Conventional amplitude-versus-offset (AVO) analysis such as linear fit
ting of isotime samples after NMO correction does not give reliable re
sults in the presence of interfering reflections or velocity errors. F
or this reason we propose a new method that is able to remove interfer
ence effects on the AVO of the target reflection and minimize the effe
cts of residual moveout. The method is based on the minimization of th
e difference between observed data and a model that includes theoretic
al descriptions of the AVO and traveltimes. This minimization is carri
ed out jointly with respect to AVO and kinematic parameters (velocitie
s and traveltimes) and requires the a priori knowledge of the propagat
ing wavelet. The kinematic parameters are given by the NMO equation ex
tended to the fourth-order term and the AVO is described as a linear c
ombination of a set of orthogonal functions. The AVO functions are der
ived from a statistical model of reflection amplitude in the presence
of velocity error. Applications to synthetic and real data demonstrate
the ability of the method to attenuate distortion effects on the AVO
of the primary reflection by interfering coherent noise. The real data
example pertains to a marine case where the primary event is contamin
ated by multiple reflections generated in the water layer and by anoth
er event, reflected from an interface deeper than the target, that mer
ges with the examined reflections at the far-offsets. This causes a su
btle tuning that distorts the AVO of the target. Our method attenuates
the effects of the multiples and discriminates the 'double event' nat
ure of the target reflection at the far-offsets.