Development of cabled seafloor seismo-geodetic network and seafloor borehole observatories in the Nankai Trough

1
Development of cabled seafloor seismo-geodetic
network and seafloor borehole observatories in
the Nankai Trough

• Eiichiro Araki and Yoshiyuki Kaneda
• Japan Agency for Marine-Earth Science and
  Technology (JAMSTEC)
• Margins SEIZE-WS 08/09/25

2
Earthquakes in the Nankai Trough
Tokai
Tonankai
DONET target area
Nankai
Philippine Sea Plate
Seismogenic Zone
M8 class earthquakes every 100-150
years. Sometimes, these areas rupture in a short
period of time.
3
To understand the recurrence patterns in next
mega thrust earthquakes
New research project
Advanced simulation for the estimation of next
mega thrust earthquakes
4
To improve the Simulation Model Method
FEM model for Southwest Japan
New research project
1) Evaluation of heterogeneous structure, thermal
structure, etc.
Mantle
PHS Slab
InSAR
2) Combination of GPS and InSAR data for slip
inversion
3) Inland earthquake model during Nankai
earthquake cycle
5
When DONET is completed in 2010
Observation site map off Tokai area
DONET
Tonankai
Tokai
DONET observation targets Large earthquakes of
M8 Tsunamis Co-seismic and interseismic ground
deformation Microearthquakes Slow earthquakes
6
Historical background Sumatra earthquake in 2004
drove us to build DONET seafloor network in
Nankai.
We conducted OBS network observation above the
rupture area ONLY 2 months after the
earthquake. But we do not know what happed
during the 2004 Sumatra earthquake. We need
comprehensive seafloor data during the main shock
(and interseismic period) to capture behavior
seismogenic plate boundary.
After Araki et al., 2006.
7
Location map of DONET seafloor and borehole
observatories
During the phase I implementation, five science
node (Node A-E) will be installed. Five nodes
give us connectivity of 40 scientific sensors
that are apart from the node up to 10 km (or 20
km optional). 30W (or 50W optional) power can be
supplied to each scientific sensor. Precision
timing (to a resolution of micro second), and
uplink data capability (20Mbps/sensor) over
fiber optic cable are provided to each
sensor. 20 seafloor sensors, and three borehole
observatories are currently on the list to
connect to the DONET network.
Red line DONET cable Pink points DONET
seafloor observatories
8
Network design to optimize ability to locate
small earthquakes
Montecarlo simulation test to assess DONET
ability to locate small earthquakes

• Procedures
• Distribute target earthquakes in grid
• 2. Compute travel time to observatories for two
  cases
• Only land observatories
• Land and DONET observatories
• 3. Add random observation noise to the computed
  travel time
• 4. Locate target earthquakes using these travel
  time data
• 5. Evaluate deviation from true location.
• See percentage of earthquakes located within 5km
• from the true location.
• Repeat 100 times for each target earthquake.

9
Network design to optimize ability to locate
small earthquakes
Hypocenter determined lt 5km accuracy
With DONET, location of earthquakes of M2 can be
determined with good accuracy in depth.
Depth determined lt 5km accuracy
Depth of Initial hypocenter set to actual depth
20km. Initial hypocenter is distributed randomly
Only with land data
Land DONET data
10
Detection of seafloor deformation by DONET
pressure gauges
????10 cm
Vertical deformation maximum at 3 cm
Error assumed in the simulation GPS 1mm (H)
3mm (V) Pressure gauges 1cm(V)
With DONET, we will be able to detect small
seafloor deformation. This is not possible with
land GPS data only.
11
Design of DONET seafloor sensor

• Need for wide dynamic range, broad frequency
  range ? targeting micro eqs., large eqs., slow
  slip events, ground deformation.
• Observation parameters Ground motion (strong
  motion accerelometer, broadband seismometer) ,
  Pressure (Quartz pressure gauge, differential
  pressure gauge, hydrophone), Temperature
• Seismic sensor will be buried ( low noise
  observation, better coupling)

12
Choosing appropriate sensors for target phenomena
DONET sensor chooses seismometers and pressure
gauges suitable to observe very small to large
earthquakes, Tsunamis, and slow ground
deformation. We need set of sensors of very
broad frequency range and wide dynamic range.
Earthquakes
Slow ground deformation
Longperiod Seismic waves
Tsunamis
Earthquakes
Slow ground deformation
Longperiod Seismic waves
13
Concept design of DONET observation sensors.
Mounted on gimbals
Seismometers are buried in the seafloor and
pressure gauges are installed in the seafloor.
14
Experiments to develop DONET sensor

• Quartz pressure gauge comparison in laboratory
• Seismometer comparison in JMA vault.
• Buried seismometer observation in the seafloor.

15
Stability test result of quartz pressure gauges
Evaluated long-term stability of quartz pressure
gauges
less than 0.1 psi drift / 150days
Hewlett Packard 2813E, Paroscientific 46K and
410K, and COM quartz pressure gauges). Test
pressure is 4000 psi (applied by dead weight
tester) and temperature is 4? in water bath.
Before May 20, experiment has been interrupted
several times.
16
Seismometer evaluation test at Matsushiro vault
of JMA.

• Sensor noise is evaluated for many type
  seismometers.
• Broadband seismometer Streikeisen STS2, Guralp
  CMG3T, Kinemetrics Cronos
• Strong motion accerelometer JAE JA-5typeIII,
  JA40G, Metrozet TSA-100S
• Geophones GS-11D, HS-1, SM-6, L-28

17
Combined seismometers for wide frequency and
dynamic range
Typical Amplitude of very low frequency
earthquakes In the Nankai Trough
From results of seismometer evaluation test in
Matsushiro Vault (JMA)
18
Comparison test observation to evaluate
installation method for DONET (on the seafloor,
on the seafloor with improved anchor design, and
buried sensor)
Dec 07-Feb 08, JAMSTEC NT07-23, NT08-04 cruise
Buried seismometer
Seafloor seismometer
Pressure Temperature
18m
150m
Seafloor broadband seismometer with improved
anchor design
Pressure
Bottom current observation
19
Comparison of buried and seafloor long-period
seismic records
Vertical noise similar to seafloor
seismometer. Horizontal noise is quieter by more
than 10 times.
Teleseismic earthquake
2008/1/22 17h 13-50 second period
Pressure 1
Pressure 2
Vertical seafloor 1
Vertical seafloor 2
Vertical Buried
Horizontal Buried
Horizontal seafloor 1
Horizontal seafloor 2
Scale of horizontal component is 8 times of
vertical Unit in m/s for seismometer and Pa for
pressure.
20
Seafloor experiment of this year (Sept 17,
2008-late Nov, 2008) Test with proto-type DONET
sensor to check installation procedure.
Buried sensor (broadband and strong motion sensor
inside)
21
Phase 1.5 connection of NantroSEIZE borehole
observatories
All of these observatories will be connected to
DONET Scientific submarine cable network to
assure long-term monitoring. The cables will be
laid out in 2009. Installation of the
observatories will take place after 2010.
Riser 2.5km hole
Riserless 1km hole below sediment layer.
Riserless 500m hole penetrating shallow splay
fault
Location of the borehole observatories are
planned within the reach (10km or 20km) of
DONET. Connection of the borehole observatories
will be funded as a part of JAMSTECs
borehole observatory project (not by DONET).
Red line DONET cable Pink points DONET
seafloor observatories
22
Seismic structure and models of strain change due
to slip in plate boundary.
seismogenic plate boundary
Zone where VLF earthquakes occur.
Proposed location of observatories
Land observation has no ability to model where
and how strain due to subducting slab is
accommodated. Physical conditions to make
plate boundary is not well understood.
23
Borehole observation system in the
NantroSEIZE Major target observation of
deformation at the seismogenic plate
boundary. Why borehole? Low noise environment,
better connection to the deformation source
(direct measurement of the fault itself in the
future). Sensors Strain (3 a
components) Pore-P (including of splay
fault) Tilt, Seismic (broadband, strong motion,
geophones) Array of thermometer,
electrodes. (Future implementation of continuous
active source experiments under
consideration). DONET connectivity Power, data
recovery, and precision timing.
24
Summary

• DONET network capability 40 sensors extended
  10km (20km) from 5 nodes. 30W power/sensor,
  precision timing, data communications.
• The Network covers area from near the trench to
  Tonankai co-seismic rupture area.
• 1) Phase 1 (- 2010)
• - 20 seafloor sensors (BB strong motion
  buried seismic sensors Quartz pressure gauge,
  DPG, hydrophone, and thermometer).
• - 15 30 km spacing between sensors capable of
  determining precise depth of micro earthquakes
  and detecting Mw6 class slow events.
• 2) phase 1.5 (-2011?)
• - 3 borehole observatories 2.5km riser hole,
  1km, 500m riserless hole with strain, pore-p,
  tilt, BBstrong motion seismic, thermometers, and
  electrodes.
• 3) Phase 2 to extend network westward. (planning)
• Extended observation items include
  electromagnetic, tilt, and real-time GPS-Acoustic
  observation.