Recently, considerable progress has been made in the field
of global travel-time tomography. Global models by Van der
Hilst et al. (Nature, 1997) and Zhou (JGR, 1996) now show
upper mantle small-scale structure that approaches the
detail imaged in regional mantle studies. In the lower
mantle these models show clear images that can be
interpreted in terms of subducted (Mesozoic) oceanic
basins.
We take the improvement one step further, mapping upper
mantle structure with detail (60-100 km) directly comparable
with that obtained in high resolution regional mantle
studies. This is achieved through the implementation of a
model parametrization with cells of variable sizes, which
(contrary to Zhou) reduces the number of unknowns with 90%.
The imaged lower mantle structure resembles that of the
aforementioned studies closely. The new model makes it
possible to study areas that have never been imaged with
such detail before and compare areas in which similar
processes take place and put them in a global perspective.
Some of the features of the model include: subduction of
slab either directly into the lower mantle or (currently)
flattening in the transition zone, indicating that the
lower mantle resists, but may not prevent, whole mantle
convection; remnants of past subduction in the lower mantle
(Tethys, Pacific and Farallon plates), sometimes connected
to heterogeneity at the CMB; deep roots of cratons
underneath, e.g. Australia and Scandinavia; rift zones in
Africa and hot upwellings under Iceland. In general, we find
high amplitude (up to 5%) and small-scale heterogeneity in
the upper mantle that gradually turns into longer wavelength
and lower amplitude (but up to 1.5%) anomalies in the
mid-lower mantle (around 2000 km), from where both amplitude
and scale increase to the CMB.