Seismic transdimensional inversion tools for flat and isotropic layered structures in the ocean (SEIS-FILO)

Abstract:

Seismological studies have approximated the Earth’s structure using the stratification of flat and isotropic layers. Such one-dimensional models are suitable for single-station measurements, including receiver functions and surface wave dispersion curves, and have been widely used to illuminate a wide range of structures from shallow basin to deep mantle structures. Recently a growing number of studies come to use more sophisticated methods (e.g., full-waveform inversion) for modeling laterally heterogeneous structures of the Earth. Nevertheless, the modeling by one-dimensional structure will remain its position because of the less computational cost and no need for dense seismic instrument deployments. Especially for the ocean area, insufficient coverage by station networks will force to use such single-station methods.

Here, I present a software that performs transdimensional inversion of receiver functions and surface wave dispersion curves, with the effects of the sea water column fully considered. State-of-the-art Bayesian probabilistic sampling known as reversible-jump Markov-chain Monte Carlo is used, where the number of model parameters (i.e., the number of layers) is automatically selected based on the data complexity. The parallel tempering technique is also implemented for efficient sampling from model space. The source codes are written in a modern Fortran programming style (i.e., objective-oriented programming) a low-maintenance and easy expansion in the future. The package also provides plot utilities written in Python3, so that users can easily visualize the inversion results.