Home Contact us Gallery Site Map Resource Centre Search About us People News Publications Education & Training Events Publications | Education and Training | Current & Recent Projects | Events | Research Opportunities

Download issue (989KB PDF) Cover Page Executive Summary Introduction Data Sources Deaths Due to Natural Hazards A Building Damage Index 20th Century Building Damage Alternative Perspectives on Damage Spatial Variation in Damage A More Refined View Discussion Conclusion Further Reading Acknowledgements

Issues in Risk Science Natural Hazards Risk Assessment: An Australian Perspective - Russell Blong Data Sources The European colonization of Australia – and its written history - began at Sydney Cove in 1788. With only 20 million people spread across 7.7 million km2, even today parts of the continent are not exactly overcrowded. As an example, Australia Post divides the country into 2,433 postcodes, each with an average population of about 8,200. The largest postcode (872 in Western Australia), had a population at the 2001 Census of 20,400; the postcode covers an area of 621,400 km2, an area significantly larger than continental France! While it could be argued that nothing much happens, from a natural hazards point of view, in Postcode 872, that was exactly the rest of the nation’s view of Canberra, the national capital – except that this view changed in January 2003. So the written history of natural hazards and their consequences is short, barely more than a hundred years for substantial parts of the country. Numerous attempts had been made to construct summary histories of tropical cyclones or bushfires both in popular accounts and in more scientific literature. Inquiries, official reports and accounts in learned journals into the causes and consequences of the more damaging events almost inevitably add perspective with summaries of earlier damaging events. Dozens of local histories also offer valuable information. The Australian Bureau of Meteorology (BoM) has maintained a severe weather database for some decades, focused on severe thunderstorms, defined more-or-less by international convention as thunderstorms producing wind gusts of 90 km/h or greater, tornadoes, hailstones greater than 20 mm in diameter on the ground, or flash flooding resulting from a rainfall with a 1-hour intensity exceeding that of the rainfall intensity with a 5-year recurrence interval. For the most part, the BoM database focuses on the last 50 or so years. Similarly, the Federal government agency Geoscience Australia maintains databases on the occurrence of earthquakes, landslides and tsunami and their consequences. These databases span the entire period since European colonization. The earliest versions of the landslide and tsunami databases were compiled by Risk Frontiers for Geoscience Australia. And then there are newspaper accounts. While a range of newspapers were examined, The Sydney Morning Herald and its forerunner The Sydney Gazette, provides an unbroken record of just over 200 years of disaster reporting, for the first hundred years and more, by correspondents, in every tin-pot settlement in the country. While parts of the paper are indexed, much of the run is not and researchers read every natural hazards-related item up to the late 1990s; the photocopied items fill nearly a dozen filing cabinet drawers. This summary of the building blocks provides insights into the strengths and weaknesses of the Risk Frontiers database. It combines reports of nine perils and the consequences of individual events – tropical cyclones, bushfires, floods, wind gusts, tornadoes, hailstorms, earthquakes, landslides and tsunamis – in an integrated framework. In the end we focused on the 20th century, endeavouring to make the record as complete as possible for the period 1900-1998. We recognize that narrowing the period of record might bias the results. As an extreme example, the greatest loss of life in any natural hazard impact in Australian history, about 400 people, was produced by the Bathurst Bay cyclone (Cyclone Mahina) in 1899. Cynics, a group that certainly includes the owners of this database, can argue that any reliance on the veracity of newspaper reports casts doubt on the value of the information in a database. Cynicism is an essential attribute for researchers, but we should make sure to extend it widely. Normally one would prefer, for example, the records of deaths compiled by the Australian Bureau of Statistics (ABS) over mere newspaper accounts of deaths as a result of natural hazards. However, the ABS data don’t always fall neatly into categories such as “floods” or “bushfires”. The ABS also preserves privacy so that data on age, sex and the locations where deaths occurred are only occasionally provided unambiguously. Interestingly, the ABS records the date the record of the death was received, rather than the date of death. Some of the perils and consequences databases are less complete than we would like. For example, the severe weather part of the database has fewer storms in the first half of the 20th century than we would expect, and we know that wind-driven hail with maximum diameters less than 20 mm (i.e. non-severe thunderstorms) that can pit motor vehicles are probably missing from the register of events. We are fairly confident that the record of human deaths in floods is substantially complete, but we are less certain that we have recorded all flood damage to buildings. We haven’t kept a record but the database took more than 20 person-years worth of effort. Certainly there are errors and omissions and perhaps some double-counting. But the database is essentially complete; conclusions drawn from the data are likely to be robust. Improving the database, say another five person-years of effort, might progress the information by a small margin, but it is unlikely to alter conclusions drawn from the information. «back to top«