SPICE-UiO homepage > Home

SPICE-Research Training Network - UiO > Polarisation anomalies > The Kerguelen hotspot

Lithospheric anisotropy on the Kerguelen hotspot track inferred
from Rayleigh wave polarisation anomalies.
Page 5 of 6
Page 1 | 2 | 3 | 4 | 5 | 6

Wavefield distortion by the local structure of the Kerguelen Plateau: isotropic and anistropic models.
Continuation of page

Wavefield distortion by local structure - Anisotropic.

Plume-like anisotropic structure.

Fig. 8: Vertical cross-section showing the orientation of pyrolite in the plume model. The length of the bars is proportional to the percentage of oriented pyrolite, which is equal to 50 per cent in the plane at 20 km depth.
(after Pettersen and Maupin, Fig. 23, GJI, 149 225-246)

Following the track of the Kerguelen hotspot as proposed by Müller et al. (1993), the hotspot was located north of the Kerguelen Isles at Cretaceous time. It is therefore interesting to study the polarisation of Rayleigh waves in models where the orientation of the minerals follows the pattern expected for plume-like structures.

We construct a very simple model of a plume: a vertical slice showing the orientation of the minerals as a function of depth is given in Fig. 8. The depth of the anisotropy is dictated by the variation of polarisation anomalies with period in our data. In the horizontal plane, the model is axisymmetric, and has 50 per cent oriented pyrolite at Moho depth (20 km).

The model shows that waves which propagate through the centre of the plume show very little polarisation anomalies. The largest anomalies are strongly elliptical and occur in two 300 km wide zones away from the plume centre. A plume of the type described here located to the North-North-East of the Kerguelen Isles (centre position is indicated by the black star in Fig. 4, page 4), such that events from Java propagate on one side of the plume centre, those from Sumatra, which show no polarisation anomaly, propagate through it, and those from the North propagate on the other side, explains also quite well the observations at PAF.


We show that deviation of the wave path from the epicentre-to-station great circle - i.e. distortion of the wavefield when it propagates from the oceanic basins onto the Kerguelen Plateau - are not sufficient to explain the observations at PAF. But anisotropy may produce polarisation anomalies similar to those observed. We show, for example, that a 300 km wide uniform, anisotropic region containing 40 per cent oriented pyrolite (equivalent to 24 per cent oriented olivine), situated to the North of Kerguelen, can explain the data quite well, provided the olivine a-axis is tilted downwards from the horizontal plane by about 45°. The anisotropy has to be situated at the top of the mantle in order to match the decrease of the polarisation anomaly with increasing period (our model has a linear decrease of anisotropy from Moho to 100 km depth).

In the simulations we have performed, the structure around the anisotropic region is isotropic. Let us note however that similar results would be obtained if it was replaced by an anisotropic structure.

This study shows that Rayleigh wave polarisation anomalies can be sufficiently strong to be detectable. They provide unique information on the orientation of mantle anisotropy. Important constraints on the model can be obtained with only one station. But the numerical modelling we present also shows that the polarisation anomalies vary rapidly laterally, showing the potential of analysing polarisation anomalies at a network of stations.


Go back

Back to Polarisation anomalies.

Page 5 of 6
Page 1 | 2 | 3 | 4 | 5 | 6

Back to Top | Previous | Contact | Home

University of Oslo
Universitetet i Oslo - Institutt for geofag / SPICE 2004