Despite more than 20 years of research to improve practices for rehabilitation of bauxite-mined land at Weipa, the goal of establishing stable, diverse plant communities in the mined areas is not always achieved. The problem is that introduced plant species sometimes fail to establish, leaving mined areas covered with native grasses, weeds or, in worse-case scenarios, devoid of vegetation. Previous field studies at Weipa suggested that a 45% reduction in surface-soil organic carbon (C) and nitrogen (N) in the first year after mining may be a key factor affecting revegetation. Organic matter stores and cycles most of the N and phosphorus (P), as well as other plant nutrients, in the shallow, highly-weathered soils covering the bauxite ore at Weipa. Since concentrations of soil N and P have been shown to limit the growth of both native and introduced plant species in undisturbed Weipa soils, the links between soil disturbance, organic matter decline and revegetation failure may be nutritional. The amount of soil disturbance is determined by the method used to remove and replace soil in mined areas, termed soil stripping and replacement. This project investigated effects of stripping and replacement and subsequent soil and plant management operations on soil organic C and N dynamics, soil nutrients, soil physical quality and vegetation establishment, through (i) a survey of existing rehabilitation areas ranging in age from 1 .5 to 22.5 years, (ii) a field experiment, (iii) an in situ soil incubation, and (iv) a CENTURY model simulation.
In the survey, I compared the two currently-practiced methods of soil stripping and replacement; ‘dual-stripping' (DS), where tandem scrapers attempt to simultaneously strip the soil in two layers (0-30 cm and 30-60 cm) then immediately replace them in sequence on an adjacent mined area, and ‘stockpile' (ST), where soil stripped with little regard to the maintenance of soil layers is stockpiled for several years before being respread on a mined area. There was more organic C in surface soils (0- 10cm) of DS ( 17.1 t C ha-1) compared with ST areas ( 14.5 t C ha-1), but all rehabilitated soils contained less organic C in the surface than undisturbed (unmined) soils (33.8 t C ha-1 ). Trends for other indices of organic and chemical fertility were similar. However, when the whole soil profile (0-60 cm) was considered, amounts of organic C were: undisturbed soil (98.2 t C ha-1) = DS soil (79.4 t C ha-1) > ST soils (61.6 t C ha-1) (p < 0.05). The apparent decline in surface-soil organic matter with DS resulted from mixing soil horizons, i.e. dilution of the higher organic C topsoil (Al horizon) with subsoil (B1 horizon), bauxite and ironstone material. With time, organic matter accumulated in the surface soil (average rate of 0.5 t C ha-1 yr-1 in DS soils). Step-wise regression of organic C, soil physical properties and indices of vegetation-system development (collected outside this project) showed that, after effects of site age were removed, soil compaction was the major factor affecting revegetation.
The field experiment compared the DS and ST methods with two additional stripping and replacement methods that represented extremes of soil management practice, viz. ‘subsoil only" (SU), where only subsoil was replaced, and ‘double-pass' stripping (DP), an alternative technique of segregating soil by layers (0-20 cm and 20-60 cm). Organic C content in the surface soil (0- 10 cm), measured soon after soil replacement, was in the order; DP (24.3 t C ha-1) ≥DS (1 4.6 t C ha-1) > ST (8.9 t C ha-1) ≥ SU (8.7 t C ha-1), compared to 28.9 t C ha-1 in undisturbed soil. There were no significant differences in organic C to a depth of 60 cm between the DS (90.6 t C ha-1), DP (105 t C ha-1) and the undisturbed soil (89.6 t C ha-1). Both surface soil and whole profile results were consistent with those of the survey, reinforcing the notion that dilution, not accelerated oxidation, accounted for the large apparent loss of surface-soil organic C from D S soils. However, soil organic C and total N did decline by 8- 13 %, averaged over all treatments, in the first two years after soil replacement.
Native grass biomass, measured one year after soil replacement in the field experiment, closely reflected treatment differences in initial surface-soil organic C quantities, with DP (3.4 t ha-1) > DS (2.3 t ha-1) > SU (0.7 t ha-1) = ST (0.7 t ha-1) (p < 0.05). The similarity of these trends indicates a concentration of grass seed in the undisturbed topsoil along with organic matter. Densities of sown plant species showed the opposite trend, viz. DP (9,600 plants ha-1) s; DS (1 2,000 plants ha-1) < SU (1 8,000 plants ha-1) s; ST (23, 000 plants ha-1) (p < 0.05). Sub-treatments in the field experiment included timing of stripping and deep-ripping, use of disc-ploughing, fertiliser rate, vegetation type (native species, introduced grass-legume pasture) and planting type (direct sown or hand-planted tubestock). Of these, disc ploughing had the greatest effect, reducing grass biomass by 3 6% and increasing sown-plant density by 50% and sown-plant diversity by 58%.
To more accurately quantify short-term effects of soil stripping and replacement on soil organic matter, I measured changes in soil organic matter in situ for 12 months after disturbance. Topsoil (0 - 15 cm), subsoil ( 15 - 60 cm) and a mixture of the two (mixed soil) were replaced into 15-cm diameter cylinders, inserted to a depth of 60 cm in an area of newly replaced soil, and kept free of vegetation. Soil replacement treatments were based on the DS (mixed soil only, and topsoil replaced at the bottom of the cylinder under subsoil), and DP (topsoil at the surface over subsoil) methods. In one year, organic C declined by 12% (subsoil), 18% (mixed soil) and 26% (topsoil), with slightly greater decline in the top layers (0 - 10 and 10 - 20 cm) of the replaced profiles. Despite a temporary increase in microbial activity during the wet season, microbial biomass C in the 0 – 10 cm layer declined by 19%, 34% and 61 % in topsoil, mixed soil, and subsoil media, respectively. These data represent the extreme scenario of organic matter decline resulting from stripping and replacement, i.e. severe soil disturbance accelerates the oxidative C loss, and the absence of vegetation means nil C and N input to the soil organic matter.
Long-term trends in soil organic C and other indices of organic fertility under current and alternative rehabilitation practices could not be discerned from either the field experiment or the survey. I used the CENTURY model to simulate long-term soil and vegetation outcomes of the four soil stripping and replacement treatments of the field trial (DS, ST, SU and DP). After 100 years simulation, predicted total organic C in the surface soils (0 - 20 cm) of each treatment had risen to a new dynamic equilibrium. Since the ‘passive' pool of recalcitrant organic C, which occupied 47% of total organic C, changed little over the simulation period, the new equilibriums differed according to initial total organic C contents. Most organic matter recovery occurred in the 'slow' fraction, although the greatest rate of change occurred in the ‘active' C pool which stabilised after 50 years at levels not dissimilar to those found in the native forest. However, the successful development of a new equilibrium was highly sensitive to the amount of N2 fixed by legumes in the system and to the degree of severity of C and N losses during fire events. Reasonable agreement was found between simulated organic C accumulation and that observed in the survey for the first 15 years of rehabilitation (r2 = 0.39 for DS and r2 = 0. 77 for ST).
I concluded that competition from volunteer grasses, not reduced quantity or quality of soil organic matter, was the key factor limiting establishment of sown-species. There was little evidence of positive relationships between indices of organic or nutritional fertility and sown-species establishment and productivity. In fact, soils with higher organic fertility fared worse in terms of density of sown-species, a coincidence of organic-enriched topsoil being the primary source of native grass seed. High grass biomass, however, enhanced recovery of surface-soil organic matter and ameliorated compacted surface soil. This amelioration may be critical for long-term vegetation success since soil compaction was identified as a factor limiting development of sown-species in the survey. My results showed that the commonly-practiced "dual-stripping' method mixed, rather than segregated soil horizons. This, however, may not be a problem in the short- to medium-term, as results of the field experiment indicated that topsoil segregation (DP) had no benefits over a mixed soil profile (DS) in terms of vegetative establishment and amelioration of soil compaction. It remains to be seen whether the presence or position of topsoil organic matter within replaced soil profiles has any influence on the long-term survival, growth arid stability of desirable plant species, either through effects on the availability of plant nutrients or through improved soil structure.