by Tom McClintock
Research Associate, Getty Conservation Institute

A pernicious impact to rock art sites, especially those that experience regular visitation or are in the vicinity of roads, is accumulation of dust. Beyond obscuring the art itself, deposited particulates can undergo a variety of diagenetic processes that lead to the formation of crusts. These layers of solidified sediment can be physically and chemically stable making their removal difficult. In the case of paintings, mobilized particulates can exacerbate existing condition issues like exfoliation or aeolian erosion. Especially with finer particulates which are hygroscopic, a layer of dust can readily attract moisture, negatively impacting paint layers below or inducing chemical alteration of an otherwise stable rock surface. This range of detriments demonstrates that, whenever possible, sources of dust should be identified and eliminated (Fig. 1).

In 2015 as part of my graduate studies in conservation and site management I was enlisted by Njanjma Rangers, an indigenous ranger group dedicated to natural and cultural resource management based in the community of Gunbalanya in Australia’s Northern Territory (Fig. 2). Their concern was that, during the dry season of May to October, traffic on the dirt road leading into the community from Kakadu National Park to the west was generating a significant amount of dust, impacting hundreds of known rock art panels within 250 meters to the road’s southeast (Fig. 3). Njanjma Rangers were interested in undertaking a study that would clearly demonstrate what everyone in the community suspected, that the road was the source of this dust, in order to lobby the Northern Territory government to seal the relevant section. The constraints were that we had almost no budget, the project area was remote, and we had no access to any specialized equipment. An additional consideration was that we wanted to make the methodology relatively adaptable and easy to replicate should other ranger groups be interested in conducting similar studies.

We had two principal research questions: 1) whether the road was the primary source of airborne dust, and 2) how far is the dust travelling, and thus how many sites are being impacted? To answer these, we devised a system of dust collection using supplies found at any standard hardware and office supply store. Using Mylar (semi-ridged sheets of clear plastic) and double-sided tape, we were able to manage sample collection in two different formats.

To measure distance dust could travel, Mylar swatches of 4cm2 with tape on one side were mounted to garden stakes stuck in the ground every 50 meters in a perpendicular line from the road (Fig. 4a & Fig. 4b). Similar swatches were placed at select sites with rock images. An additional four sampling stations were placed in a flood plain approximately 500 meters to the northwest of the road, away from off-road driving tracks, to measure ambient airborne dust conditions.

For directional sampling, 10cm diameter PVC pipe with threads on one end was cut into 6cm lengths (Fig. 5). A PVC end cap was screwed to the top of a garden stake so the pipe could be mounted atop. 12cm x 30cm sheets of Mylar were covered evenly on one side with double-sided tape, which were then wrapped around the PVC pipe. These Mylar sheaths, once mounted, were marked for magnetic north.

Both sampling methods were deployed for 1-week collection periods. When ready for collection, the exposed double-sided tape of each sample was covered with clean Mylar and labeled with permanent marker.

To measure dust accumulation, a USB microscope (widely available for ~$30 USD) was used to photograph the samples (Fig. 6). These photographs were then converted into black and white, and processed in ImageJ, a free and open-source software widely used for visual analysis (Fig. 7). By calculating the percentage of area covered (%AC) in black of the sample image, it was possible to develop a closed dataset that would yield internally quantifiable results. As can be seen in the plotted line, it was determined that %AC results for each sample stabilized after 50 images per sample (Fig. 8).

Directional sample sheets were delineated into sixteen columns corresponding to 22.5° each. %AC was calculated for each column and plotted on a rosette diagram (Fig. 9). %AC was calculated for accumulation samples and compared to the ambient dust collection samples from the flood plain.

The results demonstrate two clear conclusions. First, the directional samples consistently showed concentrations of dust in the direction of the road accounting for prevailing wind direction (measured by a directional sample placed in the floodplain), pointing to the road as the principal source of local mobilized dust. Second, when compared to the ambient collection samples placed in the floodplain, the accumulation samples all had measurably elevated amounts of dust, indicating that the dust mobilized by traffic on the road was traveling at least 250 meters (the greatest distance of a sample from the road), and rising 20 meters high (the elevation of a sample at one of the rock art sites).

Anecdotally, all these conclusions had been clearly articulated by Njanjma Rangers and other community members of Gunbalanya for years. With this simple study however, they had an evidence-based report that could be presented to officials and decision makers. By 2018, after consultations with the Northern Territory government and the Central Land Council, a decision was adopted to seal a 3km stretch of the road corresponding to the study area.

All photographs and diagrams © Tom McClintock

→ Discover more about the Rock Art Network
→ Members and affiliated institutions of the Rock Art Network