Geoscience Reference
In-Depth Information
Fig. 6.1
Principle of the inversion method.
wavelet to get back to the reflection coefficient series, and from them derive the layer
impedances. It has to assume that the starting seismic data are free from correlated
noise (e.g. multiples). Also, the wavelet present in the data has to be estimated; many
inversion methods derive the wavelet from a well tie and have to assume that it does
not change laterally away from the well. There is also an amplitude calibration to be
taken into account; real seismic traces are not directly output as reflection coefficient
values, but are scaled to give some convenient but arbitrary rms average over a trace. At
least over a limited time-gate, the ratio of reflection coefficient to trace amplitude has to
be constant if inversion is going to work. Care is needed during seismic processing to
avoid steps that might introduce artificial amplitude changes vertically or horizontally.
However, locally variable effects in the overburden (e.g. shallow gas) can reduce the
seismic energy penetrating to deeper reflectors and so reduce the reflected signal; left
to itself, the inversion process would try to interpret this as a decrease in impedance
contrast across the deeper interfaces. To remove such artefacts, a long-gate AGC may
be applied, which scales amplitudes so as to remove lateral variation when averaged
over a TWT interval of 1 s, for example.
Another issue is that the seismic traces contain data of limited bandwidth; the fre-
quencies present depend on the rock properties and the seismic acquisition technique,
but might be in the range from 5 to 50 Hz. This means that the low frequencies in
particular, which are critical for the estimation of absolute impedance values, cannot
