Slowness Driven Gaussian Beam Prestack Depth Migration for Lowfold Seismic Data

1
Slowness Driven Gaussian Beam Prestack Depth
Migration for Low-fold Seismic Data

• PhD Student Chaoshun Hu
• Supervisor Paul Stoffa
• February, 2009

2
Outline

• Introduction and Motivation
• Theory of Gaussian Beam Methods
• Details of Slowness Driven Gaussian Beam
  Migration
• Benchmark Examples
• Discussion and Conclusions

3
Introduction and Motivations
4
Introduction

• Typical Problems for Complex Media Imaging
• Poor refraction and reflection coverage of the
  primary targets (subduction plate boundary,
  accretionary prism)
• Coherent noise and multiples
• Incorrect velocity model building (for imaging
  and lithological interpretation)

5
Major Imaging Methods

• Ray Method
• Kirchhoff migration
• Gaussian beam (GBeam) migration
• Numerical Methods
• One-way wave equation (WE) migration
• Two-way reverse time migration (RTM)

(not accurate in complex media)
Too expensive for large scale crustal imaging
Kirchhoff
GBeam
WE
(Albertin et al., 2002)
6
Gaussian Beam Migration

• Gaussian Beam Migration (Hill, 1991, 2001)
• Efficient
• Accurate
• Flexible

6/37
7
Motivation 1 Subsalt Imaging Problem

• The Four Big Challenges of Subsalt Imaging
• Illumination problems Shadow zones due to high
  contrast salt
• Inaccurate velocity models
• Limitations of seismic imaging algorithm
• 4. Multiple reflections.

( Etgen, J., 2004)
7/37
8
Motivation 2 Seismic Resolution Issue
Main Lobe
hm
Side Lobe
hs
Horizontal Resolution 8 wavelength
When depth increases, velocity is increasing and
high frequency is attenuated. Therefore,
wavelength increases and seismic resolution will
be decreased.
Velocity(v)
Wavelength(?)
Frequency(f)
8/37
9
Basic Theory of Gaussian Beam Methods
9/37
10
What is a Gaussian Beam
GBeam is an asymptotic solution to
parabolic WE on ray-centered coordinate
where
is the velocity
is one of the local ray coordinates
is the distance between neighbored rays
is the curvature of the wavefront
can be calculated with Dynamic Ray Tracing
10/37
11
Plane Wavefield Synthesis with Gaussian Beams
12
Common Shot Gaussian Beam Migration
Receiver Wavefield (Local Plane Wave Beam)
Source
Image ( Image correlation of source and receiver
wavefield )
13
Flowchart Common Shot Gaussian Beam Migration
Read Velocity Model
Read common shot gathers
Calculate the traveltime and
from the
shot s to subsurface position x

Dividing the receivers to several beams. In each
beam, the seismic records are decomposed to many
Gaussian beam components by local slant stack.

For each beam, we can calculate the subsurface
travel- time and amplitude
using dynamic ray tracing from the
beam center.
For each beam, get partial images using tt( x ,
s )t( x , r ) condition
Stack and output the final image
14
Details of Slowness Driven Gaussian Beam Migration
15
Beams, Local Slant Stack, Beam Semblances and P
Gather Calculation
For single beam
If we choose the beam center for each trace
location, after the calculation of all the beam
centers, we can get a new P gather.
15/37
16
Weighting Summation in a Fresnel Zone
16/37
17
UTIG Slowness Driven GBeam Migration
17/37
18
UTIG Slowness Driven GBeam Migration
18/37
19
Simple Model OBS Migration
19/37
20
Simple Model Imaging
20/37
Acquisition Geometry
21
Simple Model Imaging
21/37
OBS Gather
22
Simple Model Imaging
22/37
Ray Parameter Gather
23
Simple Model Imaging
23/37
Conventional Single Arrival GBeam Migration
24
Simple Model Imaging
New Single Arrival Slowness Driven GBeam Migration
24/37
25
Sub-basalt OBS Imaging
25/37
26
Sub-basalt OBS Imaging
Shots
OBS
Basalt (4.5km/s)
Target
Basalt velocity model and acquisition geometry
26/37
27
Sub-basalt OBS Imaging
27/37
OBS and ray parameter gather
28
Sub-basalt OBS Imaging
(No AGC)
28/37
Conventional Single Arrival GBeam Migration
29
Sub-basalt OBS Imaging
(No AGC)
29/37
New Single Arrival Slowness Driven GBeam Migration
30
Sub-basalt OBS Imaging
(No AGC)
30/37
New Multiple Arrivals Slowness Driven GBeam
Migration
31
Real Data Imaging (Line 56 from NicStrata
Project, Courtesy of Kirk McIntosh)
31/37
32
Real Data Imaging
Conventional GBeam Depth Migration Using 1010
shots
32/37
33
Real Data Imaging
Slowness Driven GBeam Depth Migration Using 100
shots
33/37
34
Real Data Imaging
Zoomed View of Full Conventional GBeam Depth
Migration
34/37
35
Real Data Imaging
Zoomed View of Sparse Slowness Driven GBeam
Depth Migration
35/37
36
Discussion and Conclusions

• The new slowness driven Gaussian beam depth
  migration
• works fine for wide angle data like OBS
• control beams easily and efficiently
• enhance migration resolution and SNR, attenuate
  subbasalt coherent noises
• promising method for sub-basalt imaging

36/37
37
Acknowledgements

• Ross Hill
• Yue Wang
• Mrinal Sen
• Sergey Fomel
• Kirk McIntosh

• Financial Support
• UTIG TAIGER project (NSF continental dynamics
  project)
• Chevron Scholarship
• BP fellowship
• Maurice Ewing Fellowship
• Jackson School Geology Foundation

37/37