Optimisation of Source and Receiver Geometry for Ultra High Frequency Seismic Data

Summary Ultra-High Frequency (UHF) seismic data, with vertical resolutions at decimetric scales, is more commonly being implemented in industry after being used in academia on a frequent basis. Adoption of multichannel arrays rather than single channel configurations leads to improved signal to noise ratios and subsurface imaging. This allows for a wider suite of advanced processing techniques to be applied requiring constraints on the way data is positioned and processed. In conventional UHF processing the geometry allocation consists of a basic and crude setup, where databases containing lateral and layback offsets are populated assuming straight streamer configurations. In fact, streamer shapes vary from the idealised straight model during acquisition leading to issues understanding the streamer geometry in 3-Dimensions. Casting the problem of geometry as an inversion allows us to overcome the deviations from a straight streamer model in data acquisition. Understanding the influence of geometry on travel-times for direct, primary and ghost arrivals enables the optimisation of a model using a Genetic Algorithm. The results of this model can then be applied to UHF data supergathers showing that the inverted geometry has significantly improved accuracy of reflection flattening using a best-case velocity model when compared against the conventional processing.