HR: 09:30h
AN: V31F-07 INVITED     [Abstracts]
TI: Seismic Tomography and the Scale of Mantle Convection
AU: * Dziewonski, A M
EM: dziewons@eps.harvard.edu
AF: Department of Earth and Planetary Sciences, Harvard University, 20 Oxford St., Cambridge, MA 02138 United States
AB: For nearly 30 years seismic tomography has been expected to provide a definitive answer to the question of the scale of mantle convection. Yet, the results obtained so far are not totally convincing. There are aspects of the tomographic results that may favor a layered convection. Among those is the very large extent of the positive velocity anomalies in the transition zone, associated with the current zones of deep subduction; width of these anomalies can be as large as 4,000 km, indicating ponding of the subducted material. The velocity anomalies are well correlated with the topography of the 660 km discontinuity, while the topography of the 410 km is de-correlated. Studies of the tomographic power spectrum as a function of depth indicate a significant change in the pattern at about 660 km. Additional evidence comes from the distribution of the deep seismicity: besides the abrupt termination of seismicity at 700 km depth there are "earthquakes outside Benioff zones", indicating a horizontal deflection of the subducting slab. The examples of this phenomenon are widely spread, including the Kuriles, Izu-Bonin, Indonesia, Tonga and Argentina. These earthquakes are relatively rare, but often have large magnitudes. Thus, it appears that the transport of the material into the lower mantle is significantly impeded, if not disrupted. However, there are also arguments in favor of a connection between the lower and upper mantle. The principal of these, in addition to the apparent continuity of high velocity anomalies in the lower mantle with the present or past subduction, there is a very large structure in the last 500 km of the mantle dominated by degrees 2 and 3. The higher than average regions of this structure correlate well with the same wavelength expansion of the position of subduction zones some 200 million years ago. Also, the location of the slower than average regions ("superplumes") corresponds well to the occurrence of hot spots at the surface. At present there is no quantitative explanation and reconciliation of the facts described above. It appears that some episodic connection ("avalanches", for example) is needed. A partial, at least, re-circulation of the subducted material in the upper mantle may have an important geochemical signature.
DE: 1038 Mantle processes (3621)
DE: 7270 Tomography (6982, 8180)
DE: 8120 Dynamics of lithosphere and mantle: general (1213)
DE: 8124 Earth's interior: composition and state (1212, 7207, 7208, 8105)
DE: 8413 Subduction zone processes (1031, 3060, 3613, 8170)
SC: Volcanology, Geochemistry, Petrology [V]
MN: Fall Meeting 2005