Source ghost generation: observations from a dual near-field hydrophone test

Abstract

In this paper, we investigate source ghost generation using experimental data acquired using a dual-string air-gun array instrumented with two near-field hydrophone (NFH) channels per source element instead of the usual single NFH per source element. The extra NFHs enable us to solve directly for source ghosts, sidestepping the requirement for a ghost model. Experience has shown that the standard ghost model does not always represent well the physics of ghost generation for an air-gun array. In practice this model is typically parameterized via a frequency-dependent effective reflectivity, the magnitude of which often needs to be reduced more than expected for the effects of rough sea-surface scattering. The NFH recordings in this test are compared with synthetic data to highlight problems with the standard model of ghost generation and the results suggest that acoustically induced cavitation is responsible for the observed reduction in the amplitude of the ghost. We show examples of de-signature on seismic data using operators derived with and without a ghost model and discuss the merits and potential issues going forward.