Is there value in highly spatially sampled zero-offset vertical seismic profiles?

Abstract

Distributed vibration sensing (DVS) involves using a fibre optic cable to record seismic data with sub-metre spatial sampling. The results presented here, however, show that there is little point in recording highly sampled VSP data to obtain detailed interval velocity profiles due to the high sensitivity of the velocity estimation to very small errors in the pick times. Highly sampled data therefore requires the application of a pick smoothing algorithm which results in a profile that is generally consistent with that obtained from raw data with a larger spatial interval. In velocity profiles generated for DVS data, the smoothing effect of the gauge length further reduces the value of high spatial sampling. If a corridor stack is required, acceptable results, for high SNR data, can be obtained even when the minimum sampling criteria are not met if median filters are used for wavefield separation. If significant shear- or tube-wave noise is present in the data, this noise may be difficult to remove if the spatial sampling criteria is relaxed. If the SNR is low, as it generally is for DVS surveys, then acceptable corridor stacks can be obtained if the data is highly sampled.