Geophysics Chulalongkorn University

LAB 4

B+ Student name: Anh Huyen Nguyen

ID: 6072616123

Bangkok, September - 2017

 Inline 1450

View data in IL, XL, and time slice

Time slice at
3625 ms

Strong amplitude interval

Low amplitude interval

due to attenuation and
geometric spreading

This map shows time slice at 3265 ms. As can be seem Deep interval with
that there are variations horizontally in amplitude expected low frequency
though these data points have the same time value.  higher period (hence
The seismic section on the right displays data from CL wavelength)
2700
2
 1450

View data in IL, XL, and time slice (cont.) Crossline 2700

Inline 1450

The same seismic image as XL

2700 is observed: high amplitude
in the upper part, low amplitude
in the middle and low frequency
in the deepest part

3
How do you know this is marine data?
Based on the following observations:
• No data above the first strong reflectors
(down to 2.3 sec because of low velocity in
sea water). On land, a smeared image of
weathered layer should be found No data until 2.3 sec
• Seabed can be easily located with a strong
reflection

Seabed with
strong reflection

4
What is the aerial size?

• 2 straight lines along IL

and XL are drawn and
calculated length and
width of the survey are
8.99 km and 6.25 km,
respectively  56.2 km2
• A polygon is also
constructed which Polygon area: 54.46 km2
generates an area of
about 56.46 km2

5
Frequency of data Too deep for real data

2.3-4.5 sec 4.5-6 sec 6-9 sec

• Three frequency spectrums are generated with three separate subsets (captured by polygons)
• For the first spectrum in the shallow time interval (2.3-4.5 sec), frequency ranges from 15 to 75 Hz and peak at around 62 Hz
• When time increases, frequency spectrum tends to shift to the lower value. This is expected since high frequency gets absorbed
when travelling deeper into the earth. More specifically, an interval from 4.5-6 sec has a peak of 17 Hz and that for 6-to-9 sec is about
14 Hz
• Generally, this is a relatively good quality data because of the broad frequency range. However, seismic image in the middle is not
really clear due to weak reflections (processing using gain function might help). The deepest part of the survey has low frequency but
the amplitude is great
6
What is the bin size, range of IL and XL?

• Total number of traces: 37901 • Inline spacing: 30 m

• Time interval: 1.8-9 sec • Crossline spacing: 12.5 m
• Inline range: 1200-1500 (301  Bin size: 30x12.5 = 375 m2
inlines)
• Crossline range: 2500-3000 (501
crosslines)

7
What is polarity of data and phase?
• At seabed there is a strong reflection due to the significant
change in density/velocity when switching from water to solid
(increase velocity and density  increase AI). As shown by the
image, a high amplitude trough at seabed indicates that
provided data is positive polarity
• The trough with side lobes normally presents for zero-phase
wavelet

Side lobes
Strongest reflection
at seabed

changing from
seawater to
sediments: trough

Typical zero-phase
wavelet with side lobes
8
Comparison of AGC volumes to original seismic-location: IL 1450

Faults Faults
Generally, there is no significant difference
between these two volumes (original and
AGC with 500 ms window) in the upper
part at the selected location though
amplitude increases a little bit. A set of
faults can be observed in both volumes

Higher amplitude but the connectivity

is not really good  smoothing needed

Increase amplitude with

better connectivity

Original data AGC 500 ms

9
Comparison of AGC volumes to original seismic-location: IL 1450 (cont.)
• Because of a narrow window of only 100 ms, Clearer fault image
the amplitude is nearly equal for whole survey.
This leads to difficulties when it comes to
amplitude analysis because it increases with
different rates
• Again, the interval from 4.5 to 5.8 sec gains
higher amplitude but the connectivity is still
poor
• Two antithetic faults appear to be easily
detected

Poor connectivity

Original data AGC 100 ms

10
Comparison of AGC volumes to original seismic-location: IL 1450 (cont.)
• The generated data with a broad window of
1000 ms is quite similar to that of 500 ms
window: overall amplitude increases but still
cannot result in a better connectivity in the
middle part.
• The only difference is that amplitude of the
lowest part is built up significantly when using
100-ms window as input.

Anomaly high amplitude

Original data AGC 1000 ms

11
Create dip volumes

Low dip – nearly Higher dip, compared

horizontal to the same locations
in XL 2700

Low dip High dip

The low-dip volume has the same dip field as that of high-dip in the upper interval because the beds are fairly
horizontal. However, over an interval from 4.5 to 7 sec, there are slight difference between these two
volumes as dip field in high-dip volume has higher dip field.
12
Create SOF volumes + comparison to original volume
IL 1450 XL 2700

• Obviously, SOF low-dip volume gives very smooth

image of the beds, as compared to the original
volume
• Faults are observed in the original volume.
Faults However, low-dip volume cannot provide good
Original volume images of faults because there are many dim
regions around the faults. When examining closely
IL 1450 XL 2700 these locations, no offset is found. In this case, it
is not appropriate to use low-dip volume for
interpreting faults

Dim areas
SOF-Low dip volume 13
Create SOF volumes + comparison to original volume (cont.)
IL 1450 XL 2700

• Faults can be easily detected in high-dip

volume. Together with continuous reflections,
this volume is excellent in quality and can be
utilized for structural interpretation (fault and
Faults horizon)
Original volume • Generally, in terms of bed dipping there is no
significant difference between low-dip volume
IL 1450 XL 2700 and high-dip volumes because of the fairly
horizontal beds in upper region although with
careful analysis, low-dip volume shows flatter
beds than that of high-dip volume

Clear fault image 14

SOF-high dip volume
Horizon picking- 2770 ms
• Horizon starting at 2770 ms at IL 1450/XL 2700
location, piking along XL 2700  propagate using
original volume
• Time ranges from 2695 to 2870 ms
• This propagation is quite good though there are some
Discontinuous discontinuous locations. It is mainly because of strong
location
and continuous reflections in the upper part of the
survey
• When passing through fault, the contour lines start to
act strangely because the software will interpolate a
Mis-tie jump connection at fault surface. This can be
eliminated by interpreting faults first  select option:
“Stop at the fault” or create polygon  select option
“Stop at polygon”
Fault • A mis-tie location is also found  need to look closely

15
3 ms/IL versus 11 ms/IL smooth 3 traces vs
• The main difference between 11 traces
11 ms/IL these two volumes is the level of 3 ms/IL
smooth which is higher for 11
ms/IL than that of 3 ms/IL
• For high-smooth volume, the
reflections around the fault is not
strong and the opposite is found
in low-smooth one where the
fault appears to be quite clear
although some events are merged
laterally. This results in eliminating
fault offset  not appropriate for Clearer fault image
interpreting fault
• For horizon picking, low-smooth
volume should be used because
the preservation of geological
significance
Less reflection Merged events laterally
around the fault No fault offset  unreal

16
Discussion on auto picking
Auto picking using “progradation” – inline 1450

• Due to high amplitude and continuous reflection,

auto picking works well in the upper reservoir. The
generated horizon can cover nearly all survey area
with some discontinuities, as shown in the slide 15
• After smoothing, auto picking can provide better
3.5 sec horizon- 3.5 sec horizon-
original volume
propagation with less discontinuities
SOF volume
• Detail spatial distribution of these horizon will be
discussed in the next slides

4.5 sec horizon- 4.5 sec horizon-

original volume SOF volume

Poor connectivity Better connectivity

after smoothing
17
Horizon auto picking on original and SOF volumes
3.5 sec-Original volume • The 3.5-sec horizon extension using 3.5 sec-SOF volume
auto picking on original volume is
not as good as that of 2770 ms (slide
15). More specifically, an area of
around 60% of whole survey is
covered with some discontinuous
locations. There is no data in the Smooth
northeast corner of base map. This boundary
indicates that the connectivity is
No data poor and manual picking is required
in order to have better result
• Horizon auto picking using SOF
volume generates similar area
coverage with better connectivity
and geological sense. When passing
the fault, both volume have poor
reflection resulting in a relatively big
gap
Discontinuous event
through fault

18
Horizon auto picking on original and SOF volumes (cont.)
• Auto picking on original volume
for 4.5-sec horizon does not
work in this case because there
is nearly no continuity on the
map. The horizon does not
propagate in the northeast and
southwest corners which
indicates that these regions
have extremely poor data
No data quality
• Again, auto picking performing
on SOF volume yields better
result but the image is not as
good as that of 3.5-sec horizon
• Further processing steps
together with manual picking is
needed to have a reliable
interpretation
4.5 sec-Original volume 4.5 sec-SOF volume

19
Thank you!

20