Supervising Scientist Division

Supervising Scientist Annual Report 2003 - 2004: Rehabilitation

Supervising Scientist, Darwin, 2004
ISBN 0 642 24391 3
ISSN 0 158-4030

3 - Environmental research and monitoring (continued)

3.2 Rehabilitation

This programme provides advice on the implementation of mine rehabilitation strategies and assessment of rehabilitation sustainability. An important issue is the establishment of a process to bring together and gain consensus between stakeholders on agreed rehabilitation completion criteria.

Key activities in 2003-04 include:

3.2.1 Development of methods to assess revegetation success at Nabarlek minesite and their application to Ranger mine

Nabarlek is the first uranium mine in Australia to be rehabilitated under a contemporary regulatory regime and so exemplifies many issues relevant to the future rehabilitation of Ranger uranium mine.

Figure 3.1: Comparison of open grassland-shrubland dominated communities on the Nabarlek minesite with surrounding natural eucalyptus savanna woodland, September 2003

Figure 3.1: Comparison of open grassland-shrubland dominated communities on the Nabarlek minesite with surrounding natural eucalyptus savanna woodland, September 2003

The success of revegetation at Nabarlek, however, cannot be quantitatively assessed from previous studies. Hence, research started in August 2003 to: (i) develop cost-effective ground and remote sensing monitoring and assessment methods for revegetation that can be applied to Ranger; and (ii) provide a robust quantitative assessment of revegetation success based on a comprehensive characterisation of soils and vegetation across the minesite compared with adjacent natural reference sites.

Nabarlek minesite and its surrounds were zoned to encompass variations in ground surface features. Vegetation ground cover and canopy cover attributes were recorded along each transect in September 2003 (late dry season) and April 2004 (late wet season), and soil samples were taken in the dry for analysis of soil properties. High resolution Quickbird satellite images of the minesite and surrounding landscape were captured at the same time as detailed ground surveys to compare the benefits and costs of each survey method in terms of monitoring and assessing revegetation success.

A multivariate statistical model (Figure 3.2) was developed using the dry season ground survey data to explore soil-plant relationships across all sites, and provides a simple tool to help assess revegetation by using soil and plant variables as indicators of success. Soil-plant relationships for all sites are graphically presented along a successional gradient, ranging between poor vegetation-poor soil to poor vegetation-good soil on minesites, to good vegetation-good soils on unmined reference sites. Although such characterisations appear qualitative and subjective, they are underpinned by quantitative and objective criteria. Results show clearly that at least half of the minesites sampled are classified as poor vegetation-poor soil sites, indicating that progress in revegetation is unsatisfactory. Vegetation on the minesite is characterised by a dominance of grassy weeds entailing extreme fire risk due to high biomass fuel loads; acacia shrublands nearing the completion of their successional development; and a conspicuously very low density of necessary trees. The last point suggests that the abundance of trees and shrubs is unlikely to increase in the short term, and may decrease as a result of woody seedling losses due to fire.

Additionally, the low-lying evaporation pond areas have poor soil development in terms of high nutrients and proneness to waterlogging, which inhibit growth and survival of woodland plant species and are reflected in poor vegetation development.

Figure 3.2

Figure 3.2 Multivariate statistical model: Ordination (Principal Components, PC) of plant-soil relationships across all sample transects on Nabarlek minesite (solid squares) and surrounding natural reference sites (solid circles), September 2003

Hence, despite this being the ninth year since revegetation commenced at Nabarlek (in late 1995), it has been unsuccessful for at least half the minesite because of poor vegetation and soil development. Even so, the remainder of the minesite appears to be trapped in the good soil-poor vegetation characterisation because vegetation development may be perpetually constrained by a range of disturbance factors such as weeds and fire. Significant management intervention, therefore, is needed if these sites are to become self-sustaining vegetation communities analogous to reference sites.

The Nabarlek Rehabilitation Workshop (2000) felt the original revegetation objective - of blending in with the surrounding eucalypt savanna woodland - had, five years after its commencement, been unsuccessful. This position has not changed since then. Experience at other rehabilitated minesites in the tropics indicates that ecosystem resilience can be obtained in less than five years, and that an integrated landscape with land use can be obtained in ten to fifteen years. A more comprehensive assessment of revegetation success will be made with the inclusion of recent wet season survey data and the results of two Charles Darwin University postgraduate studies which examined soil-plant relationships in detail.

The Supervising Scientist Division has been developing remote sensing methods to assess rehabilitation success of minesites for several years. The major advantage of remotely-sensed data is that a complete coverage of the site and surrounding areas can be acquired, facilitating a whole-of-landscape assessment. Despite the huge sampling effort invested in the ground-based vegetation surveys, only 0.07% of the variable minesite was sampled. A key finding from ground-based surveys at Nabarlek is that 80% of the variability in all observed soil and plant data used to assess revegetation is explained by total canopy cover, itself an index of canopy density. Hence, a rapid and cost-effective means of measuring total canopy cover, such as from remote sensing captures, may provide a powerful complementary tool to help monitor revegetation success. Quickbird satellite data at 2.5 m spatial resolution, spectrally covering the visible to near infra-red, has recently become commercially available. Such data are being used to measure and compare tree canopy cover and ground cover, including weeds, on both the minesite and surrounding landscape. Preliminary results show promise in discriminating weed species and technology is now being assessed for weed mapping at the Ranger minesite.

Future studies at both Nabarlek and Ranger minesites will focus on soil-plant relationships, especially in relation to soil constraints to plant growth and survival; soil seed bank of weeds and native species; and ecological interactions between weeds, fire and native plant succession. Management options to complete revegetation at Nabarlek to the satisfaction of all stakeholders are now being considered by the Nabarlek Minesite Technical Committee and the Alligator Rivers Regional Technical Committee. Practical, pragmatic and defensible options will be explored in partnership with Nabarlek Traditional Land Owners and other key stakeholders, and will incorporate continued quantitative monitoring and assessments of revegetation success and new research knowledge on soil-plant relationships. Critical to the success of achieving any new revegetation goal is the strengthening of existing partnerships between government, industry and indigenous communities.

3.2.2 Development of improved techniques to monitor suspended sediment movement around Jabiluka

As part of a long-term study of the impact of mining at Jabiluka on the Ngarradj catchment, it is important that suspended sediment transport is monitored within the catchment. Of particular importance is the mud component of suspended sediment, as nutrients and contaminants, including heavy metals and radionuclides, are primarily transported in association with mud. Streamflow and mud concentration data have been collected within the Ngarradj catchment at sites upstream and downstream of the mine since 1998. These data have been used to develop an event-based mud transport model for the two sites to understand baseline mud movement in the catchment for future mine-related impact assessment.

Event mud loads associated with runoff events observed during the 2003-04 wet season at the stations upstream and downstream of Jabiluka were plotted against the calibrated mud transport models (Figure 3.3). The calibrated mud model is shown as the 1:1 line and the trigger levels associated with the mud transport model represent different levels of intervention by supervising authorities and the mining company. Almost all of the event mud loads observed downstream of the mine during 2003-04 are elevated above that of the 1:1 line. Event mud loads observed upstream of the mine during the year all fall well within trigger levels associated with the calibrated model and lie around the 1:1 line. In other words, elevated mud loads were observed downstream that did not occur upstream. The evidence supports the inference that the increase in mud loads at the downstream site is a result of erosion on the minesite following activities carried out to manage the site on a long-term care and maintenance basis. While it is considered that this increase in mud loads is not of major concern, given the majority of these events do not plot above the first trigger level associated with the model, the results demonstrate that the monitoring method developed by eriss is sensitive to quite small changes.

Figure 3.3

Figure 3.3 Event mud loads for 2003-04 at sites upstream and downstream of Jabiluka plotted against the fitted mud transport models (1:1 line). Associated trigger levels are indicated as dashed lines.

It is considered that this event-based monitoring technique is valid and sensitive enough to apply to mine-impacted catchment areas surrounding the Ranger mine. It is recommended that event runoff and mud concentration data are collected within the Ranger catchment areas over the next few years to calibrate mud transport models similar to that derived for the Ngarradj catchment. These models would then be used to assess rehabilitation success of the Ranger mine.