I. Sidorin et al., Dynamics of a phase change at the base of the mantle consistent with seismological observations, J GEO R-SOL, 104(B7), 1999, pp. 15005-15023
The phase change model for the origin of the D" seismic discontinuity is te
sted by comparing the results of convection modeling with seismic observati
ons. We compute a number of global dynamic models that incorporate a phase
change at the base of the mantle with different characteristics and transfo
rm the resulting temperature field and the distribution of phases to seismi
c velocities. Over 900 two-dimensional synthetic waveforms are computed for
each of the models from which S, ScS, and Scd phases are picked. The distr
ibution of the relative amplitudes and differential travel time residuals f
or these phases are statistically compared with the distribution of data fr
om four well studied regions (northern Siberia, Alaska, India, and Central
America) in a search for the characteristics of a phase transition that bes
t match these seismic observations. We find that the best fit among the mod
els considered is obtained for phase transitions characterized by a Clapeyr
on slope of similar to 6 MPa K-1 and an elevation above the core-mantle bou
ndary of similar to 150 km under adiabatic temperature or 127 GPa and 2650
K on a (P,T) diagram. Dynamic models demonstrate that the value of Clapeyro
n slope and the density difference between the phases can have significant
influence on the dynamics of plumes but probably only a minor influence on
the dynamics of subducted slabs. We find that the thermal structure of subd
ucted slabs can be important in giving rise to the seismic triplication; th
e strongest Scd arrivals in our models are observed in the area of subducti
on. The folding of the slab at the base of the mantle leads to patterns in
differential travel time distributions consistent with seismic observations
and suggests that the largest heterogeneity occurs at the top of the D" la
yer or just above it. Analysis of the spatial autocorrelation functions of
the differential travel time residuals suggests that their characteristic p
eaks reflect the patterns of slab folding and may provide constraints on th
e rheology of slabs at the base of the mantle.