Application Series 16

FIRES ARE COUNTED NATIONWIDE

In 1992, Australia produced it's first set of national fire statistics. This was the result of the combined effort of the seven metropolitan fire services and CSIRO with the use of a SIR database.

Background

The Commonwealth Scientific and Industrial Research Organisation (CSIRO) is Australia's premier research organisation with a staff of approximately 9,000. Interests vary as widely as a radio telescope assisting NASA to a research station in Antarctica.

The CSIRO division of Construction Building and Engineering, combined with its predecessor organisations, has been involved in fire testing for approximately 40 years. It first approached the concept of a national set of data for Australian Fire Incidents in 1972. Then it used the data from the New South Wales Fire Brigades and the Australian Bureau of Statistics as a basis for its knowledge on real fires. This data was analysed and extrapolated to represent the country as a whole, but as Australia has quite diverse conditions, such as three climatic zones, this data was lacking in depth and scope.

National Standards

After extensive investigation by Western Australia's Fire Brigade, NFIRS (National Fire Information Reporting System) originating in the USA, was considered an appropriate encoding system for the collection of Australian fire data and was incorporated into an Australian Standard (2577) in 1983. This work culminated in Western Australia using NFIRS codes in its reporting system in early 1984 followed a few months later by New South Wales.

This Standard was revised and re-released in 1992 and all metropolitan fire brigades and most rural services have adopted this new standard.

National Statistics

An agreement between the Australian Assembly of Fire Authorities (AAFA) and CSIRO was reached in November 1989 which meant that the first set of national fire data statistics for Australia were produced for the year 1989 - 90. It is estimated in this first attempt that approximately 75% of all fire data was captured. Shown below are two graphs produced in this publication. They are the cumulative cost of commercial structure fires by the time it takes for the first fire fighting appliance to arrive at the fire.

Problems

The problems were enormous as there are over 20 autonomous Fire Fighting Services within Australia who had minimal contact with each other regarding the collecting of fire incident data. Two major services did not encode at all, so data was unusable in this form. "blank" fields, which were missing, were confused with the valid code of "0". NFIRS code numbers had been changed by some services to suit the local conditions.

However certain data was always collected, such as the management data relating to the internal running of a service and casualty information. This data on it's own gave the number and type of fires, where the alarm was raised, times and personnel involved which gives a considerable pool of information.

SIR Database

The SIR database had to be able to supply certain key functions. It had to be able to hold large records in a fixed format and it needed to be able to hold a large number of records, currently there are 666,000 records. Initially the record length was 151 characters it is now 703 characters. Obviously with these numbers of records arriving from the brigades they had to be entered in batch mode.

The database needed to be able to do a full range of edit checks which was accomplished by every numeric field having a Valid Value. The Date and Time fields were automatically checked for legality. This way each code was guaranteed correct. Then global edits were written so that there were no conflict in fields, such as the fire brigade arriving before the fire started or having a 15 storey car fire. This group of edits was written as a boolean file and accessed prior to the records being entered in batch mode.

Having help at the end of the telephone was a great advantage. Knowing the people at SIR made it easier to explain a problem, or fax the problem and then talk it through.

The SIR database is very powerful and can do some enormous number crunching and complicated retrievals which makes it very attractive. Then once the database is written and the retrievals written to get the reports it is quite a comfortable system to use. Retrievals are easily modified to create a new retrieval.

The system is very secure from outside interference.

The English also use SIR so we could do comparisons and or assist each other.

Future

CSIRO is interested in establishing international links to have access to large data banks and different data to that existing in Australia. This will enable the researchers to investigate building materials and techniques that have been in existence in other countries, compared with the resulting fire data of those countries, to predict what would happen locally.

An example could be, where timber panelling is popular as a wall lining in the USA, in Australia it is plaster board. Hence if timber panelling became popular in Australia a pool of data already exists potentially saving lives and property.

Conclusion

It is only with knowledge that fires can be prevented and fought. This knowledge is going to come from the various experts who will best advise the myriad groups in fire protection and fire fighting. The experts will gain their knowledge from the data presented.

The data must be able to have its quality measurable and represent the total situation. In Australia we know that the fire services will deliver 100% of their data to CSIRO within the next couple of years. At which time, to the best of my knowledge, Australia and England will be the only 2 counties to achieve this 100% coverage in the world, so the database will be able to perform its function.

For more information on the use of SIR in the CSIRO contact:

Michael King
Manager, Fire Statistics Section
CSIRO Division of Building, Construction and Engineering
P O Box 310
Nth Ryde NSW 2113 Australia

Back Sir Home