Towards Routine Monitoring of Tectonic and...

Towards Routine Monitoring of Tectonic and Volcanic Deformation with Sentinel-1

Wright, Tim J (1); Biggs, Juliet (2); Crippa, Paula (3); Ebmeier, Susanna K. (2); Elliott, John (4); Gonzalez, Pablo (1); Hooper, Andy (1); Larsen, Ynvar (5); Li, Zhenhong (3); Marinkovic, Petar (6); Parsons, Barry (4);

Spaans, Karsten (1); Walters, Richard (1); Ziebart, Marek (7).1: COMET, University of Leeds, United Kingdom; 2: COMET, University of Bristol, United Kingdom; 3:

COMET, University of Newcastle, United Kingdom; 4: COMET, University of Oxford, United Kingdom; 5: Norut, Norway; 6: PPO.Labs, The Netherlands; 7: COMET, University College London, United Kingdom

INSARAP

Outline1. Why do we care about deformation observations

for tectonics and volcanoes?

2. Why is Sentinel-1 a game changer?

3. COMET plans for routine processing

INSARAP

Outline1. Why do we care about deformation observations

for tectonics and volcanoes?

2. Why is Sentinel-1 a game changer?

3. COMET plans for routine processing

INSARAP

Earthquakes with 10,000+ deaths since 1900

Figure courtesy John Elliott

Strain rate correlates with rate of earthquake occurrence

Figure reproduced with permission from Corné Kreemer, University of Reno and the Global Earthquake Model

Deformation at volcanoes is diagnostic of impending eruption/unrest

Photo: USGS

Biggs, Ebmeier et al., Nature Comms 2014

Outline1. Why do we care about deformation observations

for tectonics and volcanoes?

2. Why is Sentinel-1 a game changer?

3. COMET plans for routine processing

INSARAP

Why is Sentinel-1 a game changer?

Sentinel-1 Other SAR mission archives

1. Systematic acquisitions for

tectonics and volcanoes: “InSAR

everywhere all the time”

Haphazard acquisitions (multiple

modes, limited capacity)

2. TOPS: 250 km x 1000+ km:

Continental scale InSAR

Small areas imaged, usually less

than 100 km swaths.

3. Small perpendicular baselines,

acquisitions every 6/12/24 days,

ascending and descending -> high

coherence

Typically large perpendicular

baselines and long gaps between

acquisitions -> poor coherence

4. 20 year operational program,

designed for InSAR

Stand-alone missions not designed

for InSAR

5. Free, full and open data policy,

enables mass processing.

Restricted data access, often

commercial pricing

Why is Sentinel-1 a game changer?

Sentinel-1 Other SAR mission archives

1. Systematic acquisitions for

tectonics and volcanoes: “InSAR

everywhere all the time”

Haphazard acquisitions (multiple

modes, limited capacity)

2. TOPS: 250 km x 1000+ km:

Continental scale InSAR

Small areas imaged, usually less

than 100 km swaths.

3. Small perpendicular baselines,

acquisitions every 6/12/24 days,

ascending and descending -> high

coherence

Typically large perpendicular

baselines and long gaps between

acquisitions -> poor coherence

4. 20 year operational program,

designed for InSAR

Stand-alone missions not designed

for InSAR

5. Free, full and open data policy,

enables mass processing.

Restricted data access, often

commercial pricing

1. “InSAR everywhere, all the time” (NASA Solid Earth Science Working Group Report, 2002, NASA InSAR Workshop Report, 2004)

We can’t predict in advance the locations of future earthquakes and volcanic eruptions

1. “InSAR everywhere, all the time” (NASA Solid Earth Science Working Group Report, 2002, NASA InSAR Workshop Report, 2004)

24 August 2014 South Napa Earthquake

• Largest earthquake in California in 20 years.

• 1 death, ~160 injuries• $1 Billion costs to wine industry• Pre-earthquake Stripmap image

acquired by Sentinel-1A on 7 August (the day it reached nominal orbit)

• Post-earthquake image on 31 August, scheduled by special request.

Austin Elliott (UCDavis)

Elliott et al., EOS 2015

Source model from InSAR and GPSdata model residual

Slip UncertaintyNS NS

More in Elliott et al., Tuesday 15.30, Magellan

23 Nov 2014 Pico do Fogo Eruption

• 1st eruption in 20 yrs (1995)

• Flights to S. America diverted

• ~1500 evacuated people

• Two towns completely destroyed

Samara Donis (InVOLCAN)

David Calvo (InVOLCAN)

Observations

Ascending (20141103-20141127)

Descending (20141108-20141202)

Preferred Model

10-5 km ?Gonzalez et al., in review 2015

Thanks to ESA for expanding tectonic coverage

Original ESA mask

Current tectonic mask Ideal (?) tectonic mask

https://sentinel.esa.int/web/sentinel/missions/sentinel-1/observation-scenario

Why is Sentinel-1 a game changer?

Sentinel-1 Other SAR mission archives

1. Systematic acquisitions for

tectonics and volcanoes: “InSAR

everywhere all the time”

Haphazard acquisitions (multiple

modes, limited capacity)

2. TOPS: 250 km x 1000+ km:

Continental scale InSAR

Small areas imaged, usually less

than 100 km swaths.

3. Small perpendicular baselines,

acquisitions every 6/12/24 days,

ascending and descending -> high

coherence

Typically large perpendicular

baselines and long gaps between

acquisitions -> poor coherence

4. 20 year operational program,

designed for InSAR

Stand-alone missions not designed

for InSAR

5. Free, full and open data policy,

enables mass processing.

Restricted data access, often

commercial pricing

2. Continental Scale InSAR

Why is Sentinel-1 a game changer?

Sentinel-1 Other SAR mission archives

1. Systematic acquisitions for

tectonics and volcanoes: “InSAR

everywhere all the time”

Haphazard acquisitions (multiple

modes, limited capacity)

2. TOPS: 250 km x 1000+ km:

Continental scale InSAR

Small areas imaged, usually less

than 100 km swaths.

3. Small perpendicular baselines,

acquisitions every 6/12/24 days,

ascending and descending -> high

coherence

Typically large perpendicular

baselines and long gaps between

acquisitions -> poor coherence

4. 20 year operational program,

designed for InSAR

Stand-alone missions not designed

for InSAR

5. Free, full and open data policy,

enables mass processing.

Restricted data access, often

commercial pricing

3. Short revisit and Small perpendicular baselines -> Excellent Coherence

Creeping section of the North Anatolian Fault, 12-days

3. Short revisit and Small perpendicular baselines -> Excellent Coherence

Creeping section of the North Anatolian Fault, 24-days

3. Short revisit and Small perpendicular baselines -> Excellent Coherence

Creeping section of the North Anatolian Fault, 24-days

Typical ERS coherence (Cakir et al., 2005)

3. Short revisit -> rapid phenomena

Napa Postseismic deformation: 31 August – 12 September 2014

Afterslip model

NS

More in Elliott et al., Tuesday 15.30, Magellan

3. Short revisit -> rapid phenomena

Napa Postseismic deformation: 31 August – 24 September 2014

Afterslip model

NS

More in Elliott et al., Tuesday 15.30, Magellan

3. Short revisit -> rapid phenomena

Napa Postseismic deformation: 31 August – 6 October 2014

Afterslip model

NS

More in Elliott et al., Tuesday 15.30, Magellan

3. Short revisit -> rapid phenomena

Napa Postseismic deformation: 31 August – 18 October 2014

Afterslip model

NS

More in Elliott et al., Tuesday 15.30, Magellan

3. Short revisit -> rapid phenomena

Napa Postseismic deformation: 31 August – 30 October 2014

Afterslip model

NS

More in Elliott et al., Tuesday 15.30, Magellan

Why is Sentinel-1 a game changer?

Sentinel-1 Other SAR mission archives

1. Systematic acquisitions for

tectonics and volcanoes: “InSAR

everywhere all the time”

Haphazard acquisitions (multiple

modes, limited capacity)

2. TOPS: 250 km x 1000+ km:

Continental scale InSAR

Small areas imaged, usually less

than 100 km swaths.

3. Small perpendicular baselines,

acquisitions every 6/12/24 days,

ascending and descending -> high

coherence

Typically large perpendicular

baselines and long gaps between

acquisitions -> poor coherence

4. 20 year operational program,

designed for InSAR

Stand-alone missions, usually not

designed for InSAR

5. Free, full and open data policy,

enables mass processing.

Restricted data access, often

commercial pricing

Duration of time series (years)

0.11 2 3 4 5Le

ngt

h s

cale

of

ob

serv

atio

n (

km)

4. 20-year operational program

1 mm/yr rates over 100 km can be achieved with 5 years of acquisitions (12 day revisit)

Ability of Sentinel-1 to map tectonic strain above target threshold (1 mm/yr over 100 km)

Why is Sentinel-1 a game changer?

Sentinel-1 Other SAR mission archives

1. Systematic acquisitions for

tectonics and volcanoes: “InSAR

everywhere all the time”

Haphazard acquisitions (multiple

modes, limited capacity)

2. TOPS: 250 km x 1000+ km:

Continental scale InSAR

Small areas imaged, usually less

than 100 km swaths.

3. Small perpendicular baselines,

acquisitions every 6/12/24 days,

ascending and descending -> high

coherence

Typically large perpendicular

baselines and long gaps between

acquisitions -> poor coherence

4. 20 year operational program,

designed for InSAR

Stand-alone missions, usually not

designed for InSAR

5. Free, full and open data policy,

enables mass processing.

Restricted data access, often

commercial pricing

UK-PAF

Sentinel-1 archivemainly for

Copernicus core services

1-monthSentinel-1

rolling archive

SLCs

LiCS processing facility

processing facility

storage facility (and future public

services)

HarwellFarnborough

Sentinel-1 SAR processor

CEDA

5. “Free, full and open” data policy Mass Processing

Outline1. Why do we care about deformation observations

for tectonics and volcanoes?

2. Why is Sentinel-1 a game changer?

3. COMET plans for routine processing

INSARAP

Global Tectonics: 55 Mkm2 Ice: 5 Mkm2

Europe: 10 Mkm2

Total: 70 Mkm2

Data throughput = 0.5 TB/day [~1PetaByte over 5 years]

Geographical coverage

Processing strategy

T=0 ~1 yr (?) +12 days +24 days +36 days

Orbit Models (real time + precise)

Atmospheric Models (real time + reanalysis)

Initial time seriesInitial linear rates

Updated time seriesUpdated linear rates

Continuous, near-global processing

Interferogram(displacement)

Interferogram(displacement)

SAR images

Interferogram(displacement)

Time-series for each pixel

Strain-rate map

InSARprocessor

Time-series processor

Average velocity

map

Strain inversion

Precise orbits

Atmos. correction

Ifgmfilter

Earthquakemodels

GP

Svelo

cities

Geophysicalinversion

Volcanomodels

Work flowData

Processed data

Derivedproducts

Models

Work flow

Interferogram(displacement)

Interferogram(displacement)

SAR images

Interferogram(displacement)

Time-series for each pixel

Strain-rate map

InSARprocessor

Time-series processor

Average velocity

map

Strain inversion

Precise orbits

Atmos. correction

Ifgmfilter

Earthquakemodels

GP

Svelo

cities

Geophysicalinversion

Data

Processed data

Derivedproducts

Models Volcanomodels

INSARAP presentations today

Gonzalez poster (6), Tuesday

Bekaert, Tues 10.00, Big HallWalters, Tues 12.30, Big Hall+ Tues poster: Crippa (20)

Spaans, Thurs 16.00, Magellan

Wang, Weds 15.50, Big Hall

Elliott, Tues 15.30, MagellanHussain, Weds 11.50, Big Hall+ Tues posters: Ingleby (61), Amey (68), Lloyd (82)

Hooper, Thurs 15.20, MagellanHamlyn, Thurs 11.50, MagellanArnold, Thurs 9.00, Magellan+ Thurs posters: Biggs (76), Ebmeier (79), Gaddes (77), Gineaux (80), Bagnardi (85)

Global validation of ERA-Interim wet delay

• Major global variation in quality of ERA-I wet delay retrieval

• We can now consider the uncertainties associated with applying atmospheric corrections

Walters et al. Tuesday 12:30, Big HallSession: InSAR Theory and Techniques (2)

BAD

GOOD

Recursive Adaptive Spectral Phase filtering

RSF Filtered phase

Gonzalez et al. Tuesday Poster Session #6

5x5 boxcar Cousin based coherence

Improved pixel selection via identification of cousins

• Identify pixels with on average similar behaviour (Similar to SqueeSAR methods of Ferretti et al, 2011)

• We use mean amplitude and mean amplitude difference between master and slave

Spaans et al. Thursday 16:00, MagellanSession: Applications – Volcanoes (3)

Improved pixel selection via identification of cousins

Full interferogram Our method Small baselines

• Identify pixels with on average similar behaviour (Similar to SqueeSAR methods of Ferretti et al, 2011)

• We use mean amplitude and mean amplitude difference between master and slave

Spaans et al. Thursday 16:00, MagellanSession: Applications – Volcanoes (3)

Turkey

Iran

Using InSAR to Map Strain in Eastern Turkey

Method in Wang and Wright, GRL 2012; Turkey case study in Walters et al., JGR 2014

Using InSAR to Map Strain in Eastern Turkey

Solution from GPS Solution from InSAR

Method in Wang and Wright, GRL 2012; Turkey case study in Walters et al., JGR 2014

What will we be seeing at Fringe 2020?

• High-resolution, time-varying, global 3D maps of crustal velocities and strains.

• Automatic alert systems for volcanoes

• Complete catalogue of deformation models for continental earthquakes with M>6

• Integration with GNSS for near-real-time monitoring.

• Many scientific surprises!

INSARAP

More information on INSARAP, and download processed data at http://insarap.orgCOMET: http://comet.nerc.ac.uk

@NERC_COMET@timwright_leeds@EwFProject