Seismic hazard modelling is a multi-disciplinary science that aims to forecast earthquake occurrence and its resultant ground shaking. Such seismic hazard models consist of a probabilistic framework that models the flow of uncertainty across a complex system; typically, this includes at least two model-components developed from earth science: seismic source models, and ground motion prediction models. Although there is no scientific prescription for the length of the forecasting time-window, the most common probabilistic seismic hazard analyses (PSHA hereafter) consider forecasting probabilities of ground shaking in time windows of 30 to 50 years. These types of models are the target of this review paper. Although the core methods and assumptions of such a modelling have largely remained unchanged since they were first developed more than 50 years ago, we will review the most recent initiatives which are facing the difficult task of meeting both the increasingly sophisticated demands of society and keeping pace with advances in our scientific understanding. A need for more accurate and precise hazard forecasting must be balanced with increased quantification of uncertainty and new challenges such as moving from time-independent hazard to forecasts that are time-dependent and specific to the time-period of interest. Meeting these challenges requires the development of science-driven models which integrate at best all information available, the adoption of proper mathematical frameworks to quantify the different types of uncertainties in the source and ground motion components of the hazard model, and the development of a proper testing phase of the hazard model to quantify the consistency and skill of the hazard model. We review the state-of-the-art of the national seismic hazard modeling, and how the most innovative approaches try to address future challenges.

As part of the 2018 National Seismic Hazard Assessment (NSHA), we compiled the geographic information system (GIS) dataset to enable end-users to view and interrogate the NSHA18 outputs on a spatially enabled platform. It is intended to ensure the NSHA18 outputs are openly available, discoverable and accessible to both internal and external users. This geospatial product is derived from the dataset generated through the development of the NSHA18 and contains uniform probability hazard maps for a 10% and 2% chance of exceedance in 50 years. These maps are calculated for peak ground acceleration (PGA) and a range of response spectral periods, Sa(T), for T = 0.1, 0.2, 0.3, 0.5, 1.0, 2.0 and 4.0 s. Additionally, hazard curves for each ground-motion intensity measure as well as uniform hazard spectra at the nominated exceedance probabilities are calculated for key localities.

Seismic hazard models, commonly produced through probabilistic seismic hazard analysis, are used to establish earthquake loading requirements for the built environment. However, there is considerable uncertainty in developing seismic hazard models, which require assumptions on seismicity rates and ground-motion models (GMMs) based on the best evidence available to hazard analysts. This paper explores several area-based tests of long-term seismic hazard forecasts for the Australian continent. ShakeMaps are calculated for all earthquakes of MW 4.25 and greater within approximately 200 km of the Australian coastline using the observed seismicity in the past 50 years (1970-2019). 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. In general, all these seismic hazard models appear to be conservative relative to the observed ground motions that are estimated to have occurred in the last 50 years. To explore aspects of possible prejudice in this study, the variability in ground-motion exceedance was explored using the Next Generation Attenuation-East GMMs developed for the central and eastern United States. The sensitivity of these results is also tested with the interjection of a rare scenario earthquake with an expected regional recurrence of approximately 5,000 - 10,000 years. While these analyses do not provide a robust assessment of the performance of the candidate seismic hazard for any given location, they do provide—to the first order—a guide to the performance of the respective maps at a continental scale. This paper was presented at the Australian Earthquake Engineering Society 2021 Virtual Conference, Nov 25 – 26.

Public concerns have been raised about the potential for induced seismicity as state and territory governments lift moratoriums on hydraulic stimulation activities for the exploration and extraction of unconventional hydrocarbons. The Scientific Inquiry into Hydraulic Fracturing in the Northern Territory articulated the need for a traffic-light system “to minimise the risk of occurrence of seismic events during hydraulic fracturing operations” within the Beetaloo Sub-basin. A temporary seismic network (Phase 1) was deployed in late 2019 to monitor baseline seismic activity in the basin. Based on the data analysed herein (November 2019 – April 2021), no seismic events were identified within the area of interest suggesting that the Beetaloo Sub-basin is largely aseismic. Observations to date indicate that there is potential to identify events smaller than ML=1.5 within the basin. The recent installation of ten semi-permanent stations for continuous real-time monitoring will contribute to ongoing baseline monitoring efforts and support the implementation of an induced seismicity traffic-light system. The outcome of this study will be used to build knowledge about potential human-induced seismic activity in the region that may be associated with unconventional hydrocarbon recovery. This paper was presented at the Australian Earthquake Engineering Society 2021 Virtual Conference, Nov 25 – 26.

Here we undertake a statistical analysis of local magnitudes (ML) calculated using the two real-time earthquake monitoring software platforms use by Geoscience Australia (GA) since 2005, Antelope and Seiscomp. We examine a database of just over 10 years duration, during a period in which both systems were in operation and over 4000 earthquakes were located and magnitudes estimated. We examine the consistency of both single-station and network ML estimates of both systems, with a view toward determining guidelines for combining them into a single catalogue, as well as for determining best practice in the for the estimation of local magnitudes for regions of sparse seismic networks. Once this guidance has been developed, it is the intention of GA to re-process magnitudes for all earthquakes using a consistent approach where digital data are available and can be integrated within the currently-used SeisComP system. This paper was presented at the Australian Earthquake Engineering Society 2021 Virtual Conference, Nov 25 – 26.

The 22 September 2021 (AEST) moment magnitude MW 5.9 Woods Point earthquake was the largest in the state of Victoria’s recorded history. The ground motions were felt throughout the state of Victoria and into neighbouring states New South Wales and South Australia. Minor damage was reported in the city of Melbourne and in some regional towns close to the epicentre. This event was captured on many high-quality recorders from multiple sources, including private, university, and public stations. These recordings provide a rare opportunity to test the validity of some ground motion models thought to be applicable to the southeast region of Australia. This paper presents spectral acceleration and attenuation comparisons of the Woods Point earthquake event to some ground motion models. The results of this paper provide further evidence that the attenuation characteristics of southeastern Australia may be similar to that in central and eastern United States, particularly at shorter distances to the epicentre. This paper was presented at the Australian Earthquake Engineering Society 2021 Virtual Conference, Nov 25 – 26.

The Mwp 6.1 Petermann Ranges earthquake that occurred on 20 May, 2016 in the Central Ranges, NT, is the largest onshore earthquake to be recorded in Australia since the 1988 Tennant Creek sequence. While geodetic and geophysical analyses have characterized the extent of surface rupture and faulting mechanism respectively, a comprehensive aftershock characterization has yet to be performed. Data has been acquired from a 12-station temporary seismic network deployed jointly by the ANU and Geoscience Australia (GA), collected from five days following the mainshock to early October. Taking advantage of enhanced automatic detection techniques using the SeisComP3 real-time earthquake monitoring software within the National Earthquake Alerts Centre (NEAC) at GA, we have developed a comprehensive earthquake catalogue for this mainshock-aftershock sequence. Utilising the NonLinLoc location algorithm combined with a Tennant Creek-derived velocity model, we have preliminarily located over 5,800 aftershocks. With additional spatio-temporal analyses and event relocation, our objective will be to use these aftershocks to help delineate the geometry of the headwall rupture along the Woodroffe Thrust. These high-resolution aftershock detection techniques are intended to be implemented in real-time within the NEAC following future significant Australian intraplate earthquakes. This paper was presented at the Australian Earthquake Engineering Society 2021 Virtual Conference, Nov 25 – 26.

You may not realise it but, on average, Australia is rattled every few days by an earthquake of magnitude 3 or above. We don’t feel every small tremor that happens, but the larger earthquakes are powerful enough to cause serious damage to buildings and infrastructure, putting our community’s safety at risk.

Damaging earthquakes in Australia and other regions characterised by low seismicity are considered low probability but high consequence events. Uncertainties in modelling earthquake occurrence rates and ground motions for damaging earthquakes in these regions pose unique challenges to forecasting seismic hazard, including the use of this information as a reliable benchmark to improve seismic safety within our communities. Key challenges for assessing seismic hazards in these regions are explored, including: the completeness and continuity of earthquake catalogues; the identification and characterisation of neotectonic faults; the difficulties in characterising earthquake ground motions; the uncertainties in earthquake source modelling, and the use of modern earthquake hazard information to support the development of future building provisions. Geoscience Australia recently released its 2018 National Seismic Hazard Assessment (NSHA18). Results from the NSHA18 indicate significantly lower seismic hazard across almost all Australian localities at the 1/500 annual exceedance probability level relative to the factors adopted for the current Australian Standard AS1170.4–2007 (R2018). These new hazard estimates have challenged notions of seismic hazard in Australia in terms of the recurrence of damaging ground motions. Consequently, this raises the question of whether current practices in probabilistic seismic hazard analysis (PSHA) deliver the outcomes required to protect communities and infrastructure assets in low-seismicity regions, such as Australia. This manuscript explores a range of measures that could be undertaken to update and modernise the Australian earthquake loading standard, in light of these modern seismic hazard estimates, including the use of alternate ground-motion exceedance probabilities for assigning seismic demands for ordinary-use structures. The estimation of seismic hazard at any location is an uncertain science, particularly in low-seismicity regions. However, as our knowledge of the physical characteristics of earthquakes improve, our estimates of the hazard will converge more closely to the actual – but unknowable – (time independent) hazard. Understanding the uncertainties in the estimation of seismic hazard is also of key importance, and new software and approaches allow hazard modellers to better understand and quantify this uncertainty. It is therefore prudent to regularly update the estimates of the seismic demands in our building codes using the best available evidence-based methods and models.

This paper explores the implementation of the Natural Resources Canada’s 5th Generation national seismic hazard model as developed for the National Building Code of Canada (NBCC), within the OpenQuake-engine. It also describes the reconciliation of the differences in hazard estimates relative to the published NBCC values, calculated using GSCFRISK. Source and ground-motion input models developed for the GSCFRISK software were translated to the OpenQuake-engine format for the hazard comparison. In order to successfully undertake this process, several adjustments to the OpenQuake code were needed to mimic the behavior of GSCFRISK. This required the development of new functions for earthquake rupture scaling and ground-motion interpolation. Hazard values estimated using the OpenQuake-engine are generally in good agreement with the 2015 NBCC national-scale hazard values, with differences less than 2-3% typically achieved. Where larger differences arise, they can be rationalized in terms of differences between the behaviours of the two software engines with respect to earthquake rupture length uncertainty and maximum ground-motion integration distance.