The PI's work to date strongly suggest that one can see conductivity structure in the upper 1000km of the Earth with unprecedented resolution. The data already analyzed requires two sharp jumps near the 400 and 650km seismic discontinuities. If this result is true, it should provide a powerful constraint on the compositon and physical state of the mantle. The PI has used robust processing which removes a variety of noises, including the effects of distortion of the electric field by near surface lateral heterogeneities. This processing produces magnetotelluric (MT) impedances which are much more accurate than previously obtained. This data has been inverted using an algorithm that produces the globally flattest model fitting the data within its errors. The fact that this least structure model requires localized jumps argues strongly for their reality. The PI will very carefully examine each step of this processing to improve both the efficiency and robustness of the compu- tations. This effort will include examining other readily available data.
