Note: Where available, the PDF/Word icon below is provided to view the complete and fully formatted document
 Download Current HansardDownload Current Hansard   

Previous Fragment    Next Fragment
Thursday, 24 March 1994
Page: 2317

Senator COULTER (10.13 p.m.) —I wish to speak in this appropriation debate on the inadequacy with which Australia has been dealing with a number of natural disasters and specifically with the disaster of bushfire. For a number of years the Democrats, and, long before that, some of us in the conservation movement, have been strongly promoting the proposal that Australia should buy a fleet of Canadair water scooping aircraft which can pick up six tonnes of water in 10 seconds while in flight and can deliver that water to fires extremely quickly—indeed, a lot more quickly than any other aircraft available anywhere in the world.

  These aircraft are used extensively in Canada; they are also used right across Europe in Portugal, Spain, France, Italy and Greece. In that capacity they are, indeed, extremely effective. They are expensive aircraft. The proposals we put to government in relation to our budget proposals, which went to government earlier this week, are that in view of the expense of this aircraft it should be multi-tasked. It should be given a number of roles so that the bush firefighting role is only one role and, therefore, the costs of the aircraft for firefighting are reduced.

  Whenever we have made these proposals we have found repeatedly that there has been an objection to the purchase and the use of this aircraft on the grounds of a CSIRO study done in the early 1980s—namely, Project Aquarius. I want to lay to rest the claims that Project Aquarius was anything but a charade. It was a badly based, totally inadequate study and one which should not be given any further credibility in relation to the assessment of this aircraft. The Aquarius study was started in the early 1980s by the federal government with an amount of $3.2 million. It was hailed as Australia's most comprehensive investigation into bushfires. The project was headed by Mr Phil Cheney of the CSIRO division of forest research and the project coordinator, Dr John Nicholsen.

  A briefing was held in Canberra by the project team to provide information about the project and its phases to those organisations which were contributing in some way or other. The purpose of the project was to evaluate the following: the effectiveness of large air tankers and fire retardants for suppressing eucalypt fires of varying intensity; the effectiveness of conventional firefighting techniques under similar conditions; and the cost benefit analysis of forest and bushfire suppression in Australia. This is most important as in fact no research—I emphasise that—was carried out on bushfire suppression. The briefing stated:

Fire behaviour field studies in Western Australia, during January to March 1983, are required to measure the behaviour and characteristics of high intensity fires—

note, high intensity fires—

likely to be important in air tanker suppression operations, and to carry out some measures of the effectiveness of conventional firefighting techniques. The field studies are to be conducted in co-operation with the Forest Department of Western Australia and in conjunction with a series of experiments—

note, a series of experiments—

to study the behaviour of coalescing fires in dry fuel beds.

In the same briefing, the scientific objectives were stated as:

. . . limited to aspects of fire behaviour and fire line construction . . .

It goes on to state:

Determine the likely distribution of spot fires down wind of a 10 ha fire at different levels of fire danger.

These are important points to note. It continues:

Examine the spatial pattern of wind flow ahead of a moving front of a 10 ha fire.

Relate the increase in surface wind speed to the rate of energy release of a number of coalescing fires.

Determine the importance of particular meteorological conditions of temperature lapse and wind profile in the atmosphere to plume development and increase of surface wind speed in coalescing fires.

Determine the relationship between fire danger index, head fire intensity and the rate of fire line construction.

Measure the environmental and metabolic stresses experienced during bush fire fighting in the Australian summer.

Measure the physiological, behavioural and subjective responses to those stresses.

Determine the extent to which the stresses are reduced or aggravated by special protective clothing.

Apart from these specific objectives, a general objective is to check or modify existing empirical and theoretical models of fire behaviour, particularly with regard to the interaction of spot fires between themselves and an advancing fire line. Studies were undertaken in Western Australia and in South Gippsland in forest areas allocated by the WA forest department and the Forests Commission of Victoria. It is believed that Western Australia experienced its wettest season in 27 years and the experiments were limited and scarce.

  The South Gippsland experiment fared no better. In fact, during the whole of the experimental time, only five flights of the DC6B leased from the Canadian fire bombing company, Connair, were completed. In approximately 1985, Dr John Nicholsen, in an article written by science reporter Zena Armstrong, said:

Only five fires on three of the 16 blocks had been conducted and none of them were of sufficient intensity for any significant results. A combination of too much rain and below average temperatures in January had failed to dry out the blocks sufficiently. We were hoping for fires that had a fire danger index of 30, but none of fires that we were able to conduct reached an index of more than five or six and were easily contained by hand crew. South Australian fires in 1983 had an index of about 100 we have so little data, we cannot draw any conclusions.

I want to underscore that because that is a fair summary of the conclusions of project Aquarius on which so much of the damning indictment of the use of the Canadair CL215 as it was, CL415 as it is now, are based.

  The only research of any value to come from the work on Project Aquarius included that by Dr Graham Budd of Sydney University on firefighter stress, and some work done by Chisholm Institute of Technology under a subcontract undertaken by Mr David Packham. Packham's work included the development of the computer ASM1 model, which is designed to simulate fire suppression for a given fire and a given set of parameters. Other work by Packham included firefighter's clothing and some useful research conducted by Andrew Wilson.

  To the best of my knowledge, no further analysis of the limited data gathered in Western Australia or South Gippsland has ever been undertaken, and that must be underscored. It is believed that, in addition to the $3.2 million originally allocated, the federal government assigned a further $1.5 million, and it is also believed that a further allocation was made from CSIRO internal funds.

  It can be said that the report is an expensive computer simulation exercise, exclusive to the then Forests Commission of Victoria now called Conservation and Natural Resources. The report states:

This study has concentrated on the state of Victoria, but the model can be developed to other areas.

This was an extensive computer model exercise. Although some Country Fire Authority data was used, the study virtually concentrated on forest firefighting, both aerial bombing and fire-line construction.

  One of the major problems facing the report is that it has failed to identify and deal with the relationship between forest fires, bushfires and wild fires. The report on page 14 states:

In 1981, the former Prime Minister, Mr Malcolm Fraser requested the CSIRO to "Evaluate the Aerial Suppression of Bush Fires". However, all the practical experiments were conducted in forests, both in Western Australia and South Gippsland, including both aerial suppression and fire crew experiments.

As already mentioned, the practical data were useless for the purpose, and this study became a computer simulation study entirely. There are three distinct environments where fires can have an effect: forests, native bushland in the form of national parks, and the urban fringe bushland interface. As we saw during the New South Wales bushfires, by far the greatest losses occur in, and public attention is on, the urban fringe area—interfaces such as the Adelaide Hills and the Blue Mountains, as we saw in the recent fires. The report failed completely to adequately address firefighting requirements in these areas.

  Forest firefighting operations are, in most cases, poles apart from fighting fires on farmland and urban fringe areas. The report has, in the main, only provided for forest fire suppression by aircraft and hand-crew. Forests are managed by formally trained and academically qualified professionals.

  Foresters protect identifiable and quantifiable assets. This is obvious in the report. In discussing losses due to fires, the report has 27 pages covering in great detail timber losses, contrasted with 21 pages for water, nine pages for national parks, six for bee-keeping, four for human casualties, three for property loss and two for nutrients. On page 4 the report states:

losses were estimated. . . from the limited information available.

Forest research abounds; it can be said with reasonable certainty that the majority of bushfire research has come directly from forest research establishments or individual forest research scientists.

  The report refers on several occasions to a full analysis of data released through the National Bushfire Research Unit which is headed by Mr Cheney. To the best of my knowledge, little if any data from either Western Australia or Gippsland experiments has ever been fully or further analysed.

  Dr Joe Landsberg, who was head of forest research at CSIRO, told a House of Representatives Standing Committee on Conservation and Environment on 21 October 1983:

When I say Aquarius is coming to an end that statement needs a little enlargement. The active research program involved in Aquarius is coming to an end. I think that before we moved on seriously or before we even jumped up and down and screamed about wanting more money, we would need to spend several years—it will take as long as that—ensuring that all of the data accumulated in the course of project Aquarius has been analysed. We are using systems, as always in so much modern research; this stuff pours off a computer in digital form and disappears into magnetic tapes.

The report is seriously flawed, with assumptions, estimates and guesstimates abounding. By the author's own admission:

The (computer simulation) model AIRPRO, is large and complex, it abounds with rough approximations and simplifications of even more complex reality. The rough approximations are indeed very rough.

In the reports the words such as assumed, assumption, preferred assumption, simplest assumption, extreme assumptions, implicitly assumed, and assuming occur no less than 138 times. Estimate, estimated, underestimated, overestimated, rough estimate, broad estimate, economic estimates, and estimation occur 109 times in this report. And average, averaged and rough average occur 67 times. It really beggars one's imagination to believe that so much reliance has been placed on this study when it comes to the serious matter of the hundreds of millions of dollars that have been lost in our bushfires.

  Apart from the model structure, many of the data are subject to considerable uncertainly. There could be a divergence between the data used in the model and actual values in future years which could have a significant effect on the results, but it has not been possible to define the quantitative confidence limits. Yet again on page 35 under `Data Limitations' we read:

Qualification of all aspects of the fire suppression process is the essence of this study, but at the outset reservations about the accuracy of the data and the relationships involved should be made quite clear. In parts it may appear to be an exercise in quantifying the unquantifiable.

There is a number of reasons for this data limitation. The natural processes involved in fire behaviour and effects are highly complex and variable over space and time. Even in the course of a single fire, there may be great variation in the environment, weather conditions, fire behaviour and effects, and suppression effectiveness. In addition, this economic study concerned with a long time span and large regions of Australia face greater variability on the broader scale.

  Wide variability in the economic environment also makes it difficult to place consistent values on the effect of a given resource change. Because of these problems, research studies quantifying the factors are relatively few, covering only a small part of the full range of situations and variables ideally required. Official reporting of actual fires is often scanty. Many agencies' fire report forms did not provide space for all items of interest and often items that are on the form are not filled in comprehensively by informants. Not surprisingly when one is dealing with a bushfire, one is less concerned with filling in forms than with fighting the fire.

  For all the main factors, the best data that was reasonably available were used. Many gaps had to be filled in by interpolation, extrapolations and guesstimates based on qualitative judgments. Even when data were available for particular estimates, they frequently could not be used directly in the model because they were from an atypical situation. Ad hoc adjustments in the light of other information were therefore made.

  In seeking information, the importance of particular data items to the studies sometimes has to be traded off against the costs of acquiring them. Accordingly, the quality of the data varies through the model and there may appear to be detail in some parts that is not justified by a comparison with the rough estimates and assumptions in other parts. The model might be likened to a chain that is only as strong as the weakest link, and there were very many weak links in project Aquarius indeed.

  This means that there were two areas of serious error in the Airpro computer simulator model itself and in the data which was used to run the simulations. The economic report was completed in 1985, following the senior author's period of secondment, using the preliminary data available at the time. However, project Aquarius has, in a sense, been replaced by CSIRO's new national bushfire research unit and further analysis of the data gathered in the Project Aquarius field trials will be carried out by this unit—

  Debate interrupted.