Seismic source modelling and 3D survey parameter design, Surat Basin, Australia

Abstract

Elastic synthetic seismograms for a dynamite and Vibroseis source are used to assist the selection of source type for a 3D seismic survey in the Surat Basin. The models are based on well logs and provide key information on seismic source characteristics, which are supported by seismic field data from the study area. The main targets are Permian sands interbedded with laterally extensive coal measures. Survey parameters such as source type, maximum and minimum source-receiver (S/R) offset, and seismic bandwidth are critical in detecting these sands especially if techniques such as AVO are to be used in subsequent data interpretation. The methodology for determining the detectability of the modelled gas sand for each type of source is to subtract the seismogram for a shale model from the seismogram produced when the modelled target sand is present. These difference seismograms computed in t-p space show that with ground roll suppressed the sands can produce more than a 20% change in the modelled seismograms. This figure improves with increasing p value (or equivalent target offset) but is limited by interference from shear waves especially for a Vibroseis source. RMS amplitude attributes (within a window) are used to quantify this change. In the study area it is shown that when compared to a buried (25m) dynamite source, Vibroseis sources produce significantly larger amounts of ground roll on the near traces (offsets < 600m at target) and of shear waves on the mid to far traces (offsets > 1.4km). Furthermore, because the very strong near-normal incidence reflections from coal measures decay rapidly with offset, target sand detectability tends to increase significantly for larger offsets for both source types. For dynamite sources, detectability is relatively high for source-receiver offsets greater than about 1km, out to at least 2km. Some deficiencies of the Vibroseis source may be reduced by careful noise filtering, however the lower sample density for most 3D surveys (compared to 2D) means that such filtering techniques may be less effective. The study highlights the benefits of seismic source modelling to assist survey design.