This project (paper here) implements a joint inversion for crustal and upper mantle structure. Inversions of surface-wave phase velocities regularly use receiver functions to help constrain crustal structure, since receiver functions are sensitive to sharper structure but only give relative constraints on velocity when considered alone. This project extended this idea to the uppermost mantle with Sp receiver functions.

The figure below shows how dramatically this can help when the upper mantle features a sharp lithosphere-asthenosphere boundary. The yellow lines are what is recovered when you invert the surface wave data with the receiver function constraints on the Moho (Ps in legend). Note how the smooth model misfits the input model at ~75 km depth! Adding the constraint from Sp receiver functions make a huge improvement. The second figure, panel B, shows that in some cases you need Pn velocities (HW for head wave) to do a good job (the supplement explores why if you are curious).

That’s great, but what about the real earth?

Melt is required below the LAB

We can now recover much lower velocities in the upper mantle. Black dash lines are the minimum velocity that can be explained by two sets of laboratory experiments - anything slower and we inferred melt must be present in the mantle.

Green circles are volcanoes of Pleistocene or younger age - and volcanoes (almost) only occur where melt is required! The only real exception are the Leucite Hills at ~42N/109W, which turn out to have a very unusual composition that isn’t consistent with an asthenospheric source anyway.

But there’s more

This figure shows the slope of the Vs model below the LAB, with purple being greater. Note that this correlates very well with the Vs shown above! Where velocity is low, the velocity profile is also steeper.

Turns out, these slopes also require melt - laboratory experiments can’t explain their magnitude. Further, the hypothesis that the LAB is caused by water can’t explain the steep slopes (because the asthenosphere wouldn’t be more hydrated right at the LAB).