Open Cut » Environment
The objective of this project was to facilitate the transfer of project C24033 dispersive mine spoil rehabilitation outputs into practice. This included extending the application of the decision support tools to all mined land rehabilitation, and refinement of the package of decision support tools through a combination of user driven training and incorporation of additional results from structured trials and operational rehabilitation works to improve capacity for erosion risk modelling in order to facilitate improved rehabilitation outcomes.
Further development of the Bayesian decision support tool for evaluation of erosion risk has also sought to extend this tool for application in rehabilitation design by iterative assessment of erosion risk to identify the least-risk outcome among numerous design combinations.
Field assessment was carried out on the basis of considering erosion as the interaction of exposure to erosive energy and vulnerability, with factors fitting these two dimensions being encapsulated in the Universal Soil Loss Equation (USLE).
A number of digital environmental surveying techniques were applied to enable whole-of-site enumeration of proxy characteristics which could be statistically correlated with spatial variation in characteristics of interest. Tools applied for this purpose included soil electromagnetic (EM) surveying; drone and satellite multispectral surveying; and digital terrain mapping.
Complete spatial enumeration of site characteristics enabled each site to be gridded to establish “virtual plots”, and for site characteristics at every grid cell to be known. This, in turn, provided the basis for aggregation of data layers to enable regression analysis to statistically explain observed variation in the site parameter of interest (erosion) based on variation in potential explanatory parameters (including topsoil and spoil physical and chemical characteristics; vegetation cover; slope; slope length; and surface armouring).
This approach to site characterisation relies on data collection with a high level of spatial accuracy to ensure alignment between independently collected data layers. RTK-GPS was used for this purpose.
Detailed assessments were carried out at seven sites across the Bowen Basin and the western coalfields of New South Wales (NSW). Two sites incorporated structured trials. The remaining five sites assessed routine rehabilitation, with the area assessed being selected to capture a range in site characteristics and performance.
The report includes results from the structured trials (with repeat assessments over time) demonstrated specific effects of rehabilitation treatments.
Details of results from the routine rehabilitation sites are demonstrated in the report.
The results from detailed site assessments were used to update and refine the Bayesian belief network (BBN) model originally developed in project C24033. In this process, the original conditional probabilities of selected variables were updated to match the observed site conditions. The updated BBN model improved estimation of vegetation cover when cover was high, but overall, the updated BBN model slightly enhanced vegetation cover compared to the original BBN model. The updated BBN model produced a noticeable improvement in the prediction of surface erosion risk for all sites, predicting higher surface erosion when observed gully erosion was high (compared to the original model) and did not underestimate surface erosion for any observed scenarios. There was no case that the model failed to predict surface erosion. This indicates that the updated model might be considered as a safe model for spoil erosion assessment.
Workshop presentations to industry were held over the course of the project and held at five physical locations across New South Wales and Queensland, plus one on-line presentation. Use and application of the erosion risk model has been communicated to industry through the workshops; the recording of the on-line workshop presentation available on the project website; and the detailed “Instructions” sheet in the data entry form user interface available from the project website at: https://mine-rehabilitation.net.au/.
Each scenario explored in the model can simulate erosion outcomes for only one area at a time. However, for each area, the model can evaluate up to 20 scenarios simultaneously.
A limitation of the modelling approach adopted is that changes over time cannot be directly accommodated, however the model can be used to explore variation in erosion risk over time by varying vegetation cover specified in the input parameters based on targeted changes in cover, assuming all other parameters remain unchanged.