Surface rupture and distributed deformation revealed by optical satellite imagery: The intraplate 2016 Mw 6.0 Petermann Ranges earthquake, Australia

Ryan Gold¹, Dan Clark², William Barnhart³, Tamarah King⁴, Mark Quigley⁴, Richard Briggs¹ (¹USGS; ²Geoscience Australia; ³Department of Earth and Environmental Sciences, University of Iowa; ⁴School of Earth Sciences, University of Melbourne)

Geophysical Research Letters, 2019

Background:

As described in a previous post, we got out to the 2016 Petermann surface rupture within ~ 8 days. The only data at that point was a magnitude, some aftershocks, a focal mechanism, an ever-moving epicentre (seismologists were still refining their locations), and some cracking along a dirt road observed by Dan Clark (Geoscience Australia) during deployment of temporary seismometers. After only a few days we were forced out of the field due to rain, but thanks to a 20km hike in the wrong direction, we knew the fault was NE dipping and located at the cracking observed by Dan.

Very low rupture highlighted by shadows

Cracking along the edge of the dirt-track

Examples of the road cracking in question

As we sat waiting for the roads to dry out so we could finally go and look for the rupture, colleagues in the USA and at Geoscience Australia were busy analyzing post-earthquake satellite data. InSAR across the Petermann earthquake location became available, and showed a clear ~21 km long fault rupture at the surface, with up to 1 m of offset in the direction of the satellite. This was the first new information to reach us, and supported our field observations.

InSAR interferogram from Geoscience Australia

We finally returned to the field and started mapping the surface rupture on foot and with a drone. We returned to internet coverage to proudly send off the ~ 5km of rupture we’d mapped, and found that Ryan Gold at the USGS had mapped another 5 km of rupture, from the comfort of his office in the USA using high-resolution satellite imagery. Ryan continued to map the surface rupture remotely while we travelled on to Kununurra for more field work. Meanwhile, Dan Clark arrived at the Petermann rupture and mapped the whole 21 km length with a high-resolution GPS unit (in just two days)! He also captured helicopter imagery along most of the rupture.

Figure from the supplementary material of the paper (https://doi.org/10.1029/2019GL084926)

We’d all been in contact sharing observations and measurements, and by the end of June 2016 we had RTK GPS measurements, field photographs, drone imagery across 5 km of rupture, helicopter imagery across most of the rupture, high-resolution satellite imagery, and InSAR offset of the rupture.

Over the next 12 months, Ryan worked with Bill Barnhart to process the optical satellite imagery into pre- and post- earthquake elevation data. By differencing these elevation maps, they produced an estimate of offset across the rupture area. These data provided a third estimate of offset to compare to the InSAR derived offset, and Dan’s GPS measured field-offset. Ryan compared these offset estimates, working the results into a paper.

The paper went to review in late 2017, and was bounced back and forth between reviewers, editors and us until August 2019 when it was accepted for publication. Initially, I was invited to co-author this paper mostly because I had shared some field photographs and observations, which, at the time, hardly seemed worth a co-authorship to me. However, watching the other co-authors’ ideas get put forward, refuted, revise and incorporated, gave me the confidence to put my own ideas forward. Eventually after a few rounds of revision, I’d contributed enough ideas and observations that I felt much more confident in my co-authorship. The whole experience was a great lesson in having confidence, and how the collaboration and publication process might play out.

Key points of the paper:

• The general shape and pattern of the along-rupture offsets are comparable between all three methods, i.e. they all find maximums and minimums in the same locations
• The magnitude of vertical offset recorded by the satellite imagery DEM (specifically ICP analysis) is much higher than the vertical offset recorded in the field

• This is because the remote-sensing method captures a proportion of distributed deformation while the field measurements were predominantly spot heights taken at the rupture tip (the fault plane where it breaks the surface)
• This distributed deformation is heavily asymmetric, with the hanging-wall having a far greater amount of slip than the footwall, as is typical for a reverse/thrust earthquake

• The width of this deformation zone (e.g. the fault-zone width) also changes along strike, getting narrower and broader potentially due to fault strike and subsurface geometry
• The deformation is interpreted to result from slip being distributed into a wider zone as rupture approaches the surface, potentially due to mechanical differences due to depth or fault properties
• These results support the observation that geomorphic offset markers measured at 10s to 100s of meters across reverse faults often record higher slip estimates than the slip observed in trenches.
• Remote sensing of Petermann offsets show that these geomorphic patterns of reverse faulting may develop due to cumulative co-seismic distributed deformation patterns

The Petermann surface rupture seen by helicopter (photo credit: Dan Clark, Geoscience Australia)