Search

The changes in physical properties leading up to a volcanic eruption provide clues to processes occurring within volcanoes and may reveal premonitory signals. The Sierra Negra shield volcano, located in the Galápagos Islands, erupted on 26 June 2018 after months of continued inflation and escalating earthquake activity. We applied ambient noise interferometry to five months of continuous seismic data from 12 broadband stations to calculate crustal shear‐wave velocity changes before and during the eruption. Using the Moving Window Cross‐Spectral technique and a pre‐eruption stack of ambient seismic data as reference, we found a −0.27% decrease in velocity 17 days before the eruption in station‐pairs that pass beneath the caldera's north‐eastern sector. Sensitivity kernels suggest that the velocity changes of this precursory signal are greatest at depths corresponding to the shallow sill (∼2 km) beneath the wide caldera. Our results, considered in light of geodetic, seismicity, and petrological results, suggest that this velocity decrease is in part caused by dilatation from a ML 4.8 earthquake, and degassing after a possible magma intrusion below or at the base of the sill. The precursory velocity decrease within the sill region offers an important tool for forecasting future eruptions at Sierra Negra. Plain Language Summary: Geoscientists use seismic methods to detect changes in the characteristics of volcanoes that could indicate increasing risk of eruption. The Sierra Negra volcano, located in the Galápagos Islands, erupted on 26 June 2018. Using time variations in seismic signals mainly generated by the ocean's interaction with the shoreline, we found a decrease in the velocity of seismic waves about 2 weeks before the eruption. The decrease is largest when these waves travel through Sierra Negra's shallow magma reservoir. This velocity reduction started on June 9th, the same day as the second‐largest pre‐eruptive earthquake recorded since 2017, and ended one day before the eruption, on June 25th. A slight increase in velocity was observed on May 13th and lasted 4 days. We interpret this increase as a magma pulse that destabilizes the volcanic system. The combined effect of this magma intrusion and the magnitude 4.8 earthquake altered the volcano's state and favored the start of the eruption. The technique that we used could forecast future eruptions at Sierra Negra and other basaltic island volcanoes. Key Points: A velocity variations analysis provides important evidence of pre‐eruptive physical processes inside the volcanic edificeUsing ambient noise interferometry we obtained a precursor signal for the 2018 Sierra Negra eruption, which offers a forecasting toolThe velocity anomaly is spatially constrained to station‐pairs near the initial eruptive fissures, and is caused by extensional stress [ABSTRACT FROM AUTHOR]

The North Basin of the Malawi Rift is an active, early‐stage rift segment that provides the opportunity to quantify cumulative and recent faulting patterns in a young rift, assess contributions of intrarift faults to accommodating rift opening, and examine controls on spatial patterns of faulting. Multichannel seismic reflection data acquired in Lake Malawi (Nyasa) in 2015 together with legacy multichannel seismic data image a system of synthetic intrarift faults within this border‐fault‐bounded, half‐graben basin. A dense wide‐angle seismic reflection/refraction dip profile acquired with lake bottom seismometer data constrains sediment velocities that are used to convert fault throws from travel time to depth. Observed extension on intrarift faulting in the northern and central parts of the North Basin is approximately twice what would be predicted for hanging wall flexure, implying that the intrarift faults contribute to basin opening. The cumulative throw on intrarift faults is higher in the northern part of the rift segment than the south and is anticorrelated with throw on the border fault, which is largest in the southern part of the North Basin. This change in faulting coincides with a change in the orientation of the North Basin from a N‐S trend in the south to a NNW‐SSE trend in the north. We infer that the distribution of extension is influenced by rift orientation with respect to the regional extension direction. Almost all intrarift faults substantially offset late Quaternary synrift sediments, suggesting they are likely active and need to be considered in hazard assessments. Key Points: Intrarift faulting contributes to rift openingThe style and amount of rift faulting change along strike in North BasinNearly all intrarift faults have been active in the last ~75 kyr [ABSTRACT FROM AUTHOR]

The crustal structure surrounding the northern Malawi rift and beneath the Rungwe Volcanic Province (RVP) has been investigated using teleseismic earthquakes recorded on SEGMeNT broad-band seismic stations to determine the extent to which the crust has been modified by Cenozoic rifting and magmatism. The SEGMeNT network included 57 broad-band seismic stations deployed in northern Malawi and southern Tanzania between August 2013 and October 2015. Estimates of crustal thickness, shear wave velocity and Poisson's ratio have been obtained by modelling P -wave receiver functions using the H – k stacking method and jointly inverting receiver functions with Rayleigh wave phase velocities. These estimates are used to investigate the extent of magmatic modification to the crust, indicated by changes in Poisson's ratio, and the geometry of crustal thinning along the northern margins of the Malawi rift and beneath the RVP. The average crustal thickness for the four stations in the RVP is 39 km, the average Poisson's ratio is 0.28 (Vp / Vs  = 1.83), and the average crustal shear wave velocity is 3.6 km s–1. Although the RVP has been a site of ongoing magmatism since at least 17 Ma and is associated with a pronounced low velocity zone in the mantle, our results show little evidence that the bulk composition or thickness of the crust beneath the RVP has been significantly modified by magmatism or extension. However, Poisson's ratios of 0.29–0.31 (Vp / Vs  = 1.85–1.91) at three of the stations in the RVP, where there is also no evidence for higher Vs, may indicate the presence of partial melt in the crust. The average crustal thickness of Proterozoic terranes surrounding the northern end of the Malawi rift ranges from 38 to 42 km. For most of the terranes, average Poisson's ratios are between 0.25 and 0.26 (Vp / Vs  = 1.73–1.76), with the exception of the Irumide Belt, which has an average Poisson's ratio of 0.23 (Vp / Vs  = 1.68). The average crustal shear wave velocities for all the terranes are either 3.6 or 3.7 km s–1. These results indicate a bulk felsic to intermediate crustal composition for all terranes, consistent with previous results, and reveal that there is little, if any, crustal thinning beneath the uplifted flanks of asymmetric basins within the northern Malawi rift or beneath the RVP. Consequently, crustal thinning in the northern Malawi rift must be highly focused beneath the centres of rift basin segments, consistent with models of rift flank topography and gravity observations. [ABSTRACT FROM AUTHOR]

P and S wave tomographic models have been developed for the northern Malawi rift and adjacent Rungwe Volcanic Province (RVP) using data from the Study of Extension and maGmatism in Malawi aNd Tanzania project and data from previous networks in the study area. The main features of the models are a low‐velocity zone (LVZ) with δVp = ~−1.5–2.0% and δVs = ~−2–3% centered beneath the RVP, a lower‐amplitude LVZ (δVp = ~−1.0–1.3% and δVs = ~−0.7–1%) to the southeast of the RVP beneath the center and northeastern side of the northern Malawi rift, a shift of the lower‐amplitude anomaly at ~−10° to −11° to the west beneath the central basin and to the western side of the rift, and a fast anomaly at all depths beneath the Bangweulu Craton. The LVZ widens further at depths >~150–200 km and extends to the north beneath northwestern Malawi, wrapping around the fast anomaly beneath the craton. We attribute the LVZ beneath the RVP and the northern Malawi rift to the flow of warm, superplume mantle from the southwest, upwelling beneath and around the Bangweulu Craton lithosphere, consistent with high 3He/4He values from the RVP. The LVZ under the RVP and northern Malawi rift strongly indicates that the rifted lithosphere has been thermally perturbed. Given that volcanism in the RVP began about 10 million years earlier than the rift faulting, thermal and/or magmatic weakening of the lithosphere may have begun prior to the onset of rifting. Plain Language Summary: P and S wave tomographic models have been developed for the northern Malawi rift and adjacent Rungwe Volcanic Province (RVP) using data from the Study of Extension and maGmatism in Malawi aNd Tanzania project and data from previous networks in the study area. A low‐velocity anomaly is imaged under the RVP and northern Malawi rift. We attribute the low‐velocity anomaly to flow of warm mantle from the African superplume to the southwest of the study area, which has migrated around the side of thick Bangweulu Craton lithosphere and upwelled beneath the thinner mobile belt lithosphere to the east of the Bangweulu Craton. The observation that volcanism began in the RVP prior to the onset of rifting suggests that the lithosphere beneath the Malawi rift may have been thermally weakened prior to rifting. Key Points: Low‐velocity anomaly is imaged under Rungwe Volcanic Province and northern Malawi riftLow‐velocity anomaly is attributed to upwelling of warm mantle around side of Bangweulu Craton lithosphereLithosphere beneath the Malawi rift may have been weakened prior to rifting [ABSTRACT FROM AUTHOR]

Abstract: The southern Tanganyika Rift, within the Western rift, Africa, has earthquakes to depths of 37 km, yet few constraints exist on crustal thickness, or of early stage rifting processes in apparently amagmatic rift sectors. The aim of the TANGA14 experiment was to constrain bulk crustal properties to test whether magmatic processes modify the lithosphere in areas of deep seismicity, and the degree of lithospheric thinning. We use 11 broadband seismometers to implement receiver function analysis using H‐κ stacking, a method sensitive to crustal thickness and VP/VS ratio, to determine bulk crustal properties. Analyses include extensive error analysis through bootstrap, variance, and phase‐weighted stacking. Results show the Archean Tanzanian Craton and Bangweulu Block are characterized by VP/VS ratios of 1.75–1.77, implying a felsic bulk composition. Crust beneath the fault‐bounded basins has high VP/VS (>1.9). Anorthosite bodies and surface sediments within the region may contribute to localized high VP/VS. However, elevated VP/VS values within fault‐bounded extensional basins where elevated heat flow, hydrothermal vent sites, and deep earthquakes are observed suggest that magma may be intruding the lower crust beneath the southern Tanganyika Rift. Crustal thicknesses on/near the relatively unextended Tanzanian craton and Bangweulu Block are 41.6–42.0 km. This contrasts with the Tanganyika Rift where crustal thicknesses are 31.6 km to 39 km from north to south. Our results provide evidence for ~20% crustal thinning localized to fault‐bounded basins. Taken together, they suggest a previously unrecognized role of magma intrusion in early stage continental rifting in the Western rift, Africa. [ABSTRACT FROM AUTHOR]

The seismically and volcanically active Kivu Rift, in the western branch of the East African Rift System, is a type locale for studies of high-elevation, humid-climate rift basins, as well as magmatic basin development. Interpretations of offshore multi-channel seismic ( MCS) reflection data, terrestrial radar topography, lake bathymetry and seismicity data recorded on a temporary array provide new insights into the structure, stratigraphy and evolution of the Kivu rift. The Kivu rift is an asymmetric graben controlled on its west side by a ca. 110 km-long, N-S striking border fault. The southern basins of the lake and the upper Rusizi river basin are an accommodation zone effectively linking 1470 m-high Lake Kivu to 770 m-high Lake Tanganyika. MCS data in the eastern Kivu lake basin reveal a west-dipping half graben with at least 1.5 km of sedimentary section; most of the ca. 2 km of extension in this sub-basin is accommodated by the east-dipping Iwawa normal fault, which bounds an intrabasinal horst. Lake Kivu experienced at least three periods of near desiccation. The two most recent of these approximately correlate to the African Megadrought and Last Glacial Maximum. There was a rapid lake level transgression of at least 400 m in the early Holocene. The line load of the Virunga volcanic chain enhances the fault-controlled basin subsidence; simple elastic plate models suggest that the line load of the Virunga volcanic chain depresses the basin by more than 1 km, reduces flank uplift locally and broadens the depocentre. Not only do the voluminous magmatism and degassing to the lake pose a hazard to the riparian population, but our studies demonstrate that magmatism has important implications for short-term processes such as lake levels, inflow and outlets, as well as long term modification of classic half-graben basin morphology. [ABSTRACT FROM AUTHOR]

In rift settings, the crustal stress field is dominated by extension, which leads to rift-parallel topography and basin alignments. However in some continental rift systems, some observables of the orientation of principal stresses show substantial deviations from these patterns. Such stress field rotations are currently poorly understood and could reflect the critical role of rift magmatism in the creation of topography, the plate state-of-stress, and volcanic and tectonic processes. Yet the role of magma intrusions, crustal thinning, and rift basin and flank topography on rift zone stress field rotations remain poorly quantified. The seismically- and volcanically-active Magadi-Natron-Manyara region of the East African Rift shows a 60 ∘ local stress field rotation with respect to regional extension. Here, we test the hypothesis that such rotation is due to the cumulative effects of surface and subsurface loads (lateral subsurface density contrasts). We use analytical and calibrated numerical models of magmatic rift zones to simulate lithospheric deformation in the presence of magma bodies, crustal thinning, and topography to quantify their effect on intrusions and fault kinematics in a rift setting. Our 3D static models of a weakly extended rift suggest that surface topography influences shallow stress localization, whereas subsurface density contrasts play a larger role in lower crustal stress localization. Both patterns suggest a preferred region for melt storage beneath the rift valley. We show that the interaction between topography, crustal thinning, extension, and a pressurized magma reservoir could generate principal stress orientations consistent with the local stress rotation observed from earthquake focal mechanisms. Our results demonstrate how rift topography and the geometry of crustal thinning can guide magmatism and strain localization, highlighting the need for a three-dimensional treatment of rift kinematics. • Surface topography influences shallow stress localization. • Subsurface density contrasts influences lower crustal stress localization. • Topography, subsurface, and extension can produce local stress rotation. • Subsurface stresses are important in complete spatial understanding of rift tectonics. [ABSTRACT FROM AUTHOR]

Areas of active volcanism contain elaborate velocity structures that complicate interpretations of earthquake source mechanisms. We examine the spectral characteristics of 805 earthquakes that immediately followed a large volume basaltic dike intrusion and associated silicic flank eruption of Dabbahu volcano in the Afar Depression as recorded on near-source seismometers. We use these results to quantify the contribution of scattering and attenuation to the observed spectra of low-frequency hybrid and volcano-tectonic earthquake clusters from beneath Dabbahu volcano and around the dike zone. We find strong variations in the signal amplitude and frequency content of earthquakes recorded at stations separated by as little as 2 km, caused by preferential attenuation of high frequencies depending on the vantage point. These observations imply that there are large impedance contrasts near the cooling, solidifying, and recently intruded dike. We estimate the intrinsic absorption attenuation coefficient, Q1, and inverse scattering length, 90, averaged over a 300-sq-km area beneath Dabbahu. Our results are consistent with the highest attenuation coefficients from studies of volcanic provinces in Italy (Q-1I ≈~ 0.02, g0 ≈ 0.1 km-1 for a signal at 2 Hz). The magnitude of these two parameters indicates there are large impedance contrasts present in the area due to the recent intrusion of magma and associated fracturing. [ABSTRACT FROM AUTHOR]

A 60-km-long dyke intruded the Dabbahu segment of the Nubia–Arabia Plate boundary (Afar, Ethiopia) in 2005 September, marking the beginning of an ongoing rifting episode. We have monitored the continuing activity using Satellite Radar Interferometry (InSAR) and with data from Global Positioning System (GPS) instruments and seismometers deployed around the rift in response to the initial intrusion. These data show that a sequence of new dyke intrusions has reintruded the central and southern section of the Dabbahu segment. The first was in 2006 June and seven new dykes were emplaced by the end of 2007. Modelling of InSAR data indicates that the dykes were between 0.5 and 2 m wide, up to ∼10 km long and confined to the upper 10 km of crust. An intrusion in 2007 August was associated with a 5-km-long basaltic fissural eruption. During the new dyke injections, InSAR and GPS data show no subsidence at either of the volcanoes at the northern end of the segment, which partly fed the 2005 September dyke. Seismicity data imply that the dykes were probably fed from a source near the Ado'Ale Silicic Complex at the centre of the segment, but the lack of significant subsidence there implies that the source is very deep, or that there was minimal deflation at shallow magma sources. The new dykes are concentrated in an area where the 2005 dyke did not produce significant opening, implying that residual tensile tectonic stresses are higher in this location and are focusing the later intrusions. The sequence of dyke intrusions observed so far is similar to those seen in Iceland during the Krafla rifting episode, which lasted 9 yr from 1975 to 1984. It is likely that, with a continued magma supply, dykes will continue to be intruded until the tectonic stress is fully relieved. As observed at Krafla, eruptions are likely to become more common before the rifting episode is concluded. [ABSTRACT FROM AUTHOR]

The northern Ethiopian rift forms the third arm of the Red Sea, Gulf of Aden triple junction, and marks the transition from continental rifting in the East African rift to incipient oceanic spreading in Afar. We determine the P- and S-wave velocity structure beneath the northern Ethiopian rift using independent tomographic inversion of P- and S-wave relative arrival-time residuals from teleseismic earthquakes recorded by the Ethiopia Afar Geoscientific Lithospheric Experiment (EAGLE) passive experiment using the regularised non-linear least-squares inversion method of VanDecar. Our 79 broad-band instruments covered an area centred on the Boset magmatic segment ∼70 km SE of Addis Ababa in the centre of the northern Ethiopian rift. The study area encompasses several rift segments showing increasing degrees of extension and magmatic intrusion moving from south to north into the Afar depression. Analysis of relative arrival-time residuals shows that the rift flanks are asymmetric with arrivals associated with the southeastern Somalian Plate faster (∼0.65 s for the P waves; ∼2 s for the S waves) than the northwestern Nubian Plate. Our tomographic inversions image a 75 km wide tabular low-velocity zone ( , ) beneath the less-evolved southern part of the rift in the uppermost 200–250 km of the mantle. At depths of >100 km, north of 8.5°N, this low-velocity anomaly broadens laterally and appears to be connected to deeper low-velocity structures under the Afar depression. An off-rift low-velocity structure extending perpendicular to the rift axis correlates with the eastern limit of the E–W trending reactivated Precambrian Ambo–Guder fault zone that is delineated by Quaternary eruptive centres. Along axis, the low-velocity upwelling beneath the rift is segmented, with low-velocity material in the uppermost 100 km often offset to the side of the rift with the highest rift flank topography. Our observations from this magmatic rift zone, which is transitional between continental and oceanic rifting, do not support detachment fault models of lithospheric extension but instead point to strain accommodation via magma assisted rifting. [ABSTRACT FROM AUTHOR]

Although regional isostasy generally is associated with continental lithospheric compression and foreland basin formation, local isostatic compensation commonly is assumed in models of extensional basin formation. The assumption of negligible lithospheric strength during rifting often is justified on the basis of: (1) high heat flow and temperatures produced by elevating the lithosphere - asthenosphere boundary and (2) fracturing of the crust and lithosphere by normal faults. By modeling the development of rift basins within the Western rift system of East Africa and their associated free air gravity anomalies, we assess the role of basin-producing normal faults in modifying the flexural strength of extended lithosphere. Heat flow and seismicity data from the East African plateau region indicate that the Western rift system located on the western side of the plateau developed in old, cold continental lithosphere. These relatively narrow (40-70 km wide), but deep, basins are bounded along one side by high-angle border faults that penetrate to lower crustal levels, as indicated by seismicity data. Along the length of the Western rift system, depth to pre-rift basement and rift flank topography vary between basins from 1 to 8 km and from 1 to 2 km respectively, with deeper basins generally correlating with higher flanks. Comparison of model predictions with topography and free air gravity profiles reveals that the basin depth and the flank height in the majority of the Western rift basins studied can be explained simply by small heaves (3-10 km) across the border fault and with significant flexural strength of the lithosphere maintained during extension. Where both observed basin depth and flank height could not be reproduced, basins were located adjacent to eruptive volcanic centers active in Miocene-Recent times. In these areas, basin depth, rift flank elevation, and free air gravity anomaly may be modified by magmatic underplating of the crust. Estimates of effective elastic thickness Te obtained from our forward models vary between 17 and 38 km. These Te estimates are consistent with previous values obtained through the wavenumber domain correlation of Bouguer gravity and topography data. We find that the lithosphere beneath East Africa maintains significant flexural strength in extension despite localized, but intense, fracturing of the crust by normal faults that penetrate to lower crustal levels. Therefore the assumption of local isostasy (Te ∼ 0) appears invalid within the youthful Western rift system. [ABSTRACT FROM AUTHOR]

Three representative basins in the Western rift system of East Africa are bordered along one side by high-angle normal faults with 2- to 5-km throws (border faults). In plan view ∼100-km-long systems of linear border faults form curvilinear border fault segments bounding the East Kivu, West Kivu, and Rusizi basins. The opposite sides of these asymmetric basins are bounded by lower relief faulted monoclines or en echelon ramps. The largely unfaulted rift flanks have been uplifted 2 km above the 1.3-km-high East African plateau, with uplift narrowing basins during Quaternary time. Maximum estimates of ∼E-W crustal extension within basins are less than 25% (< 16 km), and planar border faults may penetrate the crust. The East Kivu and West Kivu basins are linked across the rift valley by a horst that serves as a hinge for subsidence in both basins. The westward tilted East Kivu and eastward tilted Rusizi border fault segments are linked along the rift by oblique-slip transfer faults that also accommodate along-axis differences in elevation. Upper Miocene-Recent eruptive volcanic centers within the comparatively high-strain interbasinal region (accommodation zone) generally coincide with the tips of border fault segments and transfer faults. The orientations of Miocene-Recent dip-slip and oblique-slip faults show little correlation with Precambrian shear zones or foliations in metamorphic basement. Differences between the East Kivu, West Kivu, and Rusizi basins in the age of initial faulting, subsidence, and age/composition of volcanic products suggest that border fault segments developed diachronously and propagated along the length of the rift. This along-axis border fault propagation and the crosscutting geometry of transfer faults contribute to the segmentation of the Western rift valley. [ABSTRACT FROM AUTHOR]

The Lake Kivu catchment in the East African Rift is subject to various geologic hazards, including frequent volcanic eruptions, earthquakes, and potential limnic overturns and degassing events. Integration of high-resolution seismic reflection data, 14C dated sediment cores, and lake-floor bathymetry reveals large axial and transverse turbidite systems in the eastern basin of the lake. The turbidites were sourced by hyperpycnal river flows during exceptional floods, and the temporal occurrence of the turbidites was climatically controlled. The turbidite record over the past ~12 k.y. is correlated with the regional paleohydrologic records from tropical East Africa. Our study suggests that flood-introduced turbidites preserved in deep lakes are indicators of hydrological changes, and that extreme floods in Lake Kivu's recent history may have triggered deep mixing events. This study also has implications for the current degassing efforts in Lake Kivu; potential geologic hazards may be triggered by extraordinary turbidity currents, and need to be considered in the design and deployment of gas extraction facilities. [ABSTRACT FROM AUTHOR]