Printer Friendly

Methods for assessing mine site rehabilitation design for erosion impact.


Surface mining is a major industry in Australia. Without proper management and design, above-grade waste rock dumps (WRDs) and voids have the potential to cause severe environmental impacts. One such impact is the pollution of waterways through erosion of post-mining landforms, including WRDs, and movement of sediment into streams and rivers. Post-mining water pollution can be reduced by designing post-mining rehabilitated landforms in such a way that erosion will be minimised. The ability to quantify sediment generation from a landform, through predictive modelling, takes the `guess-work' out of landform design. Recent research has addressed the application of hydrological, erosion, and topographic evolution modelling to mine site rehabilitation (Pickup et al. 1987; Silburn et al 1990, 1991; Evans et al 1991, 1997, 1998; Evans and Riley 1994; Evans and Loch 1996; Loch and Sell 1998; Willgoose and Riley 1998). Important parts of this process are the derivation of model input parameter values and validating the underlying assumptions of the models. There are limited data available, and different models, depending on their complexity, require different forms of site-specific data for calibration. Parameter values can be determined with confidence by rainfall simulation and rainfall event monitoring (Evans et al. 1998; Willgoose and Riley 1998), and validation studies (G. R. Hancock, G. R. Willgoose, J. R. W. Bell, K. G. Evans, D. R. Moliere, and M. J. Saynor unpubl. observ.) demonstrate that models can be applied to mine site waste rock material.

There are various models, many developed in the agricultural arena, which can be applied to a variety of erosion-management design aspects. The choice of model is controlled by a combination of factors such as likely environmental impact, cost of data collection for model parameter value derivation, process regime, and rehabilitation standards. At the less complex end of the scale, if a standard requires a maximum allowable sediment loss from rehabilitated areas, then the empirical Revised Universal Soil Loss Equation (RUSLE) (Renard et al 1994) gives a good indication of soil loss rates applicable in the initial stages of assessment and design. At the other end of the scale, the complex topographic evolution model, SIBERIA (Willgoose et al 1989), can be used to assess gully development and incision, and landform containment design. The variety of applications of erosion and hydrological modelling are discussed below.

The Chemicals, Runoff and Erosion from Agricultural Management Systems (CREAMS) (Knisel 1980), Kinematic Wave Design of Contour Bank/Waterway Systems (KINCON) (Connolly and Barton 1990), and Areal Nonpoint Source Watershed Environment Resource Simulation (ANSWERS) (Beasley et al 1980) models were used to evaluate the proposed WRD for the Coronation Hill mine in the Northern Territory (Silburn et al. 1990, 1991). The effects of runoff and erosion from batters, benches, and the re-shaped haul road channel on the WRD were simulated and the results indicated that rock-mulching and revegetating batter slopes and reducing batter slope would reduce sediment loss from the WRD.

Evans et al. (1997) derived CREAMS erodibility and surface roughness parameter values for mine spoil from Central Queensland coal mines and derived a site-specific relationship between the parameters and slope gradient. Application of the techniques described would allow refinement of sediment-loss predictions for varying slope gradients.

Evans and Loch (1996) used RUSLE parameter values to explain unexpected erosion rates from sites of different slope gradient under rainfall simulation at the ERA Ranger Mine. It was concluded that rehabilitated sites must be carefully managed and traffic restricted.

Rainfall simulation and modelling have been used at the Northparkes Mine, NSW, to assess the stability of a sound bund constructed of waste rock material. Loch and Sell (1998) used a combination of the Productivity, Erosion, and Runoff, Functions to Evaluate Conservation (PERFECT) (Littleboy et al 1992) model, the Grass Production (GRASP) (McKeon et al 1993) model, and the Modified Universal Soil Loss Equation (MUSLE) (Onstad and Foster 1975) to simulate plant growth, water balances, runoff, and erosion. Simulations showed that plant growth and surface cover peaked in October, and the rainfall erosivity peaked in December and remained high for some months. The results showed that careful selection of revegetation species should ensure cover during the peak erosivity periods. Loch and Sell (1998) also used KINCON to simulate flow for a range of channel gradients, length, and vegetative roughnesses. Results showed that only a small number of drop structures were required to drain benches between batters, and that slope length was not a major factor affecting erosion provided topsoil was used.

In cultivated areas of Central Queensland, the Groundwater Loading Effect of Agricultural Management Systems (GLEAMS) (Leonard et al 1987) erosion model was used to simulate a range of management strategies to reduce drain sedimentation through reduction of up-catchment erosion or on-farm sediment retention (Connolly et al 1999). Data were collected from a range of catchment scales, and simulation indicated that the most effective erosion-reducing strategy was a specific vegetation-irrigation combination.

With a number of successful modelling studies completed, there is now a need for `technology transfer' from the researchers to the practitioners, including regulators, mining and environmental engineers, and consultants. This paper presents methods by which modelling technology can be used in rehabilitated landform design and assessment, and extends previous studies to quantification of downstream water-quality impact. The methods are presented in the form of a case study applied to the ERA Ranger Mine in the Northern Territory of Australia (Fig. 1) but the methods have universal application. Three models are used in the assessment: SIBERIA, the Water Erosion Prediction Project (WEPP) (Flanagan and Livingston 1995), and the RUSLE. These models are described briefly, but detailed explanation of the mathematical concepts and parameter value derivation is not presented here.


Study site

The ERA Ranger Mine (Fig. 1) is adjacent to the World Heritage-listed Kakadu National Park and exploits a stratabound uranium deposit hosted by the lower member of the Early Proterozoic Cahill Formation (Needham 1988). The WRD consists of rocks from the lower member and includes carbonates, carbonaceous schists, and mica and quartz feldspar schist (Needham 1988). The waste rock is highly weatherable (Milnes 1988) and decomposes rapidly from large competent chloritic schist fragments into medium and fine gravel and clay-rich detritus. The area receives high-intensity storms and rain depressions between October and April (wet season) with virtually no rain falling during the remainder of the year (dry season). The average annual rainfall is 1480 mm.

At the conclusion of mining, ore will have been removed from two pits (Pit 1 and Pit 3) (Fig. 2). The plan for rehabilitation of tailings is that they will be stored in Pit I and Pit 3 below ground level. The final landform will be built so that only low grades of mineralised waste will be contained above ground level. The rehabilitation design must provide for the long-term containment of contaminants (i.e. over a period of a thousand years) (Wasson 1992). It must also ensure that weathering and erosion of the containment structure, in an area which experiences high rainfall intensities, do not result in the release of contaminants that would degrade the environment or aesthetics of the surrounding Kakadu National Park.


The final landform design is still being developed, and decisions on batter slopes and revegetation techniques are still to be made.

An earlier option, now discarded by ERA, was for the existing tailings dam to be rehabilitated in situ. Much study was undertaken of this so-called `above-grade' option. The outcome of the modelling studies, although no longer applicable at ERA Ranger Mine, illustrates how to assess mine site rehabilitation design for erosion impact, and has relevance to sites where contaminants may need to be stored above ground. This case study uses the above-grade rehabilitated landform (Fig. 2) proposed by Unger and Milnes (1992) as an example of the application of modelling technology.

Landform assessment

There are 3 stages (Fig. 3) that require assessment to determine if landform design meets requirements. These are landform stability, sediment delivery, and downstream water-quality impact.


An assessment of post-mining rehabilitated landform stability needs to quantify soil erosion rates and how well a structure encapsulates waste material or, in the case of ERA Ranger Mine, low-grade mineralised rock. The question is: will erosion result in gullying and expose tailings and other waste material, and allow contamination of the environment within a defined time period? The type of modelling used here can also be used to assess landform development after long-term erosion.

Once the integrity of the encapsulating structure has been confirmed, it is necessary to determine how much sediment could be delivered from the slopes of the WRD through natural catchments into downstream-receiving waterways.

The final stage is to determine whether the contribution of sediment from the rehabilitated landform to stream systems could elevate stream sediment concentrations above accepted water-quality guidelines. If the estimated impact is unacceptable, the landform should be redesigned and re-assessed. The design iterations should continue until an acceptable outcome is achieved Finally, the landform must be self-sustaining so that its integrity is retained in the future with little maintenance.

Landform stability

The model used to assess the ability of the above-grade option to encapsulate contaminants was SIBERIA. This is a sophisticated 3-dimensional (3-D) topographic evolution model simulating runoff, erosion, and deposition. It predicts the long-term evolution of channels and hillslopes in a catchment. The location and rate of development of gullies are controlled by a channelisation initiation function that is related to runoff and soil erodibility (Willgoose et al. 1989). The model solves for 2 variables: elevation, from which slope geometry is determined; and an indicator function that determines where channels exist. In this way, the evolving drainage system of a catchment can be modelled. Channel growth is regulated by an activation threshold that depends on discharge and slope gradient. A surface may initially have no gullies, but when the activation threshold is exceeded, a channel develops. The model has continued to be enhanced by G. R. Willgoose and co-workers (e.g. Moglen and Bras 1994). Although SIBERIA can model both transport- and detachment-limited sediment transport, its primary mode of use is for transport-limited environments.

Input parameter values for SIBERIA were derived (Evans et al. 1998) using rainfall, runoff, and sediment-loss data for a vegetated and ripped surface of the WRD. The use of these parameter values assumes that the whole WRD surface will be ripped and fully vegetated at the completion of rehabilitation.

SIBERIA (Version 7.05) was run on a Sun Ultra-1 Sparc workstation for the equivalent of 1000 years for the above-grade option using the derived parameters and a default gully-initiation threshold that makes gullying possible on steep batter slopes. The zero-year parameter values assume (1) that the initial surface conditions remain constant throughout the simulation period, (2) there is no temporal change in parameter values due to soil formation or ecosystem development, and (3) there is no spatial variation in parameter values over the input digital terrain model (DTM) area due to surface treatment of the WRD or the undisturbed land surface. The results can be presented in 3-D based on a 30-m grid with vertical exaggeration. The as-constructed landform after 1000 years of simulated erosion for a ripped and vegetated condition is shown Fig. 4.


Sediment transport rates were very low on the experimental site. After a 1000-year simulation, sediment movement on the landform was not obvious and cannot be seen clearly on a 3-D representation (Fig. 4). However, there is some evidence of valley development in the central depression and on the steep batter slopes. Minor deposition is visible above Pit 3 on the 1000-year output (Fig. 4). At 1000 years for the vegetated and ripped case, the maximum valley depth was 2.4 m, with the maximum deposition being 4.8 m. At 500 years, the maximum valley depth was 1.4 m, 58% of the 1000-year depth, and the maximum deposition was 2.8 m, 58% of the 1000-year deposition. For the vegetated and ripped condition, incision and deposition proceeded at a relatively constant rate, indicating that the simulated processes are almost linear with time.

A plot of the initial elevations at zero-year minus the 1000-year elevations for simulations (Fig. 5) shows erosion as positive elevation and deposition as negative. This gives a clearer indication of sediment movement. By 1000 years, valleys have formed on the vegetated and ripped surface, mostly located at the top of the steep batter slope and there is some minor incision in the central depression area.


Section A-A through the landform (Fig. 6; see Fig. 2 for section location) shows how incision through gullying can be quantified. Section A-A is taken through the tailings dam area across the top of the steep batter slope. At 1000 years, the simulations show that for a vegetated and ripped tailings cap, the maximum depth of valley incision is 2.2 m at a maximum width of [approximately equals] 60 m. This gives a side slope of the valley of about 0.073 m/m perpendicular to the direction of flow. This is a broad valley with gently sloping walls, which would not incise the encapsulated contaminants if, for example, a 5-m-deep capping layer of waste rock had been used. A final decision, with respect to the thickness of the contaminant cap, has not yet been made for erosion control.


Sediment delivery

The next step in the design or assessment process is to determine how much sediment arising from erosion on the mine site is delivered through the catchments linking rehabilitated landforms and the stream systems. When a spoil particle is eroded it is detached by processes such as rainfall impact, surface runoff, or bioturbation, but generally it is not transported directly to the water course from the upper reaches of the catchment. The particle goes through repeated cycles of detachment, transport, and deposition until, after a considerable time period, it enters the stream system. The amount of sediment ultimately delivered to a stream from a catchment is only a fraction of the sediment detached by the gross erosion of upland areas. This fraction is the sediment delivery ratio (SDR) and is defined by Robinson (1977) as `the percentage of the sediment delivered at a location on the stream system to the gross erosion of the basin', where SDR (%) is inversely proportional to catchment area and gross erosion is the sum of all forms of erosion in the contributing catchment including gully, rill, and sheet erosion.

Wasson (1992) developed an empirical relationship between sediment yield and catchment area in the ERA Ranger Mine region. As illustrated by Wasson (1992), the catchment area (all other things being equal) is expected to affect both the percentage runoff and the sediment yield, because of the greater potential for deposition of entrained sediment within the sub-catchment of larger catchments. Wasson (1992) uses the following relationship:

(1) Y = 17.0 - 1.7 ln X

where Y is the annual specific sediment yield (t/[km.sup.2].year) and X is the catchment area ([km.sup.2]). This relationship is based on 8 catchments in the region, and many of the catchments do not include disturbances of the magnitude of the ERA Ranger Mine, which would presumably give greater yields.

To undertake the assessment described here, a relationship describing the variation of SDR with area is required. The SDR values should incorporate elevated gross erosion due to disturbance within a catchment. The magnitude of the SDR for a catchment is dependent on a wide range of factors including nature, extent and location of sediment sources, slope characteristics, drainage pattern and channel conditions, vegetation cover, land use, and soil texture (Walling 1983). SDRs are catchment-specific, and there are few data that can be used to develop one for the study area. In the following analysis, the relationship fitted to data presented by Robinson (1977) and from the American Society of Civil Engineers (Anon. 1975) (Fig. 7) is used:

(2) SDR = [(8.33 - 0.51 lnA).sup.2] ([r.sup.2] = 0.996)


where A is area (ha). In this case study, the determination of sediment delivery for ERA Ranger Mine is demonstrated for one mine site catchment. This catchment is in the south-west corner of the DTM data available for the above-grade option (Fig. 2, Fig. 8). The catchment combines part of the landform with part of the catchment of an upper branch of Gulungul Creek and is referred to as the Gulungul mine site catchment (GMC). The DTM data available for this study were insufficient in that complete catchments leading to a single stream point entry could not be defined. For an analysis of the total mine site the DTM needs to be extended.


The first step in this type of analysis is to produce a contour plan of the mine site landform, adjoining catchments, and receiving streams. Individual catchments should then be interpreted and mapped onto the plan, and catchment areas determined (Fig. 2). Different areas of erosion contribution in the catchment also should be determined. In the case of the GMC, 3 different areas of erosion contribution are defined (Table 1): tailings cap, batter slope, and natural surface linking the batter slope and creek. The next step is to determine the gross erosion for each of the areas. For the ERA Ranger Mine, there have been sufficient studies to determine gross erosion through erosion-prediction modelling and this is described in the following section.

Table 1. Gulungul mine site catchment erosion rates and sediment delivery
Catchment Area Gross erosion Total catchment Sediment
portion (ha) rate erosion rate delivery
 (t/ha.year) (t/year) (t/year)

Tailings cap 40 1.2 48
Batter slope 14 34
Natural surface 62 0.6 37
Total area 116 119 42

Tailings cap

In the example chosen, the tailings cap surface comprises waste rock material covering tailings stored in a tailings dam. For this study, it will be assumed that, once rehabilitated, the tailings cap surface in the GMC will have an average slope of approximately 0.018 m/m, an average slope length of 340 m, and will be surface ripped and revegetated with 90% cover.

Input parameter values have been derived for the RUSLE for the WRD at ERA Ranger Mine (Evans and Loch 1996). The RUSLE, an empirical soil erosion model, is based on statistical analysis of erosion data collected from field plots. Application of the RUSLE is site-specific and allows for variables peculiar to an individual site. The RUSLE predicts long-term average soil losses from field areas under specific cropping and management. It cannot model deposition.

Assuming a vegetation cover factor of 0.15 for a combination of 90% vegetative cover of grasses and straw mulch (Goldman et al. 1986) and 0.3 for the control practice factor of ripping (contour strip cropping and irrigated furrows) (Mitchell and Bubenzer 1980), the RUSLE predicts an erosion rate from the tailings cap of 1.2 t/ha.years.

Batter slope

The batter slope in the Gulungul catchment has a drop of approximately 15 m, a slope of approximately 0.15 m/m, a total width along the natural surface of approximately 1421 m, and an area of approximately 14 ha. All of these values can be measured from Fig. 2. For this study, it will be assumed that the batter slope in the GMC will be rock-mulched when finally rehabilitated.

In a study of slope erosion at ERA Ranger Mine, Landloch Pty Ltd (R. Loch unpubl. observ. 1998) used the WEPP model to determine erosion per unit width of the batter slope in the rock-mulched condition. WEPP is a process-based prediction model that computes sediment transport and deposition on a landscape. WEPP divides hillslope erosion into interrill and rill erosion processes. Interrill erosion is the combination of raindrop impact detachment and lateral transport of sediment into rill flow areas. Rill erosion is the combination of detachment and transport of sediment by concentrated flows in rills (Lopes et al. 1989).

The predictions of Landloch Pty Ltd (R. Loch unpubl. observ. 1998) indicate that, for a rock-mulched batter slope with a fall height of 15 m and slope of 0.15 m/m, average annual off-slope sediment movement to the natural surface is approximately 0.024 t/m width of slope. This gives a total of 34 t/year of sediment delivered to the natural surface at the toe of the 1421-m-wide slope. Modelling assumptions applied by Landloch Pry Ltd (R. Loch unpubl. observ. 1998) were: that no flow discharged from the upper low-gradient cap surface to the batter slope (which means that erosion is due to direct rainfall runoff only), and that rill spacing on the batter slope was at 1-m intervals.

Natural surface

The natural surface in GMC is vegetated with open woodland and is part of the Koolpinyah Surface (Williams 1969), which is an undulating sandy plain covered by a skeletal soil (Needham 1988). The gross erosion from this surface can be estimated from the denudation rate. The mean denudation rate of the Koolpinyah Surface, determined from estimations by Cull et al. (1992), is 0.04 mm/year (W. Erskine and M. Saynor, unpubl. observ. 1999). Fifteen undisturbed soil cores were taken from the Koolpinyah Surface near the ERA Ranger Mine and used to determine a soil bulk density of 1.43 t/[m.sup.3]. Applying the soil bulk density to the denudation rate gave a gross erosion rate of approximately 0.6 t/ha.year.

Sediment delivery analysis summary

Erosion rates for the areas, total erosion, and sediment delivery determined using SDR from Eqn 2 are given in Table 1. Summarising the results, the GMC has a total area of 116 ha, a total erosion rate of 119 t/year, and, applying SDR of 0.35 (Fig. 7), delivers 42 t/year of sediment to Gulungul Creek at the catchment outlet of the DTM (Fig. 2).

Downstream water quality impact

Once sediment delivery has been determined, the impact of this sediment on stream water quality needs to be assessed. For this study the water-quality impact at the confluence of the outlet of the GMC and the main Gulungul Creek channel was determined (Fig. 8). This point in the catchment was chosen as an example because it is where the mine-derived sediment first enters the main stream channel. The first step in the assessment is to determine the discharge at the outlet of the GMC. This can be done using sophisticated hydrological modelling techniques, but in this case it is sufficient to determine mean annual discharge, Q ([m.sup.3]/year), from a catchment of area A ([m.sup.2]) and average annual rainfall R (m/year). This can be quantified through:

(3) Q = [C.sub.r]RA

where [C.sub.r] is the runoff coefficient. Published rainfall and runoff data (Duggan 1994) can be used to derive a runoff coefficient of 0.27 for Gulungul Creek. This is based on a catchment area of 62 [km.sup.2]. I. M. Vardavas and L. M. Cannon (1991 unpubl. observ.) calculated a runoff coefficient of 0.44 for Gulungul Creek based on a catchment area of 49 [km.sup.2], which indicates that data were collected further upstream than those of Duggan (1994). The Gulungul Creek catchment cascades from rocky escarpment country, then flows across sandy plains of the Koolpinyah Surface and enters Magela Creek in lowland areas where backflow relationships would apply. The difference in runoff coefficients is explained by the location of data collection. The larger area cited in the study by Duggan (1994) includes lowland areas of the catchment. Sedimentation is greater in lowland areas of a catchment than in rocky upland terrains. This results in a larger soil water store and lower water flow velocity for lowland areas than in upland terrains (I. M. Vardavas and L. M. Cannon 1991 unpubl. observ.). For similar vegetation and climatic conditions, smaller runoff coefficients are expected for lowland catchments (I. M. Vardavas and L. M. Cannon 1991 unpubl. observ.). Since the area cited by Duggan (1994) includes a higher portion of lowland, a smaller runoff coefficient is expected. For this case study, the mean of the values (0.36) is used for the undisturbed natural Koolpinyah Surface. This value is similar to the runoff coefficient (0.35) for the upper 600 [km.sup.2] of the Magela Creek catchment given by I. M. Vardavas and L. M. Cannon (1991 unpubl. observ.). The [C.sub.r] for the vegetated and ripped tailings cap is 0.1, and for the batter slope [C.sub.r] is 0.41 (Evans 1997). For an average annual rainfall of 1480 mm for the study area and weighting the [C.sub.r] values for each area (Table 1), the estimated mean annual discharge at the GMC outlet is 474 ML.

Dividing the annual quantity of sediment delivered to the GMC outlet (42 t) by the estimated average annual discharge at the outlet (474 ML) gives an average annual sediment concentration (89 mg/L) in the stream at that point (Table 2). The estimated annual quantity of sediment delivered to the 116-ha GMC outlet for undisturbed conditions is 23 t. This is based on a denudation rate of 0.04 mm/year, a soil bulk density of 1.43 t/[m.sup.3], and SDR of 35%. The discharge (Eqn 3) is 618 ML, using a [C.sub.r] of 0.36 for the natural surface, giving a background annual sediment concentration at that point for the undisturbed catchment of 37 mg/L (Table 2). The mine site disturbance in the GMC results in a 141% increase in the average annual sediment concentration above background at the GMC drainage channel outlet (Table 2). This increase is diluted considerably when the GMC drainage channel enters the main Gulungul Creek channel. This confluence is at a point receiving slightly less than half the total Gulungul Creek drainage (Fig. 8), which is an area of approximately 3000 ha. Based on a [C.sub.r] of 0.36, the annual discharge at this point is 15 984 ML. Using the denudation rate of 0.04 mm/year, an SDR (Eqn 2) of 18.0%, and a soil bulk density of 1.43 t/[m.sup.3], the sediment delivery at this point is approximately 309 t. This gives a background average sediment concentration in Gulungul Creek of 19.3 mg/L (Table 2). The sediment delivered to this point as a result of the mining disturbance in the GMC is the sum of the gross erosion in the disturbed 116-ha GMC (119 t) and gross erosion in the remaining undisturbed 2884 ha of the upper half of Gulungul Creek catchment resulting from denudation (1650 t) with an SDR of 18.0% applied. Average annual sediment delivery to Gulungul Creek including the disturbance in the GMC is 318 t/year. Average annual discharge is the sum of 474 ML from the disturbed GMC and 15 366 ML from the undisturbed remaining 2884 ha using a [C.sub.r] of 0.36 (Eqn 3). The average annual sediment concentration at this point as a result of mining disturbance in the GMC is 318 t [divided by] 15 840 ML = 20.1 mg/L (Table 2). This is an estimated average increase of 4.1% per year in the sediment concentration in Gulungul Creek (Table 2).

Table 2. Impacts on sediment concentration in Gulungul Creek due to mine site disturbance in the Gulungul mine site catchment (GMC)
Catchment conditions Catchment annual Catchment annual
 discharge sediment delivery
 (ML) (t)

GMC undisturbed, 618 23
 background condition
GMC with mine site 474 42
Main Gulungul Creek 15 984 309
 channel at confluence
 with GMC undisturbed,
 background condition
Main Gulungul Creek 15 840 318
 channel at confluence
 with GMC with mine site

Catchment conditions Average sediment Increase in sediment
 concentration at concentration above
 catchment outlet background
 (mg/L) (%)

GMC undisturbed, 37
 background condition
GMC with mine site 89 141
Main Gulungul Creek 19.3
 channel at confluence
 with GMC undisturbed,
 background condition
Main Gulungul Creek 20.1 4.1
 channel at confluence
 with GMC with mine site

Discussion and conclusions

The landform stability analysis showed that the proposed landform, if well vegetated, will suffer little incision by erosion during the first 1000 years after rehabilitation and that contaminants should not be exposed to the environment. The 3-D simulation allows quantification of incision and this knowledge could be used to support decisions on capping thickness. If cap thickness can be reduced, this would reduce the cost of earthworks while maintaining structural integrity. It was assumed that parameter values were constant for the total surface of the DTM. Research is required to incorporate spatial variability of parameter values (due to changes in vegetation and surface treatment) in topographic evolution simulations.

This assessment of sediment delivery demonstrates the estimation of amounts of sediment delivered to the stream system from a catchment. The analysis incorporated different conditions of the rehabilitated landform and the natural surface. Erosion estimates for the batter slope assume that erosion results from direct rainfall. If flow were to escape from the upper cap surface (shallow gradient) to the batter slope (steep gradient), then incision and increased erosion rates would occur (Evans et al. 1998; Willgoose and Riley 1998). Water must be removed from the upper surface to the lower surface in some way. Consideration should be given to creating areas of `controlled failure' where water can be discharged from the upper surface. If drop structures requiring on-going maintenance are used, then it is likely the structure will fail without regular maintenance and serious damage may result (Loch 1997). Research is required to determine the best means of removing water from upper surfaces.

The water-quality assessment indicated an increase of 141% in average annual sediment concentration in the GMC drainage channel as a result of erosion from the rehabilitated post-mining landform within the GMC. Current water-quality guidelines (ANZECC 1992) recommend that water turbidity should not increase by [is greater than] 10% through anthropogenic activities. If this is translated to sediment concentration, then the estimated increase of 141% is well above the recommended 10%, and the landform design should be reassessed. This increase could be reduced by installing sediment-trapping structures. These were not considered in the analysis. Landloch Pty Ltd (R. Loch unpubl. observ. 1998) showed that vegetating the slope would reduce sediment loss by an order of magnitude and thus reduce sediment concentration in the GMC drainage channel. Respreading topsoil should improve establishment of vegetation. The optimum depth of respread topsoil is 10 cm with a minimum of 5 cm at any point (Hannan 1995). Large clumps of vegetation cause diversion of flow and are very effective in reducing erosion (Loch 1997). Where topsoil is in short supply or of poor quality, it may be beneficial to establish `pockets' or `islands' of relatively thick topsoil and not spread it thinly over larger areas. These pockets could be used to establish vegetation cover as a potential source of propagules and organic matter for the surrounding areas. The concept of focusing on isolated stands is becoming increasingly understood in weed-control programs and could usefully be adopted in revegetation where resources, such as topsoil, are in short supply (C. M. Finlayson unpubl. observ. 1999). The increase of 4.1% in annual sediment concentration in the main Gulungul Creek channel as a result of the disturbance is within water-quality guidelines.

Such an analysis needs to be done for all the mine site catchments shown in Fig. 2 and Fig. 8. Sediment delivery and sediment concentration can then be determined at catchment outlets where they enter the main stream channel, and the impact on water quality assessed. For ERA Ranger Mine, there are at least 2 sites where impact on water quality should be assessed (Fig. 8). These sites are in: (1) Gulungul Creek, immediately downstream of the confluence of the GCM drainage channel and the main Gulungul Creek channel; and (2) Magela Creek, immediately downstream of Coonjimba Creek. Assessment of these sites is important because they are immediately downstream of where the small mine site catchments first enter the main Magela system. Of course, a thorough assessment could be made by deriving sediment concentrations downstream of the confluence of Magela Creek and Djalkmara and Georgetown creeks. Areas of `controlled failure' also need to be included.

Water quality is the key. Once effects on water quality have been determined, this knowledge can be used to re-assess landform design, if necessary, to reduce erosion by reducing slope, installing sediment traps, or increasing vegetative cover until derived downstream sediment concentrations are acceptable. The technology is available to predict impact on downstream water quality and adjust rehabilitation accordingly. There is now a need for industry and regulators to decide how far downstream or what size catchment is assessed to determine the acceptability of an impact.


The author thanks Mr Peter Waggitt and Dr Chris LeGras (Office of the Supervising Scientist), and Dr Wayne Erskine (State Forests of New South Wales), for comments on the draft manuscript and helpful discussions. The author also thanks Mr Guy Boggs (Northern Territory University), and Mr Mike Saynor and Dr Max Finlayson (Office of the Supervising Scientist) for helpful discussions.


Anon. (1975) Sediment engineering. In `Sediment sources and yields'. Chapter IV. (American Society of Civil Engineers: New York)

ANZECC (1992) Australian water quality guidelines for fresh and marine waters. National water quality management strategy. Australian and New Zealand Environment and Conservation Council.

Beasley DB, Huggins LH, Monke EJ (1980) ANSWERS: a model for watershed planning. Transactions of the American Society of Agricultural Engineers 23, 938-944.

Connolly RD, Barton N (1990). `KINCON Version 1.1. User's manual.' (Soil Conservation Research Branch, Queensland Department of Primary Industries: Brisbane)

Connolly RD, Carroll C, Francis J, Silburn DM, Simpson B, Freebairn DM (1999) A simulation study of erosion in the Emerald Irrigation Area. Australian Journal of Soil Research 37, 479-494.

Cull RF, Hancock G, Johnston A, Martin P, Martin R, Murray AS, Pfitzner J, Warner RF, Wasson RJ (1992) Past, present and future sedimentation on the Magela plain and in its catchment. In `Modern sedimentation and Late Quaternary evolution of the Magela Plain'. (Ed. RJ Wasson) pp. 226-268. Supervising Scientist for the Alligator Rivers Region research report no. 6. (Australian Government Publishing Service: Canberra)

Duggan K (1994) Erosion and sediment yields in the Kakadu region northern Australia. In `Variability in stream erosion and sediment transport', pp. 373-383. IHAS publication no. 224.

Evans KG (1997) Runoff and erosion characteristics of a post-mining rehabilitated landform at Ranger Uranium Mine, Northern Territory, Australia, and the implications for its topographic evolution. PhD Thesis, University of Newcastle.

Evans KG, Aspinall TO, Bell LC (1991) Erosion prediction models and factors affecting the application of the Universal Soil Loss Equation to post-mining landscapes in Central Queensland. In `1991 Queensland Coal Symposium Proceedings'. pp. 123-132. (Australasian Institute of Mining and Metallurgy: Melbourne.)

Evans KG, Loch RJ (1996) Using the RUSLE to identify factors controlling erosion rates of mine soils. Land Degradation and Development. 7, 267-277.

Evans KG, Loch RJ, Aspinall TO, Bell LC (1997) Laboratory rainfall simulator studies of selected open-cut coal mine overburden spoils from Central Queensland. Australian Journal of Soil Research 35, 15-29.

Evans KG, Riley SJ (1994) Planning stable post-mining landforms: The application of erosion modelling. In `Australian mining looks North--the challenges and choices'. Technical program and proceedings 1994 AusIMM Annual Conference, Darwin. Publication series 5/94. (Ed CP Hallenstein) pp. 411-414 (The Australasian Institute of Mining and Metallurgy: Melbourne)

Evans KG, Willgoose GR, Saynor M J, House T (1998) Effect of vegetation and surface amelioration on simulated landform evolution of the post-mining landscape at ERA Ranger Mine, Northern Territory. Supervising Scientist Report 134, Canberra.

Flanagan DC, Livingston SJ (Eds) (1995) `WEPP user summary--NSERL report no. 11.' (USDA-ARS NSERL: West Lafayette, IN)

Goldman SJ, Jackson K, Bursztynsky TA (1986) `Erosion and sediment control handbook.' (McGraw-Hill Book Company: New York)

Hannan LC (1995) `Mine rehabilitation. A handbook for the coal mining industry.' 2nd edn (New South Wales Coal Association: Sydney)

Knisel WG (1980) CREAMS. A field scale model for chemicals, runoff and erosion from agricultural management systems. United States Department of Agriculture, Conservation Research Report No. 26.

Leonard RA, Knisel WG, Still DA (1987) GLEAMS: Groundwater loading effects of agricultural management systems. Transactions of the American Society of Agricultural Engineers 30, 1403-1418.

Littleboy M, Silburn DM, Freebairn DM, Woodruff DR, Hammer GL, Leslie JK (1992) The impact of soil erosion on sustainability of production in cropping systems. I. Development and validation of a simulation model. Australian Journal of Soil Research 30, 757-774.

Loch RJ (1997) Landform design - better outcomes and reduced costs by applying science to above- and below-ground issues. In `Demonstrating environmental excellence 97'. 22nd Annual Environmental Workshop, 12-17 October 1997, Adelaide, South Australia. pp. 550-563. (Mineral Council of Australia: Dickson, ACT)

Loch RJ, Sell K (1998) Cost savings and improved management of rehabilitated slopes: benefits from a rainfall simulator study at the Northparks Mine. In `Environmental management. Engineering the water-environment and geo-environment'. Proceedings of the Second International Conference on Environmental Management, Australia 10-13 Feb. 1998. (Eds M Sivakumar, R Chowdhury) pp. 1187-1194. (Elsevier: New York)

Lopes VL, Nearing MA, Foster GR, Finker SC, Gilley JE (1989) The water erosion prediction project: erosion process. In `National water conference'. Proceedings of the Specialty Conference sponsored by the Irrigation and Drainage Division and Water Resources Planning and Management Division of the American Society of Civil Engineers and the Deleware Section of the American Society of Civil Engineers. (Ed TA Austin) pp. 503-510. (American Society of Civil Engineers: New York)

McKeon GM, Littleboy M, Carter J, Flood N (1993) GRASP: Grass Production Model. Draft Report, Queensland Department of Primary Industries, Brisbane.

Milnes AR (1988) Rock weathering in the waste rock dumps at the Ranger Project Area. Focus report no. 3 to Ranger Mines Pty Ltd. (CSIRO Division of Soils: Adelaide)

Mitchell JK, Bubenzer GD (1980) Soil loss estimation. In `Soil erosion'. (Eds MJ Kirkby, RPC Morgan) pp. 17-61. (John Wiley and Sons: Brisbane)

Moglen G, Bras RL (1994) Simulation of observed topography using a physically-based basin evolution model. Ralph M. Parsons Laboratory, Department of Civil and Environmental Engineering, MIT. Boston, MA. Report No. 340.

Needham RS (1988) Geology of the Alligator Rivers uranium field, Northern Territory. Bureau of Mineral Resources, Bulletin 224. Canberra, Australia.

Onstad CA, Foster GR (1975) Erosion modelling on a watershed. Transactions of the American Society of Agricultural Engineers 18, 288-292.

Pickup G, Wasson RJ, Warner RF, Tongway D, Clark RL (1987) A feasibility study of geomorphic research for the long term management of uranium mill tailings. CSIRO Division of Water Resources Research Divisional Report 87/2.

Renard KG, Laflen JM, Foster GR, McCool DK (1994) The revised universal soil loss equation. In `Soil erosion research methods'. 2nd edn (Ed. R Lal) pp. 105-24. (Soil and Water Conservation Society: Ankeny, IA)

Robinson AR (1977) Relationship between soil erosion and sediment delivery. International Association Hydrological Sciences 122, 159-167.

Silburn DM, Loch RJ, Connolly RD, Smith GD (1990) Erosional stability of waste rock dumps at the proposed Coronation Hill mine. Consultancy Series, Kakadu Conservation Zone Inquiry, Resource Assessment Commission. (Australian Government Publishing Service: Canberra)

Silburn DM, Connolly RD, Loch RJ (1991) Assessment of alternative dump designs for the proposed Coronation Hill mine. Consultancy Series no. KAC91/20, Kakadu Conservation Zone Inquiry, Resource Assessment Commission. (Australian Government Publishing Service: Canberra)

Unger CJ, Milnes, AR (1992) Rehabilitation at Ranger Uranium Mines. In `Conservation and Development issues in North Australia'. (Eds I Moffat, A Webb) pp. 221-231 (North Australian Research Unit: Darwin)

Walling DE (1983) The sediment delivery problem. Journal of Hydrology 65, 209-237.

Wasson RJ (1992) (Ed) Modern sedimentation and late quaternary evolution of the Magela Creek Plain. Research report no. 6, Supervising Scientist for the Alligator Rivers Region, AGPS, Canberra.

Willgoose GR, Riley SR (1998) Application of a catchment evolution model to the production of long-term erosion on the spoil heap at the Ranger uranium mine: Initial analysis. Supervising Scientist report no. 132, Supervising Scientist, Canberra.

Willgoose GR, Bras RL, Rodriguez-Iturbe I (1989) A physically based channel network and catchment evolution model. Ralph M. Parsons Laboratory, Department of Civil Engineering, MIT, Boston, MA. TR 322.

Williams MAJ (1969) Geomorphology of the Adelaide-Alligator area. In `Lands of the Adelaide-Alligator area, Northern Territory'. CSIRO Land Research Series no. 25, pp. 71-94.

Manuscript received 28 April 1999, accepted 29 October 1999

Kenneth G. Evans

Rehabilitation of Mine Sites, Environmental Research Institute of the Supervising Scientist, Locked Bag 2, Jabiru, NT 0886, Australia. email:
COPYRIGHT 2000 CSIRO Publishing
No portion of this article can be reproduced without the express written permission from the copyright holder.
Copyright 2000 Gale, Cengage Learning. All rights reserved.

Article Details
Printer friendly Cite/link Email Feedback
Author:Evans, Kenneth G.
Publication:Australian Journal of Soil Research
Geographic Code:8AUST
Date:Mar 1, 2000
Previous Article:Isotopic dilution methods to determine the gross transformation rates of nitrogen, phosphorus, and sulfur in soil: a review of the theory,...
Next Article:Medium-term erosion simulation of an abandoned mine site using the SIBERIA landscape evolution model.

Terms of use | Copyright © 2017 Farlex, Inc. | Feedback | For webmasters