Tools for Seismic Data Analysis
There are three main approaches to seismic analysis for predicting the lithology and fluid content of a reservoir. Each approach obtains an estimate of one of the geophysical parameters:
Velocity analysis for Vp.
AVO
analysis for SIGMA.
AAI estimation for AAI.
A full description of each analysis tool is in our
Tools Overview report.
A comparison of the different techniques to perform each type of analysis is presented in the comparison section of the Tools report. In summary:-
Each reflection on a CMP gather of seismic data follows an approximately
hyperbolic path as it is traced from one offset to the next. The hyperbola's
shape gives an indication of the normal move-out velocity of the material above the reflector. A simple mathematical relationship can translate these
normal move-out velocities to the interval velocities of the material between the reflectors. This technique has three major limitations:
- The errors involved in estimating the normal move-out velocities
(approximately 2% error).
- The error in the estimate of interval velocity rapidly increases as
the thickness of the layer decreases. For most seismic data it is impractical to estimate the velocity of layers of less than 100ms TWT.
- The assumptions used to transform normal move-out to interval
velocities (e.g. flat layers, isotropic media).
Best accuracy for interval velocity estimation: approx. 250m/s for a 100m sand, 330m/s for a 75m sand.
The amplitude of a reflection changes as the offset between the source
and receiver increases. The analysis of this variation is called AVO analysis. It provides information on the change in Poisson's ratio between the material above and below a reflecting interface. Currently the Poisson's ratio contrast is estimated by an iterative modelling analysis, where the form of amplitude behaviour on the seismic data is modelled with synthetic seismic data. However, this approach gives a non-unique estimate of Poisson's ratio. Multiple reflections on the CMP gather will increase the uncertainties.
Best accuracy for interval Poisson's ratio estimation: approx. 0.05.
The amplitude of a primary reflection on a migrated seismic section is related to the contrast in acoustic impedance of the material above and below a reflecting horizon. A simple mathematical transformation could be used to obtain the acoustic impedance of the sub-surface if the seismic data
contained all frequencies. However, the low (below 10Hz) and high (above 50Hz)frequencies are not present in seismic data. The low-frequency
information has to be estimated from geological information and an interpretation, before AAI estimates can be obtained from the seismic data. Consequently, the AAI estimates are non-unique and, their accuracy limited by the reliability of the low- frequency behaviour.
Best accuracy for AAI estimation: approx. 500 (m/s)(gm/cc)
The variation in the AAI of the sub-surface determines the appearance of the
processed seismic section. Three techniques try to establish a direct relationship between the form of the processed seismic section and a particular variation of the sub-surface lithology. They are:
- Amplitude Analysis.
- Seismic Modelling.
- Geostatistical Analysis.
This type of analysis is for
thin beds",
less than half a seismic wavelength in thickness(~25m), which are surrounded by a homogeneous medium, eg. the sand reservoirs of the Gulf of Mexico. Reflections from the top and base of the reservoir cannot be resolved for thin beds. Consequently, the thickness of the bed is estimated from the amplitude of the combined reflections. Amplitude analysis is useful, but its application is limited by the assumption that the reservoir is encased in a homogeneous medium.
Modelling is used to test the validity of different interpretations of the
Best accuracy for interval Poisson's ratio estimation: approx. 0.05.sub-surface. We generate a
synthetic
seismic section for a proposed geological model of the sub-surface and compare it with the seismic section. The geological model may be modified, depending on the match. Non-seismic information is critical to the
effectiveness of this approach to interpretation. The variety of possible
interpretations will be confusing unless the model is tightly constrained
by geological considerations.
Homogeneities in the reservoir can result in subtle variations of seismic
reflection character. Statistical analysis is used to quantify these variations and relate them to geological variations in the reservoir. Geostatistical analysis is very new to the interpretation of seismic data and will be the subject of much research in the next few years. The main limitation to the technique is that it is an empirical approach to understanding the variations seen on the seismic data. No firm relationship exists between the statistical measurements and the underlying geological variations to be mapped.
