On the use of airgun arrays for seismic refraction investigations of the crust

During the last decade our knowledge of the structure and composition of the Earth's crust has greatly increased through the application of new techniques in seismic investigations. On the one hand, numerous deep reflection studies have been carried out around the globe. On the other, refraction experiments have been improved in that observations have been made with very dense station spacings to avoid the spatial aliasing problem. A comparison of the two methods by Meissner et al. (1983) and a recent review by Mooney & Brocher (1987) showed that a combination of both methods is highly desirabie. Braile & Chiang (1986) demonstrate that the two methods contain partly different information when applied to the continental crust/mantle boundary, and supplement each other. For deep crustal refraction observations, large distances (150-300 km) between sources and receivers are necessary. This was hitherto achieved by using large explosions that provided sufficiently strong signals at the distances required. But this also limits the application, since it is not always possible to find suitable locations for firing large explosions. Furthermore, the shotpoints are separated by several tens of kilometres in normal refraction surveys, thus providing only a low multiplicity of the seismic phases. In this paper we describe results from experiments using an airgun array as the source. While the recording stations were kept fixed, the ship with the airguns moved along the profile. The speed of the ship and the firing interval control the spacing between the shotpoints. We demonstrate that airguns are highly effective and that, under favourable conditions, the signals can be detected at distances of more than 200 km from the source. To improve the quality of the data and to aid in the interpretation, the resulting dense data sets may be processed in ways that traditional data could not. Techniques that are of common use in common midpoint (CMP) reflection work can be modified to be applied to dense refraction/wide-angle reflection data, since these data are not spatially aliased with respect to frequency and wavelength. We have used slant stacking techniques to improve the signal-to-noise ratio, and have applied velocity filtering and deconvolution algorithms in the delay time-slowness (tau-p) domain to enhance the coherent phases and to suppress multiples and shear waves.