Spike recovery deconvolution: in search of a scientific field test

Abstract

In spike recovery deconvolution (SRD) (Fullagar, 1985) it is assumed a priori that the reflectivity consists of event spikes separated by zeros, and a spiky reflectivity is constructed which, when convolved with the wavelet, reproduces the seismogram to an accuracy consistent with the perceived level of noise. The principal attraction of SRD is improved resolution for which the price is increased computation relative to conventional linear deconvolution.