SRMA

2D/3D Surface Related Multiple Attenuation

Contents
Overview . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2
Input Data Preparation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2
Technical Description . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6
Requirements and Restrictions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8
Output . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8
Parameter Description Table . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9
Parameter Table . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10
General Execution Parameters - Required . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11
INPFILE - Required . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13
OUTFILE - Required . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14
VELFILE - Required . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15
APERTUR - Required . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16
NODES - Required . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18
AREA - Optional . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19
Example One . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21
Example Two . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21
Example Three . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22
Index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24

Echos Paradigm 2011.3

With Epos 4.1 Data Management

Overview
SRMA is the first of two modules designed to predict and subtract surface related multiples in 2D/3D.
SRMA performs the prediction of the multiples and SMACMS is the process where input to the original data
merged with the predicted multiple model for adaptive subtraction.
According to the theory of surface multiple attenuation, surface related multiples can be kinematically
predicted by 2D surface integration of the convolution of prestack seismic traces at possible surface multiple
reflection locations. 3D SRMA requires dense source and receiver distributions in both inline and crossline
directions. If such acquisition is available, SRMA utilizes these traces. In most marine surveys, dense
source distribution is not found. Then the process requires selecting the "most suitable" trace from the
available traces. SRMA uses the existing data along with residual moveout application to form the surface
integration over possible multiple location. The user has the option to give preference to either offset or
azimuth in this process of "suitable trace" selection.

Echos Paradigm 2011.3

With Epos 4.1 Data Management

General Use
Input Data Preparation
Multiple prediction in 2D/3D SRMA is based on the correlation of pairs of traces over the possible multiple
generation locations. This is a data driven process and for best results requires regular input data. That is
hardly ever possible in real cases. Therefore, SRMA will utilize the nearest shot(s)/receiver(s) using residual
moveout to compensate for the distances. A velocity volume used in the residual moveout applications is
one of the inputs to this process.
The following steps should be completed to make the data suitable for input to SRMA.

Removal of Direct Waves and Ghosts

The direct wave needs to be muted before SRMA; otherwise, the convolution of direct waves with primary
reflection events and multiple reflection events may produce unwanted events in the multiple model. These
unwanted events are unrealistic. They do not exist in the original input and will be added into the final
results as coherent noise after multiple subtraction; therefore, they need to be removed. Ideally, input data
should only contain primaries and surface related multiples.

Deghosting
If the guns and receivers are several meters below the free surface, they will produce ghost events. SRMA
process does not predict ghost events so they need to be removed prior to SRMA.

Source and Receiver Depth Correction

SRMA requires that both the sources and the receivers are at the free surface where multiple reflections
take place. If the sources and the receivers are several meters below the free surface, a static correction
must be applied to move them to the free surface. This can be done by calculating and applying a bulk time
shift to the traces. The time shift corresponds to the source depth and the receiver depth. If this correction
is not applied, predicted multiple events will have time shifts.

Parallel Computation
The memory, disk space, and computational time required for 3D multiple prediction is so large that it is
impossible to accomplish the task on a single computer in a reasonable time. The task must be sent to a
cluster of computers with multi-CPUs. SRMA is designed such that the entire task is split into many
individual subtasks. In each individual subtask, the multiple model for the target output locations are
computed. This is done by accessing the contributing sublines from the main input. The same data must be
read several times; therefore the contributing sublines are stored into the local disk for faster access. The
size of the local scratch disk must be large enough to store the sublines within the inline aperture of a target
output location. If not, the local disk will be recycled as needed and remote reads will be performed, which
can degrade the performance.
By default, SRMA will use ssh to connect to the slave nodes. To use rsh to connect to the slave nodes, the
user should set the environmental variable PG_SHELL to rsh. When using ssh, the user must be able to ssh
into the remote node without supplying a password or other response.

Echos Paradigm 2011.3

With Epos 4.1 Data Management

SRMA is primarily designed to be run on distributed memory systems, such as Linux clusters, consisting
of several nodes. The location where a job is started will also be the master; therefore, it should not be listed
in the MFILE of the NODES list.
SRMA is both a computationally and I/O heavy application. For that reason, there is a predefined task size
(in number of CDPs), which is equal to the quarter of number of CDPs in the inline direction aperture.
Therefore, if the size of the output is smaller than this number, a parallel job may run on a less number of
nodes than the user selected. This is to prevent I/O overload. Splitting the job into smaller chunks will not
always lead to shorter run times due to repeated read of the input data for each node. This is only a problem
when running a small job compared to the size of the available resources. The default size can be altered by
setting the environmental variable SRMA_MINPARSIZE.

Global Parameters
SRMA can be run on both 2D and 3D acquisitions. Also, for 3D acquisition, the user can elect to run the
SRMA in a 2D mode. This may be a desirable feature if the crossline dip of the survey is not considerable.
3D SRMA is a data driven process and not all the desired traces are always available. In that case, SRMA
selects the most suitable trace based on the user selected criteria. The PRIORTY option allows the user to
set offset or azimuth. Azimuth is the preferred method for selection. When OFFSET is selected, the closest
offset trace from the desired location is selected. If AZIMUTH is selected, the most suitable trace selection
is done using the azimuth of the desired trace as the criteria. In both cases, the traces are corrected using
residual moveout from the present offset to the desired offset.
By default, SRMA utilizes all the CPUs on the target nodes. MAXCPU option allows the user to limit the
maximum number of CPUs to use.

INPFILE List
INPFILE parameter is used to specify the label of the input data in PDS format. The order of the input file
must be CDP. When computing the contribution for a source-receiver location, all the data within the user
specified apertures is required on input.

OUTFILE List
OUTFILE parameter is used to label the output file. This must be a unique name and there should not be
a file associated with this name. Computation of individual crossline multiple contribution traces are done
by slave tasks on slave nodes. The order of the output file will be CDP and will contain the modeled
multiples. Each node in the parallel run will create its own output with the _PART_# extension (for example
node #1 will create _PART_1 and node #2 will create _PART_2) and write directly into those files to
prevent network traffic to the master node. These files will be merged at the end of the job with the user
selected output file label. The output file from SRMA may look like the example below where the user
selected output file label is SRMA_2D_FULLBAND.

Echos Paradigm 2011.3

With Epos 4.1 Data Management

Fig. 1
The output file contains only the modeled multiple traces. After SRMA, the user must merge (using the File
Manager) the original input file with the multiple model file output from SRMA. This file will be the input
file to SMACMS for adaptive multiple subtraction.

VELFILE List
There are no assumptions about the subsurface model. Velocities in this process are only utilized for
differential moveout to reposition the traces to the exact offset location desired in the convolutions. This is
done by using the velocities supplied by the VELFILE list. A brute velocity file will suffice for most of the
time where the differential moveouts are relatively small. For larger moveouts, a better velocity field will
produce better results. The user should supply the best velocity possible at this stage of the processing.
The velocity file is expected in PDS format and must have the same sampling interval as the input seismic
file. There is no interpolation or any extrapolation done for the velocity traces; therefore, the areal extend
of the velocity file must cover the input data area. For 3D datasets, the PDS format velocity file must be a
filled volume.

APERTUR List
SRMA is a data driven process. Therefore, the integration area for multiple modeling is limited to the
available data limits. The inline direction aperture should be as big as the maximum inline recording offset
and crossline aperture should be as big as the maximum crossline recording offset. Maximum inline offset
can be computed from number of channels times the channel interval plus the near offset distance. The
crossline direction maximum offset can be computed from the receiver cable distances.

Echos Paradigm 2011.3

With Epos 4.1 Data Management

ILSTEP and XLSTEP define the integration intervals in 2 x CDP the surface area (3D) and ILSTEP define
line integral interval in (2D). The default for ILSTEP is the CDP interval in the inline direction (constant
CDPLBLS) and for XLSTEP is the 2 x inline interval in the crossline direction (constant CDPLBLX). The
choice of these parameters will depend on the frequency content of the data and higher frequency data may
require denser integration intervals.

Fig. 2: APERUTR parameters displayed on a typical marine and land recording. The inline
direction maximum offset (ILMXOFS) is the distance from the source to the farthest channel in
the near cable. The maximum crossline direction (XLMXOFS) is the distance from the source to
the farthest cable.

NODES List
SRMA is optimized to operate on a cluster of computers, where the job is divided into subtasks, each
assigned to a slave node. NNODES describes the number of target nodes to execute where the nodes
themselves are listed in the MFILE, 1 node per line. A sample MFILE is given below.
delln10
delln11

Echos Paradigm 2011.3

With Epos 4.1 Data Management

delln12
delln13
delln14
delln15
SRMA uses all the CPUs available on each target node. Therefore it is not advisable to start more than one
task per node. For that reason, each node should be listed only once in the MFILE and the NNODES should
not exceed the number of nodes listed in the MFILE. The default for NNODES is the number of entries in
the MFILE and it cannot be larger than that.
MFILE should be a plain ASCII file listing the nodes and does not accept any commenting with special
characters.
By default, SRMA will use ssh to connect to the slave nodes. To switch to rsh connection, the user should
set the environmental variable PG_SHELL to rsh. When using ssh, the user must be able to ssh into the
remote nodes without supplying a password or other response.

AREA List
SRMA requires access to all the data within the limits described by the APERTUR list, which may be much
larger than the desired output area. The AREA option allows the user to limit the output area, either for
testing purposes or to complete missing parts. The user can specify the starting and ending inline ranges
and the starting and ending crossline ranges. The default for the area list is the limits of the input seismic
file.

Technical Description
Surface-related multiples include all multiples except those generated by totally internal reflections. They
are typically the strongest multiple events present in the seismic data and therefore need to be removed
from prestack gathers prior to any further imaging process. Surface-related multiples are classified into
different orders depending on how many times they are reflected back from the free surface. A multiple
event is called first-order surface multiple if it is reflected only once at the free surface; while a second-order
multiple is reflected twice at the free surface. In general, a N-th order multiple gets reflected N times at the
free surface. The SRMA technology is designed to attenuate these multiples in a two-step fashion. During
the first step, surface-related multiples are predicted kinematically. The predicted multiples have about the
same traveltimes as the multiples recorded in seismic data but different amplitudes and wavelet shapes.
In the second step, the surface-related multiples present in the seismic data are attenuated using predicted
multiples. The method assumes the sources and receivers have the same spatial sampling interval and they
are located at the same stations at the free surface.
This method does not require distinctive moveout differences between the primaries and multiples, and
does not require any prior knowledge of the subsurface structure, such as velocities.
The seismic data itself can be used to predict the surface-related multiples. According to the theory by
Verschuur, the multiple prediction operator can be expanded into a Taylor series. In the frequency domain,
the Taylor expansion can be expressed as
2
3
4
R
R 2
R 3
P 0 = P ------------ [ P ] + ------------ [ P ] ------------ [ P ] + ...
S ( )
S ( )
S ( )

Echos Paradigm 2011.3

With Epos 4.1 Data Management

= T 0 + T 1 + T 2 + T 3 + ...
where matrix P is the input seismic data and matrix P 0 represents multiple-free data. One column of P
represents one common shot gather for a given frequency; while one row of P represents one common
receiver gather. R is the reflection coefficient at the free surface. S ( ) is the source wavelet and is a function
of frequency. Since the source signature is usually unknown, it is set to a spike in SRMA, and R is also set
to 1. Setting S ( ) to a spike results in wavelet distortion in predicted multiple events. The effect of the
source wavelet will be accounted for by finding matching filters in SMACMS that transform the wavelet of
predicted multiples into that of the recorded multiples.
In the Taylor expansion, the terms T 0 , T 1 , T 2 , and T 3 are corresponding Taylor terms. Each term in the
Taylor series contains multiples starting from a certain order. They are obtained by multiplying the data
matrix P with lower order Taylor term. The zeroth order Taylor term T 0 = P in the expansion is the input
seismic data itself containing both the primaries and multiples of various orders. The first order Taylor
R
term T 1 T 1 = --- P P contains the first-order multiples and all other higher-order multiples except any
S

primary reflection. The second order Taylor term T 2 in the Taylor expansion contains second-order and
higher order multiples only. Note that the N-th order Taylor term in the expansion contains not only the Nth order multiples, but also all possible higher order multiples.
Each Taylor term is computed through summations of cross convolutions between input seismic traces and
lower order Taylor terms that are already computed. The input seismic data, when convolved with itself,
predicts the first order Taylor term T 1 = --- P P , which contains multiples only.
S
R

In the 3D SRMA application, only the first-order Taylor term

R
T 1 T 1 = --- P P is computed.

References
Verschuur, D. J., 1991. Surface-related multiple elimination: an inversion approach, Ph.D. Thesis, Delft
University of Technology.
Verschuur, D. J., Berkhout, A. J. and Wapenaar, C. P. A., 1992. Adaptive surface-related multiple
elimination, Geophysics 57:1166-1177.

Echos Paradigm 2011.3

With Epos 4.1 Data Management

Requirements and Restrictions

Input data must be prestack CDP ordered traces. Direct waves must be removed from the input gathers by
muting in the x-t domain and/or filtering in the frequency domain. Source-side and receiver-side ghosts
must also be removed from the input gathers if the sources and receivers are not located at the free surface.
If sources and receivers are not at the free surface, static corrections must be applied to move them to the
surface. SRMA requires that the input trace headers contain entries for the following information: CDP,
REC-X, REC-Y, SHT-X, SHT-Y and OFFSET for 2D and CDPLBLS, CDPLBLX for 3D. SRMA will use the
shot and receiver coordinates to calculate the actual distances.

Output
The output from SRMA is the multiple model. MULTNO and XFOLD headers will be created on output.
The SMACMS module should follow the modeling process.
For adaptive subtraction purpose (using SMACMS module) the output of SRMA must be merged with the
original input file using the File Manager. This will create a new virtual file that allows access to both files
at once. SMACMS on input requires gathers that have traces with two (or more) MULTNO headers. The
traces within a gather are first grouped by SUBGTHR header name if defined by the user in SMACMS, and
then by the SKEYNAM. Then the traces with MULTNO other than 0 are matched to the trace with
MULTNO 0.
The adaptive subtraction process produces better results when performed on common shot gathers. Since
DSIN allows sorting the data in any order, the user can supply the data to SMACMS in shot order. For 3D
shot gathers, header defining the streamers (cables) should be supplied as the SUBGTHR name. This will
implement multi-channel filter design in each streamer (cable) separately.

Echos Paradigm 2011.3

With Epos 4.1 Data Management

Parameter Description Table

Req

General Execution Parameters

TYPE
Multiple modeling? (2D/3D)
PRIORTY
Trace selection priority (AZIMUTH/OFFSET)
MAXCPU
Maximum number of CPUs to use on any node

Req

INPFILE
LABEL

Input File in PDS Format

Label of PDS format input file

Req

OUTFILE
LABEL

Output File in PDS Format

Label of output file

Req

VELFILE
LABEL

PDS Velocity File for Residual Moveout

Label of PDS velocity file

Req

APERTUR
ILMXOFS
XLMXOFS
ILSTEP
XLSTEP

Multiple Modeling Apertures

Maximum inline direction offset (ft or m) (constant CDPLBLS)
Maximum crossline direction offset (ft or m) (constant CDPLBLX)
Step size in inline direction (ft or m) (constant CDPLBLS)
Step size in crossline direction (ft or m) (constant CDPLBLX)

Req

NODES
NNODES
MFILE

Execution Slave Nodes

Number of nodes
Machine file listing slave nodes

Opt

AREA
ILFROM
ILTO
XLFROM
XLTO

Output Area Selection

Starting inline
Ending inline
Starting crossline
Ending crossline

Echos Paradigm 2011.3

With Epos 4.1 Data Management

10

Parameter Table

Req

1-8

9-16

17-24

25-32

33-40

*CALL

SRMA

TYPE

PRIORTY

MAXCPU

ILSTEP

XLSTEP

XLFROM

XLTO

Req

INPFILE

LABEL

Req

OUTFILE

LABEL

Req

VELFILE

LABEL

Req

APERTUR

ILMXOFS

XLMXOFS

Req

NODES

NNODES

MFILE

Opt

AREA

ILFROM

ILTO

Echos Paradigm 2011.3

With Epos 4.1 Data Management

41-48

49-56

57-64

11

General Execution Parameters - Required

1-8

9-16

17-24

25-32

33-40

*CALL

SRMA

TYPE

PRIORTY

MAXCPU

*CALL

57-64

1-8

Program name.
Columns:

TYPE

49-56

Execution directive.
Columns:

SRMA

41-48

9-16

Defines the modeling type.

Valid values:

PRIORTY

3D

For multiple modeling in 3D, supply 3D, or let the

parameter default.

2D

For multiple modeling in 2D, supply 2D.

Columns:

17-24

Type of value:

Character * 8 - Enumerated list

Default value:

3D

Defines the trace selection criteria for irregular input.

Valid values:

MAXCPU

AZIMUTH

For trace selection criteria based on azimuth, supply

AZIMUTH, or let the parameter default.

OFFSET

For trace selection criteria based on offset, supply

OFFSET.

Columns:

25-32

Type of value:

Character*8 - Enumerated list

Default value:

AZIMUTH

Specifies the maximum number of CPUs to utilize on each node. Leave blank to
utilize all the available CPUs on the target nodes.

Echos Paradigm 2011.3

With Epos 4.1 Data Management

12

Columns:

33-40

Type of value:

Integer

Lower limit:

Upper limit:

Number of CPUs on the node

Default value:

Number of CPUs on the node

Echos Paradigm 2011.3

With Epos 4.1 Data Management

13

INPFILE - Required

1-8

9-16

INPFILE

LABEL

17-24

25-32

33-40

41-48

49-56

57-64

INPFILE parameter is used to specify the label of the input data file. This file must be in PDS data set
format.

INPFILE

Operation Name.
Columns:

LABEL

1-8

Specifies the label of the input file that contain the input seismic data to SRMA.
Columns:

9-80

Type of value:

Character*72

Default value:

None

Echos Paradigm 2011.3

With Epos 4.1 Data Management

14

OUTFILE - Required

1-8

9-16

OUTFILE

LABEL

17-24

25-32

33-40

41-48

49-56

57-64

OUTFILE parameter is used to label the output file from SRMA which will contain the modeled multiples.
The file created will be in PDS format.

OUTFILE

Operation Name.
Columns:

LABEL

1-8

Specifies the output file label from SRMA.

Columns:

9-80

Type of value:

Character*72

Default value:

None

Echos Paradigm 2011.3

With Epos 4.1 Data Management

15

VELFILE - Required

1-8

9-16

VELFILE

LABEL

17-24

25-32

33-40

41-48

49-56

57-64

VELFILE parameters specify the PDS format velocity file used in the residual moveout calculations.

VELFILE

Operation Name.
Columns:

LABEL

1-8

Specifies the velocity file label in PDS format.

Columns:

9-80

Type of value:

Character*72

Default value:

None

Echos Paradigm 2011.3

With Epos 4.1 Data Management

16

APERTUR - Required

1-8

9-16

17-24

25-32

33-40

APERTUR

ILMXOFS

XLMXOFS

ILSTEP

XLSTEP

41-48

49-56

57-64

APERTUR parameters specify recording geometry. If multiple surveys are merged together use the largest
parameters. See Fig. 2.

APERTUR

Operation Name.
Columns:

ILMXOFS

XLMXOFS

ILSTEP

XLSTEP

1-8

Specifies the maximum inline direction offset of the recording geometry, ft or m. If

data from different surveys is mixed, supply the largest inline offset.
Columns:

9-16

Type of value:

Integer

Lower limit:

Upper limit:

None

Default value:

None

Specifies the maximum crossline direction offset of the recording geometry, ft or m.

Columns:

17-24

Type of value:

Integer

Lower limit:

Upper limit:

None

Default value:

None

Specifies the correlation step size in the inline direction, ft or m. Normally, this
should be 2 x CDP interval.
Columns:

9-16

Type of value:

Floating point

Lower limit:

1.0

Upper limit:

None

Default value:

None

Specifies the correlation step size in the crossline direction, ft or m. Normally, this
should be 2 x inline interval.

Echos Paradigm 2011.3

With Epos 4.1 Data Management

17

Columns:

25-32

Type of value:

Floating point

Lower limit:

0.0

Upper limit:

None

Default value:

None

Echos Paradigm 2011.3

With Epos 4.1 Data Management

18

NODES - Required

1-8

9-16

17-24

NODES

NNODES

MFILE

25-32

33-40

41-48

49-56

57-64

NODES parameters specify cluster parallelization.

NODES

Operation Name.
Columns:

NNODES

MFILE

1-8

Specifies the maximum number of nodes to start parallel tasks. On each node, all
the available computer resources will be utilized by default. The user has the
option to limit the number of CPUs with the MAXCPU option. Ideally, the file
specified in the MFILE option should list NNODES number of nodes as target. It
is not recommended to have less nodes listed in the MFILE than NNODES.
Columns:

9-16

Type of value:

Integer

Lower limit:

Upper limit:

None

Default value:

None

Specifies the file that lists the slave nodes for a parallel run.
Columns:

17-80

Type of value:

Character*64

Default value:

None

Echos Paradigm 2011.3

With Epos 4.1 Data Management

19

AREA - Optional

1-8

9-16

17-24

25-32

33-40

AREA

ILFROM

ILTO

XLFROM

XLTO

41-48

49-56

57-64

AREA parameters allows the user to limit the desired output area.

NOTE All the available data is required for the computation process.

AREA

Operation Name.
Columns:

ILFROM

ILTO

XLFROM

XLTO

1-8

Specifies the first inline label desired on output.

Columns:

9-16

Type of value:

Integer

Lower limit:

First inline defined in input seismic file

Upper limit:

Last inline defined in input seismic file

Default value:

First inline defined in input seismic file

Specifies the last inline label desired on output.

Columns:

17-24

Type of value:

Integer

Lower limit:

ILFROM

Upper limit:

Last inline defined in input seismic file

Default value:

Last inline defined in input seismic file

Specifies the first crossline label desired on output.

Columns:

25-32

Type of value:

Integer

Lower limit:

First crossline defined in input seismic file

Upper limit:

Last crossline defined in input seismic file

Default value:

First crossline defined in input seismic file

Specifies the last crossline label desired on output.

Echos Paradigm 2011.3

With Epos 4.1 Data Management

20

Columns:

33-40

Type of value:

Integer

Lower limit:

XLFROM

Upper limit:

Last crossline defined in input seismic file

Default value:

Last crossline defined in input seismic file

Echos Paradigm 2011.3

With Epos 4.1 Data Management

21

Example One

1-8

9-16

17-24

*CALL

SRMA

3D

INPFILE

CDP_FBMUTED

OUTFILE

SRMA_3D_IL1500-1600

AREA

1500

1600

APERTUR

4500

350

VELFILE

VNMO

NODES

60

25-32

33-40

41-48

49-56

57-64

C16SRMA

*END

In this example, SRMA modeling is run in 3D mode. The input file label is CDP_FBMUTED and contains
CDP sorted, first break muted prestack gathers. The output file is SRMA_3D_IL1500-1600. SRMA will be
run for the inline range from 1500 to 1600. The XLFROM and XLTO parameters are left blank, which
means the output will include all the CDPLBLX's in these inlines. The output file will have the same CDP
gathers on output as in the selected input (with the AREA list) range.
The survey was acquired with 8 streamers spaced at 100 meters apart and 358 channels at 12.5 meters
spacing with 38 meters near offset. This gives a maximum inline offset of 4500 meters and crossline offset
of 350 meters. Therefore, ILMXOFS is set to 4500 and XLMXOFS is set to 350. ILSTEP and XLSTEP uses
the defaults. The CDP spacing in this survey is 6.25, therefore, ILSTEP will be set to 12.5 and the spacing
between inlines is 12.5. 2 x 12.5 = 25 will be used for XLSTEP.
The label of the velocity file for the differential moveouts is VNMO and is an RMS velocity in PDS format.
This job will be run on 60 nodes in parallel. The nodes are listed in the c16srme file as shown below. The
job was launched from node 10.10.16.01; therefore it is not listed as one of the target nodes.
10.10.16.02
10.10.16.02
10.10.16.03
.
.
10.10.16.61
The output file from this run, SRMA_3D _IL1500-1600, will be merged with the input file,
CDP_FBMUTED. The merged file will be input into a new job with SMACMS, which will produce the
desired multiple attenuated output.

Example Two

1-8

9-16

17-24

*CALL

SRMA

2D

INPFILE

CDP_FBMUTED

Echos Paradigm 2011.3

With Epos 4.1 Data Management

25-32

33-40

41-48

49-56

57-64

22

1-8

9-16

17-24

OUTFILE

SRMA_2D_IL1500-1600

AREA

1500

APERTUR

3500

VELFILE

VNMO

NODES

20

25-32

33-40

41-48

49-56

57-64

1600

C16SRMA

*END

In this example, 2D SRMA is run on an input 3D dataset. The input file label is CDP_FBMUTED and is
CDP sorted, first break muted prestack gathers. The output file is SRMA_2D_IL1500-1600. SRMA will be
run for the inline range from 1500 to 1600. The XLFROM and XLTO parameters are left blank, which
means the output will include all the CDPLBLX's in these inlines.
The survey was acquired with 8 streamers spaced at 100 meters apart and 358 channels at 12.5 meters
spacing with 38 meters near offset. This gives a maximum inline offset of 4500 meters. Therefore, the
ILMXOFS is set to 4500.
Since SRMA is run in 2D mode XLMXOFS and XLSTEP parameters are irrelevant. ILSTEP uses the
default. The CDP spacing on this survey is 6.25 m; therefore, 2 x 6.25 = 12.5 m will be used.
The label of the velocity file for the differential moveouts is VNMO and is an RMS velocity in PDS format.
This job will be run on 20 nodes in parallel. The nodes are listed in the c16srme file as shown below. The
job was launched from node 10.10.16.01; therefore it is not listed as one of the target nodes.
10.10.16.02
10.10.16.03
10.10.16.04
.
.
10.10.16.21
The output file from this run, SRMA_2D _IL1500-1600, will be merged with the input file,
CDP_FBMUTED. The merged file will be input into a new job with SMACMS, which will produce the
desired multiple attenuated output.

Example Three

1-8

9-16

17-24

*CALL

SRMA

2D

INPFILE

INPUT2SRMA

OUTFILE

FINAL_SRMA

VELFILE

FIRSTVEL

APERTUR

4175

NODES

25-32

33-40

41-48

49-56

57-64

DC

*END

Echos Paradigm 2011.3

With Epos 4.1 Data Management

23

In this example, SRMA is run on a 2D line. The input file label is INPUT2SRMA and it contains the CDP
ordered, first break muted prestack gathers. The velocity file label for the residual moveouts is FIRSTVEL
in PDS format. The maximum recording offset of this 2D line is 4175; therefore, the ILMXOFS is set to 4175
meters. The AREA list is not supplied, which means the entire input data will be modeled. The output file
label is FINAL_SRMA. After the completion, INPUT2SRMA will be merged with the SRMA output file
FINAL_SRMA to create the input to SMACMS, the adaptive subtraction module.
The user may choose to sort the merged data to common shot (selecting the proper header in DSIN) and
running SMACMS. Adaptive subtraction works better in common shot domain.

Echos Paradigm 2011.3

With Epos 4.1 Data Management

24

Index
S
surface related multiple attenuation
SRMA 1

Echos Paradigm 2011.3

With Epos 4.1 Data Management

25