AIR Currents

December 09, 2009

December 09, 2009

Editor's Note: Twenty years ago this month, the deadliest earthquake on Australian soil occurred near Newcastle, an industrial city on Australia's east coast. The moderate-magnitude earthquake caused extensive property damage and led to the largest insurance loss1 (trended to current dollars) from a natural catastrophe in Australian history. AIR senior scientist Dr. Khosrow Shabestari and senior engineer Dr. Peeranan Towashiraporn describe the significance of the event and consider Australia's earthquake risk.

The 1989 Earthquake

On December 28, 1989, just three days after Christmas, a local news station in Australia, ABC Newcastle, interrupted its mid-morning programming to announce that there had been an explosion in the studio. Minutes later, the source of the "explosion" became evident. At 10:28 a.m. local time, a magnitude 5.6 earthquake had struck near the town of Boolaroo, 15 kilometers west of the Newcastle central business district and about 140 kilometers northeast of Sydney. Chaos erupted as bewildered workers from downtown office buildings poured into rubble-strewn streets and severed roads caused miles-long traffic jams of cars and trucks.

Despite the relatively modest magnitude, the effects were widespread; damage to buildings, facilities and infrastructure was reported over an area of 9,000 km2. The central business district in Newcastle and the nearby suburb of Hamilton suffered extensive damage. One of the hardest hit locations was the Newcastle Workers Club where a 300-ton brick retaining wall and a concrete mezzanine floor slab collapsed onto the ground-floor parking garage, causing nine fatalities and injuring dozens more. West of the city center, another three people were killed when porticos on Hamilton's Kent Hotel collapsed. Buildings as far away as Sydney suffered minor damage. In total, the event damaged more than 60,000 buildings, claimed the lives of 13 people, injured 160, and generated AUD 862 million in insured losses (1989 AUD).2

Prior to the Newcastle Earthquake, it was widely assumed that earthquakes posed little threat to urban communities in Australia and that moderate-sized earthquakes would not lead to loss of life or significant physical damage. Unfortunately, that assumption proved to be misguided.

Cat Bond Figure 1 Event 1
Figure 1. The Newcastle Workers Club suffered extensive damage (Source: City of Newcastle)
Dr. Khosrow Shabestari and Dr. Peeranan Towashiraporn

By: Dr. Khosrow Shabestari and Dr. Peeranan Towashiraporn

Seismic Hazard in Australia

The Newcastle quake, with a depth estimated at 11 kilometers, occurred below the Sydney Basin, a complex intraplate environment bordering the Newcastle Dome Basin Belt to the north and the Hunter Mooki Thrust to the east. While sources of seismicity in Australia are not entirely understood, seismologists believe that the continent's intraplate earthquakes are generated by traces of ancient geological formations, formed over the last 55 million years.

Australia is situated within the center of the Indo-Australian Plate, and is subjected to the stresses and strains from movements at the surrounding plate boundaries. As is the case in other stable regions, the earthquake activity is generally higher around the margins of the continent than in its interior. Although evidence of faulting exists in Australia's mountainous regions, most faults can only be inferred from recorded historical seismicity. The earthquake hazard in Australia's more active seismic zones—including the West Australian Wheatbelt, the Flinders Ranges of South Australia, and the Alpine region of eastern Australia—is roughly comparable to that of well-known seismic zones in the central and eastern United States.

Cat Bond Figure 1 Event 1
Figure 2. Australia's tectonic setting (Source: AIR)

Although Australia is largely considered geologically stable, the continent has experienced several moderate to large earthquakes in the not-too-distant past. The QUAKES database from Geoscience Australia, the national geological survey, contains seismic information on thousands of historical earthquakes in the Australian continent. However, this historical record is quite short. Most of these events have occurred in the last 150 years and have been located in areas of low population density; many are of magnitudes too small to be felt. At present, the largest earthquake on record was estimated at a magnitude 7.2 in Meeberrie, Western Australia, in 1941. The continent is struck by an event with a similar magnitude to the Newcastle event roughly every one to two years and researchers calculate that magnitude 6 quakes take place in Australia every five to six years.

Cat Bond Figure 1 Event 1
Figure 3. Epicentral distribution of historical earthquakes 1900-present (Source: AIR)

Implication for Seismic Hazard Estimation

In seismically active regions, such as those along plate boundaries, the spatial distribution of earthquake epicenters is well understood. Abundant historical data delineates active faults and subduction zones, and depth determination reveals even the angle at which one plate subducts beneath another. In addition, observations of the deformation of the earth's crust—made primarily by satellites using the Global Positioning System (GPS)—provide insight into earthquake occurrence rates. This surface deformation represents elastic strain. By calculating the accumulation of elastic strain, and the extent of slip that has occurred in historical earthquakes, estimates can be made as to how often such strain is—and will be—released.

In low-seismicity regions like Australia, however, surface deformation, if it is occurring at all, is extremely slow and GPS data is scarce. And with little or no expression of faults on the surface, the use of physical modeling techniques, such as kinematic models, presents challenges. Estimating the frequency and locations of future earthquakes is therefore much more reliant on the historical record. That is, the probability that an earthquake will strike in a particular location is estimated based on the premise that historical seismic activity can be used to calculate the frequency (and magnitude) of future events. Visualization techniques applied to the patterns of past earthquake epicenters (supplemented by whatever geophysical information exists) are used to define area source zones. Smoothed background seismicity is used to account for the possibility that earthquakes will occur outside these zones—where there has been little or no recorded seismic activity. Together, area source zones and background seismicity preserve the pattern of historical seismicity while allowing for the possibility that earthquakes will occur where none have been observed in the past.

Local Geology and Building Performance

In addition to highlighting the vulnerability of urbanized areas to a moderate earthquake, the Newcastle event also illustrated the impact of site conditions on the distribution and extent of building damage.

One study3 revealed a strong correlation between the presence of alluvial soils and damage; indeed, close to 80% of building damage occurred on the saturated, unconsolidated sediments often referred to as Newcastle Coal Measures. Layers of quaternary alluvium of variable thickness—ranging in depth from 8 meters to 60 meters—lie directly on top of the coal measures. Soft soils such as these can amplify ground motion, sometimes significantly; at the same time, ground shaking from intraplate earthquakes tends to attenuate, or dissipate, much more slowly than those from plate boundary events and thus, for any given magnitude, affect a much larger area. Hence, the demands on certain building types increases and damage is exacerbated.

Not surprisingly, the concentration of damage was located at sites with poor soil conditions—where shaking intensity was amplified by as much as four times that of firm soil sites. The isoseismal map below shows the distribution of Modified Mercalli values. While the earthquake's moment magnitude (5.6) was moderate, a maximum intensity of VIII was felt in pockets across a radius of roughly 80 kilometers, and swaying was reported in high-rise buildings in Melbourne, some 800 kilometers away.

Cat Bond Figure 1 Event 1
Figure 4. Isoseismal map of the Newcastle earthquake. Note the contour around Newcastle was generally of MMI VI, with small pockets as high as MMI VIII. (Source: AIR)

Data from a rock site west of the epicenter show a maximum peak ground acceleration of 0.25g. While not particularly high, most buildings were not designed to withstand even this level of ground motion. The majority of buildings that suffered major damage were constructed of unreinforced masonry (URM), a building type known for its lack of lateral resisting system. Building inspections performed immediately after the earthquake revealed extensive evidence of structural deterioration, primarily in older buildings that dated from the turn of the century.

However, even some modern buildings, including the Pasminco zinc refinery, the John Hunter Hospital (which had not yet opened) and the Newcastle Technical College experienced structural failure—highlighting the fact that even modern construction may be at risk when non-ductile elements are used on soft soil conditions.

The Risk Today

In the 20 years since this event, much has changed in Newcastle—and in Australia. The population has increased, as has the number and value of exposed properties. Offsetting the increase in exposure, a major upgrade to the Australian building code was implemented in 1993. The new code moved Newcastle from the lowest seismic zone "0" to a zone of high seismic hazard. The code was updated again in 2007 to better reflect soil conditions, refine the earthquake design spectra, and limit the use of unreinforced masonry, depending on building height and soil type.

Cat Bond Figure 1 Event 1
Figure 5. A comparison of the seismic zonation for Australia's building code from 1979 (left) to 1993 (right) (Source: AIR)

In their 2002 report, Geoscience Australia estimated that the return period of a loss similar in size to the loss from the Newcastle Earthquake is roughly every 1,500 years, or an annual exceedance probability of roughly 0.07%, which is in line with results from the AIR model. With that in mind, AIR evaluated the potential impact of a Newcastle-like event occurring in 2009. If the Newcastle earthquake recurred today—using trended exposures—AIR estimates that insured losses would be near AUD 5 billion.

However, there is a danger in focusing on a single event. While the seismic source that generated the Newcastle event is certainly a candidate, there are many other seismic source zones capable of producing large losses. The probability that a future earthquake will be exactly the same as the 1989 Newcastle event is near zero. That is why the AIR Earthquake Model for Australia includes a catalog of thousands of potential loss-producing events.

One such scenario is illustrated in Figure 6. This shows the simulated geographic distribution of losses for an M 5.6 event 40 kilometers south of Sydney—located in a zone identified as having the same hazard as the Newcastle earthquake. As such, the seismic risk of Sydney is quite comparable to Newcastle; both are located along the seismically active Sydney Basin, and a significant portion of Sydney's building stock is also composed of unreinforced masonry (URM) infill walls, which, even with newer design standards, are still highly vulnerable.

Cat Bond Figure 1 Event 1
Figure 6. Insured losses from a simulated M5.6 event near Sydney (Source: AIR)

AIR estimates that this event could result in insured losses of more than AUD 11.3 billion dollars. Losses include shake-induced building and contents damage to insurable residential, commercial, and industrial properties.

Closing Thoughts

The anniversary of the Newcastle earthquake serves as a reminder that damaging earthquakes and large losses can and do occur even in relatively stable intraplate regions. However, while the Newcastle earthquake remains the seminal event in Australia's recorded earthquake history, the record here is short—too short to rely exclusively on historical loss data to assess future earthquake losses, or where and when the next significant event may occur. By taking a probabilistic approach to catastrophe risk assessment and simulating the effects of large catalogs of potential future events, catastrophe models are perhaps most valuable in regions where uncertainty is highest.

1 Insurance Council of Australia 

2 Insurance Council of Australia 

3 Brennan. E. (1990). Geological controls on damage patterns resulting from the 1989 Newcastle earthquake. Proc. Fourth Earthquake Eng. Workshop. The Univ. of Queensland, Australia.




You’re almost done.
We need to confirm your email address.
To complete the registration process, please click the link in the email we just sent you.

Unable to subscribe at this moment. Please try again after some time. Contact us if the issue persists.

The email address  is already subscribed.