Abstract
 The heat of the Earth's formation has been slowly released through the course of its 4.5 Ga history by mantle convection and conduction at the surface.  It is important that we study the nature of this convective flow in order to understand the thermal history of the Earth, its internal composition, and the forces that drive plate motions.  Global seismic images of the mantle can provide information about convective flow. Variations in both compressional (P) wave and shear (S) wave propagation speeds, as well as a wide range of other types of data, have been used to map the mantle.  The conclusions reached by independent studies so far have shown little agreement beyond a few accepted "long-wavelength" features.  Here, a comparison of two new high-resolution models derived by independent groups shows excellent consistency, even for short-wavelength features.
 The first model uses variation in P-wave velocity and the second uses variation in shear wave velocity.  The differences in data sampling between the two studies are large, which means it is unlikely that systematic errors common to both studies contribute to the agreement in their results.    The shear wave study uses multiple bounce seismic phases to study the shallow mantle beneath regions where structures cannot be constrained by direct P- or S-wave data.  These data were fewer than those used in the P-wave study, which has excellent coverage in subduction zones especially.
 Both models show high-wave-speed linear structures in the mid-mantle beneath the Americas and southern Eurasia, as well as in other locations.  These features are associated with past subduction sites.  Upwellings are not so obvious from the seismic images.   The worst disagreement between the two models occurs in regions of poor resolution for one model or the other, especially at the bottom of the mantle.
 
Questions
The success of these two models is promising, but more studies are needed to verify that these results are real. Some of the questions that have yet to be answered:

    -What is the nature of upwellings within the deep mantle?
    -Are there truly gaps in subducted slabs within the lower mantle?
    -Do slabs continue to the core-mantle boundary, if so, how?
    -What is the cause of the apparent P and S velocity structure in the deepest mantle?
    -What do we learn about hotspots or mantle plumes from studies focused in
       hotspot locations?
    -What kinds of chemical heterogeneities are revealed by these velocity structures?
 

Images
 

Links
 
Seismic Tomography- general tomography background
Global Seismic Tomography Revisited - Grand's results are challenged
Some Recent Results of Seismic Tomography - pretty pictures
Studying the Earth's Formation - other experiments on the Earth's interior