Coastal wetland monitoring using multi-frequency polarimetric SAR - - PowerPoint PPT Presentation

coastal wetland monitoring using multi frequency
SMART_READER_LITE
LIVE PREVIEW

Coastal wetland monitoring using multi-frequency polarimetric SAR - - PowerPoint PPT Presentation

POLINSAR 2009 Coastal wetland monitoring using multi-frequency polarimetric SAR Sang-Eun PARK 1 , W ooil M. MOON 2 ,3 , Duk-jin KI M 3 , and Eric POTTI ER 1 1. University of Rennes 1, IETR, Bat 11D, 263 Avenue General Leclerc, 35042 RENNES,


slide-1
SLIDE 1

Coastal wetland monitoring using multi-frequency polarimetric SAR

Sang-Eun PARK1, W ooil M. MOON 2 ,3, Duk-jin KI M 3, and Eric POTTI ER1

  • 1. University of Rennes 1, IETR, Bat 11D, 263 Avenue General Leclerc,

35042 RENNES, France.

  • 2. Geophysics, University of Manitoba, Winnipeg, MB R3T 2N2, Canada.
  • 3. School of Earth and Environmental Sciences, Seoul National University,

Gwanak-gu, silrim-dong, Seoul 151-747, Korea

POLINSAR 2009

slide-2
SLIDE 2

2

SAPHIR

Introduction

Tidal wetlands (intertidal flats) zones of interaction between marine and terrestrial environments Highly productive fishery areas and have dynamic and diverse ecosystems. Need environmental management High development pressure (e.g. reclamation and marine pollution) Highly vulnerable to climate changes.

Quantitative estimation of the surface geophysical parameters in tidal wetlands from polarimetric SAR data sets

  • Objective:
slide-3
SLIDE 3

3

SAPHIR

Suncheon Bay

Study Area – Suncheon Bay

R A D A R S A T

  • 2

NASA/JPL AIRSAR

Suncheon City Goheung City

slide-4
SLIDE 4

4

SAPHIR

Study Area – Suncheon Bay

slide-5
SLIDE 5

5

SAPHIR

Geophysical Parameter Retrieval

( )

Ω → Θ F

Forward Scattering models

( )

Ω = Θ

−1

F

Finding the set of unknown surface parameters from polarization measurement

Independent combination of polarization measurements

= Θ = Ω

Dielectric constant Height distribution (s) Autocorrlation (l) I ntegral Equation Method Extended-Bragg Model

( )

( )

∞ =

− − =

1 2 2 2 2 2

! ) , 2 ( 2 exp 2

j z j PP j z PP

j k W I s s k k σ

[Fung et al., 1992; Fung, 1994; Wu et al., 2001 ] [Schuler et al., 2002; Hajnsek et al., 2003] Coherency m atrix ( Bragg scatter and the orientation of facets)

[ ] [ ]

) 8 exp( 1 ) 8 exp( 1 ) 2 exp( ) 2 exp( 2 ] [

2 3 2 3 2 * 2 2 2 1

⎥ ⎥ ⎥ ⎦ ⎤ ⎢ ⎢ ⎢ ⎣ ⎡ − − − + − − = σ σ σ σ C C C C C m T

s

sin slope rms θ σ =

slide-6
SLIDE 6

6

SAPHIR

Roughness Retrieval

Surface sediments can be assumed to be fully saturated with water throughout intertidal mudflats. The effect of the dielectric constant on backscatter signals can thus be neglected in the specific case of the intertidal mudflats. The intertidal mudflat is an ideal study site for the reduction of surface variables. Roughness parameters can be estimated from two independent polarization observations.

Inversion from co-pol using IEM Inversion using the extended-Bragg

⎥ ⎦ ⎤ ⎢ ⎣ ⎡ = ⎥ ⎦ ⎤ ⎢ ⎣ ⎡

− 1 VV HH

F l s σ σ ⎥ ⎥ ⎦ ⎤ ⎢ ⎢ ⎣ ⎡ + = ⎥ ⎦ ⎤ ⎢ ⎣ ⎡

− RRLL VV HH

S S G l s ρ

2 1

slide-7
SLIDE 7

7

SAPHIR

Experimental Results – AIRSAR

NASA/JPL AIRSAR PACRIM-2 (2000.09.30) L-band, Full-pol

slide-8
SLIDE 8

8

SAPHIR

, from

VV HH

s σ σ

2

, from

VV HH RRLL

S S s + ρ

0 0.4 cm

rms height

Experimental Results – AIRSAR

, from

VV HH

l σ σ

2

, from

VV HH RRLL

S S l + ρ

0 6 cm

Correlation length

rms error of the rms height 0.1 0.1 3.8 8.4 rms error of the correlation length

slide-9
SLIDE 9

9

SAPHIR

  • 2008.12.14
  • Fine Quad Pol (FQ6)
  • Ascending
  • Incidence angle:

24.5°~26.4°

RADARSAT-2

slide-10
SLIDE 10

1 0

SAPHIR

Experimental Results – RADARSAT-2

0.3cm rms height from AIRSAR rms height from RADARSAT-2

slide-11
SLIDE 11

1 1

SAPHIR

Experimental Results – RADARSAT-2

AIRSAR 2000.09.30 RADARSAT-2 2008.12.14

slide-12
SLIDE 12

1 2

SAPHIR

Experimental Results – RADARSAT-2

AIRSAR 2000.09.30 RADARSAT-2 2008.12.14

0 0.7 cm

rms height

slide-13
SLIDE 13

1 3

SAPHIR

Study Area – Kyunggi Bay

A L O S P A L S A R R A D A R S A T

  • 2
  • 2008.11.21
  • Quad Pol (FQ16)
  • Descending
  • Incidence angle:

35.5°~37.0°

  • 2007.04.13
  • Quad Pol Mode
  • Ascending
  • Incidence angle: 21.5°

SEOUL

slide-14
SLIDE 14

1 4

SAPHIR

ALOS PALSAR

Natural intertidal m udflats Reclaim ed m udflats

Surface sediments are saturated with sea water Surface water drained out Mud crack developed

Sea dike

slide-15
SLIDE 15

1 5

SAPHIR

Geophysical Parameter Inversion

Roughness Inversion for reclaimed mud flats

⎥ ⎥ ⎥ ⎦ ⎤ ⎢ ⎢ ⎢ ⎣ ⎡ + = ⎥ ⎥ ⎥ ⎦ ⎤ ⎢ ⎢ ⎢ ⎣ ⎡

− 1 2 1

| | | | α ρRRLL

VV HH V

S S F l s m

Impossible to reduce the number of surface variables Fully polarimetric inversion

Soil parameters Min. Max. Interval mV 0.1 0.6 0.01 s 0.1 cm 1.5 cm 0.1 cm l 1 cm 15 cm 0.1 cm

Simulations

slide-16
SLIDE 16

1 6

SAPHIR

Experimental Results – PALSAR

rms height Correlation length Moisture content

0 s 0.4 cm mV 0.6 0 l 12 cm Intertidal flats Reclaimed mudflat Reclaimed mudflat

slide-17
SLIDE 17

1 7

SAPHIR

Experimental Results – PALSAR

rms heights probably connected with the surface sediment texture of the intertidal flats

[Korea Ocean Research & Development Institute, 2004]

slide-18
SLIDE 18

1 8

SAPHIR

Experimental Results – RADARSAT-2

mV 0.6 0 l 10 cm Intertidal flats Reclaimed mudflat Reclaimed mudflat 0 s 0.3 cm

rms height Correlation length Moisture content rms height (PALSAR) Lim itation of m odel Calibration Missing param eter

slide-19
SLIDE 19

1 9

SAPHIR

Future Study

A L O S P A L S A R R A D A R S A T

  • 2

SEOUL TerraSAR-X TerraSAR-X

slide-20
SLIDE 20

2 0

SAPHIR

Future Study

2007.04.13 2008.11.21

PALSAR RADARSAT-2

slide-21
SLIDE 21

2 1

SAPHIR

Conclusions

Application of polarimetric SAR data for intertidal flats was investigated with inversion of the roughness characteristics of surface sediments. First results of surface roughness parameters retrieval from L-band and C-band space-borne polarimetric SAR observations. Continuous monitoring using the multi-temporal, multi-frequency space-born SAR data sets, such as ALOS PALSAR (L-band), RADARSAT-2 (C-band), and TerraSAR-X (X-band).