Kiyo Tomiyasu, Co-Seismic Slip, and the Krafla Volcano: Reflections on InSAR and Earth SciencePaul A. RosenJet Propulsion LaboratoryCalifornia Institute of TechnologySpecial Session Honoring the Achievements of KiyoTomiyasuIGARSS 2010Honolulu, Hawaii
Background
OutlineTrends in Interferometric SAR (InSAR) for Earth ScienceGeosynchronous InSAR ConceptPresentation of 90th Birthday Celebratory Plaque of Appreciation to Kiyo Tomiyasu from JPL
Interferometric SAR for Measuring Earth Surface Change
Trends in Observational Techniques for Earth ScienceFrequent sampling in timeFine spatial resolutionTime series / PS analysisExtraction of geophysical parameters automaticallyExploitation of data for rapid response to eventsAnticipated mean access timesfor upcoming systemsInterferogram stackU   Mean Access Time (Day)   ∞        4        2       1.3      1time
A multi-scale approach to InSAR time series analysisM. Simons, E. Hetland, P. Muse, Y. N. Lin & C. DiCaprioInterferogram stackUA geophysical perspective on deformation tomographyExample: Northern Volcanic Zone, Icelandtime
MotivationAssume that in the future we will have:Frequent repeats (short DT)
Good orbits with small baselines
Ubiquitous high coherenceChallenge for the future:How to deal with O(103) interferograms
How to use Cd - Invert all pixels simultaneously?1000 igramsx 1000 x 1000 pixels = 1 billion dataComputational tractabilityApproach: MInTS= Multi-scale InSAR Time SeriesTime domain: A generalized physical parameterization (GPS-like)Space domain: Wavelets – use all data simultaneously
MInTS Methodology Interpolate unwrapping holes in each interferogram where	needed (temporary) Wavelet decomposition of each interferogram	For later weighting purposes, track relative extent to which each wavelet coefficient is associated with actual data versus interpolated data Time series analysis on wavelet coefficients	Physical parameterization + splines for unknown signals - all constrained by weighted wavelet coefficients of observed interferograms Recombine to get total deformation history
Example: Iceland Northern Volcanic Zone – Instantaneous Velocity
Example: Iceland Northern Volcanic Zone – Instantaneous Velocity (nonlinear)
Summary: Iceland Northern Volcanic Zone – Instantaneous Velocity
MInTS gives us continuous time, but does not yet combine multiple LOS to get 3D displacements.  For the moment, we adopt a simple 2D reconstruction approach on a profile and neglect any rift parallel motion along the profile.  Note asymmetries.
Rift zone models compared to data
Geosynchronous SARAn approach to observing the evolution of Earth’s surface
A Geosynchronous Synthetic Aperture Radar;for Tectonic Mapping, Disaster Management and Measurements of Vegetation and  Soil MoistureIGARSS, Sydney, July 9–13, 2001Søren N. Madsen, Wendy Edelstein, Leo D. DiDomenicoJet Propulsion Laboratory, California Institute of TechnologyJohn LaBrecqueNASA Headquarters
16Previous WorkTomiyasu K.:“Synthetic Aperture Radar in Geosynchronous Orbit,” Dig. Int. IEEE Antennas and Propagation Symp., U. Maryland, 42–45, May 1978“Synthetic Aperture Radar Imaging from an Inclined Geosynchronous Orbit,” IEEE Trans. Geosci. Remote Sens. GE-21(3), 324–328 (1983)Holt, B. & Hilland, J.“Rapid-Repeat SAR Imaging of the Ocean Surface: Are Daily Observations Possible?” Johns Hopkins APL Technical Dig., 21(1), 162–169, 2000
17GeoSync SAR Orbit and Measurement DescriptionOrbit35789 km altitude (geosynchronous)60˚ inclination (not geostationary)1 day repeatInstrumentL-band SARContinuous strip mapping, interferometricScanSAR, or spotlight operation30 m diameter antenna aperture (electronically scanned array)Distributed T/R modules on membraneNadir pointed, all steering electronic (only ±8º required side to side)Radar and spacecraft bus integrated on inflatable/rigidizable structure5500 km accessible ground swath on either side of nadir100% instrument duty cycle (always in view of land)
18Operational Modes Highly FlexibleOperational modesStripmap SAR with 400 km swath width:10 m resolution @ 4–5 looksSuited for high-resolution mappingScanSAR over 5500 km swaths on either side of nadir track:50 m @ 4–5 looks  Daily continental coverageSquint-scanned SAR (beam hops to +45˚, broadside, –45˚):3–D displacement mapping of extended areas in a single day  Useful for tectonic studiesSpotlight SAR (beam dwells on single target area for long time):High resolution in azimuth, semi-continuous coverageSuitable for disaster managementHigh resolution stepped frequency SAR (step frequency within 80 MHz band on successive passes then combine coherently to get high resolution without losing SNR or increasing data rate):2 m ground range resolution, 2m azimuth resolution at far range6-10 m resolution at near rangeData rates and volumesData rate 220 Mbits/sec per 20 MHz channel2.4 TB/day with nearly 100% instrument duty cycle
19GeoSync Instrument ConceptL-band single-polarization (HH) SAR

WE3.L10.4: KIYO TOMIYASU, CO-SEISMIC SLIP AND THE KRAFLA VOLCANO: REFLECTIONS ON INSAR AND EARTH SCIENCE

  • 1.
    Kiyo Tomiyasu, Co-SeismicSlip, and the Krafla Volcano: Reflections on InSAR and Earth SciencePaul A. RosenJet Propulsion LaboratoryCalifornia Institute of TechnologySpecial Session Honoring the Achievements of KiyoTomiyasuIGARSS 2010Honolulu, Hawaii
  • 2.
  • 3.
    OutlineTrends in InterferometricSAR (InSAR) for Earth ScienceGeosynchronous InSAR ConceptPresentation of 90th Birthday Celebratory Plaque of Appreciation to Kiyo Tomiyasu from JPL
  • 4.
    Interferometric SAR forMeasuring Earth Surface Change
  • 5.
    Trends in ObservationalTechniques for Earth ScienceFrequent sampling in timeFine spatial resolutionTime series / PS analysisExtraction of geophysical parameters automaticallyExploitation of data for rapid response to eventsAnticipated mean access timesfor upcoming systemsInterferogram stackU Mean Access Time (Day) ∞ 4 2 1.3 1time
  • 6.
    A multi-scale approachto InSAR time series analysisM. Simons, E. Hetland, P. Muse, Y. N. Lin & C. DiCaprioInterferogram stackUA geophysical perspective on deformation tomographyExample: Northern Volcanic Zone, Icelandtime
  • 7.
    MotivationAssume that inthe future we will have:Frequent repeats (short DT)
  • 8.
    Good orbits withsmall baselines
  • 9.
    Ubiquitous high coherenceChallengefor the future:How to deal with O(103) interferograms
  • 10.
    How to useCd - Invert all pixels simultaneously?1000 igramsx 1000 x 1000 pixels = 1 billion dataComputational tractabilityApproach: MInTS= Multi-scale InSAR Time SeriesTime domain: A generalized physical parameterization (GPS-like)Space domain: Wavelets – use all data simultaneously
  • 11.
    MInTS Methodology Interpolateunwrapping holes in each interferogram where needed (temporary) Wavelet decomposition of each interferogram For later weighting purposes, track relative extent to which each wavelet coefficient is associated with actual data versus interpolated data Time series analysis on wavelet coefficients Physical parameterization + splines for unknown signals - all constrained by weighted wavelet coefficients of observed interferograms Recombine to get total deformation history
  • 12.
    Example: Iceland NorthernVolcanic Zone – Instantaneous Velocity
  • 13.
    Example: Iceland NorthernVolcanic Zone – Instantaneous Velocity (nonlinear)
  • 14.
    Summary: Iceland NorthernVolcanic Zone – Instantaneous Velocity
  • 15.
    MInTS gives uscontinuous time, but does not yet combine multiple LOS to get 3D displacements. For the moment, we adopt a simple 2D reconstruction approach on a profile and neglect any rift parallel motion along the profile. Note asymmetries.
  • 16.
    Rift zone modelscompared to data
  • 17.
    Geosynchronous SARAn approachto observing the evolution of Earth’s surface
  • 18.
    A Geosynchronous SyntheticAperture Radar;for Tectonic Mapping, Disaster Management and Measurements of Vegetation and Soil MoistureIGARSS, Sydney, July 9–13, 2001Søren N. Madsen, Wendy Edelstein, Leo D. DiDomenicoJet Propulsion Laboratory, California Institute of TechnologyJohn LaBrecqueNASA Headquarters
  • 19.
    16Previous WorkTomiyasu K.:“SyntheticAperture Radar in Geosynchronous Orbit,” Dig. Int. IEEE Antennas and Propagation Symp., U. Maryland, 42–45, May 1978“Synthetic Aperture Radar Imaging from an Inclined Geosynchronous Orbit,” IEEE Trans. Geosci. Remote Sens. GE-21(3), 324–328 (1983)Holt, B. & Hilland, J.“Rapid-Repeat SAR Imaging of the Ocean Surface: Are Daily Observations Possible?” Johns Hopkins APL Technical Dig., 21(1), 162–169, 2000
  • 20.
    17GeoSync SAR Orbitand Measurement DescriptionOrbit35789 km altitude (geosynchronous)60˚ inclination (not geostationary)1 day repeatInstrumentL-band SARContinuous strip mapping, interferometricScanSAR, or spotlight operation30 m diameter antenna aperture (electronically scanned array)Distributed T/R modules on membraneNadir pointed, all steering electronic (only ±8º required side to side)Radar and spacecraft bus integrated on inflatable/rigidizable structure5500 km accessible ground swath on either side of nadir100% instrument duty cycle (always in view of land)
  • 21.
    18Operational Modes HighlyFlexibleOperational modesStripmap SAR with 400 km swath width:10 m resolution @ 4–5 looksSuited for high-resolution mappingScanSAR over 5500 km swaths on either side of nadir track:50 m @ 4–5 looks Daily continental coverageSquint-scanned SAR (beam hops to +45˚, broadside, –45˚):3–D displacement mapping of extended areas in a single day Useful for tectonic studiesSpotlight SAR (beam dwells on single target area for long time):High resolution in azimuth, semi-continuous coverageSuitable for disaster managementHigh resolution stepped frequency SAR (step frequency within 80 MHz band on successive passes then combine coherently to get high resolution without losing SNR or increasing data rate):2 m ground range resolution, 2m azimuth resolution at far range6-10 m resolution at near rangeData rates and volumesData rate 220 Mbits/sec per 20 MHz channel2.4 TB/day with nearly 100% instrument duty cycle
  • 22.
    19GeoSync Instrument ConceptL-bandsingle-polarization (HH) SAR
  • 23.
  • 24.
    65 KW peaktransmit power
  • 25.
    724 kg totalinstrument mass
  • 26.
    28 KW DCinstrument powerPropulsion Modules (x2)Thin-filmSolar ArraysHorizontal booms (x12)L-band RF membrane antenna apertureSpacecraft BusTelescoping booms (x2)
  • 27.
    20GeoSync Constellations &CoverageConstellation of 10 satellites in 5 groups (2 satellites per “figure-8” ground track)Most of populated parts of Earth visible nearly continuouslyMax duration of gaps in coverage less than 2 hours for 90 % of surface3-D displacement accuracy for select target areas < 1 cm with 24 hours of observationsMaximum coverage gapMaximum 3D displacement errorRelative AccuracyMinutes
  • 28.
    InnovationNew concepts forgeosynchronous deformation observationsNew ideas in enabling technologiesSeismology from SpaceImproving Earthquake Forecasting
  • 29.
    Celebratory PlaqueToKiyo TomiyasuWithgreatest appreciation on your ninetieth birthday for a lifetime of innovation in remote sensingSigned by Charles Elachi, DirectorJet Propulsion Laboratory

Editor's Notes

  • #7 There are already several very successful methods for estimating time-variable deformation fields from stacks of radar interferograms. Today, I will provide a variation on these existing approaches. These variations are heavily influenced by a geophysical bias of how the Earth deforms. The methods we use borrow from standard practices in GPS time series analysis as well as from ideas common in seismic tomography. In essence, this is space/time deformation tomography designed to provide the best estimate of deformation everywhere at all times.