In the early nineties in the United States, research in seismic wave propagation has been eclipsed somewhat by observational studies. This trend may be neither secular nor unhealthy, because seismologists need to examine real data as well as generate synthetic data. Much can be gleaned from new data sets using existing interpretation tools. Nevertheless, as attention focusses on smaller-scale velocity features in both crustal and global studies, scattering, wave conversion and diffraction effects will impede geological inferences. Ambiguities in interpretation will persist without further advances in wave propagation theory. Recent US research has begun to address these challenges.
Aside from its own applications, activity in theoretical seismology is important for the training of savvy observational seismologists. If current trends in graduate-school enrollment continue, there will likely be fewer US-born and US-trained theoretical seismologists, so that the United States will continue to import such expertise from other countries, in the form of graduate students, professors and industrial researchers. One way to combat this tendency is to improve US undergraduate programs in geophysics. Many professors have exploited the increasingly user-friendly computer and networking interfaces to engage students in theoretical and computational projects that, two decades ago, could qualify as dissertation projects. In this context, one might hope to see an increase in the publication and distribution of flexible computer codes to use as teaching tools, especially those that visualize wave propagation effects e.g. Wysession and Shore [1994].
Acknowledgments. This work was supported by Air Force Office of Scientific Research contract F49620-94-1-0043.