Kui Zhang* and Kurt J. Marfurt

2008 AASPI Consortium annual meeting

Migration-driven velocity analysis for anisotropy/stress analysis

Motivation Our industry is acquiring more and more wide azimuth data.

Traditional seismic processing often ignores the azimuth information and implements on the full azimuth gathers.

The presence of azimuthal velocity anisotropy can degrade image quality.

The analysis of azimuthal velocity anisotropy and/or azimuthal AVO provides valuable information about fracture densities and stress orientation. Current leaders are:• Axis (NMO-based),• Weinman (migration-based).

Motivation Challenges:

• Statics caused by different elevations and the LVL zone.(short wavelength and long wavelength)

• Structural dip can bias the anisotropic analysis.• Measured elliptical NMO velocities is usually converted to

interval velocity by Dix equation. (ambiguity)

We use Kirchhoff prestack migration-driven approach to generate:

• More accurate NMO azimuthal velocities• High resolution imaging

New azimuth binning method

Vfast

Vslow

Fault plane

xi

xs

xg

(xi-xs)+

(xi-x

g)

(xi-xs)

(xi-x

g)

11°34°

56°

79°

101°

124°

146°

281°

304°

326°349°

EW

N

169°191°214°

236°

259°

S

1

1

2

23

34

4

556

67

78

8

1 km

1 2

3 4

5 6

7 8

Amplitude

Positive

Negative

0

N

ConventionalBinning

1 2

3 4

5 6

7 8

1 km

Amplitude

Positive

Negative

0

N

NewBinning

(Perez and Marfurt, 2008)

New P wave azimuthal velocity analysis approach

Coarse grid selectionNMO velocity Vnmo

Choose α and φV=Vnmo+ α cosφ

Migrate and compute

semblance

more α or φ

Yes

Scan for maximum semblance over αm and φm

migrate over αm and φm

α

+10%

0%

Maximum semblance

0 180

0 1

New converted wave azimuthal velocity analysis approach

t 1p =(

t 02 +x

si2 /v

p2 )1

/2

t 2p =(t

02 +x gi

2 /v p2 )1

/2

t 2s =(t 0

2 +x gi2 /v s

2 )1/2

t=0P P P S For P wave migration:

For PS wave migration:

gs

Ppsgp ttdm

,21 )(

gs

Spsgs ttdm

,21 )(

Vs/Vp

Maximum semblance

Acknowledgement

We thank all sponsors of AASPI consortium for their support.