Quantitative 3D+4D seismic imaging, inversion and monitoring inside the earth (I)

David Lumley
University of Texas at Dallas

Seismic waves can be generated by natural sources (earthquakes, ocean waves, ambient noise...) and by controlled manmade sources (airguns, explosives, vibroseis...).
Seismic waves propagate through the earth to first order as elastic compressional and shear waves, and create complex wavefields containing reflections, diffractions, refractions (diving, head waves), surface waves (Rayleigh, Love waves), trapped, guided and evanescent waves, and higher-order scattered and transmitted wave types.

To first order, the wave propagation velocity (travel times) and amplitudes are controlled by the compressibility, shear strength and density of the earth as a function of its heterogeneous rock physics properties (mineralogy, pore geometry, pore fluid content, pressure, temperature...), which are in turn controlled by geologic, fluid-flow and other physical processes in the earth, on diverse time scales from minutes to millions of years.

In modern seismic imaging and inversion methods, we use the seismic wavefield recorded at the earth's surface by arrays containing thousands of sensors as a boundary condition to a wave equation, and downward continue this 'receiver' wavefield in reverse-time back into the earth, and cross-correlate with a forward-modelled 'source' wavefield, to image scattering points and surfaces in the 3D earth (geologic structure, and rock/fluid heterogeneity). We can furthermore set up a variety of inverse problems, where we estimate the rock and fluid physics properties in the earth such that the modeled/simulated seismic wavefield matches the real seismic data recorded in the sensor arrays, within some misfit criteria. Since rock and fluid properties can be time-variant, for example caused by injection or withdrawal of fluids at well locations, we also use time-lapse '4D seismic' data recorded at multiple calendar times to estimate time-varying changes in rock and fluid properties, and thus image the movement of fluid-flow fronts related to changes in fluid saturation, pressure, temperature etc.

For this 3-hour IPAM tutorial lecture, I will present the basic physics and math of quantitative 3D+4D seismic imaging and inversion, and show a variety of cool real data examples from various projects around the world.

Presentation (PDF File)

Back to Computational Issues in Oil Field Applications Tutorials