A robust FWI gradient for high-resolution velocity model building

We describe a robust method to produce long-wavelength updates in gradient-based Full Waveform Inversion (FWI). The gradient is computed by applying dynamic weights in the velocity sensitivity kernel derived from impedance and velocity parameterization of the classical objective function. The new kernel implementation effectively eliminates the migration isochrones produced by the specular reflections and emphasizes the low-wavenumber components in the gradient in heterogeneous media. The new gradient is able to provide velocity updates beyond penetration depth of diving waves. We use a synthetic example to illustrate how this dynamically weighted FWI gradient successfully recovers the background velocity from pre-critical reflections. We apply the new approach to 2D and 3D dual sensor data from deep-water Gulf of Mexico. Results show how the dynamically weighted FWI gradient can combine both transmitted and reflected energy in a global FWI scheme and provide high-resolution velocity models without migration imprint in the updates.