Sensitivity kernels for a receiver function waveform misfit: shortcomings in common assumptions.

The receiver function method is widely used in seismology to study the characteristics of Earth’s major discontinuities. When applying this method, several assumptions are commonly made: (i) ray theory is valid, (ii) the incoming wave is planar, and (iii) the background velocity model is sufficiently well-known to accurately map the observations to their location in the subsurface. To test the robustness of these assumptions, we have applied the adjoint method in a full waveform framework to synthetic teleseismic receiver functions and imaged their sensitivity into the subsurface velocity parameters and discontinuities, the 660-discontinuity in particular. First, we observe a strong P-wave sensitivity to large regions within the mantle and on the discontinuity. This implies that using ray theory to map the observations to a single conversion point on the discontinuity may be insufficient. More importantly, because the kernels indicate a significant sensitivity to the far Earth wavefield, they show that the assumption of a simple, planar incoming wavefield is, therefore, inadequate. Because we observe some of the strongest sensitivity to scatterers of multiple phases that arrive within the same time-window as the P660s-phase. When applying the receiver function method, it is therefore of vital importance to be aware of the influence of scatterers or velocity perturbations both before and after conversion at major Earth discontinuities.