This commentary was first published in August 2004 in AURA Newsletter, Volume 21, Number 1, pp. 5-6.

A recipe for failure



The Dampier rock art monitoring program by the Western Australian government was initiated in response to my paper ‘The survival of the Murujuga (Burrup) petroglyphs’, in RAR 19(1). In that paper I had reported detecting quantifiable degradation of the ferruginous mineral crust in the Dampier Archipelago on whose preservation the rock art depends for its continued existence. Having measured changes in colour and crust degradation since the late 1960s, I had observed a marked deterioration since the late 1980s. In extrapolating from these empirical observations I predicted that the proposed three-fold increase in acidic gaseous emissions would cause widespread loss of colour contrast among the massive petroglyph corpus over the course of the 21st century.

I have therefore every reason to welcome the initiative of the government. Moreover, our knowledge base of the effects of atmospheric acidification on mineral patinae remains woefully inadequate, and any work that is designed to improve this is more than welcome. However, if the full potential of such an opportunity is not realised, it is just an opportunity lost, and it is for this reason that we need to examine the above proposals critically.

It seems to be agreed that the primary purpose of this exercise is to (a) determine the precise nature of the processes causing deterioration, and its secondary purpose is (b) to use this information to design ways of alleviating these effects. From a scientific perspective, the first purpose is the more valuable, because it will enrich our generic understanding of threats to rock art. However, from a purely pragmatic point of view, only the secondary purpose (b) is of ultimate interest, for the possible implementation of palliative measures. Nevertheless, it is obvious that, in order to design such measures, it is imperative that the basic research be conducted under optimal conditions. The terms of reference make it clear enough that this project will fail to deliver the optimal information we require.

The main reason for this lies in the central assumption that the deterioration processes are very slow and gradual. This is no doubt acceptable for Study 3, where colour changes are to be examined. These occur over a long period of time and the proposed four-year period may in fact be inadequate. Moreover, none of the studies is concerned with the key issue to be addressed: the physical degradation of the mineral accretions that are the very centre of this preservation problem. Instead we have monthly monitoring of four pollutants, which will offer vaguely relevant background information; we have monthly sampling of some factors of atmospheric deposition, which is much more relevant but is incompletely covered; the colour monitoring program; and simplistic accelerated weathering experiments.

Dampier is a locality of climatic extremes, where most precipitation occurs within a few weeks in the year. Gaseous emissions by themselves are not likely to dissolve the crusts, they would become operative in the presence of moisture. Unless dew has an effect, which remains unknown, we could reasonably assume that the deterioration only takes place in the course of a few weeks, essentially in the cyclon season. Obviously what we need to know more than anything else is what happens at the rock face when it rains. What I would like to know most is this: does solution take place only during rain, does it proceed evenly or does it peak, for instance in the first ten minutes of rain, when radicals in dormant solids are activated by the water? Or perhaps it takes a certain time of rainfall before the micro-erosional front becomes fully active? At this stage, we cannot claim to understand the processes of mineral accretion deterioration at all, therefore any study must commence without unwarranted assumptions. It is quite probable that most of the deterioration of the crusts is by the reaction of NO2 and rainwater in the atmosphere, yielding nitric acid, but how does its effectiveness vary as a function of duration of the precipitation event? We have no idea about these crucial details, and the most obvious aspect of the proposed study project is that it cannot possibly provide such details. The simplistic design of the project prevents the acquisition of the most important data, and instead meaningless data are likely to be collected.

It would have been much more appropriate to determine the variation of rainwater pH at the atmosphere/lithosphere interface and to plot it against time over the duration of a rain episode. The most relevant study would be a direct observation of the process under a binocular microscope, visually observing the physical mobilisation of material. The detailed recording of just one such event would tell the analyst far more than four years of unfocused and purposeless gathering of probably meaningless data. Moreover, such observation results would quickly lead to the formulation of alleviating measures. For example, if it were found that most of the degradation occurred in the early part of a rain episode and is attributable to a flushing with nitric acid, then it would be possible to avoid most damage simply by closing down emissions several hours before rain. Since most rains occur during a brief spell in the monsoon period, Woodside could easily schedule this period for annual maintenance shut-down, or for the periodic repairs or construction works entailed in the normal operation of such installations.

I am of course not suggesting that this is the correct solution, I merely use this example to show that we must not make assumptions about the relevance of specific empirical indices. All we know with certainty is that there was deterioration of the accretions since the late 1980s, because this has been monitored over decades and is quantifiable. We also know, from my work, that the micro-morphology of the accretionary deposits in the Pilbara is heavily influenced by precipitation. The core issue, then, is one of geochemistry. Emissions are the most obvious culprit, but it seems perfectly possible that factors or catalysts we have not even considered are contributing, or are even crucial. In science the solution to a problem can come even from the most unexpected direction. If we work from the assumption that only long-term average exposures are relevant, as the terms of reference in this project stipulate, we may not just be limiting the effectiveness of the project (if the short-term event peaks were the problem), we may render the entire project ineffective. And we may miss a rare opportunity to conduct badly needed basic research into this generic subject area of iron-rich mineral crust deterioration resulting from industry.

There are still other objections to the research proposal. The accelerated weathering study (‘fumigation’) is likely to lead to severely misguided pronouncements about the effects of exposure to various gases. It does not simulate the natural exposure conditions, but will extrapolate from the probably negligible effects of fumigation at multiples of concentration. If the gaseous emissions have little or no effect in the absence of water, as may be the case, what is the point of demonstrating that they have also little effect at twenty times the predicted concentration?

Of particular concern is the proposed methodology of monitoring colour changes in the accretionary ferromanganous deposits. As indicated in the above terms of reference, it is proposed to use a BYK-Gardner spectrophotometer with a 4 mm aperture opening for this purpose. Therefore the instrument proposed is the 45/0 model, catalogue number CB-6807. This instrument is designed for manufacturing processes, e.g. of plastics products. It is highly sensitive to extreme conditions of temperature and relative air humidity, and cannot be operated either above 42°C or above 85% relative humidity ¾ conditions that both occur commonly in the Dampier region. Moreover, the instrument is very imprecise, with a spectral interval of 20 nm, over a spectral range from 400-700 nm. Hence it does not even cover the full range of visible light. Indeed, Gardner call it a ‘colour guide’, a much more appropriate description than ‘spectrophotometer’. The impression that an unsuitable methodology has been proposed is reinforced by the description of how the sampling site will be re-located (not ‘relocated’): digital photographs will guide this process. Yet the baffle surrounding the aperture is of about 12 cm diameter. First, it will be physically impossible to re-locate the baffle so that the original sampling site is targeted ‘to within a millimetre’, as stated. Second, it should not be approximately re-located, or within a millimetre, it must be precisely re-located, otherwise the result can only be imprecise, if not meaningless.

What amazes me most about this methodological blunder is the fact that a vastly more precise method of colorimetry is available. It was designed precisely for measuring changes in rock art, it has been published, and has been used in the Dampier Archipelago for many years. The colorimetric method I have developed specifically for rock art is much more precise, simpler and cheaper, and I have used it not only at Dampier, but also elsewhere in the Pilbara (notably on repatinated inscriptions at Spear Hill) as well as on red pigment in Mladec Cave, Czech Republic, on petroglyphs in Saudi Arabia, and elsewhere. In this method with its spectrum of over 16 million colours the spatial sampling precision is under 100 microns. It requires no expensive American gadget and measures colour change with much greater accuracy, by relying on aliquot readings rather than single but questionable values. Moreover, because calibrated photographic records can be included to determine earlier colour changes, the duration of the monitoring period can be extended to the earliest such records we have. They are in this case of 1968, so in 2008 they will provide coverage of the highest precision for a total of forty years, whereas the results of this government endeavour will cover just 10% of that duration, at vastly inferior precision, at massive cost (many tens of thousands of dollars), and yield results which I can only reject: I remain unconvinced that the Gardner instrument is even remotely suitable for this task.

Why, then, this unfortunate choice of methodology? The answer is simple. This rock art committee does not include a single rock art researcher, or even a specialist in accretionary mineral crusts. This is despite the fact that such expertise would have been available in Perth, and amply available in the eastern states. The committee is unaware of previous attempts to measure rock art deterioration by spectrophotometry or spectrometry, which were unproductive or inconclusive, and which were the very reason why I opted for digitised colorimetry a long time ago. The exclusion from this project of rock art specialists is deliberate, as shown by the simple fact that all four studies were awarded to a single government agency that, significantly, has no previous track record in rock art research or rock art conservation. Having conducted all previous analytical work on the Dampier rock art, and having prompted this project, I find it sad that I was not consulted on any aspect of its design. Its mistakes were all avoidable.

Therein lies the problem: this is not an attempt to resolve the issue, but a political whitewash and a measure to procrastinate further. I predict that the results of this project in 2008 will be inconclusive and unreliable, and that the main finding will be that CSIRO will require further funding to continue the work. Meanwhile the government expects to continue its destruction of the Dampier rock art, bulldozing many more sites, and permitting the huge petrochemical industries to belch out ever more acidic emissions, at the rate of tens of thousands of tonnes per year. And all the while, the alternative Maitland Industrial Estate remains unoccupied.