The Influence of Seismic Scattering on Monitoring

Abstract

Seismic scattering impacts monitoring of underground nuclear tests through its effects on seismograms. For a Comprehensive Test Ban Treaty, the effects of scattering will be most pronounced for regional seismograms in the frequency range 1–20 Hz. Several methods have been used to examine scattering theoretically. Scattering from a single object shows that scattering may be divided into small-scale where objects are a wavelength or less in size, typically characterized by backscattering; and large-scale, the cause of fluctuations, blockage of phases, and deviation of ray paths. For many objects, single scattering theory treats the problem as an extension of one object by ignoring interactions between scatterers. Transport theory uses the conservation of energy, may be extended to multiple scattering, and has been particularly useful in explaining coda. A more difficult problem involves combining many scatterers with layered structure. Up to now this has mainly been treated by finite difference techniques, but other methods are under development. The effects of scattering predicted by theory, and observed in the field, are attenuation, fluctuations, wavetype conversion, and coda. One of the results of analysis of these phenomena is the characterization of small-scale earth structure as a self-affine random medium with velocity variations of 2–5% and a decreased number of scatterers at scales greater than 1–10 km. Attenuation due to small-scale scattering may be parameterized by frequency-dependent Q; an effect of large-scale structure that has received attention is Lg blockage due to zones of crustal thinning. Fluctuations reduce the coherency of seismic waves at separations of a wavelength or greater. A particular problem with transmission through structure with large-scale lateral variations is inaccuracy of locations due to distortion of ray paths. Wavetype conversion produces energy on components where there would be none in a plane layered medium. Again, a pressing problem is the generation of S waves from explosive sources, since this affects identification of sources. Coda allows investigation of attenuation and determination of source size.