1IGPP, La Jolla
2LLNL
Email:
creif@igpp.ucsd.edu
poster/oral: poster
| Many global tomographic models show some overall decorrelation of structure at the 660 km discontinuity and such a decorrelation is consistent with a weak barrier to convection at this depth. However, the tomographic models are generally made without any consideration of variable discontinuity depths and it is conceivable that omitting the 410 and 660 km discontinuities in the inversions leads to biased results. On the other hand, inversions for discontinuity topography use an existing tomographic model (built without discontinuities) to correct for the effect of structure away from the discontinuities. Clearly, tradeoffs between volume perturbations and discontinuity tomopgraphy could lead to erroneous results. To account for this tradeoff, we mix the huge data sets used in global tomography with the relatively small datasets of precursor data which are uniquely sensitive to discontinuity topography. The strength of our inversion is the combination of surface waves to constrain upper mantle structure, carefully picked long period P and S body wave data, and the long period SS and PP precursor data from the recent work of Flanagan and Shearer. We find the topography recovered on the 660 is similar to the SS precursor results, while the topography on the 410 is much more subdued, peak-to-peak amplitude of only 10km. This implies that the usual simplistic interpretation of the 410km discontinuity as the olivine/wadsleyite phase transformation may require revisiting. |