recurring slip events that occur on the deep extensions of some major faults. Their collective activation is often observed as a semi-continuous signal known as tectonic (or non-volcanic) tremor. This manuscript presents a catalog of more than 1 million LFEs detected along the central San Andreas Fault from 2001-2016. These events have been detected via a multi-channel matched-filter search, cross-correlating waveform templates representing 88 different LFE families with continuous seismic data. Together, these source locations span nearly 150 km along the central San Andreas Fault, ranging in depth from ~16-30 km. This accumulating catalog has been the source of numerous studies examining the behavior of these LFE sources and the inferred slip behavior of the deep fault. The relatively high temporal and spatial resolution of the catalog has provided new insights into properties such as tremor migration, recurrence, and triggering by static and dynamic stress perturbations. Collectively, these characteristics are inferred to reflect a very weak fault likely under near-lithostatic fluid pressure, yet the physical processes controlling the stuttering rupture observed as tremor and LFE signals remain poorly understood. This paper aims to document the LFE catalog assembly process and associated caveats, while also updating earlier observations and inferred physical constraints. The catalog itself accompanies this manuscript as part of the electronic supplement, with the goal of providing a useful resource for continued future investigations.

California↗

Rock friction under variable normal stress

This study is to determine the detailed response of shear strength and other fault properties to changes in normal stress at room temperature using dry initially bare rock surfaces of granite at normal stresses between 5 and 7 MPa. Rapid normal stress changes result in gradual, approximately exponential changes in shear resistance with fault slip. The characteristic length of the exponential change is similar for both increases and decreases in normal stress. In contrast, changes in fault normal displacement and the amplitude of small high-frequency elastic waves transmitted across the surface follow a two stage response consisting of a large immediate and a smaller gradual response with slip. The characteristic slip distance of the small gradual response is significantly smaller than that of shear resistance. The stability of sliding in response to large step decreases in normal stress is well predicted using the shear resistance slip length observed in step increases. Analysis of the shear resistance and slip-time histories suggest nearly immediate changes in strength occur in response to rapid changes in normal stress; these are manifested as an immediate change in slip speed. These changes in slip speed can be qualitatively accounted for using a rate-independent strength model. Collectively, the observations and model show that acceleration or deceleration in response to normal stress change depends on the size of the change, the frictional characteristics of the fault surface, and the elastic properties of the loading system.

Journal of Geophysical Research B: Solid Earth↗

Connecting crustal seismicity and earthquake-driven stress evolution in Southern California

Tectonic stress in the crust evolves during a seismic cycle, with slow stress accumulation over interseismic periods, episodic stress steps at the time of earthquakes, and transient stress readjustment during a postseismic period that may last months to years. Static stress transfer to surrounding faults has been well documented to alter regional seismicity rates over both short and long time scales. While static stress transfer is instantaneous and long lived, postseismic stress transfer driven by viscoelastic relaxation of the ductile lower crust and mantle leads to additional, slowly varying stress perturbations. Both processes may be tested by comparing a decade-long record of regional seismicity to predicted time-dependent seismicity rates based on a stress evolution model that includes viscoelastic stress transfer. Here we explore crustal stress evolution arising from the seismic cycle in Southern California from 1981 to 2014 using five M ≥6.5 source quakes: the M 7.3 1992 Landers, M6.5 1992 Big Bear, M 6.7 1994 Big Bear, M 7.1 1999 Hector Mine, and M 7.2 2010 El Mayor-Cucapah earthquakes. We relate the stress readjustment in the surrounding crust generated by each quake to regional seismicity using rate-and-state friction theory. Using a log likelihood approach, we quantify the potential to trigger seismicity of both static and viscoelastic stress transfer, finding that both processes have systematically shaped the spatial pattern of Southern California seismicity since 1992.

California↗

Delayed seismicity rate changes controlled by static stress transfer

On 15 June 2010, a M w 5.7 earthquake occurred near Ocotillo, California, in the Yuha Desert. This event was the largest aftershock of the 4 April 2010 M w 7.2 El Mayor-Cucapah (EMC) earthquake in this region. The EMC mainshock and subsequent Ocotillo aftershock provide an opportunity to test the Coulomb failure hypothesis (CFS). We explore the spatiotemporal correlation between seismicity rate changes and regions of positive and negative CFS change imparted by the Ocotillo event. Based on simple CFS calculations we divide the Yuha Desert into three subregions, one triggering zone and two stress shadow zones. We find the nominal triggering zone displays immediate triggering, one stress shadowed region experiences immediate quiescence, and the other nominal stress shadow undergoes an immediate rate increase followed by a delayed shutdown. We quantitatively model the spatiotemporal variation of earthquake rates by combining calculations of CFS change with the rate-state earthquake rate formulation of Dieterich (1994), assuming that each subregion contains a mixture of nucleation sources that experienced a CFS change of differing signs. Our modeling reproduces the observations, including the observed delay in the stress shadow effect in the third region following the Ocotillo aftershock. The delayed shadow effect occurs because of intrinsic differences in the amplitude of the rate response to positive and negative stress changes and the time constants for return to background rates for the two populations. We find that rate-state models of time-dependent earthquake rates are in good agreement with the observed rates and thus explain the complex spatiotemporal patterns of seismicity.

California↗

Dynamic rupture modeling of the M7.2 2010 El Mayor-Cucapah earthquake: Comparison with a geodetic model

The 2010 M w 7.2 El Mayor-Cucapah earthquake is the largest event recorded in the broader Southern California-Baja California region in the last 18 years. Here we try to analyze primary features of this type of event by using dynamic rupture simulations based on a multifault interface and later compare our results with space geodetic models. Our results show that starting from homogeneous prestress conditions, slip heterogeneity can be achieved as a result of variable dip angle along strike and the modulation imposed by step over segments. We also considered effects from a topographic free surface and find that although this does not produce significant first-order effects for this earthquake, even a low topographic dome such as the Cucapah range can affect the rupture front pattern and fault slip rate. Finally, we inverted available interferometric synthetic aperture radar data, using the same geometry as the dynamic rupture model, and retrieved the space geodetic slip distribution that serves to constrain the dynamic rupture models. The one to one comparison of the final fault slip pattern generated with dynamic rupture models and the space geodetic inversion show good agreement. Our results lead us to the following conclusion: in a possible multifault rupture scenario, and if we have first-order geometry constraints, dynamic rupture models can be very efficient in predicting large-scale slip heterogeneities that are important for the correct assessment of seismic hazard and the magnitude of future events. Our work contributes to understanding the complex nature of multifault systems.

California↗

Refining the formation and early evolution of the Eastern North American Margin: New insights from multiscale magnetic anomaly analyses

To investigate the oceanic lithosphere formation and early seafloor spreading history of the North Atlantic Ocean, we examine multiscale magnetic anomaly data from the Jurassic/Early Cretaceous age Eastern North American Margin (ENAM) between 31 and 40°N. We integrate newly acquired sea surface magnetic anomaly and seismic reflection data with publicly available aeromagnetic and composite magnetic anomaly grids, satellite-derived gravity anomaly, and satellite-derived and shipboard bathymetry data. We evaluate these data sets to (1) refine magnetic anomaly correlations throughout the ENAM and assign updated ages and chron numbers to M0–M25 and eight pre-M25 anomalies; (2) identify five correlatable magnetic anomalies between the East Coast Magnetic Anomaly (ECMA) and Blake Spur Magnetic Anomaly (BSMA), which may document the earliest Atlantic seafloor spreading or synrift magmatism; (3) suggest preexisting margin structure and rifting segmentation may have influenced the seafloor spreading regimes in the Atlantic Jurassic Quiet Zone (JQZ); (4) suggest that, if the BSMA source is oceanic crust, the BSMA may be M series magnetic anomaly M42 (~168.5 Ma); (5) examine the along and across margin variation in seafloor spreading rates and spreading center orientations from the BSMA to M25, suggesting asymmetric crustal accretion accommodated the straightening of the ridge from the bend in the ECMA to the more linear M25; and (6) observe anomalously high-amplitude magnetic anomalies near the Hudson Fan, which may be related to a short-lived propagating rift segment that could have helped accommodate the crustal alignment during the early Atlantic opening.

Journal of Geophysical Research B: Solid Earth↗

Constraining the magmatic system at Mount St. Helens (2004–2008) using Bayesian inversion with physics-based models including gas escape and crystallization

Physics-based models of volcanic eruptions track conduit processes as functions of depth and time. When used in inversions, these models permit integration of diverse geological and geophysical data sets to constrain important parameters of magmatic systems. We develop a 1-D steady state conduit model for effusive eruptions including equilibrium crystallization and gas transport through the conduit and compare with the quasi-steady dome growth phase of Mount St. Helens in 2005. Viscosity increase resulting from pressure-dependent crystallization leads to a natural transition from viscous flow to frictional sliding on the conduit margin. Erupted mass flux depends strongly on wall rock and magma permeabilities due to their impact on magma density. Including both lateral and vertical gas transport reveals competing effects that produce nonmonotonic behavior in the mass flux when increasing magma permeability. Using this physics-based model in a Bayesian inversion, we link data sets from Mount St. Helens such as extrusion flux and earthquake depths with petrological data to estimate unknown model parameters, including magma chamber pressure and water content, magma permeability constants, conduit radius, and friction along the conduit walls. Even with this relatively simple model and limited data, we obtain improved constraints on important model parameters. We find that the magma chamber had low (<5wt%) total volatiles and that the magma permeability scale is well constrained at ~10-11.4 m2 to reproduce observed dome rock porosities. Compared with previous results, higher magma overpressure and lower wall friction are required to compensate for increased viscous resistance while keeping extrusion rate at the observed value.

Washington↗

Quantifying the heterogeneity of the tectonic stress field using borehole data

The heterogeneity of the tectonic stress field is a fundamental property which influences earthquake dynamics and subsurface engineering. Self-similar scaling of stress heterogeneities is frequently assumed to explain characteristics of earthquakes such as the magnitude-frequency relation. However, observational evidence for such scaling of the stress field heterogeneity is scarce. We analyze the local stress orientations using image logs of two closely spaced boreholes in the Coso Geothermal Field with sub-vertical and deviated trajectories, respectively, each spanning about 2 km in depth. Both the mean and the standard deviation of stress orientation indicators (borehole breakouts, drilling-induced fractures and petal-centerline fractures) determined from each borehole agree to the limit of the resolution of our method although measurements at specific depths may not. We find that the standard deviation in these boreholes strongly depends on the interval length analyzed, generally increasing up to a wellbore log length of about 600 m and constant for longer intervals. We find the same behavior in global data from the World Stress Map. This suggests that the standard deviation of stress indicators characterizes the heterogeneity of the tectonic stress field rather than the quality of the stress measurement. A large standard deviation of a stress measurement might be an expression of strong crustal heterogeneity rather than of an unreliable stress determination. Robust characterization of stress heterogeneity requires logs that sample stress indicators along a representative sample volume of at least 1 km.

Journal of Geophysical Research B: Solid Earth↗

A decade of induced slip on the causative fault of the 2015 Mw 4.0 Venus earthquake, northeast Johnson County, Texas

On 7 May 2015, a M w 4.0 earthquake occurred near Venus, northeast Johnson County, Texas, in an area of the Bend Arch-Fort Worth Basin that reports long-term, high-volume wastewater disposal and that has hosted felt earthquakes since 2009. In the weeks following the M w 4.0 earthquake, we deployed a local seismic network and purchased nearby active-source seismic reflection data to capture additional events, characterize the causative fault, and explore potential links between ongoing industry activity and seismicity. Hypocenter relocations of the resulting local earthquake catalog span ~4–6 km depth and indicate a fault striking ~230 ° , dipping to the west, consistent with a nodal plane of the M w 4.0 regional moment tensor. Fault plane solutions indicate normal faulting, with B axes striking parallel to maximum horizontal compressive stress. Seismic reflection data image the reactivated basement fault penetrating the Ordovician disposal layer and Mississippian production layer, but not displacing post-Lower Pennsylvanian units. Template matching at regional seismic stations indicates that low-magnitude earthquakes with similar waveforms began in April 2008, with increasing magnitude over time. Pressure data from five saltwater disposal wells within 5 km of the active fault indicate a disposal formation that is 0.9–4.8 MPa above hydrostatic. We suggest that the injection of 28,000,000 m 3 of wastewater between 2006 and 2015 at these wells led to an increase in subsurface pore fluid pressure that contributed to inducing this long-lived earthquake sequence. The 2015 M w 4.0 event represents the largest event in the continuing evolution of slip on the causative fault.

Texas↗

Magnetotelluric imaging of lower crustal melt and lithospheric hydration in the Rocky Mountain Front transition zone, Colorado, USA

We present an electrical resistivity model of the crust and upper mantle from two‐dimensional (2‐D) anisotropic inversion of magnetotelluric data collected along a 450 km transect of the Rio Grande rift, southern Rocky Mountains, and High Plains in Colorado, USA. Our model provides a window into the modern‐day lithosphere beneath the Rocky Mountain Front to depths in excess of 150 km. Two key features of the 2‐D resistivity model are (1) a broad zone (~200 km wide) of enhanced electrical conductivity (<20 Ωm) in the midcrust to lower crust that is centered beneath the highest elevations of the southern Rocky Mountains and (2) hydrated lithospheric mantle beneath the Great Plains with water content in excess of 100 ppm. We interpret the high conductivity region of the lower crust as a zone of partially molten basalt and associated deep‐crustal fluids that is the result of recent (less than 10 Ma) tectonic activity in the region. The recent supply of volatiles and/or heat to the base of the crust in the late Cenozoic implies that modern‐day tectonic activity in the western United States extends to at least the western margin of the Great Plains. The transition from conductive to resistive upper mantle is caused by a gradient in lithospheric modification, likely including hydration of nominally anhydrous minerals, with maximum hydration occurring beneath the Rocky Mountain Front. This lithospheric “hydration front” has implications for the tectonic evolution of the continental interior and the mechanisms by which water infiltrates the lithosphere.

Colorado↗

Decaying lava extrusion rate at El Reventador Volcano, Ecuador measured using high-resolution satellite radar

Lava extrusion at erupting volcanoes causes rapid changes in topography and morphology on the order of tens or even hundreds of meters. Satellite radar provides a method for measuring changes in topographic height over a given time period to an accuracy of meters, either by measuring the width of radar shadow cast by steep sided features, or by measuring the difference in radar phase between two sensors separated in space. We measure height changes, and hence estimate extruded lava volume flux, at El Reventador, Ecuador, between 2011 and 2016, using data from the RADARSAT-2 and TanDEM-X satellite missions. We find that 39 new lava flows were extruded between 9 February 2012 and 24 August 2016, with a cumulative volume of 44.8M m 3 dense rock equivalent and a gradually decreasing eruption rate. The average dense rock rate of lava extrusion during this time is 0.31 ± 0.02 m 3 s −1 , which is similar to the long-term average from 1972 to 2016. Apart from a volumetrically small dyke opening event between 9 March and 10 June 2012, lava extrusion at El Reventador is not accompanied by any significant magmatic ground deformation. We use a simple physics-based model to estimate that the volume of the magma reservoir under El Reventador is greater than 3 km 3 . Our lava extrusion data can be equally well fit by models representing a closed reservoir depressurising during the eruption with no magma recharge, or an open reservoir with a time-constant magma recharge rate of up to 0.35 ± 0.01 m 3 s −1 .

El Reventador Volcano↗

Temporal stress changes caused by earthquakes: A review

Earthquakes can change the stress field in the Earth’s lithosphere as they relieve and redistribute stress. Earthquake-induced stress changes have been observed as temporal rotations of the principal stress axes following major earthquakes in a variety of tectonic settings. The stress changes due to the 2011 Mw9.0 Tohoku-Oki, Japan, earthquake were particularly well documented. Earthquake stress rotations can inform our understanding of earthquake physics, most notably addressing the long-standing problem of whether the Earth’s crust at plate boundaries is “strong” or “weak.” Many of the observed stress rotations, including that due to the Tohoku-Oki earthquake, indicate near-complete stress drop in the mainshock. This implies low background differential stress, on the order of earthquake stress drop, supporting the weak crust model. Earthquake stress rotations can also be used to address other important geophysical questions, such as the level of crustal stress heterogeneity and the mechanisms of postseismic stress reloading. The quantitative interpretation of stress rotations is evolving from those based on simple analytical methods to those based on more sophisticated numerical modeling that can capture the spatial-temporal complexity of the earthquake stress changes.

Journal of Geophysical Research B: Solid Earth↗

Source spectral properties of small-to-moderate earthquakes in southern Kansas

The source spectral properties of injection-induced earthquakes give insight into their nucleation, rupture processes, and influence on ground motion. Here we apply a spectral decomposition approach to analyze P -wave spectra and estimate Brune-type stress drop for more than 2000 M L 1.5–5.2 earthquakes occurring in southern Kansas from 2014 to 2016. We find that these earthquakes are characterized by low stress drop values (median ∼0.4MPa) compared to natural seismicity in California. We observe a significant increase in stress drop as a function of depth, but the shallow depth distribution of these events is not by itself sufficient to explain their lower stress drop. Stress drop increases with magnitude from M1.5–M3.5, but this scaling trend may weaken above M4 and also depends on the assumed source model. Although we observe a nonstationary, sequence-specific temporal evolution in stress drop, we find no clear systematic relation with the activity of nearby injection wells.

Kansas↗

Deformation of the Pacific/North America plate boundary at Queen Charlotte Fault: The possible role of rheology

The Pacific/North America (PA/NA) plate boundary between Vancouver Island and Alaska is similar to the PA/NA boundary in California in its kinematic history and the rate and azimuth of current relative motion, yet their deformation styles are distinct. The California plate boundary shows a broad zone of parallel strike slip and thrust faults and folds, whereas the 49‐mm/yr PA/NA relative plate motion in Canada and Alaska is centered on a single, narrow, continuous ~900‐km‐long fault, the Queen Charlotte Fault (QCF). Using gravity analysis, we propose that this plate boundary is centered on the continent/ocean boundary (COB), an unusual location for continental transform faults because plate boundaries typically localize within the continental lithosphere, which is weaker. Because the COB is a boundary between materials of contrasting elastic properties, once a fault is established there, it will probably remain stable. We propose that deformation progressively shifted to the COB in the wake of Yakutat terrane's northward motion along the margin. Minor convergence across the plate boundary is probably accommodated by fault reactivation on Pacific crust and by an eastward dipping QCF. Underthrusting of Pacific slab under Haida Gwaii occurs at convergence angles >14°–15° and may have been responsible for the emergence of the archipelago. The calculated slab entry dip (5°–8°) suggests that the slab probably does not extend into the asthenosphere. The PA/NA plate boundary at the QCF can serve as a structurally simple site to investigate the impact of rheology and composition on crustal deformation and the initiation of slab underthrusting.

Journal of Geophysical Research B: Solid Earth↗