Strain Signals Associated with the January 2007 Cascadia Episodic Tremor and Slip Event as Recorded on PBO Borehole Strainmeters

X (km)

Y (k

m)

Strain SXY component

Coulomb 3.0.04.12 02−May−2007 14:50:05 TimPatchS1.inpStrain calc. Depth: 0.01 km

Coulomb 3.0.04.12 02−May−2007 14:49:59 TimPatchS1.inpStrain calc. Depth: 0.01 km

0 50 100 150 2000

50

100

150

200

250

−1

−0.8

−0.6

−0.4

−0.2

0

0.2

0.4

0.6

0.8

1x 10−8

B004

B018

Wendy McCausland and Evelyn Roeloffs , U.S. Geological Survey, Vancouver, WA

AbstractThe Plate Boundary Observatory (PBO) was designed to study the strain fields resulting from the tectonic deformation at active plate boundaries, including The Episodic Tremor and Slip (ETS) events in the Cascadia subduction zone. As yet, it is not known whether the same source causes both the seismically-measured tremor and the geodetically-measured slow slip. In September 2005, the first ETS related strain event was recorded on two of the borehole strainmeters located on the Olympic Peninsula. During the next anticipated ETS event in January 2007, thirteen PBO borehole strainmeters were operational. Preliminary analyses of the 1 Hz data on station B018 (near Delphi, WA) show a strain event beginning around January 16, 2007. The event first manifests as a shear event, and then as E-W extension. The strain changes are roughly concurrent with the onset of strong tremor in the southern to central Puget Sound, and with geodetic changes recorded on the regional GPS network and long-base tiltmeter at Shelton, WA. Preliminary analyses of station B004 on the Northern Olympic Peninsula show a strain event that begins around January 26, 2007, later than at B018. The event first shows up as E-W compression and then as shear. The timing of the event at B004 is consistent with the migration of the tremor epicenters. The availability of 20 Hz strainmeter data will allow for better evaluation of the timing of the slow slip event than is possible from the GPS data and may reveal strain associated with the tremor.

Map showing the locations of PBO borehole strainmeters in Canada, Washington,and Oregon. Red triangles show locations of currently installed and operating stations. Blue squares are planned stations. Stations B004 and B018 are labelledbecause they are used in the analyses at right. Map modified from Unavco/PBO website.

X: -121.15Y: 48.718

B004

B018

Goal: Our goal is to incorporate the strainmeter data into geodetic models of slow slip events. To do this we need to calibrate the instruments from instrument strain to formation strain. Finally we compare the strainmeter data to simple slip models.

Gladwin Tensor StrainmeterStrainmeter measures horizontal extension on variable capacitance sensors fromwhich the full horizontal strain tensor can be determined. Three of the gaugesare oriented at 120 degrees intervals and the fourth gauge is perpendicular to one of the other three gauges. Resolution of the sensors is 1 nanostrain, the sampling rate is 20Hz and the depth of installation is between 500 and 800 feet.

Study Location:

Drawing: GTSM Technologies From UNAVCO/PBO Website

Figure showing the strain signals when the 3 evenly spaced gauges are combined into geographic coordinates: differential extension, shear and areal strain components. The timeseries have been linearized and the long-term trends have been removed. Blue time series is before the atmospheric pressure and tides have been removed.

Relating Instrument Strains to Local Strains

We want to measure elongation along an azimuth in the rock formation. If the strainmeter weren’t there we would measure the linear strain equation. But when a strainmeter is present in the ground, we get a similar equation with the factors C and D (areal and shear coupling coefficients). The coefficients reflect the coupling of the instrument and grout to the formation. D is typically larger than C, meaning the instrument is more sensitive to shear strain than areal strain. The relationship between regional strain and strain within meters of the instrument has yet not been determined. C, D are normally estimated from earth tide-induced strains.Tidal models are not accurate in this region therefore we sought another way to estimate C and D using teleseismic earthquake phases.

Figure showing a teleseismic earthquake on a broadband seismometer anda nearby borehole strainmeter. The seismogram has been rotated into radial and transverse components and converted to a displacement seismogram. The Love and Rayleigh wave arrivals are marked.

For planar Rayleigh waves we can expect the strain to vary as cosine(2*theta).Theta is the angle between the wave propagation direction and the gauge orientation. Figure above shows the expected strain equation; strainmeter data for the Rayleigh wave of an earthquake at station B018 for the different gauges; and the theoretical distribution of strain amplitude given the nominal calibration used by PBO, and the values determined at B004 and B018 based on Love and Rayleigh wave phases from 3 teleseismic events. Figure below shows the relative gauge amplitudes are plotted with respect to cosine(2*theta). As expected the values are linear and the slope of that line describes the ratio C/D.

Using the calculated values of C and D, and gauge weights determined fromthe consistency equation, we can transform the instrument strain to formation strain. Figure below shows the instrument strain values on the left vertical axes and the formation strain values on the right axes for both B004 and B018. Timeseries includes the January 2007 slow slip event.

Strainmeter Results from January Cascadia Slow Slip Event

Strainmeters are good at recording the detailed time history of the slip event.

Because the data is detrended, the net strain change is difficult to determine.

The strain changes on these two strainmeters do not overlap in time, butoccurs on B018 before B004 indicating the slip event migrated to the north.

The areal signal on both instruments is weak.

Both stations have a reversal in the shear strain, but not in the differential extension. The reversal suggests that the source moved past the station.

The sense of strain is different on the two instruments indicating they have adifferent spatial relationship to what is slipping.

Seismic Results from January 2007

Figure showing network locations by Aaron Wech and Ken Creager. Tremor generally migrates from south to north during the tremor and slow slipevent. From PNSN website.

GPS Results from January 2007

Figure on left shows the net GPS-determined displacements and the figureon the right shows a geodetic model of the slip patch based on the data inthe figure on the left.. From PANGA website. GPS data show a northward slip progression as well.

Comparing Strain Data to Simple Slip Models

B004

B018

Slip patch

We used Coulomb 3.0 (Toda, Stein, Lin, Sevilgen) to model a slip patch centered at 47.5N 122.9W with a slab depth of 40km, slab dip of 20 degrees and a total slip of 1cm. We modeled oblique slip in the opposite direction to North America- Juan de Fuca convergence direction (~55deg).

Contour Plot of Shear Strain from Model

Data from B018

Differential Extension from ModelData from B018

Areal Strain from ModelData from B018

Conclusions- Instruments

Strainmeters clearly recorded the January 2007 Cascadia slow slip event. Strainmeters record the time history of slow slip event which is useful for refining geodetic models.Strainmeters are capable of recording large teleseismic events, and moderate size local earthquakes.Ratio of Areal to Shear coupling is smaller than what PBO is using. In instrument strain, shear appears larger than it is.

Conclusions- Slip Event

B004 recorded the slip event several days later than B018, indicating that the slip event migrated from south to north, which is consistent with the migration of the tremor epicenters.There is no temporal overlap between the signal at B018 and at B004, therefore we can only model one station at a time. B013 will help fill this gap. Sense of shear is different at B004 and B018 indicating the stations have a different spatial relationship to the slip surface.The two shear components have different time histories indicating that the source geometry changed or moved.The reversal of the EN shear component at B018 indicates that the source initiated south of the station and then moved northward.

-10

-5

0

5

10

0

30

60

90

120

210

240

270

300

330

Propagation

Direction

B018

B004

PBO nominal

Propagation Direction

-150

-100

-50

0

50

100

150

1255 1275 1295 1315 1335

B018 Kuril Islands M8.3 Nov 15 2006Rayleigh Wave

nanostrain (instrument)

Seconds

CH0

CH1

CH2

CH3

Average

For Planar Rayleigh Waves

Linear Gauge Strains from

Surface Waves

nano

stra

in (

inst

rum

ent)

Shear Strain

Differential Extension

-3

-2

-1

0

1

2

3

4

-0.8 -0.4 0 0.4 0.8

B004 Rayleigh Waves

Hawaii15Oct06 C/D=0.18Kurils 15Nov06 C/D=0.14Kurils 13Jan07 C/D=0.13

Relative AmplitudeRel

ativ

e A

mpl

itude

Instrument and Formation Strain

-0.12

-0.08

-0.04

0

0.04

0.08

12/20/06 1/3/07 1/17/07 1/31/07 2/14/07 2/28/07

(εEE

+εNN

)/2

-0.12

-0.08

-0.04

0

0.04

0.08B018

microstrain (instrument)

microstrain (formation)

Jan. 14 262119

-0.2-0.1

00.10.20.30.4

12/20/06 1/3/07 1/17/07 1/31/07 2/14/07 2/28/07

εEE

-εNN

-0.02

0

0.02

0.04B018

microstrain (instrument)

microstrain (formation)

-0.2-0.1

00.10.2

12/20/06 1/3/07 1/17/07 1/31/07 2/14/07 2/28/07

εEN

-0.02

0

0.02B018

microstrain (formation)

microstrain (instrument)

-0.2-0.15

-0.1-0.05

00.05

0.10.15

0.2

12/20/06 1/3/07 1/17/07 1/31/07 2/14/07 2/28/07

εEE

-εNN

-0.04

-0.02

0

0.02

0.04B004

microstrain (instrument)

microstrain (formation)C=0.6 C/D=0.15

Jan. 26 29 31

(B004 Areal not shown -no detectable signal)

-0.1

0

0.1

0.2

12/20/06 1/3/07 1/17/07 1/31/07 2/14/07 2/28/07

εEN

-0.02

0

0.02

0.04B004

microstrain (formation)

microstrain (instrument)

mic

rost

rain

(in

stru

men

t)m

icro

stra

in (

inst

rum

ent)

microstrain (form

ation)

mic

rost

rain

(in

stru

men

t)m

icro

stra

in (

inst

rum

ent)

mic

rost

rain

(in

stru

men

t)

microstrain (form

ation)m

icrostrain (formation)

microstrain (form

ation)m

icrostrain (formation)

-0.2

-0.1

0

0.1

0.2

11/15/06 12/20/06 1/24/07 2/28/07

εEE

-εNN

B004

microstrain

-0.2

-0.1

0

0.1

0.2

11/15/06 12/20/06 1/24/07 2/28/07

εEN

B004

microstrain

-0.2

-0.1

0

0.1

0.2

11/15/06 12/20/06 1/24/07 2/28/07

(εEE

+εNN

)/2B004

microstrain

-0.2

-0.1

0

0.1

0.2

0.3

0.4

11/15/06 12/20/06 1/24/07 2/28/07

εEE

-εNN

B018

microstrain

-0.2

-0.1

0

0.1

0.2

11/15/06 12/20/06 1/24/07 2/28/07

εEN

B018

microstrain

-0.2

-0.1

0

0.1

0.2

11/15/06 12/20/06 1/24/07 2/28/07

(εEE

+εNN

)/2B018

microstrain

Instrument Strains

Areal

Shear

DifferentialExtension

mic

rost

rain

mic

rost

rain

mic

rost

rain

mic

rost

rain

mic

rost

rain

mic

rost

rain

-0.2-0.1

00.10.20.30.4

12/20/06 1/3/07 1/17/07 1/31/07 2/14/07 2/28/07

εEE

-εNN

-0.02

0

0.02

0.04B018

mic

rost

rain

(in

stru

men

t)

microstrain (form

ation)

-0.2-0.1

00.10.2

12/20/06 1/3/07 1/17/07 1/31/07 2/14/07 2/28/07

εEN

-0.02

0

0.02B018

microstrain (form

ation)mic

rost

rain

(in

stru

men

t)

-0.12

-0.08

-0.04

0

0.04

0.08

12/20/06 1/3/07 1/17/07 1/31/07 2/14/07 2/28/07

(εEE

+εNN

)/2

-0.12

-0.08

-0.04

0

0.04

0.08B018

mic

rost

rain

(in

stru

men

t)

microstrain (form

ation)