A Study of Dense Medium Scattering and Its Applications in Sea Ice Research in Ross Island, Antarctica Y. J. Lee 1, W. K. Lim 2, H. T. Ewe 1 and H. T.

A Study of Dense Medium Scattering and Its Applications in Sea Ice Research in Ross Island, Antarctica

Y. J. Lee1, W. K. Lim2, H. T. Ewe1 and H. T. Chuah1

1Universiti Tunku Abdul Rahman2Multimedia University

• Introduction• Development and Application of

the Dense Medium Phase and Amplitude Correction Theory (DM-PACT)

• Modeling of the Sea Ice and Ice Shelf

• Development of Inverse Scattering Models for Sea Ice Thickness Retrieval

• Conclusion

Contents

Introduction • Development of remote sensing

technology – satellites and synthetic aperture radar (SAR).

• Increase in use of remote sensing devices for data retrieval.

• Many applications evolving with the technology.

• Need for proper forward and inverse scattering models for such applications.

K.M. Golden, M. Cheney, K.H. Ding, A.K. Fung, T.C. Grenfell, D. Isaacson, J.A. Kong, S.V. Nghiem, J. Sylvester, and D.P. Winebrenner, “Forward electromagnetic scattering models for sea ice,” IEEE Transactions on Geoscience and Remote Sensing, Vol. 36, No. 5, pp. 1655–1674, 1998

• Many forward models were developed.• Several models utilized the Radiative

Transfer Theory (RTT), which can be written in the form (Chandrasekhar 1960):

• I, e, P, dΩ and z are the Stokes vector, extinction matrix, phase matrix of the medium, solid angle and vertical direction respectively.

Development of the DM-PACT

[1] cos dIPIdz

Ide

Chandrasekhar, S., Radiative Transfer. New York: Dover, 1960

• Early modeling assumes a homogeneous medium – scatterers in far field from one another.

• Only applies when:• Scatterers are small.• Average spacing larger than /3.

• Solution: Improvements were made to the phase matrix, P, in [1] using the Antenna Array Theory.

• The improved phase matrix, P has the following expression:

• is the dense medium phase correction factor (Chuah et al. 1996) and S is the Stokes’ matrix for Mie scatterers with the Close Spacing Amplitude Correction (Fung and Eom 1985).

[2] ||)',';,( 2

hhhv

vhvvn PP

PPSP

n

2

Chuah, H.T., S. Tjuatja, A.K. Fung, and J.W. Bredow, “A phase matrix for a dense discrete random medium: evaluation of volume scattering coefficient,” IEEE Transactions on Geoscience and Remote Sensing, Vol. 34, No. 5, pp. 1137–1143, 1996A.K. Fung, and H.J. Eom, “A study of backscattering and emission from closely packed inhomogeneous media,” IEEE Trans. Geoscience & Remote Sensing, Vol. 23, No. 5, pp. 761–767, 1985

• takes into account the close spacing effect of scattering among the scatterers.

• S relates the scattered intensities to the incident intensities of the scatterer.

n

2

• From equation [2], can further be expressed as the following:

• where• d is the average distance between scatterers• k is the propagation vector• l is the correlation length

]3[)()()(4

exp!

)(

1

1

223

22

3

3

2

22

22

qzyx

siq

sik

k

n

kakakaq

lk

d

l

qq

k

d

e

d

e

si

si

a kq

l

d

k l

qerf

qd l jk l

qrr r( ) exp Re

( / )

2 2

4 2

n

2

The improved phase matrix – Dense Medium Phase and Amplitude Correction Theory (DM-PACT).

• An analysis using the DM-PACT based on iterative solution on electrically dense medium was done (Ewe et al.).

S. Tjuatja, A.K. Fung, and J. Bredow, “A scattering model for snow-covered sea ice,” IEEE Trans. Geoscience & Remote Sensing, Vol. 30, No. 4, pp. 804–810, 1992H.T. Ewe, and H.T. Chuah, “An analysis of the scattering of discrete scatterers in an electrically dense medium,” 1998 IEEE International Geoscience and Remote Sensing Symposium Proceedings (IGARSS'98), Vol. 5, pp. 2378–2380, 1998H.T.Ewe, and H.T. Chuah, “A Study of Dense Medium Effect Using A Simple Backscattering model,” 1997 IEEE International Geoscience and Remote Sensing Symposium Proceedings (IGARSS'97): Remote Sensing - A Scientific Vision for Sustainable Development, Vol. 3, pp. 1427–1429, 1997

• A backscattering model was developed (Ewe et al.) for inhomogeneous media.

• Model incorporates:• Iterative solution for RTT• DM-PACT• IEM Model to characterize rough surface

Application of the DM-PACT

H.T. Ewe, M.E. Veysoglu, C.C. Hsu, R.T. Shin, and J.A. Kong, “Radiative Transfer Theory for Remote Sensing of Sea Ice”, 1994 IEEE International Geoscience and Remote Sensing Symposium Proceedings (IGARSS'94): Surface and Atmospheric Remote Sensing – Technologies, Data Analysis an Interpretation, Vol. 1, pp. 623–625, 1994H.T. Ewe, H.T. Chuah, and A.K. Fung, “A backscatter model for a dense discrete medium: Analysis and numerical results,” Remote Sensing of Environment, Vol. 65, No. 2, pp. 195–203, 1998Fung, A.K., Microwave Scattering and Emission Models and Their Applications, Norwood, MA: Artech House, 1994C.Y. Hsieh, and A.K. Fung, “Application of an Extended IEM to Multiple Surface Scattering and Backscatter Enhancement,” Journal of Electromagnetic Waves and Applications, Vol. 13, No. 1, pp. 121–136, 1999A.K. Fung, W.Y. Liu, K.S. Chen, and M.K. Tsay, “An Improved IEM Model for Bistatic Scattering from Rough Surface,” Journal of Electromagnetic Waves and Applications, Vol. 16, No. 5, pp. 689–702, 2002

• The backscattering model developed by Ewe et al. was applied to sea ice and ice shelf.

• Ground truth measurement was conducted at Ross Island, Antarctica to verify the model.

• Satellite images were also acquired as part of the study.

Modeling of the Sea Ice and Ice Shelf

• Backscattering model was initially tested for single layer – sea ice.

• Model configuration:Z

θsθ

Air-Sea Ice Interface

Sea Ice-Ocean InterfaceOcean (Lower Half Space)

Air Layer

Sea ice Layer

M.D. Albert, T.E. Tan, H.T. Ewe, and H.T. Chuah, “A theoretical and measurement study of sea ice and ice shelf in Antarctica as electrically dense media,” Journal of Electromagnetic Waves and Applications, Vol. 19, No. 14, pp. 1973–1981, 2005

• Simulation results:Theoretical and Measured Backscattering Coefficient Sites A – P

Back

scatt

eri

ng C

oeffi

cient

(dB

)

0

–5

–10

–15

–20

–25

–30

–35

–40

MeasuredTheoretical

A B C D E F G H I J K L M N O PSites

• A multilayer model was developed to take into account the snow cover on the sea ice (Mohan et al.).

• Extension of the backscatter model by Ewe et al.

• Results show good matching between the multilayer model, Matrix Doubling method and CEAREX measurements.

Y.J. Lee, H.J. Yap, M.D. Albert, H.T. Ewe, and H.T. Chuah, “Multiyear Field Measurement and Sensitivity Study of Radar Returns in Scott Base, Antarctica,” Proceedings for the 3rd Malaysian International Seminar on Antarctica (MISA3): From the Tropics to the Poles, Sabah, Malaysia, pp. 21–28, July 2007M.D., Albert, H.T. Ewe, and H.T. Chuah, “Understanding the Scattering Mechanism in Sea Ice and Its Relation to Remote Sensing”, Joint Scientific Committee on Antarctic Research (SCAR)-International Arctic Science Committee (IASC) Open Science Conference, St Petersburg, Russia, 8–11 July 2008

• Model configuration: Z

θ θs

Air-Snow interface

Snow-Sea Ice interface

Sea ice

Snow layer

Ocean (half space)

Sea Ice-Ocean interface

Sea Ice

-35

-30

-25

-20

-15

-10

-5

0

10 20 30 40 50 60 70

Multi-Layer Model RT VV

Matrix Doubling VV

CEAREX Like Polarized DS 9

Multi-Layer Model RT VH

Matrix Doubling VH

CEAREX Cross Polarized DS 9

Ba

cksc

att

erin

g C

oe

ffic

ien

t (d

B)

Incident Angle (Degree)

• Many inverse models were developed to retrieve sea ice thickness.

• Based on existing models, new models utilizing the RT and DM-PACT were later developed.

Development of Inverse Scattering Models for Sea Ice Thickness Retrieval

Golden, K.M., D. Borup, M. Cheney, E. Cherkaeva, M.S. Dawson, K.H. Ding, A.K. Fung, D. Isaacson, S.A. Johnson, A.K. Jordan, J.A. Kong, R. Kwok, S.V. Nghiem, R.G. Onstott, J. Sylvester, D.P. Winebrenner, and I.H.H. Zabel, “Inverse electromagnetic scattering models for sea ice”, IEEE Transactions on Geoscience and Remote Sensing, Vol. 36, No. 5, 1675–1704, 1998

• For active remote sensing:• Radiative Transfer Inverse Scattering Model

(RTISM)• For passive remote sensing:

• Neural Network with RT-DMPACT• Genetic Algorithm with RT-DMPACT

Y.J. Lee, W.K. Lim, and H.T. Ewe, “A Study Of An Inversion Model For Sea Ice Thickness Retrieval In Ross Island, Antarctica”, Progress in Electromagnetics Research (PIER), Vol. 111, pp. 381–406, 2011H.J. Yap, W.K. Lim, H.T. Ewe and H.T. Chuah, “Passive Microwave Remote Sensing for Sea Ice Thickness Retrieval Using Neural Network and Genetic Algorithm”, Proceedings in Progress In Electromagnetics Research Symposium (PIERS), Beijing, China, pp. 1229–1233, 23–27 March 2009H.J. Yap, W.K. Lim, H.T. Ewe and H.T. Chuah, “Neural Network and Genetic Algorithm Inversion for Sea Ice Thickness using Passive Microwave Remote Sensing”, 4th Malaysian International Seminar on Antarctica (MISA4): Legacy of IPY to the Tropics, Petaling Jaya, Malaysia. Poster presented. 1–3 April 2009

• Flowchart for the Radiative Transfer Inverse Scattering Model (RTISM):

- Radiative Transfer Inverse Scattering Model (RTISM)

RT-DMPACT Forward Model

Expected Backscatter Data

Initial Guessed ThicknessEstimated Sea Ice

Thickness

Backscatter Data from satellite images

Error Comparison

Error Acceptable

Optimization AlgorithmImproved Guessed

Thickness

Error Unacceptable

RT-DMPACT Forward Model

Expected Backscatter Data

Initial Guessed ThicknessEstimated Sea Ice

Thickness

Backscatter Data from satellite images

Error Comparison

Error Acceptable

Optimization AlgorithmImproved Guessed

Thickness

Error Unacceptable

• Sea ice thickness retrieval using the Radiative Transfer Inverse Scattering Model (RTISM):

1

1.5

2

2.5

0 1 2 3 4 5 6

Comparison Between the Estimated and Measured Sea Ice Thickness for 2006

Estimated Sea Ice ThicknessMeasured Sea Ice Thickness

Se

a Ic

e T

hic

kne

ss (

m)

Site

• Flowchart for the Neural Network with RT-DMPACT:

- Neural Network with RT-DMPACT- Genetic Algorithm with RT-DMPACT

• Flowchart for the Genetic Algorithm with RT-DMPACT :

• Sea ice thickness retrieval :

1

1.2

1.4

1.6

1.8

2

2.2

2.4

2.6

1 2 3 4 5 6

Comparison Between NN and GA Inversion with the Measured Sea Ice Thickness

Measured Sea Ice Thickness (m)Estimated Sea Ice Thickness from NNEstimeted Sea Ice Thickness from GA

thic

kne

ss (

m)

site

Conclusion• Improvements to the phase matrix of the

Radiative Transfer equation was done.• The technique, named the DM-PACT is able to

better model the inhomogeneous media.• A backscattering model was later developed

based on RT and DM-PACT.• Application of the study towards the remote

sensing of sea ice was done.• The backscattering model was extended to a

multilayer model.• Inverse models were later developed based on

the backscattering model to retrieve sea ice thickness.

Proof of Global Warming

Thank you for your attention.