This book describes the theory and practice of seismic interferometry,with an emphasis on applications in exploration seismology. It is written at the level where it can be understood by physical scientists who have some familiarity with the principles of wave propagation, Fourier transforms, and numerical analysis. The book can be taught as a one-semester course for advanced seniors and graduate students in the physical sciences and engineering. Exercises are given at the end of each chapter, and many chapters come with MATLAB codes that illustrate important ideas.

Correlating a pair of recorded seismic traces with one another and summing the resulting correlogram for different shot records is the basic processing step of seismic interferometry. This typically results in a new trace with a virtual source and/or receiver location, also known as a redatumed trace. The redatumed trace simulates a trace as if a real shot and/or receiver were at the new datum.

The redatuming procedure has been used by the exploration geophysics community since the early 1970s, except one of the traces in the correlation pair is computed by a numerical procedure such as ray tracing while the other trace is naturally recorded. Ray tracing, or more generally numerical modeling, uses an imperfect model of the Earth's velocity distribution which leads to defocusing errors in model-based redatuming. In contrast, seismic interferometry is free of such problems because it only uses recorded traces in the correlation. This freedom also allows one to utilize all of the events in the trace, including higher-order multiples and coherent noise such as surface waves, leading to enhanced resolution, illumination, and the signal-to-noise ratio in the reflectivity image.