We study the thermal structure around a cold deformable lithospheric slab as it sinks to the core-mantle boundary and migrates along it. We present analytical results for the steady thermal structure established by a steady but spatially varying motion. The analysis gives a time-like criterion for the thermal signature of a cold slab to persist by the time that the slab moves along the core-mantle boundary. The model is used to assess the feasibility of a purely thermal origin for some of the observed seismic reflectors near the core-mantle boundary. Calculations of the time-like criterion show that the dynamical conditions in our model, namely the velocity and the thickness of the descending slab, are hard to reconcile with observations of subduction and seismic features. Seismic reflections and refractions from anomalously fast regions above the core-mantle boundary could be explained as thermal slabs if the thickness of the slab at subduction was larger than 200 km or somewhat less if the slab did not split at the core-mantle boundary. A simple thermal model also predicts from mineral physics a certain correlation between S- and P-wave velocity anomalies, which is not observed. However, a purely thermal origin cannot be ruled out if the slab is buckling. This process could be in agreement with the observations: the amplitude of the seismic anomalies, the vertical extent of high-gradient zones and the P versus S comparisons. Chemical heterogeneities and phase transformations remain alternative or complementary explanations.
Bibliographical noteFunding Information:
This paper has benefited greatly from discussions with Michael Weber and Yariv Hamiel. We are grateful for constructive comments in review by Rob van der Hilst and Michael Wysession. This study was made possible by generous grants from the Ministry of Science, Niedersachsen (AA) and the Lady Davis Fellowship programme (CM), for which we are also grateful. [CL]
- Core-mantle boundary
- Mantle discontinuity
- Plate tectonics
- Thermal anomalies