Structure Of Upper Mantle Research Articles

Correction to: Updates to the Regional Seismic Travel Time (RSTT) Model: 2. Path-dependent Travel-time Uncertainty

The regional seismic travel time (RSTT) model and software were developed to improve travel-time prediction accuracy by accounting for three-dimensional crust and upper mantle structure. Travel-time uncertainty estimates are used in the process of associating seismic phases to events and to accurately calculate location uncertainty bounds (i.e. event location error ellipses). We improve on the current distance-dependent uncertainty parameterization for RSTT using a random effects model to estimate slowness (inverse velocity) uncertainty as a mean squared error for each model parameter. The random effects model separates the error between observed slowness and model predicted slowness into bias and random components. The path-specific travel-time uncertainty is calculated by integrating these mean squared errors along a seismic-phase ray path. We demonstrate that event location error ellipses computed for a 90% coverage ellipse metric (used by the Comprehensive Nuclear-Test-Ban Treaty Organization International Data Centre (IDC)), and using the path-specific travel-time uncertainty approach, are more representative (median 82.5% ellipse percentage) of true location error than error ellipses computed using distance-dependent travel-time uncertainties (median 70.1%). We also demonstrate measurable improvement in location uncertainties using the RSTT method compared to the current station correction approach used at the IDC (median 74.3% coverage ellipse).

Shear‐Wave Velocity Structure of the Southern African Upper Mantle: Implications for Craton Structure and Plateau Uplift

AbstractWe present a 3D shear‐wave velocity model of the southern African upper mantle developed using 30–200 s period Rayleigh waves recorded on regional seismic networks spanning the subcontinent. The model shows high velocities (∼4.7–4.8 km/s) at depths of 50–250 km beneath the Archean nucleus and several surrounding Paleoproterozoic and Mesoproterozoic terranes, placing the margin of the greater Kalahari Craton along the southern boundary of the Damara Belt and the eastern boundaries of the Gariep and Namaqua‐Natal belts. At depths ≥250 km, there is little difference in velocities beneath the craton and off‐craton regions, suggesting that the cratonic lithosphere extends to depths of about 200–250 km. Upper mantle velocities beneath uplifted areas of southern Africa are higher than the global average and significantly higher than beneath eastern Africa, indicating there that is little thermal modification of the upper mantle present today beneath the Southern African Plateau.

Constraints on the Depth, Thickness, and Strength of the G Discontinuity in the Central Pacific From S Receiver Functions

AbstractThe relative motion of the lithosphere with respect to the asthenosphere implies the existence of a boundary zone that accommodates shear between the rigid plates and flowing mantle. This shear zone is typically referred to as the lithosphere‐asthenosphere boundary (LAB). The width of this zone and the mechanisms accommodating shear across it have important implications for coupling between mantle convection and surface plate motion. Seismic observations have provided evidence for several physical mechanisms that might help enable relative plate motion, but how these mechanisms each contribute to the overall accommodation of shear remains unclear. Here we present receiver function constraints on the discontinuity structure of the oceanic upper mantle at the NoMelt site in the central Pacific, where local constraints on shear velocity, anisotropy, conductivity, and attenuation down to ∼300 km depth provide a comprehensive picture of upper mantle structure. We image a seismic discontinuity with a Vsv decrease of 4.5% or more over a 0–20 km thick gradient layer centered at a depth of ∼65 km. We associate this feature with the Gutenberg discontinuity (G), and interpret our observation of G as resulting from strain localization across a dehydration boundary based on the good agreement between the discontinuity depth and that of the dry solidus. Transitions in Vsv, azimuthal anisotropy, conductivity, and attenuation observed at roughly similar depths suggest that the G discontinuity represents a region of localized strain within a broader zone accommodating shear between the lithosphere and asthenosphere.

Constraints on the Upper Mantle Structure Beneath the Pacific From 3‐D Anisotropic Waveform Modeling

AbstractSeismic radial anisotropy is a crucial tool to help constrain flow in the Earth's mantle. However, Earth structure beneath the oceans imaged by current 3‐D radially anisotropic mantle models shows large discrepancies. Here, we provide constraints on the radially anisotropic upper mantle structure beneath the Pacific by waveform modeling and subsequent inversion. Specifically, we objectively evaluate three 3‐D tomography mantle models which exhibit varying distributions of radial anisotropy through comparisons of independent real data sets with synthetic seismograms computed with the spectral‐element method. The data require an asymmetry at the East Pacific Rise (EPR) with stronger positive radial anisotropy ξ = = 1.13–1.16 at ∼100 km depth to the west of the EPR than to the east (ξ = 1.11–1.13). This suggests that the anisotropy in this region is due to the lattice‐preferred orientation of anisotropic mantle minerals produced by shear‐driven asthenospheric flow beneath the South Pacific Superswell. Our new radial anisotropy constraints in the Pacific show three distinct positive linear anomalies at ∼100 km depth. These anomalies are possibly related to mantle entrainment at the Nazca‐South America subduction zone, flow at the EPR and from the South Pacific Superswell and shape‐preferred orientation (SPO) of melt beneath Hawaii. Radial anisotropy reduces with lithospheric age to ξ < 1.05 in the west at ∼100 km depth, which possibly reflects a deviation from horizontal flow as the mantle is entrained with subducting slabs, a change in temperature or water content that could alter the anisotropic olivine fabric or the SPO of melt.

Lateral Seismic Anisotropy Variations Record Interaction Between Tibetan Mantle Flow and Plume‐Strengthened Yangtze Craton

AbstractThe southeastern Tibetan Plateau experienced intense Cenozoic shortening and extrusion along the western Yangtze Craton, SW China, where the Permian Emeishan flood basalts were emplaced. Seismic anisotropy along a dense E–W‐trending linear seismic array across the west Yunnan is determined by shear‐wave splitting analysis to characterize the deformation associated with these processes. Interestingly, a strong N–S‐oriented anisotropic zone collocated with the Eocene–Oligocene potassic magmatism is observed between two weak anisotropic zones in the Three‐River Orogenic Zone and the western Yangtze Craton, respectively. The lateral anisotropy variation and distinct upper mantle structures collectively suggest that the western most Yangtze Craton, strengthened by the Permian mantle plume, acted as a rigid barrier to divert the Cenozoic Tibetan mantle flow into the Three‐River Orogenic Zone. Consequently, the western edge of the Yangtze Craton has been partially eroded and its anisotropic orientation was realigned to the direction of the mantle flow. This observation indicates that both the remodeled anisotropy in the lithosphere and the dynamic flow in the asthenosphere contribute to the mantle anisotropy at the lateral lithosphere‐asthenosphere transition, which is usually hard to seismologically resolve from the vertical lithosphere‐asthenosphere boundary.

Seismicity in the central and southeastern United States due to upper mantle heterogeneities

SUMMARYSources of stress responsible for earthquakes occurring in the Central and Eastern United States (CEUS) include not only far-field plate boundary forces but also various local contributions. In this study, we model stress fields due to heterogeneities in the upper mantle beneath the CEUS including a high-velocity feature identified as a lithospheric drip in a recent regional P-wave tomography study. We calculate velocity and stress distributions from numerical models for instantaneous 3-D mantle flow. Our models are driven by the heterogeneous density distribution based on a temperature field converted from the tomography study. The temperature field is utilized in a composite rheology, assumed for the numerical models. We compute several geodynamic quantities with our numerical models: dynamic topography, rate of dynamic topography, gravitational potential energy (GPE), differential stress, and Coulomb stress. We find that the GPE, representative of the density anomalies in the lithosphere, is an important factor for understanding the seismicity of the CEUS. When only the upper mantle heterogeneities are included in a model, differential and Coulomb stress for the observed fault geometries in the CEUS seismic zones acts as a good indicator to predict the seismicity distribution. Our modelling results suggest that the upper mantle heterogeneities and structure below the CEUS have stress concentration effects and are likely to promote earthquake generation at preexisting faults in the region’s seismic zones. Our results imply that the mantle flow due to the upper-mantle heterogeneities can cause stress perturbations, which could help explain the intraplate seismicity in this region.

Seismic Structure, Gravity Anomalies and Flexure Along the Emperor Seamount Chain

AbstractThe Hawaiian‐Emperor seamount chain in the Pacific Ocean has provided fundamental insights into hotspot generated intraplate volcanism and the long‐term strength of oceanic lithosphere. However, only a few seismic experiments to determine crustal and upper mantle structure have been carried out on the Hawaiian Ridge, and no deep imaging has ever been carried out along the Emperor seamounts. Here, we present the results of an active source seismic experiment using 29 Ocean‐Bottom Seismometers (OBS) carried out along a strike profile of the seamounts in the region of Jimmu and Suiko guyots. Joint reflection and refraction tomographic inversion of the OBS data show the upper crust is highly heterogeneous withPwave velocities <4–5 km s−1, which are attributed to extrusive lavas and clastics. In contrast, the lower crust is remarkably homogeneous with velocities of 6.5–7.2 km s−1, which we attribute to oceanic crust and mafic intrusions. Moho is identified by a strongPmParrival at offsets of 20–80 km, yielding depths of 13–16 km. The underlying mantle is generally homogeneous with velocities in the range 7.9–8.0 km s−1. The crust and mantle velocity structure has been verified by gravity modeling. While top of oceanic crust prior to volcano loading is not recognized as a seismic or gravity discontinuity, flexural modeling reveals a ∼5.0–5.5 km thick preexisting oceanic crust that is overlain by a ∼8 km thick volcanic edifice. Unlike at the Hawaiian Ridge, we find no evidence of magmatic underplating.

Crust and Upper Mantle Structure of the South China Sea and Adjacent Areas From the Joint Inversion of Ambient Noise and Earthquake Surface Wave Dispersions

AbstractIn this study, we built a high‐resolution 3D S‐wave velocity model of the South China Sea (SCS) and adjacent areas by the joint inversion of surface wave dispersions retrieved from ambient noise and earthquake data. We first measured Rayleigh wave phase velocity dispersions from the waveforms recorded by 94 broadband seismic stations. Subsequently, we used a generalized least squares inversion scheme with spatially varying resolution to perform 2D tomographic inversion at different periods. Lastly, we performed 1D inversion at different grid points and obtained the 3D S‐wave velocity model in the depth range of 15–250 km. Our model reveals a prominent high velocity (high‐V) anomaly beneath the SCS basin and shows that the lithosphere beneath the SCS basin is about 70–80 km thick. We observe a prominent low velocity (low‐V) anomaly in the asthenosphere beneath the SCS basin, and we interpret this as evidence for the ongoing partial melting of the asthenosphere. An obvious high‐V anomaly is also revealed beneath the Khorat Plateau (KP) in the depth range of 80–150 km, suggesting that the KP has a cold and thick lithosphere and has suffered limited destruction in the Cenozoic era. In the northern part of Kalimantan Island (KI), a prominent low‐V anomaly is shown in the upper 60 km of the lithosphere revealing the upwelling of the asthenosphere. Our findings shed lights on the tectonic evolution of the SCS and adjacent areas.

S‐Velocity Mantle Structure of East Asia From Teleseismic Traveltime Tomography: Inferred Mechanisms for the Cenozoic Intraplate Volcanoes

AbstractClusters of Cenozoic volcanoes are found in Korea and Northeast China away from plate boundaries, but the mechanisms responsible for their formation is still unknown. To resolve this issue, we derive an S‐wave velocity model of the upper mantle structure beneath East Asia using 131,947 S‐wave relative traveltimes from 974 earthquakes recorded at 1,948 seismic stations in China, Korea, Taiwan, and Japan. Our results show that low‐velocity anomalies beneath Mt. Baekdu (Changbaishan) and Jeju Island extend to the bottom of the mantle transition zone (MTZ) through slab gaps in the stagnant Pacific slab. On the other hand, low‐velocity anomalies beneath the Datong volcanic field are observed to connect to the edge of the stagnant Pacific slab, which may be attributed to subduction‐triggered upwelling or deep dehydration from the hydrous MTZ perturbed by subducting slabs. Furthermore, tilted high‐velocity anomalies are found all the way down to the MTZ beneath the Great Xing'an Range and the Songliao basin, which may indicate delaminated lithosphere. The intraplate volcanoes north of the Songliao basin, such as the Wudalianchi and Halaha volcanoes, are located just above the delaminated bodies, thereby implying that decompression melting after delamination may be responsible for the intraplate volcanoes. Therefore, the dominant origin of intraplate volcanoes in East Asia may be subduction‐related hot upwelling, complemented by shallow origins such as edge‐driven convection and lithospheric delamination.

Mantle Dynamics Beneath the Sichuan Basin and Eastern Tibet From Teleseismic Tomography

AbstractThe cratonic Sichuan Basin is located east of the Tibetan Plateau, and is surrounded by mountains that have undergone complex deformation and uplift since the Cenozoic. Imaging mantle structure is important for understanding its formation, and to date most models suggest a deep cratonic root underlies the basin, blocking eastward extrusion of lithospheric material beneath Tibet. Here, we obtain detailed upper mantle structure from teleseismic tomography in the region utilizing travel time data from earthquakes recorded at 506 seismic stations, including 25 new stations in the poorly sampled Sichuan Basin. Contrasting to previous models, we show eastward and southeastward dipping high‐velocity anomalies extending eastward ∼150–400 km into the upper mantle from the Sichuan Basin. We suggest, the southeastward subduction of the Yangtze Block occurred in the Mesozoic and may be reactivated in the Cenozoic, with the relatively thin and weak lithosphere to the east of the Sichuan Basin prone to deformation in response to the eastward growth of the Tibetan Plateau. A west‐dipping high‐velocity anomaly beneath eastern Tibet is interpreted as delaminated lithosphere. This delamination may accelerate the development of the Xianshuihe fault zone and the horizontal extrusion of the Tibetan Plateau. Beneath the East Qinling orogen, the eastward extrusion of the plateau material is not obvious suggesting limited horizontal lithospheric extrusion is present north of the Sichuan Basin.

Updates to the Regional Seismic Travel Time (RSTT) Model: 2. Path-dependent Travel-time Uncertainty

The regional seismic travel time (RSTT) model and software were developed to improve travel-time prediction accuracy by accounting for three-dimensional crust and upper mantle structure. Travel-time uncertainty estimates are used in the process of associating seismic phases to events and to accurately calculate location uncertainty bounds (i.e. event location error ellipses). We improve on the current distance-dependent uncertainty parameterization for RSTT using a random effects model to estimate slowness (inverse velocity) uncertainty as a mean squared error for each model parameter. The random effects model separates the error between observed slowness and model predicted slowness into bias and random components. The path-specific travel-time uncertainty is calculated by integrating these mean squared errors along a seismic-phase ray path. We demonstrate that event location error ellipses computed for a 90% coverage ellipse metric (used by the Comprehensive Nuclear-Test-Ban Treaty Organization International Data Centre (IDC)), and using the path-specific travel-time uncertainty approach, are more representative (median 82.5% ellipse percentage) of true location error than error ellipses computed using distance-dependent travel-time uncertainties (median 70.1%). We also demonstrate measurable improvement in location uncertainties using the RSTT method compared to the current station correction approach used at the IDC (median 74.3% coverage ellipse).

Structure of the crust and upper mantle beneath the Bransfield Strait (Antarctica) using P receiver functions

The Bransfield Strait is a tectonically active back-arc rift basin located between the South Shetland archipelago (SSI) and the Antarctic Peninsula (AP). We use teleseismic P-wave receiver functions from twelve seismic stations in order to improve the current understanding of the crustal and upper mantle structure of the region. We calculate the thickness and average Vp/Vs ratio of the crust, the thickness of the Mantle Transition Zone (MTZ) and their spatial variability along the study area. Our analysis reveals a homogeneous crustal thickness in the AP, averaging 33.5 ± 1.2 km. The VP/VS ratio increases from east to west, with an average of 1.78 ± 0.04 typical of continental crust. In the SSI results are highly heterogeneous. The minimum crustal thickness is observed in Deception Island (~15 km), near the rift axis, while the crust thickens towards the edges of the basin to ~30 km. High crustal VP/VS ratios (>1.95) in the archipelago correlate with a shallow Moho and may be linked to the presence of melts in the crust, in accordance with the active volcanism in the region. We detected an underplating layer and the Moho of the Phoenix slab beneath the SSI. There is evidence of a low velocity zone (LVZ) in the mantle wedge under the SSI and also under the AP, suggesting there is widespread partial melting of the upper mantle. We infer that the Phoenix slab dips steeply, as we observe a thickened MTZ under the AP.

ГЕОЭЛЕКТРИЧЕСКИЙ РАЗРЕЗ ЛИТОСФЕРЫ ХАНКАЙСКОГО МАССИВА ПО ПРОФИЛЮ С. ПОЖАРСКОЕ–П. ШКОТОВО ПО ДАННЫМ МАГНИТОТЕЛЛУРИЧЕСКИХ ЗОНДИРОВАНИЙ

The geoelectric section of the lithosphere is constructed down to a depth of 150 km based on the results of MT soundings carried out along the profile crossing the Khanka Massif from north to south. The crust and upper mantle structure of the massif is investigated. It has been found that the Khanka Massif is divided into the northern and southern parts bordered by the northern restriction of the Spassk Zone. A higher resistivity observed in the upper part of the section and the upper mantle is typical of the southern part of the massif. The northern part is characterized by two extended low-resistivity areas located beneath the southern boundaries of the Matveevka and Nakhimovka zones that penetrate from the upper mantle up to the surface. A comparison of the results obtained with the seismic prospecting and gravimetric data has shown their direct or indirect consistency.

Crustal and upper-mantle structure of the Eastern Caribbean and Northern Venezuela from passive Rayleigh wave tomography

We explore the shear-wave lithospheric velocity structure of the Eastern Caribbean and Northern Venezuela using ambient noise tomography with stations deployed around the study area. We construct cross-correlation functions from continuous seismic records, and measure phase velocities of fundamental-mode Rayleigh waves. These velocities are further projected onto 0.6°x0.6° phase velocity grids for each period between 5 s and 50 s. The pseudo-dispersion curve at each grid point is inverted for 1D shear velocity profiles by using a Markov Chain Monte Carlo scheme. The interpolated 3D velocity model shows that the mean shear velocity of the Eastern Caribbean lithospheric mantle is lower than the global average, which is in agreement with values reported in other large igneous provinces. We interpret that low velocities in the lithospheric keel are associated with an anomalous composition and/or an elevated thermal state; this gives the Caribbean plate a high buoyancy that determines the subduction polarities in the region. The results also indicate that: (a) the mantle beneath Northern Venezuela retains compositional anomalies related to extension processes of different ages; (b) the overriding of the Caribbean plate by the Great Antilles arc seems to be much slower than previously suggested; and (c) the localized volcanism in the center of the Lesser Antilles arc is related to asthenospheric flow through the tear induced on the subducted slab by major strike-slip faults.

Seismic images of crust and upper mantle in Central Iran (from Jiroft to Ashtian) via teleseismic waves

A detailed knowledge of the thickness of crust and upper mantle structure is important for understanding a plate tectonics and geodynamics in the region. We use body wave for detecting details of the subsurface structure. The information in this research is collected from a seismic linear profile that extends across the Sanandaj-Sirjan metamorphic zone in seismic states of Central Iran and Zagros. We compute P receiver functions to investigate crustal and upper mantle discontinuities. We use teleseismic events (mb ≥ 5.5, 30° < Δ < 95°) registered between 1996 and 2018 and recorded at 10 short-period stations with 3 components and 17 broadband stations with high signal to noise ratio. The observed depth of Moho in the study area is approximately 50 km and rises to 70 km at the end of the seismic linear profile beneath Sanandaj-Sirjan zone. In Central Iran, depths discontinuities in the transition zone are shown by the reference model of deviation, which can be attributed to the convergence of Arabian plate with the Central Iran plateau. Also, the study area was identified as geothermal susceptibility by SUNA and this observation was confirmed.

A Recent Tear in Subducting Plate Explains Seismicity and Upper Mantle Structure of Southern Costa Rica

AbstractPast studies of southern Costa Rica have generated a multitude of tectonic scenarios to account for different data sets. Flat slabs, detached slabs, and slab windows have been proposed to address the uplift of the Cordillera de Talamanca (CT), cessation of volcanism, and absence of deep seismicity beneath southern Costa Rica. In this study, we investigate the crust and the upper mantle along the southwest flank of the CT using the receiver function methodology. We observe two regional positive P‐to‐S converted pulses at delay times of ∼2–4 s and ∼5–8 s. The first likely represents a gradational crust‐mantle boundary of the upper plate. The second represents a similar impedance increase ∼50–60 km deep that extends from central Costa Rica to Panama. Compared to well‐located seismicity, this boundary is offset to the NE from the Cocos plate Benioff zone beneath northern CT, and remains observable through a gap in seismicity farther to the southeast. This offset makes it difficult to interpret this feature as related to the presently subducting lithosphere. Instead, we propose that the 50–60 km deep boundary marks the Moho of a lithospheric fragment left behind under the CT in the course of Panama Triple Junction migration through Costa Rica over the last 10 Ma. Our interpretation accounts for the geophysical, geochemical, structural, and geomorphic observations in the literature explaining the complex geodynamic scenario observed in southern Costa Rica.

Updates to the Regional Seismic Travel Time (RSTT) Model: 1. Tomography

A function of global monitoring of nuclear explosions is the development of Earth models for predicting seismic travel times for more accurate calculation of event locations. Most monitoring agencies rely on fast, distance-dependent one-dimensional (1D) Earth models to calculate seismic event locations quickly and in near real-time. RSTT (Regional Seismic Travel Time) is a seismic velocity model and computer software package that captures the major effects of three-dimensional crust and upper mantle structure on regional seismic travel times, while still allowing for fast prediction speed (milliseconds). We describe updates to the RSTT model using a refined data set of regional phases (i.e., Pn, Pg, Sn, Lg) using the Bayesloc relative relocation algorithm. The tomographic inversion shown here acts to refine the previous RSTT public model (rstt201404um) and displays significant features related to areas of global tectonic complexity as well as further reduction in arrival residual values. Validation of the updated RSTT model demonstrates significant reduction in median epicenter mislocation (15.3 km) using all regional phases compared to the iasp91 1D model (22.1 km) as well as to the current station correction approach used at the Comprehensive Nuclear-Test-Ban Treaty Organization International Data Centre (18.9 km).

The Crust and Upper Mantle Structure of the Southern Sikhote Alin from Magnetotelluric Sounding and Density Modeling

Based on the results of correlation of the geoelectric and density sections constructed using three-dimensional inversion of three profiles located in the southern Sikhote Alin, the distribution regularities were revealed in the electric and density parameters, the relationships between them were established, and new data on the crust and upper mantle structure of the region were obtained. A relationship between certain density and electric inhomogeneities within the upper mantle and the Late Cretaceous subduction processes was proposed.

Lithospheric Boundaries and Upper Mantle Structure Beneath Southern Africa Imaged by P and S Wave Velocity Models

AbstractWe report new P and S wave velocity models of the upper mantle beneath southern Africa using data recorded on seismic stations spanning the entire subcontinent. Beneath most of the Damara Belt, including the Okavango Rift, our models show lower than average velocities (−0.8% Vp; −1.2% Vs) with an abrupt increase in velocities along the terrane's southern margin. We attribute the lower than average velocities to thinner lithosphere (~130 km thick) compared to thicker lithosphere (~200 km thick) immediately to the south under the Kalahari Craton. Beneath the Etendeka Flood Basalt Province, higher than average velocities (0.25% Vp; 0.75% Vs) indicate thicker and/or compositionally distinct lithosphere compared to other parts of the Damara Belt. In the Rehoboth Province, higher than average velocities (0.3% Vp; 0.5% Vs) suggest the presence of a microcraton, as do higher than average velocities (1.0% Vp; 1.5% Vs) under the Southern Irumide Belt. Lower than average velocities (−0.4% Vp; −0.7% Vs) beneath the Bushveld Complex and parts of the Mgondi and Okwa terranes are consistent with previous studies, which attributed them to compositionally modified lithosphere resulting from Precambrian magmatic events. There is little evidence for thermally modified upper mantle beneath any of these terranes which could provide a source of uplift for the Southern African Plateau. In contrast, beneath parts of the Irumide Belt in southern and central Zambia and the Mozambique Belt in central Mozambique, deep‐seated low velocity anomalies (−0.7% Vp; −0.8% Vs) can be attributed to upper mantle extensions of the African superplume structure.

Evolutionary full-waveform inversion

SUMMARY We present a new approach to full-waveform inversion (FWI) that enables the assimilation of data sets that expand over time without the need to reinvert all data. This evolutionary inversion rests on a reinterpretation of stochastic Limited-memory Broyden–Fletcher–Goldfarb–Shanno (L-BFGS), which randomly exploits redundancies to achieve convergence without ever considering the data set as a whole. Specifically for seismological applications, we consider a dynamic mini-batch stochastic L-BFGS, where the size of mini-batches adapts to the number of sources needed to approximate the complete gradient. As an illustration we present an evolutionary FWI for upper-mantle structure beneath Africa. Starting from a 1-D model and data recorded until 1995, we sequentially add contemporary data into an ongoing inversion, showing how (i) new events can be added without compromising convergence, (ii) a consistent measure of misfit can be maintained and (iii) the model evolves over times as a function of data coverage. Though applied retrospectively in this example, our method constitutes a possible approach to the continuous assimilation of seismic data volumes that often tend to grow exponentially.