Impact of fire on the Weipa Bauxite, northern Australia

Abstract

More than half of western Cape York Peninsula experiences fire every dry season, and the effects of this on the bauxite are twofold: gibbsite is dehydrated to boehmite or alumina and Fe-oxyhydroxides are converted to maghemite. Effects are most significant on the earthen materials of termite nests, particularly those coating the trunks of the common Eucalyptus tetrodonta (Darwin stringybark), and where dead trunks, branches and roots have burnt. Fire-induced dehydration of ooliths in termite nests is suggested as the source of the high-boehmite redsoil in the Weipa Bauxite deposit.

Key Words:

• boehmite
• Cape York
• fire
• γ-alumina
• gibbsite
• maghemite
• redsoil
• regolith

Introduction

Taylor et al. (2008), reported a decrease with depth through the bauxite in the boehmite:gibbsite ratio, in the concentration of poorly-diffracting material and in magnetic susceptibility at Jacaranda, Andoom, and suggested that these changes with depth were the result of fire. Increasing boehmite content toward the surface is not uncommon in bauxite deposits from seasonal tropical regions (Tardy et al. 1991; Paquet & Clauer 1997) and this distribution of boehmite and gibbsite is attributed to high surface temperatures and dry conditions during the dry season allowing gibbsite to dehydrate to boehmite, or establishing thermodynamic conditions where boehmite is the crystallising phase rather than gibbsite.

Records of fire on Cape York date back to 1623, with detailed records available at least since the early twentieth century (Crowley & Garnett 2000). On average almost 60% of the region is now burnt annually, with most fires occurring relatively late in the dry season (September and October) (Felderhof & Gillieson 2006). Figure 1 shows the fires that burned during 2005 in central Cape York.

Figure 1 Fire scar map for fires during 2005 on central Cape York (courtesy Chris Devonport, Ecobyte Systems Pty Ltd and CRC for Tropical Savannas).

Williams et al. (2004) measured the temperature in the soil during low-intensity fires in eucalypt savannah at Cape Cleveland near Townsville. The surface reached a temperature of about 200°C but there was no temperature increase below a depth of about a centimetre. Late dry-season fires were more intense and heated the top 5 mm to 200°C with slightly elevated temperatures reaching to 30 mm. Similar effects can be expected in the eucalypt forests of western Cape York Peninsula.

In the Highlands of New South Wales, Raison et al. (1986) recorded temperatures as high as 700°C below the leaf litter, but the maximum at 20 mm depth was only 94°C. Tunstall et al. (1976a) found that a grass fire in central north Queensland could heat the first metre of the trunks of trees to over 600°C, with average maximum temperatures of about 250°C. From these observations it seems there is sufficient heat in a passing bushfire to significantly alter the mineralogy of both surface soil and tree-sheathing termite nest material.

Under a large burning log or in a tree stump, much higher temperatures can be expected. Tunstall et al. (1976b) measured the temperature and soil moisture changes under a large structured pile of burning logs 10 m long, 2 m wide and 1.4 m high. Soil-surface temperatures at the centre of the pile reached 800°C, and immediately below the pile centre at a depth of 40 cm soil temperatures reached 100°C. Maximum temperatures reached at least 250°C at a depth of 5 cm at the centre of the pile and on the surface up to a metre from the pile edges. At the centre of the pile the soil lost almost all its moisture to a depth of 30 cm and over half its moisture at the surface at a distance of 1 m from the pile. Such an intense fire can certainly change bauxite mineralogy.

There is a voluminous literature on the effect of heat on gibbsite and boehmite because of the technological importance of the various dehydration products. Brindley & Choe (1961) found that fine-grained gibbsite (<1 μm) dehydrated to χ-alumina, whereas coarser gibbsite formed boehmite. Battacharya et al. (2004) using 1.5 μm synthetic gibbsite found it started to transform to boehmite at 250°C and the conversion was complete around 350°C. At about 500°C, χ-alumina formed from the boehmite. Kogure (1999) transformed very thin gibbsite crystals by electron beam heating to χ-alumina which after further heating transformed to a spinel phase, either η- or γ-alumina. Although there appears to be a variety of intermediate alumina species produced when gibbsite or boehmite are heated (Sato 1987), most authors find that gibbsite either dehydrates to χ-alumina, or to boehmite which itself transforms to γ-alumina.

Under experimental conditions synthetic boehmite dehydrates to γ-alumina, with the temperature of transformation dependent on particle size. Tsukada et al. (1999) found 5 nm boehmite crystals transformed at 450°C, whereas 20 nm material transformed at about 500°C. Wilson (1979) reported 80–90% conversion after firing at 400°C for one week.

Materials and Methods

Samples of bauxite and of the soil over bauxite were taken from regions that had recently experienced fire, as evidenced by the immediate presence of burnt wood and charcoal. These included bauxite surrounding a burnt root, Grunter pit, East Weipa, termite nest material from inside a hollow, fallen, tree branch, termite nest material from the outside fire-blackened bark on E. tetrodonta, and bauxitic gravel on the surface at The Breakaway, east of Weipa, following a fire (localities are given in Eggleton et al. 2008). Samples of gibbsitic bauxite were held at 40°C for 36 months, at 60°C for 15 months and at 110°C for 10 months in laboratory furnaces.

Results

Prolonged heating can be expected underneath and within burning fallen logs. XRD of termite nest material from inside a fallen burnt log ( Figure 2a, termite species unknown) showed the major broad peaks of γ-alumina and yielded a mineralogical analysis: poorly-diffracting material (largely γ-alumina) 86%, hematite 7%, quartz 4%, anatase 3%, gibbsite 2%.

Figure 2 (a) Burnt log with termite earth within, Andoom. (b) Burnt root in bauxite. Bench is 3 m high.

Burning roots cause deeper parts of the bauxite to be heated ( Figure 2b). Simple field examination of the pisoliths from a root burn in the Grunter pit, East Weipa showed that most had vitreous cores, indicative of nanocrystalline alumina (Tilley & Eggleton 1996). XRD analysis of the bauxite at this root burn yielded 31% poorly-diffracting material, 21% gibbsite, 18% kaolinite, 15% quartz, 7% boehmite and 3% hematite. Individual pisoliths have a hard vitreous core and a cortex which is softer and has a dull or earthy lustre. The composition of selected fragments of vitreous material, though with some cortices adhering, is 60% poorly-diffracting material, probably alumina, 20% gibbsite and 10% kaolinite. The cortices, again with adhering vitreous material, contain 45% poorly-diffracting material, 37% gibbsite and 13% kaolinite.

Significant heat probably results from scorching the bark of E. tetrodonta by even low-intensity fires. Many of these trees are sheathed with bauxite ooliths and fines as a result of the constructions by Schedorhinotermes spp., and the heat from the burning bark might be sufficient to convert gibbsite to boehmite. Such a sheath at Andoom in an area that had not been recently burnt had 26% gibbsite, 20% boehmite, whereas the earth around the trunk of a recently burnt tree had 11% gibbsite, 29% boehmite.

Samples heated in the laboratory to 40°C for 36 months, at 60°C for 15 months and at 110°C for 10 months showed no weight loss.

Discussion

Taylor et al. (2008) noted a decline in magnetic susceptibility with depth through the bauxite at Jacaranda, Andoom and that some pisoliths were sufficiently magnetic to be picked up by a hand magnet. As has been commonly reported (Fitzpatrick 1985; Ketterings et al. 2000; Grogan et al.2003), maghemite is formed from ferrihydrite, goethite or hematite during bush fires via a reduction/oxidation sequence involving organic matter and the iron minerals.

Exactly paralleling the magnetic susceptibility decline with depth is the change in boehmite:gibbsite ratio (Taylor et al. 2008). Surface temperatures in the tropics can reach 50°C during the heat of the day (Shirtliff 2007) but such high temperatures do not extend to any significant depth. Laboratory heating to 60°C had no apparent effect on gibbsite over a period of 15 months. Three mineralogical changes have been identified to be the result of fire in the Weipa bauxite: the production of maghemite, the dehydration of gibbsite to boehmite, and the conversion of both gibbsite and boehmite to various forms of alumina. The major effect, in terms of amount of material altered is the production of boehmite; locally, alumina may be the main alteration product, as for example in the sampled root burn. Taylor et al. (2008) showed that the boehmite content of the bauxite decreases with depth, and suggested that the boehmite content of each size fraction of the bauxite could be partly explained as a thin boehmite-rich coat over a more gibbsitic pisolith. We here suggest that fine boehmite, produced by fire at the surface, washes down into the bauxite to become part of the matrix that eventually coats the pisoliths. Larger boehmitic particles, being stronger than gibbsitic particles, survive transport by rainwash better and accumulate in topographic lows as redsoil.

We do not consider that the boehmite at Weipa crystallised in preference to gibbsite simply in response to low water activity caused by tropical heat (Tardy 1997) nor by climate-induced dehydration of gibbsite (Tardy et al.1991). In terms of distance from the heat of the sun, redsoil and the rest of the bauxite are essentially identical, and there seems to be no reason why only ooliths in redsoil should be susceptible to boehmitisation as a result of solar effects. On the other hand ooliths that have experienced fire and been boehmitised either in the topsoil or in tree-sheathing termite nest material, might readily be transported into lower parts of the plateau and accumulate as redsoil. It is noteworthy in this context that the bauxites of the Los Pijiguaos deposit of Venezuela (Soler & Lasaga 2000; Meyer et al. 2002) apparently contain no boehmite despite having developed in a seasonally dry climate similar to that of Weipa. The forest at the Los Pijiguaos deposit is rainforest, and possibly has not experienced frequent fire: certainly satellite images of the region show no fire scars such as are abundant on western Cape York Peninsula. By contrast, the bauxites of Sangaredi, Guinea are similar to those at Weipa in having increasing boehmite toward the surface (Boski & Herbosch 1990). The climate there is similar to the Weipa climate, having 5 months of no rainfall followed by about 2 m of rain in the wet season, but unlike Los Pijiguaos, fire is common on the savannah shrub and grasslands of the bauxite plateaux.

In some places poorly-diffracting material, now identified as largely tohdite, η- or γ-alumina, comprises as much as 20% of the upper parts of the bauxite. Poorly-diffracting material is produced by intense heating of gibbsite or boehmite pisoliths, ooliths and fines. Poorly-diffracting material-rich pisoliths are dense and very hard, and are resistant to breakage. They could readily accumulate near the surface by a transport process that winnowed the more fragile particles. The increase in poorly-diffracting material around 2 m in the Grunter bauxite profile (Taylor et al. 2008) provides evidence for crude stratification in the deposit. The presence of both kaolinite, which breaks down above 450°C, and γ-alumina, which forms between 450° and 500°C, places some constraints on the temperatures reached inside the burnt log sampled.

Eggleton & Taylor (2008) showed that the biota of the Weipa Plateau have a homogenising effect on the bauxite deposit. Opposing homogenisation is the effect of fire. Fire dehydrates gibbsite and goethite, most notably near the surface, but burning roots may penetrate through the bauxite. As is obvious from the stands of eucalypts across the plateau and as evidence from the mine pit faces shows, no part of the bauxite has escaped the influence of tree roots and consequently of fire as it slowly burns down dead tap roots. Very probably the production of boehmite or alumina is reversed over time in the lower, wetter parts of the bauxite, but the overall effect is one of increasing dehydration toward the surface. The presence of abundant boehmite toward the top of the deposit is of profound significance in the production of alumina, because it requires a higher temperature process to dissolve it in caustic than does gibbsite. The presence of γ-alumina and other poorly crystalline forms probably also affects processing but in a manner not recognised.

Although we present data that demonstrate the occurrence of the temperatures required to convert minerals by heat, none take time into account. Individual fires are relatively quick, but roots can burn for months. Similarly experimental results do not demonstrate that multiple heating to bushfire temperatures may be important, yet most surfaces on the Weipa Plateau are burnt on a regular basis over relative short time spans. Until such experiments replicate natural conditions more closely we believe their results are only indicative. Similar observations by others on deep weathering profiles and bauxite leads us to believe fire is important in the formation of maghemite, boehmite and poorly-diffracting material.

Conclusion

Frequent wildfires across the Weipa Plateau dehydrate the gibbsite to boehmite or alumina in the surface layer of soils, in bauxite fines used by termites for above-ground nest construction, and at depths through the bauxite when dead roots burn.