The Marmousi2 Elastic Model and Synthetic Data

1
The Marmousi2 Elastic Model and Synthetic Data

• SEG Annual Meeting
• October 2003

Gary Martin, GX Technology Robert Wiley and Kurt
Marfurt, AGL/University of Houston
2
Outline

• Introduction
• The Marmousi2 model
• Marmousi2 acquisition
• Marmousi2 synthetic data
• Processing and migration examples
• Conclusion

3
Introduction

• This work has been performed as part of the Next
  Generation Numerical Modeling and Imaging
  project.
• New suite of shared seismic models
• Update the popular and very useful Marmousi model
  using modern computing power.

4
The Marmousi2 Model
5
Original Marmousi

• Geologic Model
• Created in 1988 by IFP
• Based on real geology from West Africa (Angola)
• Complex model (160 layers)
• 9.2 km x 3 km
• Simulation
• Acoustic simulation using a second order (in time
  and space) finite difference algorithm
• 4m grid size, frequencies up to 55Hz
• Maximum offset 2.5 km

6
Marmousi2

• Model
• Larger model (2X length)
• Deep water (450m)
• Longer offsets (up to 17km)
• Hydrocarbons with AVO effects
• Stratigraphic features

7
Marmousi models
8
Lithology and Features
9
Layer Properties

• Determined by lithology, fluid content, and
  original Marmousi P-wave velocity.

• Lithologies
• Sand
• Shale
• Marl
• Salt

• Fluids
• Water
• Gas
• Oil (varying GOR)

10
P-wave Velocity

• Extracted from original Marmousi model
• Modifications for
• Hydrocarbon saturated layers
• Salt velocity, reduced to 4500m/s
• Velocity parameterization
• Constants
• Vertical velocity gradients
• P-wave velocity is the source for both density
  and S-wave transforms

11
P-wave Velocity
km/s
1
2
3
4
5
12
S-wave Velocity

• Determined directly from P-wave velocity using
  Greenburg and Castagna (1992)
• Sand Vs 0.804Vp 856
• Shale Vs 0.770Vp 867
• Limestone Vs 1.017Vp 0.055Vp2 1030
• Marl
• 30 limestone, 70 shale
• Using mixing laws and effective medium theory

m/s
13
S-wave Velocity
km/s
0
1
2
3
14
Density

• Determined directly from P-wave velocity using
  Castagnas (1993) transforms
• Sand r 0.2736Vp.261
• Shale r 0.2806Vp.265
• Limestone r 0.3170Vp.225
• Marl
• 30 limestone, 70 shale
• Using mixing laws and effective medium theory

g/cm3 and m/s
15
Density
g/cm3
1.0
1.5
2.0
2.0
2.5
3.0
16
Hydrocarbons

• Gas and oil emplaced into some sand layers.
• 5 gas sands
• 3 oil sands
• 1 oil and gas cap

• Fluid substitution used to provide realistic
  properties that will be useful for AVO
  calibration.
• Seismic responses verified using simplified
  models and ray-tracing.

17
Data Acquisition
18
Marmousi2 Acquisition

• Acquisition objectives
• Elastic simulation.
• Simultaneous streamer, OBC, and VSP acquisition.
• Multi-component recording for OBC and VSPs.
• Record Wave-front snapshots.
• Reflecting boundary at free surface.

19
Acquisition Geometry
20
Source

• An Ormsby wavelet with frequencies up to 80 Hz,
  (5-10, 60-80).

Near field
Far field
21
Computation Considerations

• Grid size
• To ensure adequate sampling (to prevent numerical
  problems) a grid size of 1.25m was required.

Marmousi grid 4m x 4m
Marmousi2 grid 1.25m x 1.25m
22
Computation

• Algorithm
• E3D National Labs. elastic modeling code
• 2nd order time, 4th order space finite difference
  elastic modeling
• Computation Time
• Sun 6800 Starfire system at UH
• 20 CPUs for 5 months
• 4 ½ CPU days per shot
• Total hours 70,000 8 CPU years

23
Marmousi2 Synthetic Data
24
Synthetic Data

• Wavefront snapshots
• Synthetic Shot Records
• Streamer
• OBC
• VSP

25
Wave Propagation Example
2
Shot 2 Complex Geology, x9000
26
Shot 2 Complex Geology
time
0
27
Shot 2 Complex Geology
time
0.25
28
Shot 2 Complex Geology
time
0.50
29
Shot 2 Complex Geology
time
0.75
30
Shot 2 Complex Geology
time
1.00
31
Shot 2 Complex Geology
time
1.25
32
Shot 2 Complex Geology
time
1.50
33
Shot 2 Complex Geology
time
1.75
34
Shot 2 Complex Geology
time
2.00
35
Shot 2 Complex Geology
time
2.25
36
Shot 2 Complex Geology
time
2.50
37
Shot 2 Complex Geology
time
2.75
38
Shot 2 Complex Geology
time
3.00
39
Shot 2 Complex Geology
time
3.25
40
Shot 2 Complex Geology
time
3.50
41
Shot 2 Complex Geology
time
3.75
42
Shot 2 Complex Geology
time
4.00
43
Shot Record Examples
1
2
44
Shot 1 Streamer and OBC Pressure
-4 Offset (km)
8 -4 Offset (km)
8
0
Streamer pressure
OBC pressure
t (s)
4
45
Shot 1 OBC Velocity components
-4 Offset (km)
8 -4 Offset (km)
8
0
Vx
Vz
t (s)
4
46
Shot 1 VSP records at offset500m
3.5 Z (km) 0 3.5
Z (km) 0 3.5 Z (km)
0
0
t (s)
4
pressure
Vx
Vz
47
Shot 2 Streamer and OBC Pressure
-4 Offset (km)
8 -4 Offset (km)
8
0
Streamer pressure
OBC pressure
t (s)
4
48
Shot 2 OBC Velocity components
-4 Offset (km)
8 -4 Offset (km)
8
0
Vx
Vz
t (s)
4
49
Shot 2 VSP records at offset500m
3.5 Z (km) 0 3.5
Z (km) 0 3.5 Z (km)
0
0
t (s)
4
pressure
Vx
Vz
50
Data Sets

• Four datasets are available from UH
• Full dataset (all offsets, all components)
• Acoustic dataset (all offsets)
• A 6km acoustic towed streamer subset
• Wavefront snapshots
• A Marmousi2 website is being created
• www.uh.edu/Geosciences/Marmousi2 ??

51
Processing and Migration Examples
52
Processing and Migration Examples

• Using the 6km acoustic streamer subset
• Minimum processing
• Correct for source/receiver depths
• Shift data by 72ms (correct for wavelet)
• NMO/Stack
• Poststack time migration
• Poststack depth migration
• Prestack time migration
• Prestack depth migration

53
Simple NMO/Stack
54
ProMAX Stolt PoSTM
55
ProMAX Kirchhoff PoSDM
56
PrimeTime Kirchhoff PreSTM
57
Primus Kirchhoff PreSDM
Shortest path
58
Primus Kirchhoff PreSDM
Max energy
59
Optimus Wave Equation PreSDM
60
Conclusion
61
Conclusion

• A complex structure 2D elastic model with
  realistic hydrocarbons and stratigraphic features
  has been created.
• A high quality elastic finite difference
  synthetic data set has been created.
• Model and data are freely available.

62
Conclusion

• Uses
• AVO calibration
• Velocity estimation
• Migration algorithm evaluation
• Multi-component and VSP evaluation
• Multiple suppression, etc.
• Limitations
• Model and data are 2D.

63
Acknowledgements

• GX Technology Corporation
• Don Larson, GX Technology
• Aline Bougeois, IFP
• Fred Hilterman, GDC