Heterogeneity, anisotropy, and scattering in upper and 
lower mantle boundary layers

Ed Garnero
February 18, 2005

ABSTRACT

 Seismologists have long characterized Earth's lowermost mantle as far from 
simple homogeneity.  Models based on seismic data include sharp horizontal 
discontinuities, strong anisotropy, thin ultra-low velocity zones (ULVZ), and 
sharp vertical boundaries between normal and low velocity mantle. The scales of 
heterogeneity span several orders of magnitude (a few to 1000's of km). This 
diversity of structures is beginning to be understood in terms of geodynamics 
and mineral physics, with dense partial melts causing the ULVZs, a 
post-perovskite solid-solid phase transition producing regional layering, and 
the possibility of large-scale variations in chemistry, such as distinct and 
dense thermo-chemical piles which can guide plume initiation near its 
boundaries.

 Three deep mantle regions will be highlighted in this presentation: the 
lower mantle beneath South Africa, the Pacific, and the Caribbean.  The first 
two represent low shear velocities that correlate with surface hot spots, the 
latter is high velocity and underlies subduction.  Seismogram record sections 
show travel time offsets and waveform complications that require nearly 
vertical walls separating the anomalously slow structure from normal mantle.  
Data sampling the African superplume are best fit by ~1200 km wide low 
velocities extending up to 1000 km above the CMB.  The clearest detection of 
ULVZs may be near the edges of these low velocity structures, consistent 
with plume genesis there. Many seismic data and modeling examples will be
presented. Seismic results are reconciled with thermochemical convection 
calculations with temperature- and depth-dependent rheology, that employ
surface boundary conditions consistent with Earth's recent plate recent
plate history. Initial calculations predict that dense piles develop with
remarkable resemblance to the geometry of present day seismic low velocity 
provinces providing us with first-order thermochemical models that may be 
used to better understand lower mantle heterogeneity. In addition, we find
that the observation of plumes statistically residing at the edges of large,
negative seismic anomalies may be geodynamically supported. 

 Our recent investigation of upper mantle phenomena appear similarly 
complex.  Analyses include migration of the SS precursor wavefield, and slant 
stacking of short period array data that reveal scattered energy preceding PP.  
Array analyses upper mantle triplications with short period P wave show promise 
at mapping fine scale structure near the priniciple upper mantle phase 
boundaries.  These findings support the notion of significant short scale 
heterogeneity in Earth's mantle near all important boundary layers.