Ice-RiftSeismicityinAntarctica:AnAnalogforUnderstandingTidallyGeneratedSeismicActivityonEnceladusandEuropaKira G. Olsen1,2, T. A. Hurford2, N.C. Schmerr3, K.M. Brunt2,3

1Howard University, Washington DC, 2 NASA Goddard Space Flight Center, Greenbelt, MD, 3University of Maryland, College Park, MD kira.olsen@nasa.gov

Background & Motivation

Data & Methods

Key Findings

References

Seismicity Likely Tensional Events

Improving Our Understanding of Icy-Ocean Worlds

Future Work

Seismicity Occurs When Rifts are in Tension The icy crustal shells of Enceladus and Europa contain abundant

evidence of brittle-deformational processes in the form of rifts, ridges, and other surface features thought to be formed by tidal processes1.

Brittle deformation within ice, such as fracturing and slip events, will generate seismic waves that contain information about the orientation and magnitude of stress within the ice, timing and location of ice failure, and more2,3.

On-ice seismic data collected by future lander missions to icy-ocean worlds (including the Dragonfly mission to Titan) will enable study of fracturing, stress state, and interior structure4. In preparation for such missions, thorough understanding of the character of ice-generated seismicity on Earth is needed.

Thick ice shelves in Antarctica offer a robust terrestrial analog to icy shells, as ocean tides (rather than wind) supply the dominant stress that actively deforms surface features, including rifts.

Geometry of an Antarctic ice shelf

Grounding Line

Tidal Flexure

Ice Sheet Ice Shelf

100's of km

We identify over 20,000 icequakes within two years of continuous broadband seismic data collected by 9 seismometers operating on the Ross Ice Shelf in West Antarctica from 2014 through 20165.

Prioir to event detection we deconvolve the seismograph frequency-response function from the seismic data to obtain displacement seismograms, and filter the data to 5-20 Hz.

We detect icequakes using an automated short-term average/long-term average approach, coupled with manual detection review. The majority of the icequakes were recorded on only one seismic station.

We determine the surface-wave arrival azimuth for each detected event by analyzing the polarization of Rayleigh waves, using a single-station technique6.

The majority of tidally linked brittle failure appears to occur at rifts. Seismometers located > 10 km from a rift record few icequakes (< 10/day). Seismometers within 10 km of an active rift record abundant seismicity (10-100 icequakes/day).

Icequake timing correlates with falling tide and low temperatures at both rift WR4 and WR6, suggesting that brittle failure preferentially occurs when tensional stresses dominate. These findings are consistent with results from a previousstudy in the region12.

Antarctic ice shelves

Falling Tides

Low Temperatures

0 π/2 π 3π/2 2π

Tidal Phase

0

500

1000

1500

2000

2500

3000

3500

Iceq

uake

Cou

nt

DR14DR13DR15

1

0

1

Ver

tical

Tid

al d

ispl

acem

ent [

m]

A focused deployment of seismometers and GPS instruments later this year to the Ross Ice Shelf, Antarctica will answer outstanding questions about rift deformation.

Multiple seismic stations in close proximity will allow for precise icequake location. Depth extent of seismicity will also be explored.

This upcoming deployment will also be used to investigate the magnitude range of icequakes and implications for stress release along rifts.

Future work will also explore differences between rifts actively undergoing extension and those that are not.

Seismometers close to the rifts record the highest numbers of icequakes when the tide is falling. This relationship suggests that the icequakes are tensional-failure events.

GPS observations collected on either side of Nascent-Iceberg Rift show that the rift widens and contracts ~0.1 m during each tidal cycle10. Similarly measurable displacements likely occur at rifts WR4 and WR6.

Figure: Temperature data from Gill Weather Station. Number of icequakes per day. Amplitude of noise values in a frequency band centered on 5 Hz (3 - 7 Hz); smoothed over seven days.

Figure: Map showing along-flow ice strain rates7 overlain on a 2014 MODIS mosaic image of the Ross Ice Shelf8. Inset shows polarization direction of Rayleigh waves for all icequakes recorded at the three seismic stations closest to a rift. Red triangles show locations of seismometers in XH network. Rift names from Walker et al., 2013.

1Greeberg et al. (2002) Review of Geophysics, 40, 1-34.2Vance et al. (2018) Astrobiology, 18, 37-53.3Panning et al. (2018) JGR: Planets, 123, 163-179.4Stahler et al. (2018) JGR: Planets, 123, 206-2325XH Network, IRIS Data Management Center.6Baker & Stevens (2004) Geophysical Research Letters, 31, L09611.7Alley et al. (2018) Journal of Glaciology, 64(244), 321-332.8Haran et al. (2018), updated 2019. MOA2014, Image Map, Version 1. Date Accessed: 3/10/2020.

The largest number of icequakes per day at the three stations close to the rifts occur at cold temperatures (~ -40˚C and below).

Seismic noise levels in the frequency band used to detect icequakes (5 - 20 Hz) are relatively constant throughout the year, suggesting that the low numbers of icequakes detected per day in the austral summer (October - March) do not result from a detection bias.

Figure: Tidal displacement calculated at the location of DR14 using the OSU TPX09-atlas model of Egbert & Erofeeva, 2002.

Upcoming DeploymentWinter 2020

Calculations of tidal stress on Enceladus show the south polar tiger stripe rifts undergoing alternating periods of tension and compression13.

We will use the seismic catalog produced from this analog study to calculate stress state surrounding the tidally active rifts within the Ross Ice Shelf, and will produce the first observationally constrained model of stress evolution surrounding an interior, ice-locked rift.

This analog work will inform the next generation of tidal models for Enceladus and Europa.

Some rifts with clear surface expressions on the ice shelf appear not to be actively deforming based on seismic and strain-rate observations. This suggests that some mapped features on icy worlds may similarly no longer be actively deforming.

0 5 10 15 20 25 30 35 40Distance to Rift [km]

0

4000

8000

12000

16000

Tot

al N

umbe

r of

Iceq

uake

s

DR14

DR13

DR15

DR07

DR10

RS03DR06

DR05DR04

9Walker et al. (2013) Journal of Geophysical Research: Earth Surface, 118, 2343-2364.10Hurford & Brunt (2014) EPSL, 401, 275-283.11Egbert & Erofeeva (2002) Journal of Atmospheric and Oceanic Technology 19.2, 183-204.12Olinger et al. (2019) Geophysical Research Letters, 46, 6644-6652.13Hurford et al. (2007) Nature, 447, 292-294.

Europa

Enceladus

Figure: (left) Map of ice velocities in Antarctica modified from Rignot et al. (2011). The ice shelves have the highest ice velocities and are seen in purple on the map. (top) Diagram of an ice-shelf environment, not to scale.

100 km

N

DR04

RS03

DR07DR10

DR06

DR05

DR14

DR15

DR13

Gill Weather Station

Rift WR6

Rift WR4

Nascent-Iceberg Rift

Rift WR2

Photo Credit: Scripps Institution of Oceanography Ice Shelf Vibrations Group. Seismic station being installed on the Ross Ice Shelf in 2014. Seismometers were buried beneath ~1 m of firn.

Figure: Example of a detected icequake. (top) Vertical-component seismogram with event onset (red line) and offset (blue line) times identified by the detection algorithm. (bottom) Short-term average/long-term average amplitude values of the seismic trace shown above (short-term window = 2 s, long-term window = 100 s). Red dashed line shows threshold value, above which a detection is recorded.

Seismometer + co-located GPS

GPS

N10 km

MODIS Image

Alexandria sulcus

Cairo sulcus

Baghdad sulcus

Damascus sulcus

90º

270º

0º180º

70º S, 225º W : north–south fault

Apocentre Pericentre

Fraction of orbit from pericentre

1.5

Com

pres

sion

Str

ess

(10

5P

a)

Tens

ion

1

0.5

–1

0.25 0.5 0.75 0.0 0.25–1.5

–0.5

0

0º W

90º W

180º W

270º W

80º S

70º S

Sub-spacecraftlongitude

Predicted for 24 April 2007

50100

Figure: From Hurford et al., 2007. Top left: Stress evolution across a north-south fault. Top right: Cassini image of tiger stripe rifts. Bottom right: Snapshot of calculated stress state, black is compression, green is tension, yellow indicates switch from compression to tension.

Figure: Map of geophysical instrument deployment plan for 2020 field season.

Figure: Number of events at a station vs. station-to-rift distance.DR14

N = 11074

DR13N = 2887

DR15N = 4374