Surface-Wave Dispersion and Receiver Function
In this tutorial, we demonstrate the use of BayesBay to infer a 1-D shear-velocity model from the joint inversion of surface-wave dispersion curves and a receiver function.
Background
Surface waves (Rayleigh and Love) are seismic waves generated by earthquakes and ambient seismic noise that propagate along the Earth’s surface. These waves exhibit dispersion in vertically heterogeneous media, meaning their propagation velocity depends on frequency. Different frequencies sample different depths, thus providing information about the Earth’s elastic properties at various depths.
Receiver functions are seismic waveforms derived from teleseismic earthquake recordings, used to investigate discontinuities in the Earth’s crust and upper mantle. These discontinuities generate converted seismic waves (P-to-S conversions), which manifest as complexity in recorded waveforms. By deconvolving horizontally and vertically polarised seismic waveforms, one isolates these converted waves, obtaining the receiver function.
Both surface-wave velocities and receiver functions are primarily sensitive to subsurface shear-wave velocity (\(V_S\)), though they are also sensitive to compressional-wave velocity (\(V_P\)) and density (\(\rho\)).
The problem of finding a subsurface model that best fits observed dispersion curves and receiver functions is inherently nonlinear and non-unique, with multiple distinct models potentially explaining the data equally well. Typically, the Earth’s interior is discretised into layers overlying a homogeneous half-space, each characterised by thickness, \(V_S\), \(V_P\), and \(\rho\). In this tutorial, we implement a flexible discretisation via Voronoi tessellation using the bayesbay.discretization.Voronoi1D class, enabling the number of layers itself to be inferred from the data.
This tutorial comprises: