Seismic noise‐reduction techniques for use with vertical stacking: An empirical comparison

Abstract

A systematic comparison of a wide range of noise‐reduction techniques applied to a single data set collected for this purpose was attempted. The study includes a comparison of the relative benefits of noise‐reduction schemes for very different noise levels. The comparisons use computer simulations of field techniques for noise reduction as applied to deep crustal, 48-fold VIBROSEIS® data. The noise‐reduction techniques involve amplitude scaling (diversity stacking and automatic gain control), noise editing (the use of fixed‐gain and self‐updating noise‐rejection systems with and without buffered memory for recursive editing), and reduced numbers of recording bits (mantissa‐only and sign‐bit‐only formats). The effectiveness of each noise‐reduction procedure is assessed by a study of its effects on noise levels seen on source‐point gathers and on both true‐amplitude and amplitude‐balanced common‐midpoint (CMP) stacks. For low levels of ambient noise, true‐amplitude CMP stacks can be substantially improved by the appropriate noise‐reduction techniques, but CMP stacks incorporating gain control before CMP stacking show only minor improvements. In contrast, when a high level of ambient noise is present, both true‐amplitude and gain‐controlled CMP stacks can be greatly improved by appropriate noise‐reduction processing before vertical stacking. Of procedures involving zeroing of noise bursts, self‐updating noise‐rejection systems were more effective than fixed‐gain noise‐rejection systems for all the conditions simulated here. A system incorporating recursive editing techniques is still more effective at low levels of ambient noise, but this system tends to edit recorded traces too severely at very high noise levels. Mantissa‐only and sign‐bit‐only recording give very similar results, and show an effect comparable to that given by self‐updating editing systems. Diversity stacking produced significant noise reduction in all conditions studied, and may be the most widely applicable and most generally useful of the noise‐reduction methods studied here. ®Trade and service mark of Conoco Inc.