S14A-01
Whole Mantle 1-D Structure From Short-period Body Waves
Modeling Earth's anelastic structure as well as elastic structure is important for several reasons: (1) the depth dependence of attenuation and the shear-to-bulk Q ratio constrain the physical state of the deep Earth including its melt content, (2) attenuation can be a strong indicator of temperature variations because they have a larger effect on attenuation than on elastic velocity, and (3) attenuation causes physical dispersion of seismic velocities, which must be taken into account when interpreting travel time data. However, attenuation studies have proven challenging because of the typically large scatter in attenuation measurements and the difficulty in separating out source and elastic propagation effects from the intrinsic attenuation signal. We describe a new 1-dimensional Q model for short-period body waves. It is modeled from a dataset of 15,000 differential t* measurements of teleseismic P and S waves recorded in broadband seismograms. The S waveform is synthesized from the observed P wave and then cross-correlated to the observed S wave. The t* that gives the best correlation coefficient provides an estimate of attenuation along the ray path. To avoid biases from multipathing and other propagation path effects, we perform the cross- correlation only on the first half swing of the waveform. An advantage of our dataset is that it is little affected by the source-time function or instrument response, since the P and S waves are recorded at the same station from the same event. We invert our t* measurements for a 3-layer Qμ model (2 in upper mantle and 1 in lower mantle). Our new Q model has higher Q values (less attenuation) compared to existing Q models derived from longer period datasets, especially in the lower mantle. This is consistent with frequency dependence of Q as has been suggested both from laboratory experiments and previous seismic observations. The attenuation is strongest in the upper mantle and we examine regional variations in Q by computing station and event terms (i.e., by averaging the t* residuals). These results show correlations with tectonics that are generally consistent with regional attenuation studies.
S14A-02 INVITED
A Seismically Sharp Lithospheric Base Persisting to the Lowermost Mantle Beneath the Caribbean
The detailed structure of lower mantle slabs is not well known and it is crucial to examine the fate of subducted slabs in the mid- and lower mantle, since it is closely related to the question of mantle convection pattern and mass flux across the mantle transition zone. To address this problem, we applied a newly developed seismic wave migration method to broadband data recorded by Californian seismic networks from South American earthquakes. Out-of-plane seismic P wave reflections which propagate from the source are reflected at seismic heterogeneities and arrive at the stations, are observed beneath the Cocos plate. The reflection locations extend through Earthfs lower mantle down to the core-mantle boundary (CMB), and coincide with the edges of tomographically mapped high seismic velocities. The observed energy is well explained by underside reflections off a sharp contrast between the base of the former depleted harzburgitic Nazca plate lithosphere and the viscously entrained more fertile underlying lherzolitic former asthenosphere. We also detect weaker reflections corresponding to the apparent top of the slab, which may arise from the boundary between the Nazca plate and the overlying former basaltic oceanic crust. Mass flux across the 670 km discontinuity of the former oceanic asthenosphere, lithosphere and basaltic crust is thus supported, intimately linking processes and structure at the top and bottom of Earthfs mantle by large scale mantle flow.
S14A-03
Seismic velocity anomalies in the D" layer beneath Hawaii
Global shear velocity tomographic models show two large-scale low velocity structures, so called superplumes, in the lower mantle, under Africa and under the mid-Pacific. Sharp lateral velocity changes have been documented at some borders of the superplumes. Here, we report the prominent postcursors to the Sdiff waves which sample the D" layer beneath Hawaii. The tomographic models show that this region corresponds to the northern boundary of the Pacific superplume. The events and stations are located in Papua New Guinia and in southern U.S, respectively. We first measured travel time anomalies of S and Sdiff phases with respect to PREM. They correlate well with the Vs anomaly distribution of the D" layer shown by the tomographic models. Depending on whether the paths mostly sample the D" layer within or outside of the superplume, the anomalies change by about 10 seconds within a 20 degrees change of azimuth. The result indicates strong lateral Vs heterogeneities in this region, such as the ones reported for other borders of the superplumes. The postcursors show the following features: 1) they arrive 20 to 40 seconds after the Sdiff waves 2) the arrival times strongly depend on station azimuths, rather than epicentral distances. The later phase arrives later for northern stations at smaller azimuths and earlier for southern stations at greater azimuths. This feature is in contrast to the main Sdiff phase which progressively delays to the south. Consequently the first and second arrivals come closer along southern paths than along northern paths. We applied beamforming in order to find the incident angle and direction of the second arrival. The incident angle is almost the same between the first and second arrivals but the incident azimuth is 7 degrees southward relative to the azimuth of the first arrival. There are similar and dissimilar points between this postcursor and the previously reported Sdiff splitting which sample the southern borders of the Pacific and African superplumes (Wen, 2001, To et al 2005). The newly observed postcursor is well separated from the main Sdiff phase and its amplitude is as large as the main phase, whereas the postcursor from either southern Pacific or Africa has smaller amplitude and overlaps in time with the main phase. By conducting waveform modeling for several simple structures, we discuss the origin of the later phase from the D" layer beneath Hawaii.
S14A-04 INVITED
Scattering of PKKP and P'P' - a tool for mapping fine-scale mantle heterogeneity
Much of what we know about the Earth's deep internal structure at scale lengths around 10 km comes from teleseismic observations of 1 Hz scattered waves. These waves originate when high-frequency surface or body waves interact with fine-scale volumetric heterogeneity or rough boundary layers. Previous studies used scattered energy related to PKP, PKiKP, and Pdiff to identify and map the small-scale structure of the mantle and core. These studies elucidate the mineral-physical and geochemical constitution of the Earth and can help answer questions such as the existence of melt in the Earth's deep interior, the location of geochemical reservoirs, and the fate of subducted crustal and slab material. We use observations of scattered energy related to the core phases PKKP and P'P' to study fine-scale mantle heterogeneities. PKKP and P'P' are maximum travel-time phases with respect to perturbations at their reflection points (either at the core-mantle boundary or at the surface, respectively). Therefore, deep-mantle scattered energy arrives as precursors to the main phases where it can be observed free from the contaminating contribution of crustal scattering. Additionally, the scattered energy arrives over a large distance range in a quiet time window. Scattered PKKP and P'P' waves are sensitive to scatterers at different depths within the mantle and hold promise for development of both one dimensional scattering profiles and three-dimensional maps of fine-scale heterogeneity in the Earth's interior.
S14A-05 INVITED
PKPPcP or P'P' precursor from a 785km discontinuity?
Precursors of P'P' around 80 degrees are very useful for studying discontinuities in the Earth, particularly the 410km and 660 km discontinuities. P'P' precursors from deeper discontinuities, however, can be contaminated by the PKPPcP (or PcPPKP) phase, which is strongest around the antipode and extends to 275 degrees according to geometric ray theory. Because P velocity in the outer core is much lower than that in the lower mantle, tunneling effects are prominent for distances larger than 250 degrees, and PKPPcP becomes PKPPdiff for distances larger than 275 degrees. We first apply normal mode summation to study the amplitude of PKPPcP vs distance, which predicts PKPPcP to be a strong phase for distances well beyond 280 degress at 0.1HZ. Then we apply full wave technique to modeling of PKPPcP's amplitude at distances from 210 to 310 degrees for 0.1 HZ---2 HZ, and find that PKPPcP is strong phase for distances around 280 degrees and 0.3 Hz, but is very weak at 1Hz. The frequency dependence of PKPPcP amplitude agrees well with the P'P' precursor observations of le Stunff et al. (1995), which were best observed in a low frequency band (0.01 to 0.3 Hz), disappearing at higher frequency (around 1HZ). Thus the P'P' precursor are better interpreted as PKPPcP, instead of reflection by a 785km discontinuity beneath Africa. The relatively strong amplitude of PKPPcP beyond 280 degrees makes P'P' precursors less applicable for discontinuities deeper than 660km.
S14A-06 INVITED
Reflectors Here, Reflectors There ... Reflectors Everywhere?
Much of global seismology focuses on the direct waves (P and S), with less consideration for reflected waves from major interfaces like SS (reflected at the surface), and even less attention given to smaller amplitude reflected waves such as S660S (the underside shear-wave reflection off the 660-km discontinuity), which provide valuable constraints on mantle discontinuity structure. These phases are generally of such low amplitude that waveform stacking is required to increase their signal-to-noise. Here, we examine SS and ScS precursors respectively for bottom and topside mantle reflections using all the available data in the IRIS FARM database. Previous studies have shown that the globally coherent mantle reflectors are only imaged near 410, 520 and 660 km. However, the possibility remains that regional reflectors at other depths may be common, but they are too intermittent or have too much topography to appear in global data stacks. We search for robust regional reflectors by performing common bouncepoint stacks at 1 km depth intervals and 400 km lateral spacing. We apply weights based on the reflection angle and the signal-to-noise ratio of the direct SS and ScS phases and use bootstrap resampling to constrain the reliability of the imaged reflectors. We report on observed major and minor global reflectors as well as robust regional anomalies and compare our results with those obtained in other studies.
S14A-07
Imaging Delamination Beneath Tien Shan Using Multiple-ScS Reverberations
Mantle reflectivity mapping through the use of multiple-ScS reverberations has proven to be very effective, especially for the transition zone discontinuities. However, most studies of this kind have examined paths crossing ocean basins and few have sampled sub-continental mantle (e.g. Sipkin and Revenaugh, 1994). In large part, this is explained by the fact that the relatively thin and homogeneous oceanic lithosphere coupled with lower upper mantle Q result in much less scattered wave interference than the thick, heterogeneous lithosphere and high Q of the continents. But given sufficient deep seismicity, the method does provide useful results. Here we apply it to the Tien Shan region using recordings captured by the PASSCAL Tien Shan Experiment. Convergent features such as crustal shortening and magmatic underplating have resulted in extremely complex lithosphere (Vinnik et al., 2004). As with all previous applications of multiple ScS reverberation mapping, we identify the 410-km and 660-km discontinuities. For paths connecting events in Hindu Kush and the Tien Shan network, we observe several other reflectors, including a shear impedance decrease at approximately 110 km depth, which coincides well with receiver function imaging of the base of the seismic lid; and a large shear impedance increase at 300 km depth. The latter, which we dub S, is 50% larger than the 660 km discontinuity and is by far and away the largest reflector we have ever seen in the upper mantle in 25 years of looking. The feature is robust: it appears in all subsets of the data, is not affected by changes in processing parameters and can be observed in individual seismograms (unlike most other mantle reflectors). Having determined that it is real, it remains to understand its origin. A reasonable explanation is a piece of delaminated lithosphere that has broken off due to extreme lithospheric thicknening. Delamination has previously been suggested by Chen et al. (1997) from tomography and the thin seismic lid. If correct, the delaminated material must either geometrically focus the shear reverberations or have a highly unusual velocity contrast with surrounding material likely requiring some decompression melting in material rising to fill the void.
S14A-08
Large-scale Shear Velocities Beneath Hotspot Locations: New Observations and Travel Time Synthesis
The existence, lateral dimension, and depth of mantle plumes beneath hotspots have been issues of contentious debate in the past two decades. To a large extent, the difficulty lies in the insufficient data resolution in the transition region (400-1000 km) between the upper and lower mantle beneath major hotspot locations. In this study, we report a new type of observation, the ray parameter variation of SS precursors, to image large-scale mantle shear velocities from the surface down to 1200 km beneath 17 major hotspots. We significantly improve the resolution by a High-resolution Radon transform method that utilizes time-domain inversions to simultaneously constrain differential times and ray parameters. Perturbations of S410S-SS ray parameters present the most revealing observations of hot thermal anomalies in this study. We identify both positive and negative jumps in travel time curves (hence ray parameters) for rays bottoming beneath the majority of the hotspot locations; these anomalous jumps are not observed at non-hotspot locations. Through careful modeling of the observed ray parameter jumps using 2-D finite-difference ray tracing, and accounting for depth, width, sign and strength of the velocity columns, we are able to unequivocally categorize the mantle beneath the aforementioned 17 hotspots to have low seismic velocities at depth 1) comparable to transition zone (400-670 km) depths, or 2) down to 900-1200 km or more. While our data and modeling strategies do not enable us to probe the source of hotspots beyond 1200-km depth, they do have major implications for the discussion of shallow vs. deep-rooted hotspots around the world. For example, the Hawaii and Macdonald hotspots exhibit strong type 1 characteristics, while the Azores and Canary hotspots are describable by type 2 mantle velocity variations.