Ant community response to management practices on rehabilitated mine sites

In Australia, there is a requirement to rehabilitate land altered by mining operations. The recovery of such landscapes can be long and protracted due to the often large disturbance caused by mining. It is important to monitor ecosystem recovery to ensure it is progressing towards the pre-determined end-use. In Queensland, monitoring criteria are predominantly based on vegetation, soil physico-chemical conditions and landform stability. To ensure recovery is efficient, management practices are often conducted to improve individual aspects of the rehabilitated ecosystem. These management procedures typically target vegetation parameters (such as increasing biodiversity) and can cause secondary disturbances to the system. Generally, minimal attention is paid to the impact of such secondary disturbances on other biota. Ants have been widely used as bioindicators to assess the impact of land management practices and, in particular, the rehabilitation of mine sites. However, there is limited published research on the influence of secondary disturbances on ant communities at recovering landscapes such as those presented post-mining. The question thus arises as to whether ant communities in these systems are more sensitive to subsequent management practices compared to those occurring on land not previously disturbed by mining. In this research, ant communities were examined as bioindicators of the assessment of management practices or secondary disturbances on two rehabilitated mines. Firstly, at a rehabilitated sand mine on North Stradbroke Island, the differences between ant communities at sites subjected to clearing, prescribed burning and/or wildfire were studied. Unmined habitats were also examined to determine whether the post-rehabilitation disturbances were facilitating ant succession to a composition similar to that in surrounding natural areas. Data from approximately two years after the last secondary disturbance showed that these sites contained a simple ant community composed primarily of non-specialised species, typical of disturbed but unmined habitats. Despite this basic community structure, many specialist and late successional ants were also present. Data collected at three years post-disturbance showed that ant composition at managed rehabilitation sites increased in similarity to the unmanaged rehabilitation site, suggesting the impacts of the management practices were short-term. Regardless of this, multivariate analyses indicated that the ant communities in the rehabilitation areas remained substantially different to those in the unmined native habitat. Therefore, the various management practices imposed did not facilitate convergence to the native ant composition. This may be because not enough time has elapsed since rehabilitation or it may be due to differences in vegetation composition and habitat structure. It is recognised that it may be unrealistic to assume that ant communities at the rehabilitated sites will converge with those in unmined areas due to the intrinsic nature of the biophysical alterations resulting from mining. At a second location in central Queensland (Norwich Park Coal Mine), the ant communities on rehabilitated pastures with different grazing histories were investigated and compared to the ant fauna of nearby unmined sites. Results showed that there was minimal disparity in the ant characteristics assessed between rehabilitated pasture with no, low, medium and high levels of cattle grazing. However, the cattle had been removed from paddocks approximately two years prior to ant sampling, possibly allowing sufficient time for both vegetation and ant communities to restructure. A rehabilitated pasture that had continuous very low grazing up until ant sampling, showed a composition similar to the other rehabilitation sites, but with additional species that were otherwise found only at the unmined reference sites. Analysis of vegetation and habitat variables suggested that the composition of ants was influenced by the cover of sward-forming grasses, particularly buffel. Such cover decreased the amount of bare ground and insolation (affecting thermophilic species) and increased vegetation complexity at ground level (influencing mobility of large species). Habitat alterations created by management practices are suggested to be the primary influence on the ant fauna examined in this research. Ant composition was shaped by habitat structure at both ground level and in the mid- to upperstorey, a response similar to that often observed following disturbances on unmined land. This may indicate that the ant communities in the studied post-mining rehabilitated areas have similar resilience to ants on unmined land subjected to the same management procedures. Additionally, trends in ant parameters were not precisely reflected in vegetation variables, suggesting that more than vegetation surveys in rehabilitation monitoring programs are required for a comprehensive understanding of rehabilitation progression and success. In addition, this thesis examined aspects of ant sampling methodology. Although there has been a call for standardisation for many years, this has not been successfully achieved. Therefore, a component of this project included an extensive literature review and evaluation of sampling methods, leading to recommendations for future standardisation. A methodology trial to determine an appropriate density and sample size of pitfall traps was also conducted. The resulting methodological and procedural advancements provided by this body of work should validate the need to foster greater attention on improving the efficiency and accuracy of ant research