Earthquake Sensation: integrating GPS and inertial sensors

1
Earthquake Sensationintegrating GPS and
inertial sensors
CENS - UCLA July 25, 2003

• Kenneth W. Hudnut, Ph.D.
• Chief, So. Calif. Earthquake Hazard Assessment
  Project
• Earthquake Hazards Team
• U. S. Geological Survey

2
Southern California is the nations most
dangerous place for earthquakes - why?
Silver linings State-of-the-art earthquake
monitoring arrays Many of the worlds best
researchers work within this natural laboratory
3
San Andreas fault

• 35 mm/yr slip rate
• gt70 of total plate boundary motion
• 1685, 1812, 1857 eqs
• Big Bend compression
• 1971 Sylmar (M 6.7)
• 1994 Northridge (M 6.7)
• SoCal is now heavily wired (much like Japan
  Taiwan)
• 150 BB CISN stations
• 250 SCIGN stations
• Catalog SCEC CMM3

4
Improving hazard assessment

• Temporal variations do occur
• Clustering (e.g., Basin Range, ECSZ, Asia)
• Discrepant geological and geodetic rates
• Sequences involving fault interaction (e.g.,
  Joshua Tree - Landers - Big Bear - Hector Mine
  Anatolian system, etc.)
• Implement robust research findings into hazard
  assessment
• Variability does not mean predictability

Courtesy Anke Friedrich
5
GPS
6
SCIGN Data Products

• 1st Year
• Combined time
• series (1996-2002)
• 3rd Year
• Real-time earth-
• quake response
• 5th Year
• Resolve rates on
• primary LA basin
• faults (and others)

7
(No Transcript)
8
New methods and data integration

• Precise topographic mapping of surface ruptures
  and active fault scarps
• slip models for prehistoric events
• rapid assessment of surface slip and damage
  patterns after large events
• Requires precise integration of GPS INS for
  flight navigation

1957 Gobi-Altai earthquake surface rupture
9
High resolution topography along surface rupture
of the October 16, 1999 Hector Mine, California
Earthquake (Mw7.1) from Airborne Laser Swath
Mapping

• Hudnut, K. W. (USGS), A. Borsa (UCSD),
• C. Glennie (Aerotec, LLC) and J.-B. Minster (UCSD)

Bulletin of the Seismological Society of
America Special Issue on the Hector Mine
earthquake (2002) http//pasadena.wr.usgs.gov/off
ice/hudnut/docs/
10
Airborne laser swath mapping (ALSM)

• precise topographic mapping of surface ruptures
  and active fault scarps

Airborne platform navigation must be highly
precise and requires high-rate GPS data

• representation of actual fault ruptures recorded
  and preserved in unprecedented detail

11
Geolocation Vectors and Error Sources
Vector from CMearth to GPS phase center Magnitude
directional errors both are stochastic, time
and location variant.
Vector from GPS phase center to laser Magnitude
error is constant if no airframe flexing.
Directional error due to constant and
time-varying biases in INS.
Vector from laser to ground footprint Magnitude
error due to timing, instrument and atmospheric
delays. Directional error from constant mirror
mounting offsets and time-varying biases in
reporting of scan angles (both pitch and roll).
Note additional errors due to imperfect synchroni
zation of GPS, INS, mirror scan and laser firing
times must be modeled and removed as well.
12
Exploded ordnance (crater)
Lavic Lake Roll Pitch Maneuvers
pitch maneuvers
13
Maximum slip section of the 1999 Hector Mine
eq. surface rupture
Photo by Keith Stark (SCIGN)
14
Estimating slip on max. slip segment ofthe
fault
15
(No Transcript)
16
(No Transcript)
17
(No Transcript)
18
(No Transcript)
19
Laser Scanof the San Andreas
Proposal to the NSF EarthScope Science RFP Prof.
Mike Bevis, PI (OSU)
Requires high-rate (1 Hz) GPS data from
SCIGN sites along fhe fault special care with
IMU-INS
20
GPS Fault Slip Sensor

• K. Hudnut, G. Anderson, A. Aspiotes,
• N. King, R. Moffitt, K. Stark
• (all at USGS-SCIGN, Pasadena CA)

APEC symposium Proc. Paper, Fall AGU poster, and
paper in preparation for Bulletin of the
Seismological Society of America http//pasadena.
wr.usgs.gov/office/hudnut/slipsensor/
21
Early Warning
The speed of light gtgt the speed of sound
EmergencyResponse
Seismic and GPS Stations
Utilities
Transport- ation
(e.g Wu Teng, BSSA, 2002 Allen Kanamori,
Science, 2003)
22
(No Transcript)
23
Real-Time GPS Network - Enhancing SCIGN

• On 15 November 2002, first-ever GPS fault slip
  sensor deployed across San Andreas fault at
  Gorman, Calif.
• Augments seismic early warning system - resolves
  the observational deficiency inherent with
  inertial sensors that cannot discern tilt from
  acceleration
• Upgrade SCIGN telemetry
• DSL, frame relay
• Radio repeaters, WiGate and dedicated links
• Data buffering
• Augment SCIGN real-time acquisition and
  processing system
• Implemented sub-daily processing (4 hr) for 100
  SCIGN stations (down from 24 hr)
• Implementing multiple real-time streaming GPS
  processors (commercial software)

24
Lone Juniper Ranch and Frazier Park High School
First prototype GPS fault slip sensor
Spans the San Andreas fault near Gorman,
California
25
Cleaned-up test results
Why is real-time GPS processing noisy and less
robust than post-processing? Ambiguity
resolution, multipath, atmosphere and clock
errors - what can be done?
26
Upgrading SCIGN telemetry
Low cost options such as frequent FTS dial-up,
radio nets, and DSL Development testing of near
real-time GPS precise processing, etc.
27
Conclusions

• GPS and inertial sensor integration for
  high-accuracy applications is practical in SoCal
  because SCIGN can be readily upgraded to support
  centimeter-level accuracy in real-time

• Earthquake applications of GPS/INS integration
  are
• societally relevant - significant economic impact
  and consequences would result from technological
  innovations
• scientifically exciting for the very dynamic SCEC
  research community - major breakthroughs in
  earthquake source physics are likely to result
  from collaborations