Those publicly funded databases that

Abstract

We reformulate the usual approach to estimation of air-gun signatures from near-gun records as a least-squares inversion. We show that this has advantages compared to the commonly-used iterative method; we are able to more accurately treat the motion of the air bubble from each gun as the bubble moves away from the hydrophones, we are able to relax the constraint in the current method that the number of recording hydrophones must equal the number of guns in an array, and we are also able to relax the constraint that hydrophones must be placed near to a gun. We show by a singular value analysis, however, that when additional hydrophones are deployed the derived signatures are likely to be most accurate when the hydrophones are close to the guns (e.g. not in a mini-streamer or similar arrangement). Example directional far-field signatures derived by our approach compare well with modeled signatures and are suitable for use in processes such as shot-by-shot signature deconvolution.