Development of a high-resolution shallow seismic reflection system.

The aim of this study was to investigate the applicability of the seismic reflection method for exploration at shallow depths ( 100m) and to develop a high-resolution system capable of achieving such objectives. This would have a wide range of applications including engineering scale surveys and hydrogeological studies. The system used comprised high-frequency response equipment coupled with a data-logging microcomputer. Trial surveys were made at two sites in the Bardon Hill area (Charnwood Forest) where Triassic sediments (0-100m thick) overlie a Precambrian basement, providing large acoustic impedance contrasts. Site 1 was characterized by a thin (1-2m) weathered layer, shallow water table (2m) and a firm topsoil while in Site 2 the weathered layer was 2 to 4 times as thick as that in the first site, the water table was deeper and the soil was cultivated. The variations in near-surface material strongly affected the quality of the results. Detailed comparisons were made of a number of different modifications of the signal production, data acquisition and data processing aspects of the seismic reflection system. Among the most important factors needed to produce the most successful reflection system was the nature of the seismic energy source: the 'Buffalo gun' proved most effective in providing high frequency energy required for this scale of survey; modifications to it in the course of this study further increased its effectiveness. The high-frequency (100Hz) geophone and the nature of its coupling with the ground were crucial in extracting and recording high-resolution data. Another important element in the system was the field microcomputer, useful for storing data and providing immediate quality control of data in addition to being a cheap processor. Both preliminary processing on a microcomputer and a standard processing package on a mainframe computer were used. The effectiveness of all types of process was dependent on the quality of the seismic field records. Preliminary processing was adequate to produce a satisfactory image of subsurface geology provided that field data were of high quality (in Site 1) and a number of programs were developed to enable this. The common mid-point stacking technique provided an improved image of shallow subsurface structure by increasing the signal to noise ratio and enhancing reflections. The most useful advantages of using the standard processing package were in improving the continuity and increasing the resolution of the reflections by the application of residual statics and deconvolution, respectively. The dependence of signal quality on variation in physical properties of the near-surface material were analysed in detail. The thickness of the weathered layer and the firmness of the topsoil were the most important factors affecting the transmission and recording of high frequencies. The reflection data can be directly compared with both seismic refraction results and borehole logs obtained along the same survey lines. The reflection data provided a significantly more accurate model of the sub-Triassic surface than did the refraction data; in addition, they provided information on the internal structure of the Triassic strata. The resolution and accuracy of the reflection data demonstrate the applicability of this method in engineering geophysical investigations. Further refinement of this system may be effected by investigating the use of lower-energy sources and more powerful field microcomputers with additional software.