** Next:** Attempts to invert FEAVO-affected
** Up:** Vlad: Velocity estimation for
** Previous:** Introduction

Beyond the mere ``historical'' interest, previous work on defining and
describing the FEAVO effect can help answer some potential objections
to the proposed approach.
For example, one objection might be that FEAVO effects are randomly
distributed.
Therefore, their expression in AVO pickings is the same
as that of random noise, and can be treated the same
way.
White et al. (1988)
showed that focusing and defocusing
caused by a random distribution of small velocity anomalies will not
lead to the canceling of the FEAVO effects by each other.
Moreover, they prove that
even if the
distribution of velocity anomalies is random, the probability that
FEAVO effects appear is not randomly distributed with the
distance from the shot.
Therefore, FEAVO presence cannot be equated
with random noise in AVO picking, since
FEAVO-causing anomalies are not randomly distributed in real data.
In the case of
shallow anomalies the distribution is influenced by depositional
patterns [the ``Rotten alligators'' in Claerbout (1985)], and
in the case of deep-origin FEAVO, by the presence of faults
Hatchell (2000). Besides, the amplitude of the FEAVO
effects is often so large and spatially extended that it renders AVO
analysis impossible.

** Next:** Attempts to invert FEAVO-affected
** Up:** Vlad: Velocity estimation for
** Previous:** Introduction
Stanford Exploration Project

11/11/2002