Ho Times Attributes

Time Attributes and Their Application
In Reservoir Characterization
Time Attributes and Their Applications in Reservoir Characterization
By : Sigit Sukmono 1
1. Dip and Azimuth
¾ The principle on how to calculate the dip and azimuth is quite
simple as shown in Figure 1.
¾ Basically the dip and azimuth are the magnitude and direction
of the time gradient vector; from a local reference, which
computed on every horizon sample which is interpreted.
¾ The computations are carried out by fitting a plane though
adjacent data points and posting the calculated values at the
central data point (Figure 1).
¾ The dip values can be expressed in degrees or radians, or
more commonly, in miliseconds per meter.
¾ Dip and azimuth values are normally displayed on separate
maps. These maps should be studied independently because
faults, which affect the mapped horizon do not necessarily
show up clearly.
¾ This is illustrated in the following examples.
¾ Figures 2 shows the time map of a horizon representing the
top of a reservoir (horizon A), whereas Figure 3 and 4 show
the corresponding dip and azimuth display.
¾ When both maps are compared, the azimuth map clearly
provides a better definition of the faulting than dip map.
T
azimuth
Dip
Figure 1. The principle of dip and azimuth computation
Figure 2. Time structure map of horizon A
(Rijks & Jauffred, 1991)
Figure 3. Dip map of horizon A (Rijks & Jauffred,
1991)
Figure 4. Azimuth map of horizon A (Rijks & Jauffred,
1991)
¾ The explanation for the difference in the clarity of faulting
between the two attributes can be explained as follows. A
fault will be clearly defined on the azimuth map when the dip
direction of the fault plane is opposite to the dip direction of
the beds. It will be poorly expressed when the fault plane dip
direction is similar to that of the horizon.
¾ A fault will best expressed on the dip map when the dip angle
of the fault plane is notably different from the horizon dip,
and will be poorly expressed when the dip angle is close to
that of the horizon.
¾ Figures 2 shows the time map of a horizon representing the
top of a reservoir (horizon A), whereas Figure 3 and 4 show
the corresponding dip and azimuth display.
¾ When both maps are compared, the azimuth map clearly
provides a better definition of the faulting than dip map.
2. Shaded Relief (Artificial Illumination)
¾ In a shaded relief display, a grid is produced in which the

values are proportional to the brightness of the reflection
when “illuminated” by a light source. The principle is
somewhat similar to looking at the map of a seismic marker
which would be illuminated by the sun shining from a certain
direction at a given elevation.
¾ Figure 5 illustrates the principle. Depending on the direction
of illumination, fault F will appear as a dark (A) to a bright
(B) lineament.
¾ Rijks and Jauffred (1991) gave an example of the application
of this method to the case of Amerveen gas field. The
principle reservoir here is the Rotliegend Sandstone (Figure
6).
¾ The southeast ward hading fault of this graben is not
illuminated and appears as a black line-up. The other fault
plane, hading in the opposite direction, is brightly illuminated
and appears as a white line-up (Figure 7).
¾ The throws of these faults are very small, less than 10 ms.
The graben shows only as two faults on the dip map with the
relief effect (Figure 8)
¾ Other differences can be seen between shaded relief and dip
display. The fault trend in the western part of the area, which
appears to be a continuous event on the dip map, has in fact a
distinct en echelon configuration on the shaded relief map
(the white arrow). This observation is of great interest as it
hints at fault block communication.
¾ It can be observed that how clearly the faults are imaged,
either un the dip or azimuth display is very much depends on
the orientation of the faults to the bedding plane.
¾ To overcome the problem of the difference in detectability
between one display and another, dip and azimuth map can be
displayed in combination.
¾ Figure 9 shows a combined dip and azimuth map of the
reservoir referred to above. The different colors in the wheel
represent the azimuth values while color intensity is an
indication of the dip angle (the steeper the dip, the darker the
color).
¾ This map shows that the synthetic faults throwing to the
southwest show in dark blue, and the antithetic faults
throwing to the northeast are in deep orange.
Intentionally left blank
16
Figure 5. Shaded
relief-artificial
illumination principle
(Rijks & Jauffred,
1991)
Figure 6. Seismic display
showing the Rotliegend
sandstone reservoir
(Rijks & Jauffred, 1991)
Figure 7. Shaded relief map of
Annerveen Field (Rijks &
Jauffred, 1991)
Figure 8. The dip map of the
Annetveen Field (Rijks &
Jauffred, 1991)
0 1 km
Figure 9. The combined dip-
azimuth map of the Nun
River Field (Rijks & Jauffred,
1991)
3. Similarity
Degree of similarity or dissimilarity can be used to enhance the

lateral changes of seismic data due to changes of geological
conditions. Several major algorithms are available to compute
this parameter and followings is the discussion on each
algorithm.
3.1. The Manhattan Distance / Semblance
¾ The Manhattan distance is a statistical measure of the

dissimilarity. In the continuity cube, the main similarity
measurement is in the cross-correlation, which compare the
center trace with neighbor traces by multiplying samples from
the center trace times those of a neighbor trace within a
window and normalizing by the autocorrelations of these two
traces.
¾ The resulting values will vary between –1 and +1 where, the
most discontinuous data, such as faults, resulted in number
close to - 1.
¾ The center trace window is compared to the target trace
window using a Manhattan distance and this Manhattan
measurement is output.
¾ The Manhattan distance is the sum of the absolute value of
the sample difference between the window. This sum
(numerator) is divided by the sum of the absolute values of
each sample of the two traces within the specified window.
¾ The resulting values are numbers between 0 and + 1, where
the most discontinuous data, such as faults, is closer to + 1.
Figure 10 illustrates the example on how to compute this
Manhattan distance.
¾ The number of neighbor traces to be compared with the
center trace generally varies between one to maximum of
eight surrounding traces. The length of data intended to be
covered in analysis can also be determined in the sliding
comparison window.
Figure 10. An illustration on how
the Semblance coefficient can
represent the geologic conditions
(Landmark, 1999)
¾ Semblance values are calculated for each center reference
trace to neighbor trace pair within the scan pattern.
¾ From all of these measurement, a single target neighbor trace
is selected. A second calculation computes Manhattan
distance dissimilarity using the central trace and the target
trace in the comparison window.
¾ This Manhattan distance value is output for the middle
sample of the center trace window.
¾ The low similarity values can result from the following
causes :
- Near vertical faults or low-angle faults
- Contrasts in seismic character due to stratigraphic or
lithologic change
- Highly dipping reflectors
- Lack of reflector (e.g. salt)
- Poor quality data
¾ Fault with detectable vertical throw is generally associated
with a narrow zones of low similarity. The something occurs
also if there is an abrupt contrast in seismic character due to
stratigraphic or lithologic changes such as in the case of
channel sands.
¾ Gradual stratigraphic contrast, such as these associated with
transgressive sequences will produce broad regions of
moderate values. Highly dipping pattern will result in broad
regions of low similarity zones with poor data quality of lack
of reflectors (e.g. salt structures) can also produce broad
region of low similarity.
¾ There are two methods generally used to determined which
windows on the neighbor traces will be used in the semblance
calculations. The first option uses unconstrained semblance to
scan the center trace to each neighbor trace. The second
option uses slant or dip-scan stacks to compute a best fit plane
orientations.
¾ The Unconstrained method measures data similarity via a two
trace semblance analysis. For each center trace, data within a
fixed time window is compared to data in a sliding time
window on a neighbor trace.
¾ The best semblance value computed during the sliding
process is retained for that neighbor. This process is repeated
for each selected neighbor.
¾ The planar dip constrained method uses semblance analysis to
perform dip-scant. Slat or dip-scan, stacks using 8 neighbor
traces are generated over a range of dips in two perpendicular
directions, and the semblance values are calculated.
¾ Analysis of these semblance value yields the best orientation
of a plane. Where this plane intersects the neighbor traces,
then it defines the center of the window of samples used for
each trace.
¾ Semblance, using the fixed window on the center trace and
the newly defined fixed window on each selected neighbor
trace is computed and retained.
¾ After each selected neighbor trace has been scanned and the
best semblance value for each neighbor retained, the target
trace for evaluating the Manhattan distance dissimilarity can
be selected between the one has the highest semblance
(Evaluation statistic-maximum) with the center trace, or the
one with lowest semblance with the center-trace (Evaluation
statistic-minimum).
¾ Finally, a Manhattan distance dissimilarity value is computed
using the center reference trace, the target neighbor trace, and
the value output at the middle sample of the center trace
window.
¾ These attributes are useful for detecting lateral trace changes
caused by faulting or stratigraphic
Time Attributes change.
and Their Applications in Reservoir Characterization
¾ A data set of Manhattan distance dissimilarity values will
have a range of to 1. High values are areas of highest
dissimilarity and may indicate areas of rapid change such as
fault. average or median values gives a view of the
“representative” data similarity in the area.
¾ A low Manhattan distance values indicate a very uniform
geology (low dissimilarity).
Figure 11. An illustration
of how the dissimilarity
analysis enhance the fault
images (Landmark, 1999)
Figure 12. An example of
trace selection using
semblance (Landmark,
1999)
¾ However, a low semblance or high Manhattan distance values
do not always indicate structural or stratigraphic changes.
They may result from a low signal-to-noise ratio or poor data
quality.
¾ Figure 11 shows an example of how the above algorithm
selecting the target trace using semblance. Figure 12 shows
how the dissimilarity attributes is computed using Manhattan
distance.
3.2. Unconstrained
Option
¾ In the unconstrained option, the semblance is computed as
follow :
k =t + N / 2
∑ k k
(G + H + d ) 2
S (t, d ) = k =t − N / 2
(1)
k =t + N / 2

2 ∑(Gk + H k + d )
2
k =t −N / 2 
¾ Where S is the semblance values computed from any trace
couples, G and H are the line t and dip d, and N is the number
of samples in the comparison windows.
¾ The computation is repeated for every possible dip value
within the comparison limit. The highest semblance value
obtained is retained for latter selection of the target trace.
3.3. Planar Dip Constrained
¾ This option removes planar dip before computing

semblance along the plane directions or dips can.
¾ In the line dip is computed by finding the maximum
semblance of the eight adjacent trace windows over a
range of sample shift. The time for maximum shift is
computed from a parabolic interpolation of semblance
values nearest the maximum.
¾ The same process is used for cross line dip. The sample
indices (possibly fractional) for each trace relative to the
center sample on the center trace can be computed by using
the equation of a plane.
¾ The window of data on the center trace may be directly
compared using semblance, with the corresponding parts of
the selected adjacent traces and the results stored for target
trace selection.
¾ The same equation applies as for the unconstrained option
above, except the dip value d in S (t,d) has been determined
for each trace by the dip scan process.
¾ Generally there is also an option to set the scan pattern which
will determine the number and arrangements of adjacent
traces to be scanned by semblance analysis with the current
center trace.
¾ Figures 13-14 illustrate the common scan pattern. For
instance the “line –z traces” will compare the current active
trace with the two adjacent in line, cross line, or diagonal
traces. On the other hand, “Ell-z Traces” will compare the
current trace with an adjacent inline and cross line trace.
¾ The length of trace data ( in ms) need to be determined in
each semblance calculation. The window can range from 0
ms (single-sample multiplication) to the full trace length.
¾ Typically, the window is set between 20 to 100 ms to ensure
that adequate data is included to estimate trace similarity. On
output, one-half of the window length will be zero at the
beginning and end of the race.
¾ Shorter windows is a very sensitive measure of trace
dissimilarity, but it is also very liable to biasing from noise
effects. Therefore, a short window should be used only for the
data that has high signal-to-noise ratios.
¾ Longer windows will reduce the risk of noise effects, but
because of the high degree of statistical averaging, may
obscure the subtle dissimilarities
Figure 13. An illustration on the algorithm used for computing the Manhattan Distance (Landmark, 1999)
Figure 14. Types of common scan patterns (Landmark,
1999)
Exercise 1
Figure E-1. Dip/azimuth time slice map at 1000 ms. : interpret the strike-dip
Exercise 2
Figure
15
Figure E-2. Time-sliced semblance volume. Dark blue represents high semblance, white is low semblance :
delineate reef, back-reef and lagoonal seismic facies
Exercise 3
M N
S
slice
2000
ms
Figure E3. Cosinus phase section of traverse path N-S through M well. Interpret the reef, fore-reef, back-
reef and barrier reef seismic facies
Exercise 4
Figure E4. Variance cube time-sliced at 2000ms. Black is high S

dissimilarity, white is low similarity. Interpret the reef, fore-
reef, back-reef and barrier reef seismic facies. Position of line
S-N in exercise 3 is given.
Exercise 5
high
low
50
45
40
35
Variance x Amplitude 30
G ro s s S a n d
25
20
15
10
0
0 0.02 0.04 0.06 0.08 0.1 0.12 0.14 0.16 0.18 0.2
RMS of Variance x RMS of Amplitude
Kross Korelasi Linear (Kross Korelasi)
MLD-02
MLD-03
Figure E5. 3D visualization of variance x amplitude. Interpret the extent of paleochannel, paleohigh and
the thickness of the sand within the channel.
Bibliography
Baxes, G. A., 1994. Digital image processing, principles and

applications, New York, John Wiley and Sons, 452 p.
Barteussen, K. A. and Ursin, B., Approximate Computation of
the Acoustic Impedance from seismic data, SEG 51 st meeting
Dalley, R.M., July 1989. Dip and Azimuth Displays for 3-D
Seismic Interpretation. First Break.
Dabeye, H.W.J. and Van Riel, P., Lp-norm deconvolution
Geophysical Prospecting, Vol.38, No. 4 (May 1990)
Galbraith, J.M. and Millington, G.F., Low frequency recovery in
the inversion of seismograms, CSEG Calgary, 1978
Jain, A.K., 1989. Fundamentals of digital image processing.
Englewood Cliffs, N.J, Prentice Hall, 4569 p.
Levy, S. and Fullagar, P.K., Reconstruction of a sparse spike
train from a portion of its spectrum and application to high-
resolution deconvolution Geophysics, Vol.46, No.9 (September
1981).
Oldenburg, D.W. Scheuer, T. and Levy. S., Recovery of the
acoustic impedance from reflection seismograms. Geophysics,
Vol.48, No.10 (October 1983).
Pratt, W.K., 1991. Digital image processing. New York, John
Wiley and Sons, 698 p.
Russ, J.C., 1995. The image processing handbook, 2nd edition.
Boca Raton, Fl, CRC Press, 674 p.
Sheriff, R.E., 1991. Encyclopedic dictionary of exploration
geophysics, 3 rd ed. Tulsa, SEG, 376 p.
Swan, A.R.H., and M. Sandilands, 1995. Introduction to
Geologic data Analysis. Cambridge, MA, Blackwell Science, 446
p.
Yilmaz, O., 1987. Seismic data processing. Tulsa, SEG, 526 p.

Ho Times Attributes

Uploaded by

Copyright:

Available Formats

Ho Times Attributes

Uploaded by

Document Information

Original Description:

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Ho Times Attributes

Uploaded by

Copyright:

Available Formats

Time Attributes and Their Application

Time Attributes and Their Applications in Reservoir Characterization

Time Attributes and Their Applications in Reservoir Characterization

Time Attributes and Their Applications in Reservoir Characterization

Time Attributes and Their Applications in Reservoir Characterization

Figure 1. The principle of dip and azimuth computation

Time Attributes and Their Applications in Reservoir Characterization

Time Attributes and Their Applications in Reservoir Characterization

Time Attributes and Their Applications in Reservoir Characterization

Time Attributes and Their Applications in Reservoir Characterization

Time Attributes and Their Applications in Reservoir Characterization

Time Attributes and Their Applications in Reservoir Characterization

¾ In a shaded relief display, a grid is produced in which the

Time Attributes and Their Applications in Reservoir Characterization

Time Attributes and Their Applications in Reservoir Characterization

Time Attributes and Their Applications in Reservoir Characterization

Time Attributes and Their Applications in Reservoir Characterization

Time Attributes and Their Applications in Reservoir Characterization

Time Attributes and Their Applications in Reservoir Characterization

Time Attributes and Their Applications in Reservoir Characterization

Time Attributes and Their Applications in Reservoir Characterization

Time Attributes and Their Applications in Reservoir Characterization

Time Attributes and Their Applications in Reservoir Characterization

Degree of similarity or dissimilarity can be used to enhance the

Time Attributes and Their Applications in Reservoir Characterization

¾ The Manhattan distance is a statistical measure of the

Time Attributes and Their Applications in Reservoir Characterization

Time Attributes and Their Applications in Reservoir Characterization

Time Attributes and Their Applications in Reservoir Characterization

Time Attributes and Their Applications in Reservoir Characterization

Time Attributes and Their Applications in Reservoir Characterization

Time Attributes and Their Applications in Reservoir Characterization

Time Attributes and Their Applications in Reservoir Characterization

Time Attributes and Their Applications in Reservoir Characterization

Time Attributes and Their Applications in Reservoir Characterization

Time Attributes and Their Applications in Reservoir Characterization

Time Attributes and Their Applications in Reservoir Characterization

Time Attributes and Their Applications in Reservoir Characterization

Time Attributes and Their Applications in Reservoir Characterization

Time Attributes and Their Applications in Reservoir Characterization

¾ This option removes planar dip before computing

Time Attributes and Their Applications in Reservoir Characterization

Time Attributes and Their Applications in Reservoir Characterization

Time Attributes and Their Applications in Reservoir Characterization

Time Attributes and Their Applications in Reservoir Characterization

Time Attributes and Their Applications in Reservoir Characterization

Time Attributes and Their Applications in Reservoir Characterization

Time Attributes and Their Applications in Reservoir Characterization

Time Attributes and Their Applications in Reservoir Characterization

Time Attributes and Their Applications in Reservoir Characterization

Figure E4. Variance cube time-sliced at 2000ms. Black is high S

Time Attributes and Their Applications in Reservoir Characterization

RMS of Variance x RMS of Amplitude

Kross Korelasi Linear (Kross Korelasi)

Time Attributes and Their Applications in Reservoir Characterization

Baxes, G. A., 1994. Digital image processing, principles and

Time Attributes and Their Applications in Reservoir Characterization

Time Attributes and Their Applications in Reservoir Characterization

You might also like