Mapping the Tonga Slab

Publication Status is "Submitted" Or "In Press: 
LDEO Publication: 
Publication Type: 
Year of Publication: 
1991
Editor: 
Journal Title: 
Journal of Geophysical Research-Solid Earth and Planets
Journal Date: 
Aug 10
Place Published: 
Tertiary Title: 
Volume: 
96
Issue: 
B9
Pages: 
14403-14427
Section / Start page: 
Publisher: 
ISBN Number: 
0148-0227
ISSN Number: 
Edition: 
Short Title: 
Accession Number: 
ISI:A1991GA40700014
LDEO Publication Number: 
Call Number: 
Abstract: 

We applied residual sphere analysis to 25 intermediate- and deep-focus earthquakes in the Tonga subduction zone and mapped variations in slab strike and dip, depth extent and deep lateral advection at all latitudes along the arc. As in the northwest Pacific, travel time data from intermediate- and deep-focus events are well modeled by simple thermal slab structures and a thermal coefficient of velocity (partial v(p)/partial T) of -0.5 +/- 0.2 m/s/-degrees-K (i.e., peak velocity perturbations of about 3-7%). In all regions where the seismicity extends to depths of 600 km or more, the shortest thermal slab models that cannot be rejected at the 90% confidence level reach depths of 750 km (northern Tonga) to 850 km (southern Tonga). These models, however, require a value of partial v(p)/partial T = -1.5 m/s/-degrees-K, 3 times the value suggested by laboratory studies. If the magnitude of partial v(p)/partial T is constrained to be less than -1.0 m/s/-degrees-K, the minimum penetration depth varies from 800 km to 900 km. The travel time data are equally consistent with slab structures that deform at the base of the upper mantle, thickening by a factor of 3, and models that contain no advective thickening. Inferred slab structure varies considerably throughout the subduction zone, and the nearly ubiquitous match between the observed seismicity and detailed undulations in slab strike and dip is evidence that residual spheres do reflect near-source heterogeneity. Travel times in southern Tonga (21.5-degrees-S-27-degrees-S) are consistent with, and in some cases require, slab models that bend to a steeper dip at depth. This phenomenon is also observed in parts of the northwest Pacific but is not required in central (19-degrees-S-21.5-degrees-S) and northern (13-degrees-S-19-degrees-S) Tonga. The residual spheres of both intermediate- and deep-focus events suggest a sharp bend in slab strike at 25-degrees-S. This feature coincides with a kink in strike observed at all depths in the seismicity and occurs at the intersection of the trench and the Louisville Ridge. Northern Tonga is the locus of considerable complexity in deep slab structure. Seismicity, strain rate and travel time data are consistent with (although they do not uniquely require) a slab structure in northern Tonga which curves sharply to the west at the base of the upper mantle but which strikes to the north at shallower depths and in the lower mantle. The contorted segment would have begun subducting roughly 10 m.y. ago during a major plate reorganization. In northernmost Tonga, a decrease in the maximum depth of seismicity and inferred slab structure provides evidence for southward shear of deep material relative to the shallow slab. Compared to data from the northwest Pacific, where travel time observations also suggest slab penetration below the deepest earthquakes, Tonga seismicity and residual spheres show more complexity at depth and may indicate more deformation internal to the slab. These observations argue against convective models which are rigorously stratified at 650 km depth, but they are consistent with a discontinuity in mantle properties, such as an increase in viscosity or a small percent variation in chemistry, if it merely resists and deforms the descending slabs without acting as a complete barrier to flow.

Notes: 

Ga407Times Cited:39Cited References Count:104

DOI: