Magnetotelluric Investigations of the Kilauea Volcano, Hawaii

1
Magnetotelluric Investigations of the Kilauea
Volcano, Hawaii

• G. Michael Hoversten1
• Erika Gasperikova1
• Greg A. Newman2
• Jim P. Kauahikaua3
• 1 Lawrence Berkeley National Laboratory
• 2 Sandia National Laboratories
• 3 US Geological Survey, Hawaiian Volcano
  Observatory

2
Outline

• Project Objectives
• Survey Background
• The Magnetotelluric method
• Data Quality
• Modeling and Inversion
• Conductivity Structure
• Conclusions

3
Objectives

• Produce detailed 3D images of the electrical
  conductivity structure beneath Kilaueas east and
  southwest rift zones
• use this to infer the location of melt - partial
  melt zones and the structure of the rift systems
• Push the envelope of large-scale 3D
  Magnetotelluric forward and inverse modeling in
  the presence of rugged topography and coastal
  effects

4
MT Survey Location
Active Vents over the last 20 years

• 70 80 spatially distributed sites
• are planned for 3D inversion
• 33 sites acquired in August 2002

• Site access on the East Rift Zone
• is by helicopter
• Site access in the forest is by foot

5
Magnetotelluric Method
Induction Coil
Covered Electrodes

• Electromagnetic impedance Z(x,y,f)E(x,y,f)/H(x,y,
  f)
• function of the earth conductivity structure
• Magnetic Induction coils measure orthogonal
  magnetic fields
• 1m long, 20 lb each, 10 lb battery
• must be buried for wind noise
• 200 m grounded electric dipoles measure
  orthogonal electric fields
• Cu-Cu2S electrodes

6500 ft on Mauna Loa
2500 ft on East Rift Zone
6
Data Quality

• Design recording time 48 hours
• 36 hours was more common
• pickup deploy time helicopter schedule
• Overall quality was good
• noise in the dead band around 10 sec. period
• Second phase will aim for 72 hour deployments
• Frequency range (200 - 0.0002 Hz) translates to
  100 m to 20 km depth sensitivity in models
• below 0.01 Hz (100 sec) must use 3D to account
  for island effects

Best data at 72 hour recording
Average data at 36 hour recording
7
Modeling and Inversion

• 3D numerical model of Hawaii surrounding sea
  floor
• 14 Million cells
• use 512 nodes on MP cluster
• 3D topography costal effects depart from 2D
  below 100 sec period
• 2D inversions use 0.01 to 100 seconds data
• 3D inversion tests are underway to determine
  optimal grid and domain size
• extremely computationally intensive (CPU
  memory)
• 3D is where the data interpretation is going in
  the coming year

8
Modeling and Inversion

• Sites were grouped into lines for 2D inversion
• E in-line H perpendicular to line were used for
  inversion (least affected by out-or-plane
  effects)
• Frequency range used for 2D limited based on 3D
  modeling of topography and coastal effects
• 0.01 to 100 sec. period
• depth sensitivity depends on conductivity (5-10
  km)
• Regularized inversion (smoothing)
• sharp contrasts bleed into background
• termination of conductors at depth does not mean
  the structure terminates, just loss of data
  sensitivity

9
Conductivity Structure

• Depth location of Vp/Vs anomaly (Dawson et al.)
  ties to high conductivity beneath southern
  Kilauea caldera
• high fracturing and/or partial melt
• High conductivity zones 2-4 km below surface in
  the east
• High conductivity zone 1-6 km below surface in
  the west
• depth sensitivity limited by 100 sec cut off in
  2D inversions

North
10
Conductivity Structure

• South dipping conductors south of caldera ties to
  low Vp/Vs anomaly
• deep conductor to 8 km may be expression of
  Summit magma reservoir (Ryan 1988)
• Deep, high resistivity ties to Haslinger et al.
  high velocity
• MT (100sec cut) loosing sensitivity at 10km
• Deep low Vp NE trend from Haslinger et al. ties
  to deep high conductivity zone beneath Puu Oo
• Depth miss-tie at line intersection
• this area is at junction of rift zones data is
  more 3D than elsewhere

North
11
Conductivity Structure

• Deep conductor seen on all western sites around
  Puu Oo appears to dip to the south
• fading of conductivity with depth is smoothing
  effect of inversion due to loss of information
  (100sec cut)
• Single site (J30) between east west limits
  information on the connection between East Rift
  and the Caldera area
• Additional MT sites between East and South-West
  rift will be acquired in 2003

North
J30
12
Conclusions (1)

• High quality MT data can be acquired on high
  resistivity, recent lava flows
• use of remote magnetics facilitates data
  collection where access is time weight limited
• Extreme topographic and oceanland contact
  effects can be modeled in 3D
• accounted for in 2D inversions by limiting
  frequency range
• modeled explicitly in 3D inversion
• 2D inversions of data gt 100 sec period yield
  conductivity structures that tie with published
  velocity models and correlate with recent
  volcanic activity
• conductivity velocity values indicate shallow
  partial melt zones beneath the south-southwest of
  Kilauea caldera

13
Conclusions (2)

• A near vertical conductive structure was imaged
  at the approximate location of the summit magma
  reservoir (Ryan 88)
• use of data below 100 sec in 3D should extend the
  depth images
• The deep high conductivity zone to the west
  beneath Puu Oo and the active vents is larger
  than expected relative to the conductive zones
  images at the summit magma reservoir
• this may be an artifact of 2D inversion and/or
  sparse data
• additional sites in 2003 and further scrutiny
  during 3D inversion and modeling
• MT offers a tool to complement other geophysics
  for imaging the structure of the volcanic rift
  systems on Kilauea
• we expect 20 km depth of investigation with
  full data set
• Future expanded site coverage and 3D inversion
  should provide even better definition of Kilauea
  structures

14
Acknowledgements

• Support for this work was provided by the U.S.
  Department of Energy under Contract No.
  DE-AC03-76SF00098.
• We thank Electro Magnetic Instruments for
  contributing 9 prototype MT24 systems, without
  which this work would not have been possible
• We thank the USGS Hawaiian Volcano Observatory
  for logistical and personnel support
• We thank the USGS Menlo Park for financial and
  staff support