Surface-rupturing historical earthquakes in Australia and their environmental effects: new insights from re-analyses of observational data

 

Tamarah R. King¹, Mark Quigley¹, Dan Clark² (¹School of Earth Sciences, University of Melbourne; ²Geoscience Australia)

Geosciences, 2019 

Background

From May – July 2019 I didn’t really see my friends, I postponed my birthday, and I barely spoke to family. I was completely focused on writing a review paper of historic Australian surface rupturing earthquakes.

I’d spent years coming back to these eleven surface ruptures in QGIS wondering why they didn’t seem to resemble ‘faults’ as I thought they should (i.e. planar structures). At some point I’d digitised historic maps of the ruptures and some of the historic offset data, with the idea of trying to find answers to why they had such odd shaped surface ruptures. But my research was focused on other things, so I put these ideas and questions aside. Then in May 2019, My supervisor Mark was invited to contribute to a special edition on environmental effects, and suggested I could collate data from Australian surface ruptures. The deadline was two months away. Seemed easy?

In those two months, I systematically found and read every journal article, government report, conference abstract, newspaper article, geological map, and blog post on these surface ruptures. I digitised every map, every table of data, and every graph I could find. I poured over every photograph to see if I could identify previously unreported environmental effects. I collated available seismological data and observation. I found geological maps, geophysical maps, seismic lines, boreholes, drill holes, and well holes. I noted patterns between ruptures, similarities, differences. My head was devoid of anything but surface rupture data for two months.

Me, June 2019

By the end of June, I’d almost finished collating the data but I hadn’t started writing the paper, and it was due. We got an extension, I started trying to tie everything into a single paper. It didn’t work. So instead of writing the paper, I wrote reports on each earthquake. This took…a while. The final seven reports are approximately 30,000 words total. My co-authors were overwhelmed to say the least! But they persevered. They proof-read every report, and fixed them up as if they were individual manuscripts. Then, Mark and I sat down to decide what the final paper would look like.

I forced 30,000 words into 8 tables: summary data, literature, regional geology, geophysics and basement, seismology, surface measurements, environmental effects, and paleoseismology. I summarised these tables in the text, and wrote a discussion highlighting some key take-aways. I submitted the paper, and published the reports on EarthArxiv to serve as stand-alone records of each event, in the hope of saving anybody from having to do what I’d done again.

50 years of field work, and tens of thousands of words of data, in a single table

This paper overwhelmed my life for a few months. But by the end I’d synthesised the seismology, geology, geophysics, surface observations, sub-surface extrapolations, implications for seismic hazard models, and tectonophysics of these ruptures. While I wouldn’t recommend this experience to anybody, it’s nice to have satisfied my scientific curiosity.

Key points

• All of these earthquakes initiated on, and propagated along, pre-existing bedrock structures
• The structures were well aligned for reverse rupture relative to the direction of maximum compressive stress in the crust

• There is evidence of absence for prior earthquakes on many of these faults, implying a lack of recurrence OR incredibly long recurrence
• In the absence of recurrence, erosion-rates provide a proxy measure for scarp residence time in the landscape, which in turn provides a proxy measure of slip-rate

• Scientists are fallible. I found multiple examples where surface measurements were incorrectly transcribed, or focal mechanisms were incorrectly interpreted, or other minor errors.
• These errors have consequences. Incorrect measurements are built into scaling relationships, fault mechanism are categorised incorrectly in global databases, erroneous slip-rates feed into global estimates for intraplate faulting. None of these are catastrophic life-changing issues, but they’re still problems that we should be conscious of when collating databases.

• Australian surface ruptures are poorly described by many scaling relationships

• Australian Precambrian crust has a propensity towards very shallow moderate magnitude earthquakes (less than 7 km hypocentre depth) which increases the likelihood of surface rupture (and shaking intensity)
• In this paper I derive new net-slip measurements based on field measurements of vertical and lateral offset, and my best estimates of along-rupture dip (from field measurements, focal mechanisms, and fault inversions)

Honestly, these are just a handful of key-points. It’s 30,000 words condensed into 26 pages after all! The paper’s open access, go read it if you want more.

The follow up to this work comes from Haibin Yang, another PhD student at the University of Melbourne who took my newly derived net-slip measurements and ran some really cool analyses on them. I want to finish by sharing that in the process of analysing my net-slip values, Haibin identified multiple calculation errors in my net-slip values (specifically for the Cadoux rupture). I point this out to show that while I found many errors in other people’s work, my own work is not immune! I was quite defensive when Haibin pointed out these errors. But, I quickly realised that mistakes happen, and the only way forward was to fix them, and learn from them. Haibin’s work is based off updated, and thoroughly double-checked, net-slip values.