Karakterisasi Letusan Merapi menggunakan Data SAR (Synthetic

Aperture Radar)

Saepuloh et al., (2013), Interpretation of ground surface changes prior to the 2010 large eruption of Merapi volcano using ALOS/PALSAR, ASTER TIR and gas emission data, Journal of Volcanology and Geothermal Research, Vol. 261, pp. 130-143.

Saepuloh et al., (2010), SAR- and gravity change-based characterization of the distribution pattern of pyroclastic flow deposits at Mt. Merapi during the past ten years, Bulletin of Volcanology, Vol. 72, No. 2, pp. 221-232.

Asep Saepuloh, Dr. Eng. FITB, ITB

Content

1. An overview of Mt. Merapi Eruption 1996-2006

2. ALOS/PALSAR Observed 2010 Eruption

3. Eruption precursor using time series of D-InSAR

and ASTER TIR

Interaction of Microwave Signal to Surface

𝑃𝑅 = 𝑃𝑇 𝜎0𝐴𝐺2𝜆2

4𝜋 3𝑅4

The radar equation

Dielectric permittivity

Incident angle

Roughness funct.

σ0 = 4𝑘4ℎ02cos4𝜃𝑖 𝛼 2𝜔

αℎℎ =𝜇𝑟 − 1 𝜖𝑟𝜇𝑟 − sin2 𝜃𝑖 + 𝜇𝑟 sin2 𝜃𝑖 + 𝜇𝑟

2 𝜖𝑟 − 1

π 𝜇𝑟 cos 𝜃𝑖 + 𝜖𝑟𝜇𝑟 − sin2 𝜃𝑖

4

α𝑣𝑣 =𝜖𝑟 − 1 𝜖𝑟𝜇𝑟 − sin2 𝜃𝑖 + 𝜖𝑟 sin2 𝜃𝑖 + 𝜖𝑟

2 𝜇𝑟 − 1

π 𝜖𝑟 cos 𝜃𝑖 + 𝜖𝑟𝜇𝑟 − sin2 𝜃𝑖

4

Magnetic permeability

Sm

all

Per

turb

atio

n M

odel

• The Synthetic Aperture Radar (SAR) data were used.

• The superiority of SAR data means that they can provide periodic Earth observations regardless of the time or weather.

Study area

http://www.stanford.edu/

1. An Overview of the Last Decade of Mt. Merapi Eruption

• Ground check

- Geological observation - Rock sampling collection

Objectives and Data Used

Data used:

1. JERS-1

2. RADARSAT-1

Advantage: Sun-synchronous orbit with moderate incidence angle (39º-37º)

Limitation: Ten years observation covering four times of eruption (1996-2006)

JERS-1

RADARSAT-1

CSA

Objectives:

Detect and characterize the

pyroclastic flow deposits (P-

zone) each eruption period

Temporal SAR Data

Pair selection of SAR data to detect pyroclastic flow deposits with each eruption. A black triangle denotes a major eruption. A period of high

seismic activity is shown by a gray bar.

Intensity images of JERS-1 and RADARSAT-1 SAR data around Mt.

Merapi at a descending mode.

The P-zones are detected by comparing the two SAR data before and after events.

The two data with same path and raw reduce the geometrical error.

te-1

te+1

P-zone

te

Temporal Pattern Analysis of the P-zones

Schematic of the P-zone Parameters: A= Distribution area = Perimeter, D= Flow distance α= Included angle

γ= Collapse direction

Each selected pair is used to detect P-zone.

Then, the parameters of each detected P-zone are calculated and analyzed.

Temporal change of P-zones extracted from seven SAR data pairs. Thin arrows stand for

collapse directions of each P-zone.

Overlay of seven P-zones to highlight the temporal change of

distribution pattern.

2. ALOS/PALSAR Observed 2010 Eruption

Advanced Land Observing Satellite

Communication antenna

Solar Paddle PALSAR

PRISM

AVNIR-2

Item Specification1270 MHz / 23.6 cm

Chirp band width

Image modes

Single polarization (HH or VV)

Dual pol. (HH+HV or VV+VH)

Quad-pol. (HH+HV+VH+VV)

Off-nadir angle

Variable: 9.9 – 50.8 deg.

(inc. angle range: 7.9 - 60.0)

Look direction Right

Yaw steering ON

Swath width

70 km (single/dual pol.@41.5°)

30 km (quad-pol.@21.5°)

Ground resolution~ 9 m x 10 m (single pol.@41.5°)

~ 19 m x 10 m (dual pol.@41.5°)

~ 30 x 10 m (quad-pol.@21.5°)

~ 71-157m (4 look) x 100m (2 look)

Data rates

Orbit cycle 46 days

Centre frequency

28 MHz (single polarisation)

14 MHz (dual, quad-pol., ScanSAR)

ScanSAR (HH or VV; 3/4/5-beam)

ScanSAR: 20.1-36.5 (inc. 18.0-43.3)

350 km (ScanSAR 5-beam)

Rg (1 look) x Az (2 looks)

240 Mbps (single/dual/quad-pol)

120 or 240 Mbps (ScanSAR)

Mission:

To provide maps for Japan and other countries including those in the Asian-Pacific region (Cartography). To perform regional observation for "sustainable development", harmonization between Earth environment and development (Regional Observation). To conduct disaster monitoring around the world (Disaster Monitoring). To survey natural resources (Resources Surveying). To develop technology necessary for future Earth observing satellite (Technology Development).

Backscattering Intensity Before and After Eruption

2007.9.12

2010.11.15

Extracted Pyroclastic Flow

Deposits

Pyroclastic Flow Deposits 1996-2010

The eruption of Mt. Merapi in November 2010 has a different characteristic in comparison with the last decade of eruption. The coverage area of the

pyroclastic flow deposits is about 7 times larger than the eruption in 2006.

3. Eruption precursor using time series of ALOS/PALSAR and ASTER TIR

Three point of view:

A. Deformation

B. Lava dome shape

C. Surface thermal

D. EDM measurement

A. Deformation

23 data pairs with short temporal baseline were used to generate interferograms.

Pair

Acquisition Time Baseline

Path/Row Off nadir

angle Mode

1st 2nd Perpendicular

(m)

Temporal

(days)

1 2007/6/8 2006/12/6 2598 184 431/703 34.3 Ascending

2 2007/9/8 2007/1/21 979 230 431/703 34.3 Ascending

3 2007/10/24 2007/6/8 344 138 431/703 34.3 Ascending

4 2007/12/9 2007/9/8 140 92 431/703 34.3 Ascending

5 2008/1/24 2007/10/24 180 92 431/703 34.3 Ascending

6 2008/3/10 2007/12/9 188 92 431/703 34.3 Ascending

7 2008/4/25 2008/1/24 172 92 431/703 34.3 Ascending

8 2008/6/10 2008/3/10 553 92 431/703 34.3 Ascending

9 2008/7/26 2008/4/25 162 92 431/703 34.3 Ascending

10 2008/9/10 2008/6/10 78 92 431/703 34.3 Ascending

11 2008/10/26 2008/7/26 742 92 431/703 34.3 Ascending

12 2008/12/11 2008/9/10 109 92 431/703 34.3 Ascending

13 2009/1/26 2008/10/26 32 92 431/703 34.3 Ascending

14 2009/6/13 2008/12/11 292 184 431/703 34.3 Ascending

15 2009/10/29 2009/1/26 432 276 431/703 34.3 Ascending

16 2009/12/14 2009/6/13 56 184 431/703 34.3 Ascending

17 2010/1/29 2009/10/29 141 92 431/703 34.3 Ascending

18 2010/3/16 2009/12/14 216 92 431/703 34.3 Ascending

19 2010/6/16 2010/1/29 342 138 431/703 34.3 Ascending

20 2010/9/16 2010/3/16 218 184 431/703 34.3 Ascending

21 2010/11/1 2010/6/16 329 138 431/703 34.3 Ascending

22 2010/12/17 2010/9/16 96 92 431/703 34.3 Ascending

23 2011/2/1 2010/11/1 36 92 431/703 34.3 Ascending

Improvement of perpendicular baseline

Local Limitation of InSAR ALOS/PALSAR

Ascending and descending pairs are limited

Need assumption to obtain three displacement component

No GPS measurement since 2006 up to 2010

High tropospheric disturbance

Meteorological measurement is not available

Large amount of interferogram is required

Deformation is not constant over time

Need advanced method to separate the deformation and tropospheric effect

A. Deformation

Pair-wise Logic

Pair-wise logic method (Massonnet and Feigl, 1995) was

applied to the 24 pair data to reduce the atmospheric phase

delay from atmosphere.

The two interferograms with shared data were paired.

The atmospheric signal during that acquisition will

contaminate the displacement signal in both interferograms.

The addition of both interferograms caused doubled

deformation signal and removed the atmospheric signal

(Hanssen, 2001).

A. Deformation

Pair-wise Logic

P-16

date 1- date 2

S-16

date 1- date 3

+ +

-

-

P-17

date 2- date 3

- -

+

+

Unwrapped Interferograms 2006-2010 after pair-wise logic applied

Inflation in 2008 and 2010

The deformation do not follow the topographical pattern: low contribution of tropospheric delay.

Inflation 2008

EDM : ~10 mm/day

InSAR: ~1.5 mm/day

Inflation 2010

InSAR : ~3 mm/day

The fringes discontinuity indicated the distribution of the new volcanic products at early eruption stage.

The inflation signal was detected in 2008 and the eastern flank shows the deformation

highly toward satellites in 2010.

2008

2010

LOS

GPS results show strong deformation especially at the eastern from the summit about 20 cm. Two arc-lines are local normal fault. The deformation signal follows the local fault system.

Early Field Deformation Measurement

(Subandriyo et. al., 2006)

GPS and EDM measurement in 2006

LOS deformation in 2010

B. Dome Shape

The seed fill method (Revol and Jourlin, 1997) was applied

to detect surface changes of lava dome

The backscattering intensity data between 2006 and 2010

were used

The seed fill method defines a simplest single linkage scheme

such that the pixel’s seed P and neighborhood pixel P' are

considered as related ℜ if their gray levels f(P) and f(P') are

under threshold value α and connected to eight neighborhood

pixel ∁8 as follows:

Seed-fill method

𝑃 ℜ 𝑃′ = 𝑡𝑟𝑢𝑒, 𝑖𝑓 𝑓 𝑃 − 𝑓 𝑃′ ≤ 𝛼 ⋀ ∁8= 1

𝑓𝑎𝑙𝑠𝑒, 𝑂𝑡ℎ𝑒𝑟𝑤𝑖𝑠𝑒

B. Dome Shape Seed fill time table

IKONOS image of lava dome on May 10th, 2006 was used as one of references for the “seed” location

(Bulletin of Global Volcanism Program BGVN 32:02)

No Acquisition

Time

Off nadir

angle Mode

D-1 2007/6/8 34.3 Ascending

D-2 2007/9/8 34.3 Ascending

D-3 2007/12/9 34.3 Ascending

D-4 2008/1/24 34.3 Ascending

D-5 2008/4/25 34.3 Ascending

D-6 2008/6/10 34.3 Ascending

D-7 2008/9/10 34.3 Ascending

D-8 2008/10/26 34.3 Ascending

D-9 2008/12/11 34.3 Ascending

D-10 2009/1/26 34.3 Ascending

D-11 2009/10/29 34.3 Ascending

D-12 2009/12/14 34.3 Ascending

D-13 2010/3/16 34.3 Ascending

D-14 2010/6/16 34.3 Ascending

D-15 2010/9/16 34.3 Ascending

D-16 2010/11/1 34.3 Ascending

D-17 2010/12/17 34.3 Ascending

D-18 2011/2/1 34.3 Ascending

B. Dome Shape Seed fill in red portions

Roughness changes of the lava dome prior to the 2010 eruption

2010.09.16 2010.11.01

σ0 = 4𝑘4ℎ02cos4𝜃𝑖 𝛼 2𝜔

Considering time series, the surface Roughness is remaining

C. Surface Thermal

The 30 scenes of ASTER TIR

data during nighttime and 2

scenes during daytime

observations were used.

The surface temperature T can

be extracted from thermal

radiance as follows:

𝑇 =2𝜋ℎ𝑐2

𝜆ln𝜏𝜀ℎ𝑐𝜆−5

𝜋𝑅𝑘+ 1

No Date Time Path/Row

T-1 2006/9/3 15:10:30 UT 230/489

T-2 2006/9/10 15:16:37 UT 231/489

T-3 2006/9/19 15:10:21 UT 230/489

T-4 2006/10/12 15:16:15 UT 231/489

T-5 2006/11/29 15:16:26 UT 231/489

T-6 2007/5/24 15:16:50 UT 231/489

T-7 2007/7/4 15:11:00 UT 230/489

T-8 2007/8/28 15:17:19 UT 231/489

T-9 2007/9/6 15:11:06 UT 230/489

T-10 2008/1/26 03:05:45 UT 120/182

T-11 2008/5/26 15:17:23 UT 231/489

T-12 2008/6/4 15:11:14 UT 230/489

T-13 2008/6/20 15:11:15 UT 230/489

T-14 2008/7/6 15:11:14 UT 230/489

T-15 2008/9/15 15:17:23 UT 231/489

T-16 2008/9/24 15:11:11 UT 230/489

T-17 2009/1/21 15:17:54 UT 231/489

T-18 2009/5/13 15:17:57 UT 231/489

T-19 2009/6/14 15:17:38 UT 231/489

T-20 2009/6/23 15:11:20 UT 230/489

T-21 2009/6/30 15:17:24 UT 231/489

T-22 2009/8/1 15:17:18 UT 231/489

T-23 2009/8/10 15:11:07 UT 230/489

T-24 2009/8/17 15:17:17 UT 231/489

T-25 2009/9/27 15:10:54 UT 230/489

T-26 2009/11/5 15:17:11 UT 231/489

T-27 2010/7/19 15:16:57 UT 231/489

T-28 2010/8/20 15:16:46 UT 231/489

T-29 2010/8/29 15:10:37 UT 230/489

T-30 2010/9/14 15:10:38 UT 230/489

T-31 2010/11/1 15:10:27 UT 230/489

T-32 2010/11/15 03:05:20 UT 120/182

AS

TE

R T

IR t

ime

table

ASTER Image Database for Volcanoes

http://igg01.gsj.jp/vsidb/image/Merapi/aster_p1.html

C. Surface Thermal

The hot spots around the summit as an indicator of thermal activity.

The radiance of hot volcanic products termed as hot spots saturates the TIR detector.

The ASTER TIR band 13 was assigned to calculate surface temperature from thermal radiance data.

The criterion of band selection is based on the highest absolute accuracy (≤1 K) and the highest radiometric resolution of measured value (0.23 K) among TIR bands (Fujisada et al., 1998).

The least affected by any atmospheric characteristics (Vaughan et al., 2010).

2006.11 2008.01

2010.11

Location of six available Electronic Distance Measurement (EDM)

arround the summit of Mt. Merapi.

Three stations from six reflectors were used:

RK: 2007.1.5~2010.10.26

RB: 2009.4.4~2010.10.26

RJ: 2010.4.1~2010.10.25

The measurement frequency is once per day in average.

The changes of the distance is the objective of the measurement.

D. EDM measurement

RK reflectors:

High attenuation in 2008 and 2010

D. EDM measurement

-1 m/yr -0.1 m/yr

-1.2 m/yr +0.1 m/yr

RB reflectors:

Shortening distance at January 2010

RJ reflectors:

Increasing distance prior to the peak eruption

The LOS Displacement Rate

The Line of Sight (LOS) displacement rate from InSAR at six locations shows that the significant uplifting occurs five times from 2007 to 2010.

A B

C

The Lava Dome Changes

The coverage area of the D-zones supposed to be the growth of the lava dome fractures as detected by ALOS/PALSAR and CO2 gas volume prior to

the eruption at the summit.

The Surface Temperature

The Line of Sight (LOS) displacement rate from InSAR at six locations shows that the significant uplifting occurs five times from 2007 to 2010.

W E

W E

InS

AR

mea

sure

men

t

Shaded map

Four region of volcanic system (Scandone et al., 2007): the Supply, Storage, Transport, and Eruptive system.

The four systems are sufficient for configuration of magmatic system at Mt. Merapi.

The Supply system serves as deep reservoir which is located in about 8.5 km depth (Beauducel and Cornet, 1999).

The Storage system serves as shallow reservoir which is located in about 2-3 km depth (Ratdomopurbo and Poupinet, 2000)

Interpretation

C B C

Conceptual model (Saepuloh et al., 2010 Bull. Volcanol)

Acceleration Rate

C B A

Temporal cross-section from the summit (distance=0) to the NE flank shows different acceleration of deformation supposed to be the flumes prior to the eruption.

Need 2 years optimum

pressure – dome failure..(?)

Deformation Precursor The inflation phenomena might

indicate the pulsatory magma ascent in which individual magma batches detached from the Supply system (Gardner et al., 1998).

The eruptive system probably connected with storage system due to high pressurized magma after the barrier system reached the maximum pressure limit.

2008

2010

The magma batch detached from the deep reservoir and stored in 2.6 km depth.

Penelitian Berjalan: Pendeteksian rekahan beresolusi tinggi dengan data SAR

Range

Depression angle

Ascending Orbit Descending Orbit

Weak

Radar

ground

range

image

o Dual SAR observation

were used: Ascending

satelite heading from South

toward North and

Descending in vice versa.

o Provide surface

information in two look

directions for the same

object.

Strong Strong

Backscattering Intensities before Eruption

The Summit

Yogyakarta 5 km

N The Summit

Yogyakarta 5 km

N

pal_091029asc pal_090807des

Backscattering intensities data before eruption in Ascending

(left) and Descending (right) show the ground surface in two

different angles of view.

LFD before Eruption

The Summit

Yogyakarta 5 km

N The Summit

Yogyakarta 5 km

N

pal_091029asc pal_090807des

Linear Features Density (LFD) map before eruption in

Ascending (left) and Descending (right) related to the density

of faults and/or fractures.

Backscattering Intensities after Eruption

The Summit

Yogyakarta 5 km

N The Summit

Yogyakarta 5 km

N

pal_110201asc pal_101105des

Backscattering intensities data after eruption in Ascending

(left) and Descending (right) show the ground surface

including new pyroclastic flow deposits.

P-Zone P-Zone

LFD after Eruption

The Summit

Yogyakarta 5 km

N The Summit

Yogyakarta 5 km

N

pal_110201asc pal_101105des

Linear Features Density (LFD) map after eruption in

Ascending (left) and Descending (right) related to the density

of faults and/or fractures.

Tota LFD Before and After Eruption

The Summit

Yogyakarta 5 km

N

Before After

The Summit

Yogyakarta 5 km

N

Total Linear Features Density (LFD) map after and before

eruptions shows the increment of fraulted and/or fractured

zones.