Short range geoacoustic inversion Short range geoacoustic inversion - - PowerPoint PPT Presentation

Yong-Min Jiang1 , N. Ross Chapman1 and Peter Gerstoft2

Short range geoacoustic inversion with a vertical line array Short range geoacoustic inversion with a vertical line array

1University of Victoria, Victoria, BC, Canada 2Marine Physical Laboratory, Scripps Institution of

Oceanography, La Jolla, CA, USA

Work supported by ONR Work supported by ONR Work supported by ONR

Jiang, Chapman and Gerstoft ASA155, Paris, France, 29 June – 4 July 2007 MPL

Jiang, Chapman and Gerstoft ASA155, Paris, France, 29 June – 4 July 2007 MPL

Objective: Objective:

Travel time geoacoustic inversion in a time varying ocean environment using broadband signal and vertical line array

• Experimental site & geometric setting up
• The time varying environment
• The inversion procedure
• Energy function, forward model and inversion algorithm
• Comparison of the inversion results with geophysical survey
• Conclusions

Outline: Outline:

Experimental site & set up Experimental site & set up

Jiang, Chapman and Gerstoft ASA155, Paris, France, 29 June – 4 July 2007 MPL

15 m 25 m 35 m 45 m 55 m 65 m

Jiang, Chapman and Gerstoft ASA155, Paris, France, 29 June – 4 July 2007 MPL

The time varying oceanic SSP The time varying oceanic SSP

start end

Jiang, Chapman and Gerstoft ASA155, Paris, France, 29 June – 4 July 2007 MPL

The effect of SSP on signal amplitude The effect of SSP on signal amplitude

Jiang, Chapman and Gerstoft ASA155, Paris, France, 29 June – 4 July 2007 MPL

The effect of SSP on arrival time The effect of SSP on arrival time

Travel time inversion Travel time inversion

Jiang, Chapman and Gerstoft ASA155, Paris, France, 29 June – 4 July 2007 MPL

15 m 25 m 35 m 45 m 55 m 65 m

The difficulties of reflection coefficient inversion

• Signal amplitude fluctuation, micro-path

interference in the direct arrivals

• We need to treat each source depth data

separately because of time varying SSP

Travel time inversion procedure

• experimental geometry at SD = 65 m:

WD, R, SD and VLA tilt

• at other source depths, invert for

water column SSP

• sediment sound speed and layer thickness

at all source depths

Environmental parameterization Environmental parameterization

Jiang, Chapman and Gerstoft ASA155, Paris, France, 29 June – 4 July 2007 MPL

• Water column SSP is included in the

inversion in terms of EOF

• SSP casts measured on JD239 +

2 SSPs measured before and after the experiment

• Fit only the thermocline region in

the inversion

• Only the first 4 EOF coefficients are

included in the inversion

Geoacoustic parameterization Geoacoustic parameterization

Jiang, Chapman and Gerstoft ASA155, Paris, France, 29 June – 4 July 2007 MPL

Source depth: 25 m Time (ms) H yd ro p h

• n

e n

• .

(a) dire ct arriva l

S R B R SRB R BR S R

20 40 60 15 13 11 09 07 05 03 01

Source depth: 35 m Time (ms) H yd ro p h

• n

e n

• .

(b) direc t a rriv al

S R B R SRB R BR S R

20 40 60 15 13 11 09 07 05 03 01

Source depth: 45 m Time (ms) H yd ro p h

• n

e n

• .

(c) dire ct arriva l

S R B R SRB R BR S R

20 40 60 15 13 11 09 07 05 03 01

Source depth: 55 m Time (ms) H yd ro p h

• n

e n

• .

(d) dire ct arriva l

S R B R SRB R BR SR

20 40 60 15 13 11 09 07 05 03 01

Source depth: 65 m Time (ms) H yd ro p h

• n

e n

• .

(e) direc t a rrival

S R B R SRB R BR SR

20 40 60 15 13 11 09 07 05 03 01

Geoacoustic parameterization is based on the resolvable layers in the data

Geoacoustic parameterization continued… Geoacoustic parameterization continued…

Jiang, Chapman and Gerstoft ASA155, Paris, France, 29 June – 4 July 2007 MPL

Energy function, forward model and inversion algorithm Energy function, forward model and inversion algorithm

Jiang, Chapman and Gerstoft ASA155, Paris, France, 29 June – 4 July 2007 MPL

∑∑

= =

− =

L H

N i N j j i j i

m T t m E

1 1 2 , ,

)) ( ( ) (

Energy function: Forward model: Based on ray theory. Inversion algorithm: Optimization algorithm - ASDE (Adaptive Simplex Differential Evolution) Embed downhill simplex (local search) into differential evolution (global search)

Example of SSP inversion result Example of SSP inversion result

Jiang, Chapman and Gerstoft ASA155, Paris, France, 29 June – 4 July 2007 MPL

Travel time fit for SSP inversion Inverted SSP for SD = 25 m data

Example of geoacoustic inversion result Example of geoacoustic inversion result

Jiang, Chapman and Gerstoft ASA155, Paris, France, 29 June – 4 July 2007 MPL

Travel time fit for I-layer geoacoustic inversion Travel time fit for 2-layer geoacoustic inversion

Comparison with geophysical survey data Comparison with geophysical survey data

Jiang, Chapman and Gerstoft ASA155, Paris, France, 29 June – 4 July 2007 MPL

Source depth Parameter 25 m 35 m 45 m 55 m 65 m layer thickness (m)

21.5 20.6 22.3 22.3 20.4

sound speed (m/s)

1609 1597 1600 1600 1584

TWT (ms)

26.7 25.8 27.9 27.9 25.8

Source depth layer I layer II sound speed layer thickness sound speed layer thickness 55 m

1581 m/s 11.9 m 1602 m/s 9.1 m

65 m

1584 m/s 14.5 m 1611 m/s 8.4 m

TABLE I. Travel time inversion results for one layer

• ver half space model

TABLE II. Travel time inversion results for two-layer

• ver half space model

WP19, ~ 230 m

From John Goff

Summary Summary

• Water column SSP has great impact on geoacoustic
• inversion. SSP in terms of EOFs included in the inversion

( geometric, geoacoustic parameters)

• The limitation
• Future work

Quantify the uncertainty of the estimates

Acknowledgements Acknowledgements

• Office of Naval Research: for sponsoring the research
• Dr. William Hodgkiss from MPL (acoustic data)
• Dr. John Goff from Institute of Geophysics, University of Texas

at Austin (geophysical chirp seismic reflection data)

Jiang, Chapman and Gerstoft ASA155, Paris, France, 29 June – 4 July 2007 MPL