Welcome to GEO's Tohoku-oki Event Supersite Website
A message from GEO's task lead
Dear Supersites Scientist,
Let me express my deepest condolences to the people of Japan
affected by this tragic earthquake and tsunami. This disaster is the
first time that multiple space agencies openly provide SAR data for
scientific use. We have to report to the agencies on how the data and
this website was used. Please take some time to provide your feedback
by writing to supersites_feedback@unavco.org. Particularly valuable are
anecdotes on how this website facilitated the understanding of the
tectonic processes in Japan, how it made your work more efficient, and
how it was used for event response, outreach and education. We want to
use your experience to argue for open data provision in the future. We also
welcome suggestions for improvement.
Falk Amelung
Lead, GEO's Geohazard Supersites Task
University of Miami
Earthquake, Friday, March 11, 2011 at 05:46:23 UTC, 38.322° N, 142.369° E
24.4 km depth Magnitude 8.9 (USGS)
Topic links
SAR
Focal Mechanism
GPS
Ground Motion
Source Slip
Seismicity
Visible
Links
If you have data or results that you would like to post on this webpage, please e-mail Falk Amelung,
the Task Leader of GEO's Supersite initiative (famelung@rsmas.miami.edu) or Susanna Gross (supersites@unavco.org).
For up-date-information about new satellite imagery acquired and analysis of the seismic data please visit
the Japanese Space Exploration Agency and the
Earthquake Research Institute at the University of Tokyo (ERI).
SAR Data
Envisat SAR scenes to match the first two tasked tracks have been ordered.
Envisat:
ESA is tasking track 74 frames 2763-2889, tracks 139, 182, 189, track 347 frames 2763-2907, and track 412 frames 681-734.
We would like to expand the tasking plan to include track 297 frames 721-795.
ERS: Postseismic ERS2 data will be available with the new 3 day orbital repeat phase, but ERS-2 data have to be repatriated to ESA and will have delayed availability.
Angela Hung of CSRSR National Central University, Taiwan and Michael Caruso of CSTARS made a special project of getting our postseismic ERS2 delivered as quickly as possible.
ERS-2 precise orbit files:
Daily ERS-2 precise orbit files based on Laser Range data (PDS) can be
retrieved from the anonymous FTP server dgn6.esoc.esa.int under the directory
japan_ers2. This directory contains daily ERS-2 PDS files which are two days
long. The generation of a new PDS file will happen each day at 09:00 UTC were
the latest day that will be available will be two days in the past. On each
day the last 5 days of PDS orbits will be updated to take into account SLR
data that can arrive with a delay of more then two days.
ESA SAR Data can be downloaded from a index or from ESA's
Earth Watching web site. The ESA download server was originally used for Haiti, and
it requires no username or password.
ALOS PALSAR:
JAXA has kindly made ALOS PALSAR Data available without a password direct from the supersites server from the ALOS directory
Desired ALOS PALSAR scenes:
| Ascending Data: |
|---|
| Path | Frames |
|---|
| 400 | 790-800 |
| 401 | 760-800 |
| 402 | 760-800 |
| 403 | 730-800 |
| 404 | 700-770 |
| 405 | 690-750 |
| 406 | 700-740 |
| Descending Data: |
|---|
| Path | Frames |
|---|
| 56 | 2800-2900 |
| 57 | 2790-2910 |
| 58 | 2800-2900 |
| 59 | 2790-2910 |
It would be ideal if ALOS ScanSAR Tracks: 61-62-63 Frames: 2800 and 2850 could also be acquired.
The first ALOS-1 PALSAR acquisition over Japan after the earthquake, on ascending path 409, is in a beam mode not suitable for InSAR because the radar look angle is 46.6 degrees. It does cover the area of Sendai.
This map from the JAXA AUIG catalog search shows the coverage. It says the acquisition time was at 13:12 on 2011/03/13, but does not explain the time zone used.
The AUIG catalog also shows two paths in the Observation Plan, ascending paths 401 on 3/15 and 404 on 3/20. These are planned for the standard look angle of 34.3 degrees that should allow InSAR analysis. It does not show yet the ScanSAR that may be acquired earlier on a descending track. This map shows the planned paths.
SAR Data can be downloaded from a ESA SAR data index
TerraSAR-X:
DLR and Astrium GEO-Information Services support research on the Sendai Earthquake event
on 2011-03-11 in a joint action by providing data of the German radar satellite
TerraSAR-X.
TerraSAR-X will acquire co-seismic data and a post-seismic time series of 3 months
duration, covering the area around Sendai in ascending (orbit 156, strip 9, incidence
35°) and descending (orbit 42, strip 10, incidence 37°) orbits in stripmap mode.
The data will be made available free of charge via the GEO Supersites mechanism. First
datasets are now accessible and additional datasets will be uploaded regularly every 11
days throughout the months of April, May and June.
Additional data of this area can be accessed via DLR's science service system
for scientific purposes and via the Astrium GEO-Information Services
TerraSAR-X Services for commercial purposes.
Radarsat-2 and Cosmo-Skymed: At this time these sensors are not available for scientific use by the Supersites community.
SAR Results
SAR co-seismic
Masanobu Shimada of JAXA contributed the first SAR interferogram, using PALSAR, showing the coast just north of the epicenter.
An English translation of their website is linked to the thumbnail above. Follow this link for the original Japanese.
Manabu Hashimoto, Yo Fukushima and Youichiro Takada of the Disaster Prevention Research Institute, Kyoto University prepared the PALSAR interferogram above. Please follow this link for more information.
Interferometric phase by the ARIA team at JPL and Caltech. Produced from
ALOS PALSAR data (ascending path 401) spanning 2010/10/28 - 2011/03/15.
One color cycle represents 11.8 cm of range change. You can also download the
KMZ file.
Azimuth (~N170 heading, horizontal) co-seismic ground displacement
associated with the 2011 Sendai Earthquake, calculated from Multiple
Aperture InSAR (Bechor and Zebker 2006; Jung et al., 2009) using the
ALOS PALSAR dataset. The data shows up to 1.5 m of displacement,
consistent with GPS measurements; some area of decorrelation can be
observed as well as processing artefact (azimuth strikes) likely related
to the different polarisation between the two granules used. The kmz
file can be
downloaded here
(Jeong-Won Park, U. of Leeds)"
Marco Chini and Christian Bignami have processed three strips composed
of 13 (track 347, descending orbit), 8 (track 074) and 6 (track 189)
Envisat frames over Honshu island (the result is in the framework of
"Tohoku-oki INGV Team", lead by Salvatore Stramondo). The interferogram
covers the whole Honshu island and the epicentral region and a large
coastal area. The whole strips have been unwrapped to measure the Line
Of Sight surface movement. The maximum displacement reaches about 2.5 m
relative to a reference point within the entire frames located nearby
the southern boundary.
They provided a map overview
of the interferogram, as well as detailed views of
Fukushima , Chiba and Sendai.
Envisat ASAR coseismic interferogram from descending track 189 (west),
347 (middle), and 74 (east) processed by JPL/Caltech ARIA project using
ROI_PAC. Time spans of the interferometric pairs are 2011/03/10 -
2011/04/09 (west), 2011/02/19 - 2011/03/21 (middle), and 2011/03/02 -
2011/04/01 (east), respectively. One color cycle represents 50 cm of
motion in the radar line of sight (approximately west at 41 degrees from
the vertical). ESA VOR orbits were used for pre-quake scenes and POR
orbits for post-quake scenes. The seismicity plot is from USGS Lastest
Earthquakes. KMZ file
Thanks to data available from the GEO Geohazards Supersite initiative,
TRE has been able to perform an analysis to extract ground displacement over the area
affected by the Japanese earthquake occurring on 11th March 2011. Using
ENVISAT S6 data, descending track 347, a co-seismic interferogram was
produced using two radar satellite images from 16th February and 21th
March. The processed area is approximately 650x75 sqkm. Each visible
fringe (a complete colour cycle ranging from red, through the colour
spectrum, back to red again) refers to ground movement on the order of
2.8cm in the satellite line of sight (LOS). For more information, please see
their website.
Michele Crosetto sent this coseismic deformation field, estimated from Envisat ASAR data from the
descending track 347, processed by the Institute of Geomatics. The
coseismic estimate was derived using the data coming from three frames,
acquired on 19 February 2011 and 21 March 2011. The maximum relative
deformation over the covered area is 1.69 m. It is worth emphasising
that this is not an absolute deformation estimate. The minimum distance
from the study area to the epicentre is about 100 km. The background
image comes from Google Earth. It is worth noting that in the low
coherence areas, characterized by a low density of observations, this
solution might suffer local aliasing errors.
Guangcai Feng, Mi Jiang, Lei Zhang and Chisheng Wang (InSAR group lead by
Prof. Xiaoli Ding, LSGI/PolyU , The Hong Kong Polytechnic University) have
processed two Envisat strips and two PALSAR strips covered east of Japan.
The coseismic interferogram of 2011 Tohoku earthquake presented have been
corrected by GPS measurements provided by the ARIA team at JPL and Caltech
(V0.3).
PALSAR coseismic interferogram from descending track 56 was processed by
Jin Woo Kim, the Ohio State University under
Dr. C. K. Shum with the aid of Dr. Zhong Lu in the USGS. Data are
acquired on 20 November 2010 and 7 April 2011, and each color cycle
represents 11.8 cm of motion in the line-of-sight direction.
ALOS PALSAR differential interferogram 28102010-15032011 (Bp =1459,9 m
and dt =138 days) of the Tohoku-oki M9.0 (USGS NEIC) earthquake. No
baseline refinement was applied due to displacement gradients over the
entire scene. Residual orbital phases might be included. Displacement up
to -3.46 m along the line-of-sight was obtained relative to the selected
reference (square). This was contributed by Issaak Parcharidis of the
Earth Observation Team at Harokopio University, Greece.
Davide Oscar Nitti of Geophysical Applications Processing in Bari, Italy prepared
several interferograms from Envisat track 347. For the complete set of interferograms, please see their website .
Displacements around Sendai estimated from TerraSAR-X Data using
Correlation Techniques
The arrows in the image show the ground displacements caused by the
devastating M9.0 earthquake near Sendai on the 11th of March 2011. The
background image is coloured using amplitude differences of the two
acquisitions and clearly shows the extent of damages caused by the
earthquake and the following tsunami-wave.
The image was generated from two TerraSAR-X images acquired on
20.10.2010 and 12.03.2011.
Applying correlation techniques, the displacement-map could be derived.
Unit-Vectors at the bottom right of the image represent 1.46 m in ground
range (approx. east-west) and 2.0 m in azimuth (approx. north-south)
direction. Under the assumption of purely horizontal displacement of the
land mass, a value of up to 3 m can be observed.
The amplitude and directions of the displacements agree with values
achieved by various groups evaluating data from permanent GPS-stations
This image was generated by:
Nestor Yague-Martinez,
Christian Minet,
Wael Abdel-Jaber, and
Michael Eineder
ALOS PALSAR interferogram subset from descending track 56 processed by
JPL/Caltech ARIA project using ROI_PAC. Time spans of the
interferometric pair is 2010/11/20 - 2011/04/07. Best-fit quadratic
surface was estimated and removed from this interferogram subset. This
removes most of the signal from the M9.0 main shock and the M7.9
aftershock on 2011/03/11 and also orbit errors. The remaining
deformation is largely caused by shallow aftershocks, including two
M5.3-6.1 aftershocks on 2011/03/19 (south blob) and four M5.8-6.0
aftershocks on 2011/03/23 (north blob). One color cycle represents 11.8
cm of motion in the radar line of sight (approximately west at 38
degrees from the vertical). Preliminary orbits were used.
KMZ file
ALOS PALSAR interferogram subset from ascending track 403 processed by
JPL/Caltech ARIA project using ROI_PAC. Time spans of the
interferometric pair is 2011/03/03 - 2011/04/18. Best-fit quadratic
phase was estimated and removed from this interferogram subset. This
removes most of the signal from the M9.0 main shock and the M7.9
aftershock on 2011/03/11 and also orbit errors. The remaining
deformation is largely caused by series of shallow aftershocks of M5.3 -
6.1 and M6.6 on 2011/04/11. One color cycle represents 11.8 cm of motion
in the radar line of sight (approximately east at 38 degrees from the
vertical). Preliminary orbits were used.
Please also see the the KMZ file.
SAR post-seismic
ERS-2 post-seismic interferogram from ascending track 24, processed by
JPL/Caltech ARIA project using ROI_PAC. Data acquired on 21 March 2011 and
27 March 2011. One color cycle represents 2.8 cm of motion in the radar
line of sight. Precise PDS orbits from ESA based on Laser Range data were used
for processing. The observed phase includes the combined contributions
from tropospheric differential phase delay variation and post-seismic
deformation. KMZ file
Sven Borgstrom and Valeria Siniscalchi have processed three ERS2
interferograms (ascending track 24, InSAR pairs 02.04.2011-05.04.2011 , 02.04.2011-08.04.2011, and
08.04.2011-11.04.2011 above) using ROI_PAC in the framework of the
"Tohoku-oki INGV Team", lead by Salvatore Stramondo. Precise PDS orbits
from ESA were used for data processing, with help from Eric Fielding and
Piyush Shanker Agram of JPL/Caltech.
The authors decided to focus their attention on the period of the 7.4M EQ
of April 7: apparently no clear deformation signals were pointed out in
the coseismic interferogram (02.04.2011-08.04.2011), probably masked by
the atmospheric contribution.
Sven Borgstrom and Valeria Siniscalchi have processed three more ERS2
interferograms from the ascending track 24 ( 26.4.2011-29.4.2011,
29.4.2011-2.5.2011 and 5.5.2011-8.5.2011, shown above) using ROI_PAC and precise PDS
orbits from ESA in the framework of the "Tohoku-oki INGV Team", lead by
Salvatore Stramondo. The three days ERS2 revisiting time means a chance
to get almost continuous monitoring, besides the possibility of
extracting the co-seismic deformation field with the post-seismic
deformation filtered out.
Simona Zoffoli contributed these COSMO-SkyMed co-seismic and post-seismic interferograms of the Tohoku
Earthquake obtained by Italian Space Agency in the framework of the
SIGRIS system.
SIGRIS was developed under funding from the Italian Space Agency (ASI)
and is operated by INGV. The coordinator is Dr. Stefano Salvi.
Shortly after the quake the Opera and SIGRIS teams
coordinated with ASI to program new COSMO acquisitions over the area,
for damage assessment and ground deformation mapping.
The first acquisitions (March 12-13) were aimed at damage/flood
mapping, and no co-seismic interferometric data were acquired.
First co-seismic interferometric data sets were acquired on March 14
CSK_co-sis_interf.jpg, and have been used by the INGV team to produce
the results summarized below.
Unfortunately the most recent pre-event acquisitions dated July 2010.
This causes strong temporal decorrelation in the interferograms
The COSMO-SkyMed interferograms are for the most part decorrelated,
due to the 8 months temporal separation of the images. Where coherence
is maintained deformation fringes can be seen, as shown in Fukushima.jpg
and csk_917_sud.jpg.
Each colour fringe represents a permanent co-seismic ground
displacement of 1.5 cm along the satellite line of sight. The fault
dislocation causing the magnitude 9 earthquake generated a large
permanent deformation of the ground in an area of several thousands of
square kilometers in Northern Japan.
The one-day post-seismic interferogram csk_post_13-14.jpg of the
Sendai area is affected by high spatial frequency atmospheric artifacts,
but limited post-seismic deformation (~1 cm) can be appreciated in the
low frequencies.
We are programming new, high frequency COSMO-SkyMed acquisitions over
most of the area, and expect to be able to map post-seismic deformation
with great detail in the following months.
We expect that COSMO-SkyMed data will be also very useful for local
deformation monitoring, for instance on seismically triggered
gravitational deformations.
Issaak Parcharidis and Micheal Foumelis of Harokopio University of Athens contributed the TerraSAR-X interferogram above, showing post-seismic deformation.
Co-seismic interferogram calculated using COSMO-SkyMed images of May 22,
2010 and March 14, 2011. Perpendicular baseline ~65 m. Image location in
inset. Processing by Cristiano Tolomei. Please see their report (in PDF format) for
more interferograms, displacement maps and error maps.
Focal Mechanism
GPS Data
Thomas Dautermann of DLR sent four hours of 20 Hz GPS data, from the Chofu site (Hatanaka compressed and zipped) for the community.
Co-seismic displacement from GPS by GEONET, reference provided by Prof. Hashimoto
Mohamed Chlieh of the IRD - Geoazur has prepared a preliminary pre-seismic interplate coupling
inversion from 224 pre-seismic GPS measurements of the Japanese Network (period 2001-2011). The
GPS measurements are plotted relative to the Amurian Plate. The fault geometry was taken to be
consistent with USGS Fault solution and Harvard CMT. Dashed line may indicate the seismic source.
Interseismic coupling distribution based on GEONET GPS data from 1996-2000
and estimated using a three-dimensional block model of the combined
offshore/onshore fault network
(Loveless and Meade, JGR, 2010)
. Coupling is estimated on a mesh of triangular dislocation elements based on the
three-dimensional slab geometry of Furuse and Kono (2003). Overlain on the
coupling contours is the finite fault slip distribution (250 cm slip
contours) from Gavin Hayes.
Ascii file with GPS vectors
If you use these vectors, please give the following credit: "Preliminary GPS displacement data (version 0.1) provided by the ARIA team at JPL and Caltech. All Original GEONET RINEX data provided to Caltech by the Geospatial Information Authority (GSI) of Japan."
Yusaku Ohta in Tohoku University in Sendai, Japan calculated the AOBL (2Hz sampling) coordinate by kinematic analysis relative to USUD (IGS) using RTKLIB Ver. 2.4.0. (Takasu et al. 2010).
The sample interval is 1Hz because of USUD is recording 1Hz data.
This baseline analysis includes base station USUD displacement by seismic wave, but still shows a clear coseismic displacement mainly EW component.
Computation of the GPS displacement waveforms applying
VADASE methodology
to the IGS high-rate data (1 Hz)
recorded at MIZU (140 Km from earthquake) by Mara Branzanti, Gabriele Colosimo, Mattia Crespi,
Augusto Mazzoni
-- VADASE Team Universita' di Roma "La Sapienza" Area di Geodesia e Geomatica
MIZU.pdf
USUD.pdf
See also a similar plot for DLR GPS station JA01.
station map
comparison with observations
Simon Banville, a Ph.D. student working with Prof. Richard Langley at the University of
New Brunswick, has analyzed the high-rate 1-Hz data from two stations of the
International GNSS Service affected by the Tohoku-oki earthquake: MIZU at Mizusawa and USUD
at the Usuda Deep Space Tracking Station. His analysis approach uses PPP or precise point
positioning, a technique that requires meticulous modeling of all of the phenomena
affecting GPS measurements to reveal station displacements with precisions approaching a
few centimetres. Inputs include high-precision satellite orbits and clocks, which, in
this case, were provided in the form of the "ultra-rapid products" generated at Natural
Resources Canada and graciously provided by Yves Mireault. The displacements shown in the
plots in terms of latitude, east longitude, and height are with respect to the station
coordinates before the occurrence of the earthquake.
Also see a similar plot for station USUD
The MIZU ground motion plots show the station moving 3.5 meters (11.5
feet) ESE in the first minute of shaking, and
then recovering back to a final position 2.3 meters (7.6 feet) ESE from
the pre-earthquake location, during the following two minutes. As late
as 2 minutes after the first shock, waves with an amplitude of almost
0.25 meter (9 inches) and a period of 5-10 seconds can be seen.
Rongxin Fang of Wuhan University sent GPS waveforms computed with PANDA software which was also used in
processing GPS data after the WenChuan earthquake. The station MIZU displacement is shown above.
Also see a similar plot for station USUD Two movies, showing MIZU and USUD were made with PANDA software. GPS seismic
waves results by Prof. Chuang Shi and his group members, Rongxin Fang ,
Kejie Chen, Min Li and Weiwei Song. PANDA(Positioning And Navigation
Data Analyst) software package is developed at GNSS Research Center of
Wuhan University including two main modules, that is Precise Orbit
Determination(POD) and Precise Point Positioning(PPP). Firstly, we use
its POD module to obtain the high precision satellite orbits and clocks,
then the PPP module is used to process 1Hz GPS data to obtain GPS
seismic waves.
James Johnson, Chris Rocken and Ted Iwabuchi of GPS Solutions have processed 7 1Hz GEONET stations using PPP and presented their waveforms on
their web page. 5 meters of offset are shown on station 0171 on the their second page.
Ronni Grapenthin of the University of Alaska has computed kinematic solutions for 7 stations with Gipsy and presented waveforms on
this web page.
James Johnson of GPS Solutions processed data from the Hiratsuka buoy off the coast near Tokyo. A one-meter tsunami first arrives at the Hiratsuka buoy at about 6:35 UTC. It is interesting to note the change in buoy motion at about 5:50 UTC (most visible in the 11-sec running average) which appears to coincide with the arrival of seismic waves. The shown PPP solution was post processed with the real-time global orbit and clock product from the VERIPOS/APEX service. For maps and further details, including true realtime solutions for stations further from the epicenter, see
the GPS Solutions Tohoku-oki web pages.
The animation above shows coordinate variations every 1 sec. relative
to the pre-earthquake a priori coordinates. The color scale is fixed for
the entire event, but north, east, height components are scaled
individually. The deformation is sped up by a factor of 24, showing 24
seconds of motion every second of movie time. For more movies and
additional information, please see GPS Solutions web
site.
1Hz GEONET data provided by the Geospatial Information Authority of Japan (GSI) via Nippon GPS Data Services Company (NGDS), were post-processed in real-time mode with realtime orbits and clock corrections from the VERIPOS/APEX global service. GPS Solutions' RTNet software was used for the PPP analysis and is also used for APEX service.
Comparison between high-frequency GPS and accelerometric signals in the Tokyo area. The
GPS station JAE1 (Thomas Dautermann, DLR, supersites) is located 4km away from the Knet
accelerometric station TKY006. GPS signal has been processed by J.M. Nocquet (Geoazur,
Nice, France) to obtain ground displacement. The figure shows the comparison between this
GPS signal - twice differentiated - and the accelerometric signal, in the [0.005Hz -
0.125Hz] range. Part of the slight differences can be imputed to the small location
differences. This observation confirms that both types of data will be useful to better
describe the source process of the 2011 Japan earthquake. (authors J.M Nocquet & M. Vallée, Geoazur, Nice,
France)
Final ground displacement after the 11 March 2011 earthquake in Japan,
in meters, with
contours, and colored by
value, from the GSI GPS data set provided by
the
ARIA team at JPL
and Caltech.
Computations and display were done at UNAVCO. For details please see
this report (in PDF).
Ground Motion
Stacked real-time 1 Hz displacement waveforms for more than 100 CRTN
stations in the southern California region from Yehuda Bock, Brendan Crowell
and Diego Melgar of Scripps Orbit and Permanent Array Center. They show the
S-wave at about 1400 seconds and the Love wave at about 2000 seconds after
the onset of the M=9.0 Honshu earthquake.
PDF with contact information This is a random sampling of groundwater gauges in South Florida, all showing tremors after the disaster in Japan. (Click on the image for a larger version.) The anomalies began 34 minutes after the earthquake and lasted almost two hours.
Mainshock Slip Distribution
Takeo Ito, Kazuhiro Ozawa, Tsuyoshi Watanabe and Takeshi Sagiya of
Nagoya University have inverted for the slip distribution on a curved
plate interface using coseismic GPS vectors from the Geospatial
Information Authority of Japan. The above plot shows the slip
distribution, which extends about 400km N-S and reaches a maximum of
about 10 meters. For more details, please see their PDF document.
Yagi Yuuzi and Naoki Nishimura of Tsukuba university have created a source model by fitting seismic waveforms.
For more information please see their web page (in Japanese) or the English translation.
Fault slip model projected onto the surface with topography and aftershocks.
Waveform modelling by Gavin Hayes of the
USGS. Yellow circles are foreshocks.
Fault slip model projected onto the surface with plate boundaries and aftershocks.
GPS and Waveform modelling by
Shengji Wei (Caltech), Anthony Sladen (Geoazur) and the ARIA group (Caltech-JPL).
Lingsen Meng and Jean-Paul Ampuero (Caltech) have prepared movies of the rupture propagation.
Fault slip model projected onto the surface inverted from GPS data by Rongjiang Wang of the GFZ. To see the
modeled GPS vectors, click on the thumbnail image above.
Fault slip model projected onto the surface inverted from GPS data by Andreas Hoechner, Andrey Babeyko and Stephan Sobolev of the GFZ. To see the
modeled GPS vectors, click on the image above.
Slip distribution of the 2011 M9.0 Tohoku-chiho Taiheiyo-oki
Earthquake. Arrows indicate the motion of the upper plate.
GPS modeling by Fred Pollitz.
A preliminary highly smoothed model of coseismic slip for
the March 11 main shock. Red and yellow vectors show horizontal (left)
and vertical (right) predicted and observed displacements respectively.
The triangulated model geometry follows the megathrust as close as
possible. The model is derived with a least squares inversion using
Laplacian damping, with strike slip components damped more heavily than
down dip components. Damping parameters were chosen by sense of smell.
The model is constrained by both GPS displacements and 12 open ocean
DART buoy tsunami records distributed around the Pacific. The model has
a moment magnitude of 9.0, although this is somewhat arbitrary given our
ignorance of the elastic structure. GPS offsets are derived from 5
minute epoch time series processed by the ARIA team at JPL and Caltech.
Displacements due to the Mw 7.9 aftershock have been isolated and
removed. This model was generated by M. Simons, F. Ortega, J. Jiang, A.
Sladen, and S. Minson at Caltech as part of the ARIA project. All
orginal GEONET RINEX data provided to Caltech by the Geospatial
Information Authority (GSI) of Japan.
Preliminary model of coseismic slip of the March 11, 2011 Tohoku
Earthquake. The fault model has variable dip and strike. We sample the
cross-section of the subduction zone geometry analysis carried out by
Gavin Hayes
The fault plane is divided into 240 subfaults of 25x25 km. We use the
sea-level recordings at 7 deep-sea Bottom Pressure Gauges in open sea
(DART)
provided by NOAA .
We use GPS data processed at INGV Rome using JPL's GIPSY-OASIS software,
the kinematic precise point positioning strategy and JPL's Rapid orbit
and clock products. The coseismic displacements are calculated as a
simple difference of the position estimates averaged 15 minutes before
and after the main shock excluding the first 5 minutes during the most
intense ground shaking. All original GEONET RINEX data provided by the
Geospatial Information Authority (GSI) of Japan. The inversion method is
based on
a global search technique.
We add smoothing and seismic-moment-minimization constraints to the slip
distribution. The rupture front is assumed to be circular and
propagating at constant speed, fixed at 1.0 km/s, which however needs to
be further investigated. This model was generated by F. Romano, A.
Piatanesi, S. Lorito, and N. D'Agostino at Istituto Nazionale di
Geofisica e Vulcanologia in Rome (Italy). Tsunami modeling performed
with
COMCOT at
CASPUR HPC Center in Rome (Italy).
Preliminary inversion of seismic moment release along the 2011 Japan
earthquake source. Distribution of seismic moment release along the modeled source . Total
seismic moment release is 3.95e22 Nm, corresponding to Mw = 9.03. This
model naturally implements the seismic source in terms of seismic
moment, so that the moment magnitude estimate is not biased by
slip-moment conversions.
Daniele Melini member of the "Tohoku-oki INGV Team" (lead by Salvatore Stramondo) obtained
a preliminary model of seismic moment release of the March
11, 2011 Japan earthquake by modeling GPS displacements with a spherical
self-gravitating deformation model (Melini et al., GJI 2008). The model
assumes a PREM rheological layering.
The fault plane geometry is fixed to USGS seismological estimates.
Preliminary GPS time series are provided by the ARIA team at JPL and
Caltech. All original GEONET RINEX data provided to Caltech by the
Geospatial Information Authority (GSI) of Japan.
Stress Change and Earthquake Hazard
Shinji Toda (DPRI, Kyoto Univ.), Ross Stein (USGS) and Volkan Sevilgen (USGS) have computed stress changes on
faults throughout Japan due to the Tohoku mainshock. Faults brought closer to failure are shaded red on the
above plot. For higher resolution images and many other plots, please see their pdf. The same work is presented on Shinji Toda's web page (in Japanese). English translation
Tsunami
This animated gif, which plays when you click on the image above,
comes from the
Earthquake Research Institute, University of Tokyo, Prof. Takashi
Furumura and Project Researcher Takuto Maeda.
Researchers at the GFZ have made a model of the maximum amplitude and arrival time of the tsunami using their own source model.
Seismicity
Aftershocks from IRIS earthquake browser plotted with IDV, background seismicity in blue.
Charles Ammon of Penn State University plotted major historic seismicity with aftershocks and the mainshock (largest red circle).
He also has a source slip model and several movies on his website.
Aftershocks drawn with Jules Verne Voyager.
Stuart Wier made a 6MB animated gif, 3D view of the aftershocks using the IDV visualization software.
The colored points are the main shock and 350 aftershocks observed in the first 60 hours after the mainshock.
These are colored and sized by magnitude. The orange dot is the mainshock. The blue dots are all located
earthquakes in the region over magnitude 5, before 1 March 2011; more than 3000 earthquakes. All data is from
the IRIS Earthquake Browser.
Visible and SAR damage mapping
Arii Motofumi of Mitsubishi Space Software detected floating objects in
PALSAR images taken March 13. He provided images from frame 740 and frame 720 and spreadsheets that list coordinates and size estimates for frame 720 and frame 740 .
DLR has created many detailed maps of the tsunami damage. Please see their
Tohoku-oki page
In the frame of Japan's disaster Astrium GEO Information Services (Spot
Image) made SPOT and FORMOSAT-2 Satellite Imagery over Japan available
to the scientific community. The data is available for download
from the Spot Image website
The low resolution data from MODIS on the Terra satellite has captured
an excellent image of the flooding
The area of flooding is so large it is many pixels for MODIS, which has
at best 250 m resolution.
GeoEye before and after photos
EOAU and Hatfield Consultants have made damage maps using Radarsat-2 data. The small thumbnail above is linked to a much larger image showing tsunami damage in Miyagi Prefecture. They created a similar map for Iwati Prefecture, and also offer lower resolution versions of the Miyagi and the Iwati maps. The Canadians have also given us the original images upon which the Sendai map was based, before and after the earthquake and tsunami.
Rapid mapping of Earthquake and Tsunami damage for the International Charter of Space and major disasters
Paul De Fraipont of SERTIT has contributed damage maps made from two Spot-5 images tracks acquired over the East coast of
Japan in support of the International Charter. These maps were processed and analysed by SERTIT rapid mapping service
team. Maps were uploaded into SERTIT server, respectively on Sunday 13th
early in the morning for 200 km of coast around Sendai, and Tuesday 15th of
March 2011 for global damage map over
the North-East Honshu 200 km of coast.
The East Honshu coast map is linked to the thumbnail image above. There
is an updated version made Monday the 14th available on the SERTIT web server.
Radar amplitude images of the coastline affected by the earthquake
and tsunami near the town of Sendai. Data were acquired by
the ALOS satellite of the Japanese Space Agency (ScanSAR mode,
descending orbit 60, Jan 21, 2009 - Mar 14, 2011). Changes in the radar
backscatter outline areas affected by inundation and earthquake damage.
Unfortunately, the orbital baseline is too large for interferometry.
Contributed by Yuri Fialko, SIO/UCSD.
Urs Wegmüller, Charles Werner, and Maurizio Santoro , all of Gamma
Remote Sensing AG, CH-3073 Gümligen, Switzerland made maps of damage
based upon the coherence loss observed in Envisat InSAR analysis of
track 347. Coherence loss between a comparable pre-seismic pair
(20101121_20110219, -23m, 90days) and a co-seismic pair (20110219
20110321, 306m, 30days). For 30 day and longer interval pairs over this
site the coherence is mainly high (> 0.5) for urban area. Damage from
the earthquake and the tsunami cause decorrelation in the co-seismic
pair. In severely damaged urban areas the coherence reduces from a high
coherence (> 0.5) in the pre-seismic pair to a low value which results
in large coherence loss values (> 0.3). Damage over vegetated areas is
not well recognized in the coherence loss between the coherence is
already low in the pre-seismic pair. To display the coherence loss a
color scale between 0.1 (green) and 0.5 red) is used. The image
brightness is the backscattering of 20110219. Red color is a clear
indication of damage in urban areas. For more details please see their report (in PDF format)
and their RGB kmz
file and their coherence loss
kmz file.
TerraSAR-X damage map between 2010.10.20 and 2011.03.12. ASTER 1-arcsec
GDEM (product of METI and NASA) was used for geocoding. Red (2010.10.20,
before earthquake), Green (2011.03.12, after earthquake), and Blue
(Difference of backscattering coefficient of two TerraSAR-X images).
Purple-colored area is inundated by Tsunami. Please also see another
version of this image with inudated areas shown in red, and
a similar map made from Palsar ScanSAR data, and and a detailed
view of the area around Sendai both with and without inundated areas
shown in red, by thresholding
the difference of backscattering coefficient between two dates.
This work was done by Jin Woo Kim under prof. C. K. Shum at the Ohio
State University with the aid of Dr. Zhong Lu in the USGS.
Alessandro Piscini, Stefania Amici and Malvina Silvestri completed the
analysis of inundated areas (in blue) along the coast of Honshu (the
result is in the framework of "Tohoku-oki INGV Team", lead by Salvatore
Stramondo). Unsupervised and supervised algorithms have been applied for
soil classification using ASTER and EO1-Hyperion satellite images in the
visible spectral range. (a) pre-tsunami ASTER( 15m/pixel); (b) March 13
2011 map from EO1-Hyperion data( 30m/pixel); (c) and (d) March 14 and 19
2011 map from ASTER. (for ASTER images we thank Dave Pieri from JPL and
Maria Fabrizia Buongiorno from INGV)
Overview map of tsunami inundation extent of the northeast coast of
Honshu derived from MODIS Aqua images (250-m spatial resolution)
acquired on February 23, 2011 and March 13, 2011. Produced by the
E-DECIDER team under the International Charter and contributed by Margaret Glasscoe, JPL.
Please see this zipfile for several more maps.
Stefania Amici has processed EO1-Hyperion data and ASTER to identify
combusting area (the result is in the framework of "Tohoku-oki INGV
Team", lead by Salvatore Stramondo). SWIR color composite RGB indicate
three different combusting area ( 1, 2, 3). The ASTER TIR data acquired
on March 2011 has been used to validate data: two strong active
combustion features on ASTER image (3b) are present on the
EO1-Hyperion day after image (3a). (ASTER TIR image, courtesy D.Pieri,
JPL).
Links
|
Corporate Contributors
|
