Practical approaches to maximizing the resolution of sparker seismic reflection data

Sparkers are a type of sound source widely used by the marine seismic community to provide high-resolution imagery of the shallow sub-bottom (i.e., < 1000 m). Although sparkers are relatively simple, inexpensive, and high-frequency (100-2500 Hz) sources, they have several potential pitfalls due to their complicated and unpredictable signature. In this study we quantify the source characteristics of several sparker systems and develop a suite of simple processing approaches for both single channel and multi-channel sparker data. In all cases, the results show improved vertical resolution and reflection coherency. Correcting for small static variations in multi-channel seismic (MCS) data is a critical first step to preserve the broad frequency content during stacking, and to reduce the shot-to-shot variability of outgoing and incoming signals. Application of predictive deconvolution to static-corrected, post-stack traces suppresses short-path multiples and restores the latent high-resolution reflection patterns. However, if shot-to-shot source signatures are recorded directly, pre-stack deterministic deconvolution followed by post-stack predictive deconvolution produces the most robust results. Processing sparker data without broadband techniques results in less confident or completely missed interpretations when compared to the broadband equivalent. If processed correctly, marine sparker data can provide exceptional sub-bottom imagery that rivals other more repeatable marine seismic sources (e.g., high-frequency air-guns).