Pre-Stack Wavefrom Inversion: Current Status And Future Directions

Pre-stack seismic waveform inversion is a highly<br>challenging task. The non-linearity and nonuniqueness<br>together with compute intensive forward<br>modeling often make the problem intractable. Much<br>progress has been made in the last decade in the<br>application of global and local optimization methods<br>in model parameter estimation, data fitting and<br>uncertainty estimation. In isotropic media, we found<br>that a hybrid multi-stage waveform inversion<br>approach based on prior travel time and AVO<br>analysis in the plane wave (delay time – slowness or<br>τ−p) domain is fairly robust and computationally<br>efficient. Much of current research is focused on<br>including the effects of transverse isotropy on AVO<br>and pre-stack analysis. Unlike isotropic media, P<br>wave normal moveout analysis in the delay time-ray<br>parameters (τ−p) domain reveals that only two<br>anisotropy parameters, namely, elliptic velocity (αel)<br>and k can be derived successfully in a weak<br>transversely isotropic medium.