John C. VanDecar

Tectospheric structure beneath southern Africa

P-wave and S-wave delay times from the broadband data of the southern Africa seismic experiment have been inverted to obtain three-dimensional images of velocity perturbations in the mantle beneath southern Africa. High velocity mantle roots appear to extend to depths of at least 250 km, and locally to depths of 300 km beneath the Kaapvaal and Zimbabwe cratons. Thick roots are confined to the Archean cratons, with no evidence for similar structures beneath the adjacent Proterozoic mobile belts. The Kaapvaal craton was modified ca. 2.05 Ga by the Bushveld magmatic event, which affected a broad swath of cratonic mantle beneath and to the west of the exposed Bushveld Complex. The mantle beneath the extended Bushveld province is characterized by seismic velocities lower than those observed in regions of undisturbed cratonic mantle. The mantle beneath the Limpopo Belt, an Archean collisional zone sandwiched between the Kaapvaal and Zimbabwe cratons, exhibits a cratonic signature.

Upper mantle seismic velocity structure beneath Tanzania, east Africa: Implications for the stability of cratonic lithosphere

The assertion of cratonic stability put forward in the model for deep continental structure can be tested by examining upper mantle structure beneath the Tanzania Craton, which lies within a tectonically active region in east Africa. Tomographic inversions of about 1200 teleseismic P and S travel times indicate that high-velocity lithosphere beneath the Tanzania Craton extends to a depth of at least 200 km and possibly to 300 or 350 km. Based on the thickness of mantle lithosphere beneath Archean cratons elsewhere, it appears that the mantle lithosphere of the Tanzania Craton has not been extensively disrupted by the Cenozoic tectonism in east Africa, thus corroborating the assertion of cratonic stability in the model for deep continental structure. The presence of thick, high-velocity structure beneath the Tanzania Craton implies relatively low temperatures within the cratonic mantle lithosphere, consistent with relatively low surface heat flow. The thick cratonic keel is surrounded by low seismic velocity regions beneath the east African rifts that extend to depths below 400 km. Our models show a shear velocity contrast between the cratonic lithosphere and the uppermost mantle beneath the eastern branch of the rift system of about 5% to 6%, but from resolution experiments we infer that this contrast could be underestimated by as much as a factor of 1.5. We attribute about half of this velocity contrast to the depleted composition of the cratonic keel and the other half to thermal alteration of upper mantle beneath the rifts. Low-density structures that may be required to provide buoyant support for the elevation of the Tanzania Craton must reside at depths greater than about 300–350 km.

Mantle seismic structure beneath the Kaapvaal and Zimbabwe Cratons

We present results of seismic tomography for a broad region of southern Africa using data from the seismic component of the Kaapvaal Project, a multinational, multidisciplinary experiment conducted in the late 1990s. Seismic images provide clear evidence of mantle structures that mimic the surface geology across the region and provide important constraints on subcrustal structure associated with Archean cratons. Specifically, a thick (~250 to 300km) mantle keel exists beneath the Kaapvaal craton; a slightly thinner (~225 to 250km) keel exists beneath the Zimbabwe craton and parts of the Archean Limpopo mobile belt. Mantle velocities lower than surrounding regions are evident across a broad swath beneath the surface expression of the Bushveld Complex, a ~2.05 Ga layered mafic intrusion. These reduced velocities may be due to mantle refertilization during intrusion of Bushveld magmas, or they may be caused by a thermal perturbation of more recent origin, perhaps related to the ~183 Ma Karoo magmatic event.

Upper mantle P-wave speed variations beneath Ethiopia and the origin of the Afar hotspot

The Afar hotspot has long been attributed to one or more thermal upwellings in the mantle, in particular starting thermal plumes characterized by a head that spreads laterally beneath the lithosphere, and a tail. New P-wave tomography images of the upper mantle beneath Ethiopia reveal an elongated low wave speed region that is deep (>400 km) and wide (>500 km). The location of the low wave speed anomaly aligns with the Afar Depression and Main Ethiopian Rift in the uppermost mantle, but the center of the anomaly shifts to the west with depth. The shape, depth extent, and location of the low wave speed anomaly is not consistent with a starting thermal plume presently beneath the hotspot. Instead, the anomaly suggests that the hotspot may be the surface manifestation of a broad mantle upwelling connected to the African Superplume in the lower mantle beneath southern Africa.

P and S velocity structure of the upper mantle beneath the Transantarctic Mountains, East Antarctic craton, and Ross Sea from travel time tomography

P and S wave travel times from teleseismic earthquakes recorded by the Transantarctic Mountains Seismic Experiment (TAMSEIS) have been used to tomographically image upper mantle structure beneath portions of the Transantarctic Mountains (TAM), the East Antarctic (EA) craton, and the West Antarctic rift system (WARS) in the vicinity of Ross Island, Antarctica. The TAM form a major tectonic boundary that divides the stable EA craton and the tectonically active WARS. Relative arrival times were determined using a multichannel cross-correlation technique on teleseismic P and S phases from earthquakes with mb ≥ 5.5. 3934 P waves were used from 322 events, and 2244 S waves were used from 168 events. Relative travel time residuals were inverted for upper mantle structure using VanDecars method. The P wave tomography model reveals a low-velocity anomaly in the upper mantle of approximately δVp = −1 to −1.5% in the vicinity of Ross Island extending laterally 50 to 100 km beneath the TAM from the coast, placing the contact between regions of fast and slow velocities well inland from the coast beneath the TAM. The magnitude of the low-velocity anomaly in the P wave model appears to diminish beneath the TAM to the north and south of Ross Island. The depth extent of the low-velocity anomaly is not well constrained, but it probably is confined to depths above ∼200 km. The S wave model, within resolution limits, is consistent with the P wave model. The low-velocity anomaly within the upper mantle can be attributed to a 200–300 K thermal anomaly, consistent with estimates obtained from seismic attenuation measurements. The presence of a thermal anomaly of this magnitude supports models invoking a thermal buoyancy contribution to flexurally driven TAM uplift, at least in the Ross Island region of the TAM. Because the magnitude of the anomaly to the north and south of Ross Island may diminish, the thermal contribution to the uplift of the TAM could be variable along strike, with the largest contribution in the Ross Island region. The tomography results reveal faster than average velocities beneath East Antarctica, as expected for cratonic upper mantle.

Seismic imaging of a hot upwelling beneath the British Isles

The Iceland plume has had an important influence on vertical motions in the North Atlantic. The convecting mantle in this region contains a large-scale low-velocity seismic anomaly, which correlates with a long-wavelength gravity high and bathymetric feature. This suggests that an arm of plume material has extended, or is extending, from Iceland, in a direction perpendicular to the Mid-Atlantic Ridge. Here we present the results of a detailed teleseismic traveltime study that reveals the high-resolution morphology of this low-velocity anomaly beneath the British Isles. Our images provide insights into the nature of plume-lithosphere interactions. The low-velocity anomaly imaged in this study correlates geographically with a region of high gravity anomalies and high topography that was associated with Paleogene magmatism and phases of epeirogenic uplift during the Cenozoic Era. There is evidence that the distribution of British earthquakes is also related to the low-velocity anomaly. The low-velocity anomaly is interpreted to represent hot material from the original Iceland plume head that became trapped beneath thinned regions of lithosphere ca. 60 Ma.

Subduction of the Chile Ridge: Upper mantle structure and flow

We deployed 39 broadband seismometers in southern Chile from Dec. 2004 to Feb. 2007 to determine lithosphere and upper mantle structure in the vicinity of the subducting Chile Ridge. Body-wave travel-time tomography clearly shows the existence of a long-hypothesized slab window, a gap between the subducted Nazca and Antarctic lithospheres. P-wave velocities in the slab gap are distinctly slow relative to surrounding asthenospheric mantle. Thus, the gap between slabs visible in the imaging appears to be filled by unusually warm asthenosphere, consistent with subduction of the Chile Ridge. Shear wave splitting in the Chile Ridge subduction region is very strong (mean delay time ~3 s) and highly variable. North of the slab windows, splitting fast directions are mostly trench parallel, but, in the region of the slab gap, splitting fast trends appear to fan from NW-SE trends in the north, through ENE-WSW trends toward the middle of the slab window, to NE-SW trends south of the slab window. We interpret these results as indicating flow of asthenospheric upper mantle into the slab window.

Source-side shear wave splitting and upper mantle flow in the Chile Ridge subduction region

The actively spreading Chile Ridge has been subducting beneath Patagonian Chile since the Middle Miocene. After subduction, continued separation of the faster Nazca plate from the slow Antarctic plate has opened up a gap—a slab window—between the subducted oceanic lithospheres beneath South America. We examined the form of the asthenospheric mantle flow in the vicinity of this slab window using S waves from six isolated, unusual 2007 earthquakes that occurred in the generally low-seismicity region just north of the ridge subduction region. The S waves from these earthquakes were recorded at distant seismic stations, but were split into fast and slow orthogonally polarized waves at upper mantle depths during their passage through the slab window and environs. We isolated the directions of fast split shear waves near the slab window by correcting for upper mantle seismic anisotropy at the distant stations. The results show that the generally trench-parallel upper mantle flow beneath the Nazca plate rotates to an ENE trend in the neighborhood of the slab gap, consistent with upper mantle flow from west to east through the slab window.

Implications of spatial and temporal development of the aftershock sequence for the Mw 8.3 June 9, 1994 Deep Bolivian Earthquake

On June 9, 1994 the Mw 8.3 Bolivia earthquake (636 km depth) occurred in a region which had not experienced significant, deep seismicity for at least 30 years. The mainshock and aftershocks were recorded in Bolivia on the BANJO and SEDA broadband seismic arrays and on the San Calixto Network. We used the joint hypocenter determination method to determine the relative location of the aftershocks. We have identified no foreshocks and 89 aftershocks (m > 2.2) for the 20-day period following the mainshock. The frequency of aftershock occurrence decreased rapidly, with only one or two aftershocks per day occuring after day two. The temporal decay of aftershock activity is similar to shallow aftershock sequences, but the number of aftershocks is two orders of magnitude less. Additionally, a mb ∼6, apparently triggered earthquake occurred just 10 minutes after the mainshock about 330 km east-southeast of the mainshock at a depth of 671 km. The aftershock sequence occurred north and east of the mainshock and extends to a depth of 665 km. The aftershocks define a slab striking N68°W and dipping 45°NE. The strike, dip, and location of the aftershock zone are consistent with this seismicity being confined within the downward extension of the subducted Nazca plate. The location and orientation of the aftershock sequence indicate that the subducted Nazca plate bends between the NNW striking zone of deep seismicity in western Brazil and the N-S striking zone of seismicity in central Bolivia. A tear in the deep slab is not necessitated by the data. A subset of the aftershock hypocenters cluster along a subhorizontal plane near the depth of the mainshock, favoring a horizontal fault plane. The horizontal dimensions of the mainshock [Beck et al., this issue; Silver et al., 1995] and slab defined by the aftershocks are approximately equal, indicating that the mainshock ruptured through the slab.

Inversion of body-wave delay times for mantle structure beneath the Hawaiian islands: results from the PELENET experiment

Abstract As an experiment to assess the nature of resolvable seismic velocity anomalies in the mantle surrounding the Hawaiian Islands, we operated a temporary network – known as PELENET – consisting of seven broadband portable seismometers on the islands of Kauai, Molokai, Maui, and Hawaii during 1995–1999. Here we report the inversion of body-wave delay times across the network, including arrivals at the Global Seismic Network station KIP on Oahu and the Ocean Seismic Network Pilot Experiment south of Oahu, to determine the three-dimensional seismic velocity structure of the upper mantle of the region. The consistent feature observed in both P- and S-wave inversions is a low-velocity anomaly beneath Maui and Molokai, which we suggest may reflect a zone of secondary melting in the asthenosphere downstream from the hotspot locus. The inversions do not resolve a cylindrical low-velocity plume in the upper mantle beneath the island of Hawaii, but resolution tests indicate that this outcome could be the result of the sparse and nearly linear distribution of stations combined with the incomplete azimuthal coverage of earthquake sources. A determination of the detailed three-dimensional structure of the upper mantle beneath the Hawaiian hotspot will require a simultaneous deployment of both ocean-bottom and land seismometers.