2D Synthetic Wave Propagation Forward Modeling Toward High-Resolution Tomography of Mount Fuji

High-resolution imaging of volcanic structures relies on accurate forward modeling of teleseismic
wavefields. In general, full waveform inversion requires global wavefield simulations, which are compu-
tationally expensive. To reduce this cost, localized waveform inversion or “box” waveform tomography
approaches have been developed. These methods compute the global wavefield once and inject the re-
sulting displacement and stress fields at the boundaries of a local region of interest. However, even
this initial global simulation can remain computationally demanding and depends on the choice of a
background model, such as one-dimensional or three-dimensional long-wavelength tomographic models.
Monteiller et al. (2021) have shown that this step can be significantly sped up by substituting the full
wavefield with a single plane wave, which is a valid approximation for teleseismic earthquakes. Moti-
vated by these results and in the context of Mount Fuji, we further simplify the forward modeling by
considering small regions, typically 100 × 100 × 50 km, a scale at which incoming teleseismic P waves
are expected to be close to planar at the boundaries. We present a series of synthetic experiments in
a two-dimensional Cartesian framework, placing seismic sources at various locations outside the study
region. We simulate wave propagation in background media defined by one-dimensional Earth models,
onto which two-dimensional slab-like perturbations are added. From the resulting wavefields, we esti-
mate slowness vectors and apparent velocities at the boundaries of the box and quantify deviations from
an ideal planar arrival, which allows us to assess the validity of the plane wave approximation. Thus,
we provide an efficient forward modeling strategy for high-resolution imaging of Mount Fuji and similar
volcanic systems. More generally, our results enable us to define practical criteria under which plane
wave injection can be justified.