Full waveform inversion

Full waveform inversion estimates subsurface properties by iteratively adjusting a wave-based Earth model until its simulated seismic records match observed data.

Wave equation: A partial differential equation describing how waves propagate through space and time. Forward simulations solve this equation to predict seismic recordings from a proposed Earth model.
Seismic wave: An elastic disturbance that travels through Earth materials and can be recorded at the surface or downhole. These propagating disturbances carry the waveform information that inversion fits.
Velocity model building: The estimation of spatial seismic velocities used to represent the subsurface. Full waveform inversion can refine velocity models used for imaging and depth positioning.
Travel-time tomography: An imaging method that infers subsurface structure from seismic-wave travel times. It fits arrival times rather than the complete recorded waveforms.
Low-frequency seismic data: Seismic measurements containing long-period wave energy that constrains broad subsurface variations. Low frequencies can reduce cycle skipping and provide a stable starting point for inversion.
Forward modeling: The calculation of predicted observations from a specified model and physical laws. Each inversion iteration predicts seismic data by propagating waves through the current subsurface model.
Seismic reflection: A seismic wave returned from a boundary where subsurface properties change. Reflected arrivals encode contrasts that help constrain subsurface structure.
Subsurface imaging: The reconstruction of underground structure and properties from geophysical measurements. Full waveform inversion is one approach to turning seismic observations into subsurface models.
Migration (seismic imaging): A seismic processing method that maps recorded reflections to their likely subsurface locations. Migration images reflectors, while full waveform inversion updates a physical model through data fitting.
Multi-scale optimization: An optimization strategy that solves a problem progressively across scales or resolutions. Starting with low-frequency or smoothed information can help the inversion reach useful solutions.

Linked from 3 pages