Sites recording the extinction or extirpation of tropical–subtropical and cool–cold temperate rainforest genera during the Plio–Pleistocene aridification of Australia are scattered across the continent, with most preserving only partial records from either the Pliocene or Pleistocene. The highland Lake George basin is unique in accumulating sediment over c. 4 Ma although interpretation of the plant microfossil record is complicated by its size (950 km2), neotectonic activity and fluctuating water levels. A comparison of this and other sites confirms (1) the extinction of rainforest at Lake George was part of the retreat of Nothofagus-gymnosperm communities across Australia during the Plio–Pleistocene; (2) communities of warm- and cool-adapted rainforest genera growing under moderately warm-wet conditions in the Late Pliocene to Early Pleistocene have no modern analogues; (3) the final extirpation of rainforest taxa at Lake George occurred during the Middle Pleistocene; and (4) the role of local wildfires is unresolved although topography, and, elsewhere, possibly edaphic factors allowed temperate rainforest genera to persist long after these taxa became extinct or extirpated at low elevations across much of eastern Australia. Araucaria, which is now restricted to the subtropics–tropics in Australia, appears to have survived into Middle Pleistocene time at Lake George, although the reason remains unclear. <b>Citation:</b> Macphail Mike, Pillans Brad, Hope Geoff, Clark Dan (2020) Extirpations and extinctions: a plant microfossil-based history of the demise of rainforest and wet sclerophyll communities in the Lake George basin, Southern Tablelands of NSW, south-east Australia. <i>Australian Journal of Botany </i>68, 208-228.

<p>The hazard factors in every version of AS 1170.4 since 1993 have been based on a seismic hazard map published in 1991. In this paper I statistically test the validity of that 1991 map. <p>Two methods are used to calculate the hazard for 24+ sites across Australia. Firstly, for each site I calculate how many standard deviations (?1) separate the 1991 hazard value from the calculated PSHA value. Secondly, the magnitude frequency distribution (MFD; i.e. a and b values) is adjusted so that the calculated hazard matches the 1991 hazard value. The number of standard deviations (?2) in the MFD that separate the adjusted MFD differs from the best estimate MFD is subsequently calculated. The first method was applied using four seismic source models (AUS6, DIM-AUS, NSHM13 and these combined), while the second method used NSHM13 only. The average number of standard deviations was calculated from the best 20 of the 24 sites. These statistics are considered a test the validity of the 1991 map. The two methods using five models in total all give similar results. The 1991 map is found, on average, to overestimate the hazard by 3 standard deviations. This suggests that the 1991 map is best described as a 95th+ percentile map. <p>Practitioners using this map, whether for setting building standards or assessing insurance exposure, need to be conscious that the seismic design values are not scientifically valid relative to modern mean probabilistic seismic hazard assessments.

Modern geodetic and seismic monitoring tools are enabling study of moderate-sized earthquake sequences in unprecedented detail. Here we use a variety of methods to examine surface deformation caused by a sequence of earthquakes near Lake Muir in Southwest Western Australia in late 2018. A shallow MW 5.3 earthquake near Lake Muir on the 16th of September 2018 was followed on the 8th of November by a MW 5.2 event in the same region. Focal mechanisms produced for the events suggest reverse and strike slip rupture, respectively. Recent improvements in the coverage and observation frequency of the Sentinel-1 Synthetic Aperture Radar (SAR) satellite in Australia allowed for the timely mapping of the surface deformation field relating to both earthquakes in unprecedented detail. Interferometric Synthetic Aperture Radar (InSAR) analysis of the events suggest that the ruptures are in part spatially coincident. Field mapping, guided by the InSAR results, revealed that the first event produced an approximately 3 km long and up to 0.5 m high west-facing surface rupture, consistent with slip on a moderately east-dipping fault. Double difference hypocentre relocation of aftershocks using data from rapidly deployed seismic instrumentation confirms an easterly dipping rupture plane for the first event. The aftershocks are predominantly located at the northern end of the rupture where the InSAR suggests vertical displacement was greatest. The November event resulted from rupture on a NE-trending strike slip fault. Anecdotal evidence from local residents suggests that the southern part of the September rupture was ‘freshened’ during the November event, consistent with InSAR results, which indicate that a NW-SE trending structural element accommodated deformation during both events. Comparison of the InSAR-derived deformation field with surface mapping and UAV-derived digital terrain models (corrected to pre-event LiDAR) revealed a surface deformation envelope consistent with the InSAR for the first event, but could not discern deformation unique to the second event.

<div>In mid-2022 two paleoseismic trenches were excavated across the Willunga Fault at Sellicks Hill, ~40 km south of Adelaide, at a location where range front faulting displaces a thick colluvial apron, and flexure in the hanging wall has produced an extensional graben. Vertical separation between time-equivalent surfaces within the Willunga Embayment and uplifted Myponga Basin indicate an average uplift rate of 40 m/Myr since 5 Ma across the Willunga fault at the trench location, equivalent to a slip rate of 57 m/Myr across a 45° dipping fault. </div><div> The field sites preserve evidence for at least 4-5 large earthquake events involving approximately 6.9 m of discrete slip on fault planes since the Mid to Late Pleistocene. If the formation of red soil marker horizons in the trenches are assumed to relate to glacial climatic conditions then a slip-rate of 26-46 m/Myr since the Mid Pleistocene is obtained. These deformation rate estimates do not include folding in the hanging wall of the fault, which is likely to be significant at this site as evidenced by the existence of a pronounced hanging wall anticline. In the coming months, the results of dating analysis will allow quantitative constraint to be placed on earthquake timing and slip-rate, and a structural geological study seeks to assess the proportion of deformation partitioned into folding of the hanging wall.</div><div> The 2022 trenches represent the most recent of ten excavated across this fault. Integration of the 2022 data with those from previous investigations will allow fundamental questions to be addressed, such as whether the Willunga fault ruptures to its entire length, or in a segmented fashion, and whether any segmentation behaviour is reflected in local slip-rate estimates. Thereby we hope to significantly improve our understanding of the hazard that this, and other proximal Quaternary-active faults, pose to the greater Adelaide conurbation and its attendant infrastructure.</div> This paper was presented to the 2022 Australian Earthquake Engineering Society (AEES) Conference 24-25 November (https://aees.org.au/aees-conference-2022/)

<div>This Geoscience Australia Record contains technical data and input files that, when used with the Global Earthquake Model’s (GEM’s) <em>OpenQuake-engine</em> probabilistic seismic hazard analysis software (Pagani<em> et al.</em>, 2023), will enable end users to explore and reproduce the 2023 National Seismic Hazard Assessment (NSHA23) of Australia (Allen<em> et al.</em>, 2023b). Output data, as calculated by Geoscience Australia using Version 3.16.1 of the <em>OpenQuake-engine</em>, are also provided. This report describes the NSHA23 input and output data only and does not discuss the scientific rationale behind the model development or the development of the NSHA23 earthquake catalogue. These details are provided in Allen<em> et al.</em> (2023b) and (Allen<em> et al.</em>, 2024), and respective references therein. The NSHA23 provides estimates of seismic hazard for the six Australian states and two mainland territories. However, it does not provide updated hazard factors for Australia’s Antarctic and other offshore territories (e.g., Christmas Island, Cocos Island, Heard Island, Lord Howe Island, Macquarie Island and Norfolk Island).</div>

One of the key challenges in assessing earthquake hazard in Australia is understanding the attenuation of ground-motion through the stable continental crust. There are now a handful of ground-motion models (GMMs) that have been developed specifically to estimate ground-motions from Australian earthquakes. These GMMs, in addition to models developed outside Australia, are considered in the 2018 National Seismic Hazard Assessment (NSHA18; Allen et al., 2017). In order to assess the suitability of candidate GMMs for use in the Australian context, ground-motion data forom small-to-moderate Australian earthquakes have been gathered. Both qualitative and quantitative ranking techniques (e.g., Scherbaum et al., 2009) have been applied to determine the suitability of candidate GMMs for use in the NSHA18. This report provides a summary of these ranking techniques and provides a discussion on the utility of these methods for use in seismic hazard assessments in Australia; in particular for the NSHA18. The information supplied herein was provided to participants of the Ground-Motion Characterisation Expert Elicitation workshop, held at Geoscience Australia on 9 March 2017 (Griffin et al., 2018).

People in Australia are surprised to learn that hundreds of earthquakes occur below our feet every year. The majority are too small to feel, let alone cause any damage. Despite this, we are not immune to large earthquakes.

In plate boundary regions moderate to large earthquakes are often sufficiently frequent that fundamental seismic parameters such as the recurrence intervals of large earthquakes and maximum credible earthquake (Mmax) can be estimated with some degree of confidence. The same is not true for the Stable Continental Regions (SCRs) of the world. Large earthquakes are so infrequent that the data distributions upon which recurrence and Mmax estimates are based are heavily skewed towards magnitudes below Mw5.0, and so require significant extrapolation up to magnitudes for which the most damaging ground-shaking might be expected. The rarity of validating evidence from surface rupturing palaeo-earthquakes typically limits the confidence with which these extrapolated statistical parameters may be applied. Herein we present a new earthquake catalogue containing, in addition to the historic record of seismicity, 150 palaeo-earthquakes derived from 60 palaeo-earthquake features spanning the last > 100 ka of the history of the Precambrian shield and fringing extended margin of southwest Western Australia. From this combined dataset we show that Mmax in non-extended-SCR is M7.25 ± 0.1 and in extended-SCR is M7.65 ± 0.1. We also demonstrate that in the 230,000 km2 area of non-extended-SCR crust, the rate of seismic activity required to build these scarps is one tenth of the contemporary seismicity in the area, consistent with episodic or clustered models describing SCR earthquake recurrence. A dominance in the landscape of earthquake scarps reflecting multiple events suggests that the largest earthquakes are likely to occur on pre-existing faults. We expect these results might apply to most areas of non-extended-SCR worldwide.

Instrumentally observed earthquakes sequences typically show clusters of earthquakes interspersed with periods of quiescence. These ‘bursty’ sequences also have correlated inter-event times (‘long-term memory’). In contrast, elastic rebound theory forms the basis of the standard earthquake cycle model, and predicts large earthquakes to occur regularly through cycles of strain accumulation and release (periodicity). In this model the conditional probability of future large earthquakes is reduced immediately following fault rupture, and inter-event times are independent. Here we use the burstiness and memory coefficient metrics to characterize more than 100 long-term earthquake records. We find that large earthquake occurrence on the majority of Earth’s faults is weakly periodic and does not exhibit long-term memory; earthquakes occur more regularly than a random Poisson process although inter-event times are variable. In contrast, clustering occurs in slowly deforming regions (annual rates < 2 x 10-4), and is not explained by elastic rebound theory. <b>Citation:</b> Griffin, J. D., Stirling, M. W., & Wang, T. (2020). Periodicity and clustering in the long‐term earthquake record. <i>Geophysical Research Letters</i>, 47, e2020GL089272. https://doi.org/10.1029/2020GL089272

<div>This paper explores several area-based tests of long-term seismic hazard forecasts for the Australian continent. Using the observed seismicity, ShakeMaps are calculated for earthquakes that are expected to have generated moderate-to-high levels of ground shaking within continental Australia in the past 50 years (January 1972 through December 2021). A “composite ShakeMap” is generated that extracts the maximum peak ground acceleration “observed” in this 50-year period for any site within the continent. The fractional exceedance area of this composite map is compared with four generations of Australian seismic hazard maps for a 10% probability of exceedance in 50 years (~1/500 annual exceedance probability) developed since 1990. </div><div>In general, all these models appear to forecast higher seismic hazard relative to the ground motions that are estimated to have occurred in the last 50 years, with the most recent hazard model yielding a fractional exceedance area most similar to the target 1/500 annual exceedance probability. The sensitivity of these results to various modeling assumptions was tested by exploring an alternative ground-motion characterization model that forecasts higher overall ground-shaking intensities.&nbsp;The sensitivity of these results is also tested with the interjection of a rare scenario earthquake with an expected regional recurrence of approximately 8,700 years. While these area-based analyses do not provide a robust assessment of the performance of the candidate seismic hazard for any specific location given the limited independent data, they do provide—to the first order—a guide to the performance of the respective maps at a continental scale</div> <b>Citation:</b> Allen TI, Ghasemi H, Griffin JD. Exploring Australian hazard map exceedance using an Atlas of historical ShakeMaps. <i>Earthquake Spectra</i>. 2023;39(2):985-1006. doi:10.1177/87552930231151977