The Victorian Earthquake Hazard Map


This report summarises the development of a new Probabilistic Seismic Hazard Analysis (PSHA) for Victoria called the Victorian Earthquake Hazard Map (VEHM). PSHA provides forecasts of the strength of shaking in any given time (return period). The primary inputs are historical seismicity catalogues, paleoseismic (active fault) data, and ground-motion prediction equations.

A key component in the development of the Victorian Earthquake Hazard Map was the integration of new geophysics data derived from deployments of Australian Geophysical Observing System seismometers in Victoria with a variety of publicly available datasets including seismicity catalogues, geophysical imagery and geological mapping. This has resulted in the development of a new dataset that constrains the models presented in the VEHM and is also is provided as a stand-alone resource for both reference and future analysis.

The VEHM provides a Victorian-focussed earthquake hazard estimation tool that offers an alternative to the nationally focussed 2012 Australian Earthquake Hazard Map (Burbidge 2012). The major difference between the two maps is the inclusion of active fault location and slip estimates in the VEHM.

There is a significant difference in hazard estimation between the two maps (even without including fault-related seismicity) due primarily to differences in seismicity-analysis. These issues are described in the discussion section of this report, again resulting in a higher fidelity result in the VEHM. These differences make the VEHM a more conservative hazard model.

The VEHM currently exists as a series of online resources to help assist those in engineering, planning, disaster management. This is a dynamic dataset and the inputs will continue to be refined as new constraints are included and the map is made compatible with the Global Earthquake Model (GEM) software, due for release in late 2014.

The VEHM was funded through the Natural Disaster Resilience Grants Scheme. The NDRGS is a grant program funded by the Commonwealth Attorney-General’s Department under the National Partnership Agreement on Natural Disaster Resilience signed by the Prime Minister and Premier. The purpose of the National Partnership Agreement is to contribute towards implementation of the National Strategy for Disaster Resilience, supporting projects leading to the following outcomes:

  1. reduced risk from the impact of disasters and

  2. appropriate emergency management, including volunteer, capability and capacity consistent with the State’s risk profile.

Earthquakes, hazard and damage in Australia

Effects of Earthquakes

The effects of earthquake shaking depend on the amplitude of the motion, the frequency content, and the duration. Amplitude is determined by magnitude and distance. Frequency content is determined by the earthquake magnitude and stress change, with small earthquakes giving dominantly high frequency motion, and increasing magnitudes give an increasing proportion of energy at decreasing frequencies (longer periods). Duration of earthquake motion is determined by the earthquake magnitude, and is comparable with the rupture duration (less than 1 second for magnitudes less than 5.0 and greater than 10 seconds for magnitudes larger than 7.0).

Earthquakes in Australia

Australian earthquakes occur in a ‘stable’ continental region (SCR), so are infrequent compared to plate boundary settings. In a typical region an event is only felt on average each 5 to 10 years. The whole continent experiences about 600 recorded events each year, with 2 events of M \(>\) 5 (Leonard 2008). Earthquakes in all regions of Australia are distributed over many faults, but with few longer than 100 km. It follows that there is relatively low maximum magnitude, probably Mw 7.2 to 7.5, limited by the thickness of the seismic zone and the length of active faults.

The record of seismicity in continental Australia is heterogeneous. A number of distinct zones of seismicity have been defined across the Australian continent. One of these, known as the South Eastern Seismic zone, corresponds broadly with the southern part of the Eastern Highlands, extending into southwest Gippsland. Compared to other areas of Australia, seismicity in this region has been consistently elevated in the previous decade, and seems to be controlled by the arrangement of dense, highly interlinked fault networks with typically short fault lengths.

Almost all Australian earthquakes are in the upper crust, from the surface to a depth of about 20 km. Moderate magnitudes can cause damage, such as Newcastle, 1989, ML 5.6, which caused about A $3 billion damage. While the orientation of the stress field in Australia is well constrained, variations in its magnitude are not as well understood. Stress is almost always horizontal compression, and reverse (thrust) faults therefore predominate. Ruptures tend to have high stress drop, giving high frequency, high acceleration and short duration motion.

Engineering and Hazard

Earthquake effects on structures and people are minimised by building to an earthquake code. In practice, this means buildings that will not collapse, even if they are badly damaged by the earthquake. Building codes use risk criteria, which usually specify the average return period of earthquake ground motion that should not interrupt the operation of the structure, and the longer period that should not cause collapse of the structure.

Many earthquake building codes, such as the Australian Earthquake Loading Code AS1170.4-2007, adopt the 500-year earthquake as the criterion. This means that a building designed to last 100 years will have a 20% chance that its design motion will be exceeded during its lifetime. This does not matter very much in active areas on tectonic plate boundaries where the 500-year earthquake is almost as large as the largest credible earthquake. However, in relatively stable continental regions such as Australia, the 500-year earthquake is quite small, and will give much lower level of motion than an earthquake with a magnitude that will recur at intervals of thousands of years. Whether the traditional PSHA approach described below, is adequate to meet these large, infrequent events is not well understood.

The 2012 Australian Earthquake Hazard Map has been recommended as a replacement to the current earthquake loading code AS1170.4-2007.

Probabilistic Seismic Hazard Analysis

Probabilistic Seismic Hazard overview

Probabilistic Seismic Hazard Analyses (PSHA) (Cornell 1968, McGuire 1995) produce estimates of the probability of exceeding various levels of ground motion (intensity measures) for a given location and time interval. The primary aims are to produce seismic hazard curves—usually drawn with the annual rates of exceedance on the vertical and increasing values of intensity measure (e.g. values of PGA) on the horizontal. Alternatively, time intervals of interest are selected (e.g. 500 years) in which case a map of expected values of the intensity measure can be made.

The standard PSH methodology estimates hazard by summing the contributions from all potentially damaging earthquakes in the region, the primary steps are as follows, e.g. (Baker 2008):

  1. Define all earthquake sources capable of producing significant events.

  2. Characterize the earthquake frequency-magnitude relationship (the rates at which earthquakes of various magnitudes are expected to occur).

  3. Characterize the source-site distance distribution.

  4. Compute the ground motion (distribution) as a function of earthquake magnitude, distance; i.e. the conditional probability.

  5. Combine all probabilities (and uncertainties) using the total probability theorem.

The application of traditional PSHA approaches requires estimation of the properties of the seismic source zones, which are often determined by subjective judgments that may be different in various studies. In low-seismicity areas such as Australia, the earthquake occurrence is often modelled as a spatio-temporal Poisson process, i.e. earthquakes are memoryless and spatially random, with rates that are completely described by the Gutenberg-Richter relationship (no ‘characteristic earthquakes’). The seismicity rates and frequency-magnitude distribution (b-values) are calculated from historical seismicity or from geologically derived fault slip rates (more commonly the former). The maximum credible earthquake (\(M_{max}\)) is imposed given empirical deductions from past earthquakes and estimated maximum fault length. Given that SCRs tend to have highly clustered seismicity, it is common to subdivide the region into sub regions know as area source zones, in which it is supposed that the spatial-Poisson conditions are approximately valid.

In the last few decades, fault slip rates have increasingly been used to constrain earthquake recurrence relationships and inform hazard maps (Pace 2006). In Australia, this technique has been incorporated into a couple of previous studies (Brown 2004, Somerville 2008) . It is particularly useful in regions like Australia, however, where seismicity is relatively infrequent, historical records were likely derived from sparse networks and there are numerous geological features present that indicate recent seismic activity and that can be dated using a variety of techniques. Inclusion of fault sources represents a key point of difference between the current study and the latest national hazard map by Geoscience Australia.

Recently, PSHA has also come under criticism from a number of scientists (Klügel 2012, Stein 2012). Criticisms range from a lack of model testing, failures of the prescribed maximum credible magnitude in past PSHA (e.g. in the case of Tohoku) to the assumption of Poisson statistics, and even inherent problems with energy conserva