Thermal convection is the motor of Earth dynamics and therefore is the link
between plate motions, hotspots, seismic velocity variations in the mantle
, and anomalies of the gravity field. Small scale mantle anomalies, such as
plumes, do, however, generally escape detection by tomographic methods. It
is attempted to approach the problem of detection in a somewhat statistica
l manner. Correlations are sought between spherical harmonic expansions of
the 'fields' under study: the hotspot distribution, mantle velocity variati
ons, gravity, heat flow. Using spherical harmonic representations of global
fields implies integration and averaging over the whole globe. Thus, altho
ugh relationships may remain masked in the space domain by a multitude of e
ffects, tendencies may become visible in the spectra or in appropriate aver
ages.
The main results are the following: There is a significant long wavelength
(n=2,3) negative correlation between the hotspot density and the P-wave vel
ocity variation in the lower mantle. Positive hotspot density of degree 2 t
o 9 generally correlates with low seismic velocity in all depths of the upp
er mantle and with positive gravity. This fits well with plume-type convect
ion. These results are also confirmed regionally for a number of individual
mid-ocean ridges and hotspots. The hotspot density and the free air anomal
ies are distinctly positive above regions of low velocity extending to grea
t depth. The effect is not distinct at ridges with shallow velocity anomali
es. In a general way, we suggest that the antipodal upwellings (Pacific, Af
rica) are divided by downwelling currents around the shrinking Pacific. Pla
te boundaries can easily move away from their past connections with the dee
per mantle. Small scale plume currents seem to be depicted in the hotspot e
xpansion. (C) 1999 Elsevier Science Ltd. All rights reserved.