Professor Russell Blong

NHRC is kindly sponsored by:
Swiss Re Australia
QBE Insurance
Benfield Greig Australia
Guy Carpenter & Co

In this issue:

Perils, Postcodes & Risk Accumulation Zones

The NHRC has undertaken to produce a product that will provide up-to-date information on the occurrences and consequences of natural perils in Australia. The project is funded by the Insurance Foundation, and will be completed by the end of the Year 2000. Its primary purpose is to assist in the formulation of underwriting strategies, the control of risk assessment and risk rating and the management of portfolios.

  • The 1954 earthquake in Adelaide - the same magnitude as the 1989 Newcastle earthquake - is well remembered. What larger earthquakes have occurred in South Australia and what were the consequences?
  • What parts of Sydney were affected by the 1 January 1947 hailstorm? Was it more damaging than that of 18 March 1990?
  • Were the 1939 Black Friday fires more severe than the 1983 Ash Wednesday fires? Were the same areas affected?
There is no simple yet authoritative source of information on natural perils in Australia. The Insurance Council of Australia has maintained valuable records of catastrophe payments made on insured events for the last 31 years, but this record is short compared to the return periods of many natural perils. In order for meaningful data on the magnitude, frequency and occurrence of, say, tropical cyclones to be determined, a record of up to 100 years length is required. Additionally, many of the perils that damage insured property in Australia are complex, in that damage can result from more than one agent. There is a wealth of information available on the above, but the data are scattered, often unverifiable, and differing sources are difficult to reconcile and impossible to relate to present day insured risks, urban areas, postcodes and ICA accumulation zones.

This project aims to remedy that situation, and to take much of the guesswork out of rating natural perils. The information that the NHRC have undertaken to compile will provide up-to-date statistics integrating the experience of several perils, allow comparisons between events or areas and produce relevant data on a postcode or Risk Accumulation Zone basis.

The first product will contain historic and recent information on the incidence and consequences of natural perils in Australia from 1900-2000. It will include event summaries and analyses, damage indices, insurable tangible damage, risk assessments, maps, illustrations, newspaper clippings and references to other literature. The records will be simple to search for individual perils, dates, postcodes, urban areas, risk accumulation zones or combinations of these.

  • Click on a postcode, request data on one or more perils, or enter a year or a range of years. Search routines will deliver lists of events and their consequences. Request a damage summary. Verbatim accounts, photographs, maps and statistical data will be provided (where possible). A Damage Intensity Index will be provided for each impact, summarising the estimated degree of damage.
The second product will contain an indicative risk-rating scheme identifying, for each urban area, postcode and ICA Risk Accumulation Zone, the relative importance of each natural peril. The rating scheme will be particularly relevant to household, motor vehicle and commercial portfolios.
  • Click on an ICA Risk Accumulation Zone. Obtain a history of peril impacts in the selected area, with each event labelled with an intensity index, or a peril rating for each area of interest, peril by peril, or an integrated peril rating. Make comparisons, for any peril, with other postcodes or Risk Accumulation Zones.
Workshops will be held after the completion of each of the products, to showcase them to the insurance industry and to enable users to gain the maximum benefit from them.

Stay tuned for information updates!

For further information contact:
Laraine Hunter

Phone: +61 2 9850 9684
Fax: +61 2 9850 9394


Towards an Integrated Approach to Bushfire Risk Assessment at the Urban Fringe Using GIS

Bushfires are endemic to the Australian continent, and cause extensive losses every year. Particularly for the area of urban fringes where bushlands are heavily affected by human activities and ecological and physical factors, bushfires carry the greatest disaster potential to the surrounding properties and lives. In this study the challenges of undertaking an integrated bushfires risk assessment fall into three major components: data integration, multi-criteria modelling and integrated objective decision making. By using these components in an integrated fashion, it is anticipated that thorough bushfire risk assessment at urban peripheries will be achieved.

As we already know, a large group of factors contribute to the occurrence of bushfires, including meteorological conditions; vegetation fuel loads; topographical variations; the design of residential structures and human activities. Heterogeneous data integration is the first step in fully assessing the proposed risk. A collection of census data, remotely sensed imagery, digital terrain models and other geographical themes, such as road and vegetation types, is included. GIS provide a platform to collect spatially-referenced data for further spatial analyses. Additionally, some non-spatial data can be included in the GIS databases. More importantly, exploring ways of combining possible data or how they can be interactively integrated is crucial. For example, discrete census data after building zone-disaggregated surface models are superimposed by land use maps to approximate spatial distributions of certain indicators; and census data as a priori probabilities to assist the remotely sensed imagery classification. Therefore, not only does this step form the comprehensive site-based data set, but it also provides insights into further and deeper use of them in order to facilitate risk assessment.

Given a rich set of spatial and non-spatial data, the multi-criteria evaluation (MCE) techniques are capable of coupling these data sets. GIS-based MCE approaches place bushfire risk assessment in an integrated context. There are a number of classic methods of MCE, with either ingeniously selected algorithms or successfully applied practices. However, most of them are derived from non-hazards related projects; thus the key is to develop rational process-oriented MCE for bushfires risk assessment. Based on the review of a series of fire growth models, an integrated approach which can effectively delineate the critical areas will be proposed. Moreover, the advantages and disadvantages of the various MCE models will be further elaborated with their respective practical significance. For the practical implementation of the MCE models, Graphical User Interfaces (GUI) which link the GIS software and the MCE need to be constructed.

The successful integration of physical, ecological and census data is fundamental to understanding bushfire risk. However, the ultimate aim is to make scientific decisions and to take preventive measures in order to mitigate potential bushfire damage. The objective decision-making relevant to risk modelling is based on the quantification of high quality and high reliability data, the validated models and the uncertainties inherited. It is argued that the full treatment of errors and uncertainties involved in the process of risk modelling is an essential part of bushfire risk assessment. By doing so, the actual quantification of the environmentally and/or anthropologically critical factors at the investigated area could be identified and the integrity of risk assessment can be guaranteed.

For futher information contact:
Keping Chen
Phone: +61 2 9850 8433

Earthquake Damage Analysis - A Joint Approach  

Australia lies in a low seismic zone but the ML 5.6 Newcastle earthquake of 1989 caused 12 fatalities, around 160 injuries and A$1091million insured damage [1996 $]. Copious amounts of data on this event have been collected by various institutions, these data providing insights into the damage such events cause in the Australian situation. On January 7th 1990 the UK-based Earthquake Engineering Field Investigation Team (EEFIT) surveyed the damage, photographing 600 building sites and recording damage information. Their field report was released in 1991.

The NHRC has also spent considerable time investigating the aftermath of the Newcastle earthquake. Our interest is in the damage to buildings and insurance claims, with the aim of producing earthquake loss curves that are more directly applicable to Australia. With the assistance of Benfield Greig Australia and a number of Australian insurance companies, a new set of loss curves was developed and used in our 1997 Sydney PML Study.

In early 1997, Dr Robin Spence, of Cambridge Architectural Research and Dr Andrew Coburn, of CARtograph, visited the NHRC and discussed the work they had been doing with the data contained in the EEFIT report. As their book "Earthquake Protection" (Wiley 1992) indicates, they are interested in the engineering of buildings to withstand earthquakes. The ensuing discussion led to a joint research project, which combined data in our overlapping databases. The first step in the process was to re-photograph the 600 properties originally photographed by EEFIT in downtown Newcastle (Hunter St) and Beaumont and Lawson Streets, Hamilton - the areas severely effected by the æquake. This gave a valuable record of how the earthquake effected not only repairs but to some extent the re-gentrification of the Beaumont Street area. From 10-19th Nov 1997 Laraine Hunter worked with Robin Spence, in Cambridge, in order to join the two databases.

The combined database now contains information on the 600 properties including:

  • photographs post-earthquake and eight years later
  • property location details, geocoded onto a street map (using MAPINFO)
  • building age
  • EEFIT damage scale
  • council damage scale
  • soil data
  • soil depth
  • indication of repair costs
Currently most of the data on cost of repairs is for domestic dwellings, but we hope to include some details for commercial buildings. 

The EEFIT data contained a damage scale where D0= Undamaged, D1= Slight Damage, D2= Moderate Damage, D3= Heavy Damage, D4= Partial Destruction and D5 = Collapse. Figure 1 shows the number of properties of each EEFIT construction type per damage category. Figure 1 indicates the dominance of brick construction in this area and shows that the majority of buildings appeared undamaged. Brick construction had the highest percentage of damaged buildings, and dominated the D3 category. Very few properties fell into the D4 and D5 categories.

Figure 2 compares the EEFIT damage ratio with the number of insurance claims and the Claim/Sum Insured (SI) ratio. Of particular interest is the high number of claims in the D0 category (no damage). This is probably due to the limitations of the initial damage assessment, which considered only external damage. Presumably these claims related to internal damage of a minor nature, as the average Claim/SI ratio is relatively low. As previously noted, few claims fell into the D4 and D5 categories but these had high Claim/SI ratios, as expected. Interestingly, Category D1 had fewer claims than D2 but a higher Claim/SI ratio. This is probably due to a higher proportion of brick claims in the D1 category, increasing the repair cost.

This initial analysis provides a greater insight into the relationship between construction type and damage. Future work with CARtograph may investigate the building age/damage relationship or the effect of soil properties on damage, but this will require improved insured loss. data.

For further information contact:
Laraine Hunter
Phone: +61 2 9850 9684
Fax: +61 2 9850 9394



Under an agreement recently signed, the copyright and intellectual property residing in the well-known flood-damage modelling program ANUFLOOD has been transferred from The Australian National University to Macquarie Research Ltd.

ANUFLOOD was developed during the 1980s and early 1990s principally by David Ingle Smith [known to everyone as Dingle] and Mark Greenaway at the Centre for Resource and Environmental Studies at The Australian National University. Dingle has long been at the forefront of flood hazard studies in Australia and beyond, with Mark as the computer genius behind it all. Dingle's recent retirement provided the opportunity for NHRC to take over the development of ANUFLOOD.

ANUFLOOD is an inter-active program designed to assess tangible urban flood damage. Input information required by ANUFLOOD includes a building-by-building description of location, ground and floor heights, construction material, value, house size, number of storeys and so on.

Flood frequency input to ANUFLOOD uses a listing of flood stages expressed as probabilities. Flood slope can be allowed for in two ways and a range of flood frequency curves can be used where necessary.

Flood stage-damage curves are provided for three sets of residential property with a further set for commercial property subdivided by size and susceptibility of contents to flood damage.

In the last twelve months, following damaging floods in Coffs Harbour [New South Wales] and more recently in Townsville [Queensland] and Katherine [Northern Territory], the insurance industry has shown renewed interest in flood insurance in Australia. Although flood insurance for domestic properties has long been offered by some insurers in Queensland and the Northern Territory and is often available for commercial properties everywhere, flood is excluded as an insured peril on most residential properties. A strong body of opinion, fueled no doubt by conflicts with policy holders over the differences between flood and storm water damage [the latter is normally covered], indicates a growing interest in the industry in providing flood insurance more widely.

The availability of flood maps for many areas in New South Wales makes adequate risk rating easier there than in some other areas. However, in almost all areas too little attention has been paid to the potential impact of floods with magnitudes greater than once in 100 years. While flood frequency issues are likely to provide the greatest challenges to risk rating, further work is required to devise ways of obtaining the necessary portfolio information in cost-effective ways. We intend also to reach arrangements with insurers for the collection of additional stage-damage data.

In the next few months NHRC staff will begin to upgrade ANUFLOOD with a Windows 95 front end and modifications to ensure that the program is suited to the needs of insurers with an interest in risk rating for flood insurance purposes. Dingle Smith will continue to work with NHRC staff as a consultant.

The agreement between ANU and Macquarie Research Ltd gives NHRC the right to change the name of ANUFLOOD. We aren't in a hurry, but send us some suggestions!

For further information contact:
Russell Blong
Phone: + 61 2 9850 8397
Fax +61 2 9850 9394


nhq homepage
nhrc homepage