Inversion Fundamentals Case Study
Inversion Fundamentals Case Study
GeoConvention 2020 1
Figure 1a shows the stacked seismic data with horizon interpretation and Vclay logs. Within the
reservoir, between wells B and C, there is a break in the light-blue horizon, with one possible
interpretation shown. The horizon slice (Figure 1b), 20 ms above the base of the reservoir,
indicates that this possible interpretation is plausible; consistent amplitudes at the wells on either
side of a higher amplitude zone. In fact, only the logs below the light-blue horizon indicate that
something more may be going on, with an increase in Vclay below the horizon at well C.
Careful analysis of the amplitude variations, taking into account the structural changes, and
integrating the well-log data are all crucial components of classic interpretation. These
approaches should not be abandoned, rather they should be assisted by the inversion process,
which considers exactly the same factors in an analytical manner.
Prestack Inversion
As discussed in the introduction paper, inversion transforms prestack seismic amplitude changes
into models of P-impedance, S-impedance, and density. To achieve this outcome, inversion uses
the following components in its calculations:
Component Purpose
- Prestack seismic gathers - Amplitude changes with incidence angle according to properties
- Wavelet - Accounts for the effects of reduced resolution of boundaries
- Reflectivity equations - Relate seismic amplitudes to impedances and density
- Well impedances/density - Provide the correct magnitude of values as a starting point
Processes that are carried out as part of the inversion are done with the goal of making these
components more closely meet their purpose. For example: noise attenuation on the prestack
seismic gathers ensures that amplitude changes are a result of the rock properties, rather than
unpredictable sources; selecting an appropriate window for the wavelet extraction is done to
match the frequency content of the zone of interest, rather than data above or below.
The various steps for the inversion are illustrated using the seismic data example discussed. The
final volumes of P-impedance, S-impedance, and density can be transformed into a multitude of
other elastic parameters. The density section corresponding to the seismic stack shown in Figure
1a is shown in Figure 2. It is now apparent that while there does appear to be a discontinuity
GeoConvention 2020 2
Figure 3. a) Histogram of density values from McMurray well logs, coloured by Vclay. The higher-clay
points are centred at higher densities, making density a useful attribute for facies distinction. b) A
crossplot of MuRho vs. density showing well points coloured by facies. The choice of crossplot axes
provides a good separation of the different facies types.
between wells B and C, the properties on either side of this discontinuity are not, in fact, similar.
Well C shows as having high density, while the discontinuity between the wells is comparatively
low. It is now a valuable function of the interpreter to explain what these differences mean in a
geological context.
Geological Classification
The critical link between geological properties and elastic properties is found through either well
analysis or rock-physics modelling. In this example we will limit ourselves to well data, which
matches elastic properties from the sonic, dipole sonic, and density logs with geological properties
from other petrophysical curves and analyses. Each reservoir can have unique characteristics,
and for the McMurray, density is a key property for distinguishing between the different facies with
varying amounts of clay content.
Figure 3a shows a histogram of density values for all well-log points in the McMurray reservoir
interval. The histogram points are coloured by Vclay and it is apparent that the points with high
clay content tend to have higher densities. By identifying multiple attributes that distinguish
geological properties, useful interpretive crossplots can be created. Figure 3b shows a crossplot
of MuRho vs. density, where the points are coloured by facies. There is a good separation of the
sands, IHS facies, and muds. These crossplots of well data provide a template for interpreting
the equivalent data from seismic inversion.
The seismic inversion volumes are converted to include a range of elastic properties, including
MuRho. Along with density, these points are crossplotted and compared with the well template
(Figure 4a). In this case, interpretive cutoffs are applied to the data, corresponding to the log
GeoConvention 2020 3
Figure 4. a) The same crossplot shown in Figure 3b with seismic points overlain. The seismic and well
data correlate well, allowing for the well facies to be interpreted on the equivalent seismic points. b) A
horizon slice extracted 20 ms above the base of the reservoir through the classified seismic volume. The
facies interpretation shows a more clear picture of the geology than from amplitudes alone (Figure 1b).
facies. A time slice matching Figure 1b is shown through the classified inversion data in Figure
4b. Here is it much more apparent that the properties at wells B and C are more distinct.
Conclusions
Seismic inversion, specifically prestack AVO inversion, is a useful tool for extracting quantitative
rock properties from the seismic amplitudes. By calculating properties such as P-impedance, S-
impedance, and density through the inversion process, a number of useful elastic properties can
be determined. While useful on their own, the elastic properties are more functional when
interpreted into a geological context based on well or rock-physics templates.
Using an example from the McMurray oil sands, I demonstrate the steps needed to perform an
AVO inversion and where special consideration to parameters is important. The results of the
inversion are then compared to the conventional options for interpretation. While consistent
with one another, the advantage of the interpreted inversion is that the geological context is
more immediately apparent.
Acknowledgements
Suncor Energy Inc. is gratefully acknowledged for permission to use the dataset and for
contributions to the case study project.
GeoConvention 2020 4