Below Water Table Mining, Pit Lake Formation, and Management Considerations for the Pilbara Mining Region of Western Australia

Abstract

Located in northern Western Australia, the Pilbara is the highest productivity region for iron ore and other metal mining in Australia. As elsewhere, mine closure guidelines typically require post-closure landforms to be safe, stable, non-polluting and sustainable here in the long-term. I reviewed the primary literature, including international, national and state government guidelines and regional case studies for mine closure and related socio-environmental topics, to understand the key risks and management strategies needed to achieve these broad expectations for below water table (BWT) mining. Many BWT open cut mining projects will result in pit lakes in this region, many of which will be very large and will degrade in water quality with increasing salinisation over time. As an arid region, risks are dominated by alterations to hydrology and hydrogeology of largely unmodified natural waterways and freshwater aquifers. Although remote, social risks may also present, especially in terms of impacts to groundwater values. This remoteness also decreases the potential for realising practicable development of post-mining land uses for pit lakes. Explicitly considered risk-based decisions should determine closure outcomes for BWT voids, and when pit backfill to prevent pit lake formation will be warranted. However, maintaining an open pit lake or backfilling a void should also be considered against the balance of potential risks and opportunities.

1. Introduction

As for most mature mining jurisdictions, mine closure guidelines in Western Australia require post-closure landforms to be physically safe, geotechnically stable, geochemically non-polluting, and sustainable in the long-term [1,2]. Jurisdiction-specific guidelines usually depend on site-specific characteristics, and stakeholder values, such as long-term end uses for the water resource, as well as accounting for regional water quality and ecological tolerances and any pre-existing disturbances. However, all pit closure planning requires the development of closure objectives/outcomes and associated completion/success criteria to demonstrate achievement of these [3,4]. For instance, although pit lakes are novel ecosystems not previously present in the landscape [5], they must still meet a water quality standard that does not present an unacceptable risk to receptors (e.g., human, livestock, and ecological uses) and the broader environment [6,7,8]. Pit lakes are further increasingly expected to demonstrate consideration of, if not also provide, opportunity as a post-mining land use (PMLU) [9,10,11,12,13].

I reviewed the primary literature, including international, national and state government guidelines and regional case studies for mine closure and related socio-environmental topics, to understand the key risks and management strategies needed to achieve these broad expectations for below water table (BWT) mining. The review prioritised peer reviewed journals. To this end, Google Scholar searches were made using the key words “Pilbara”, “mining”, “mine”, “pit lakes”, “mine lakes”, and “closure planning”. Due to recent substantial changes in the scale of mining in the Pilbara, I also strove to source information from articles from after the year 2000. I supplemented journal articles with conference proceedings, which were also used due to the paucity of the primary literature on mining in the Pilbara, especially those from the regular industry conferences “Australian Acid and Metalliferous Drainage (AMD) Workshop, “Mine Closure”, “Life of Mine”, “International Mine Water Association (IMWA)”, and the “International Conference on Acid Rock Drainage (ICARD)”.

Below Water Table Mining

The large and numerous current and proposed mines of the Pilbara region of Western Australia (WA) are forecast to increasingly extend into deposits below the water table (BWT). Mining BWT often requires extensive dewatering to access ore, and not all pits can be backfilled to prevent pit lakes forming due to the availability of materials and cost [14]. Upon cessation of mining, dewatering ceases and the water table recovers to a post-mining level that is expected to develop into a pit lake. Climate is the most important factor of the hydrologic processes associated with a pit lake [15]. In general, surface hydrologic processes (e.g., direct precipitation, evaporation, and surface water runoff, including occasional stream or river inflows) are defined by the regional climate and interact with groundwater processes to form a simple water balance budget for the pit lake.

Following groundwater rebound and dissolution of the cone of depression, the pit lake begins to fill with water and groundwater inflow into the pit initially increases as the inflow area increases. Later, groundwater seepage into a pit slows as the change in the pressure head decreases [16]. As a result, total inflow into a pit lake is expected to gradually decrease as the open pit fills, while total outflow is expected to increase due to increased evaporation from the greater surface area of the pit lake area. At some stage, total inflow approximates total outflow and the water level in an open pit will reach equilibrium, albeit responding dynamically to changes in seasonal precipitation and evaporation rates. Water level fluctuations may occur, e.g., due to occasional cyclones and associated surface water inflows. Groundwater inflows are generated from precipitation recharge and tend to buffer short-term climatic changes, but long-term climatic changes will be reflected in groundwater inflows.

2. Pilbara Context

The Pilbara region is defined variously under different purposes but is located in Western Australia, 1200 km north of Perth, and extends over some 300,000 km² (Figure 1). The region is almost exclusively Crown land, with freehold land generally concentrated along the coastline [17]. Predominant land uses are mining and pastoral, although the areas to the north and north-east of the Pilbara are too dry to support the grazing of livestock and include few settlements. Three national parks have also been proclaimed in the region.

2.1. Climate

The Pilbara region consists of a climate classified as an arid hot desert (BWh) by the Köppen–Geiger system [18]. Very hot summers (up to 50 °C) last from October to April and mild to cool winters (near 0 °C) occur from May to September. Summer rainfall occurs from either tropical cyclones or thunderstorms, while winter rainfall is typically from low pressure trough systems [19]. Average annual rainfall in this highly active mining region is low while evaporation can be very high [20]. Monthly evaporation also significantly exceeds rainfall throughout the year and seasonally ranges from around 150–200 mm per month during the dry season, and up to 450 mm in December and January during the wet season.

Sporadic and intense thunderstorms are typical for the region from January to March, and an annual average of two tropical cyclones make landfall, resulting in daily rainfall amounts of up to 200 mm over a 24 period [21]. Flooding is enhanced when multiple tropical lows occur within a few weeks of each other [22].

2.2. Geology

The Pilbara includes a wide variety of ancient sedimentary and metamorphic rocks [23]. The iron ore deposits for which the region is perhaps best known have mostly been developed by later, local enrichment of sedimentary banded iron formations.

In the Pilbara region, iron ore is mined from three major stratigraphic entities, namely the Turee Creek Group, the Hammersley Group (the mineralised banded iron formation of the Hamersley Group and the Brockman or Marra Mamba iron formations) and the Fortescue Group, or from eroded and redeposited material, e.g., channel iron deposits. Much mining may be considered an essentially benign activity from a geochemical perspective, not least because the ironstones are mined at shallow depth and from natural exposures over large parts of the ancient terrain [24].

Most deposits are ore bodies that existed under reducing conditions but with a genesis not related to sulphide mineralisation. Supergene-enriched banded iron formation (BIF) or detritals that are above the water table and have been exposed to long term weathering are unlikely to contain sulphides within the ore. However, sulphide-bearing shale or lignite may be inter-bedded with or lie stratigraphically below the ore body.

The local geology of each deposit varies; however, the most problematic formations with potentially acid-forming rocks are considered to be the black shale units, such as those found in the Mount McRae Shale (MCS) and Brockman Iron and Marra Mamba Iron Formations which present, to some extent, at most mine sites in the region [25]. Owing to the discreet occurrence of lignite in the region, little is presently known about its geochemical properties [26]. The MCS hosts mostly black shale, and it is overlain by the Dales Gorge Member of the Brockman Iron Formation, which is often exposed in pits where the lower portion of Dales Gorge Member is mined. MCS if unoxidised, is known to contain pyrite. Both lignite and shale lithologies are known to host sulphide mineralisation when unoxidised [27].

2.3. Mining

Mining in the Pilbara region dates back more than 100 years to the discovery of gold at several localities, and the proclamation of the Pilbara Goldfield in 1888. Today, the Pilbara region produces a substantial proportion of the value of mine products in WA, being host to the largest iron ore deposits in the world, which are the predominant economic product of the region [24].

The Goldsworthy iron ore deposit (Figure 2) was initially discovered by the then Mines Department in the 1950s, and mine production commenced at Mt Goldsworthy in 1962 [23]. Identification of many of the present-day mining operations soon followed, with the discovery of iron ore deposits at Mt Tom Price in 1962 closely followed by the Mt Whaleback, Pannawonica, and Paraburdoo orebodies. Further substantial mine developments since the late 1980s, including the commencement of mining at Jimblebar, Channar, Marillana Creek, Marandoo and Brockman No 2 Detritals, and Yandicoogina.

2.4. Ecology

The Pilbara region in remote north-western Western Australia is one of the oldest land surfaces on Earth. With a diversity of geography from ancient and vast sand deserts to the north and east, and highly metamorphosed rocks to the south, it has long been regarded as one of Australia’s centres of biological diversity, endemism, and refugia [28]. The region is an area of exceptionally high biotic diversity and endemism and harbours taxa genetically distinct from neighbouring regions, despite close geographical proximity of populations [29]. This is especially so at the eastern and southern margins, where habitat gradients are pronounced. Genetic divisions may also relate to major drainage divides and geological discontinuities associated with eastern and western Pilbara terrains. The Pilbara landscape supports one of the richest reptile assemblages in the world [30]. Although no mammal species are endemic to the Pilbara bioregion, it is also home to some of Australia’s most charismatic vertebrate species, including the northern quoll (Dasyurus hallucatus), mulgara (Dasycercus spp.), and greater bilby (Macrotis lagotis).

2.5. Mining Below the Water Table in the Pilbara

Across all commodities, there were already more than 2000 mine voids and more than 150 mines operating BWT in Western Australia early in the millennium. Many were located in the Pilbara and Goldfields regions [31]. Approximately three quarters of Pilbara iron ore resources were already located BWT [23]. There are now an estimated 670 open pits in the Pilbara that have the potential to become pit lakes in the future [17] (Figure 2). The final dimensions of the Whaleback Mine, the world’s largest iron ore mine, will measure 5.5 km long, 1.5 km wide, and 500 m deep [32], and is expected to form a pit lake [33]. Consequently, the major issue for the Pilbara arid iron ore mines continues to be the potential impact of mines which extend BWT on regional groundwater quality [34].

Pilbara pit lakes are typically either ‘throughflow’ lakes or terminal sinks [31,35,36]. Often due to their location transecting dynamic groundwater systems, throughflow lakes both have groundwater inflows, then seep pit lake waters back into regional aquifers. In areas of net evaporation, like the Pilbara, pit lake evaporation rates often exceed water inflow rates, causing the pit lake to function as a hydraulic ‘terminal sink’ [36]. If the steady-state pit lake elevation stabilises below the surrounding pre-mining groundwater level, the pit lake then becomes a terminal sink, with no water released into the environment through seepage into the groundwater system, although a terminal pit lake may still discharge to ground and even surface waters during extreme rainfall events where freeboard is insufficient to retain flood flows to the pit void.

Some Pilbara pit lake types can also occasionally act as ‘flowthrough’ (sensu [37,38]), by receiving, and then discharging, local surface waters during episodic events, such as storms, or where a pit intercepts a floodplain. For instance, a ‘terminal sink’ pit lake may still overflow during extreme rainfall events where the freeboard is insufficient to retain flood flows to the pit void.

2.5.1. Water Quality

Although salinity typically increases closer to the coast, in general, the iron ores of the Pilbara form good aquifers with potable or near-potable water quality, barring often high concentrations of nitrate [39,40]. However, acid and metalliferous drainage (AMD) may be generated following the disturbance of these geologies, increasing solute concentrations and decreasing the pH of the water it contacts [41]. AMD arises predominantly from the oxidation of pyrite, FeS2, present within carbonaceous shale units interbedded with the sedimentary ironstone, and as localised secondary mineralisation. Disseminated pyrite and secondary pyritic veining may also be abundant in lignite occurring locally in the overburden. The precise chemical character of the AMD, and, thus, the potential pollution risk to the environment, differs from that generated by other types of mining. While pyrite oxidation can generate highly acidic, low pH, sulphate- and iron-rich solutions, characteristic of AMD, the dissolved metal(loid) concentrations are generally neither as elevated, nor as varied, as in drainage formed during the mining of other commodity types. AMD associated with Pilbara iron ore mining is found to have traits similar to those of coal deposits, where biophile elements, such as the metalloids arsenic and selenium, and the metals mercury and cadmium, are notable contaminants of potential concern (COPCs) [27].

The chemogenesis of a pit lake is a function of many geochemical processes that can occur within a pit lake. Each of these processes contributes to long-term pit lake water quality. Reactions within the damaged rock zone surrounding the pits can occur from blasting and other impacts. The damaged rock zone of a pit lake includes the shell of highly fractured rock or sediment around the pit perimeter and also talus at the base. Mineral reactions, but especially dissolution of soluble mineral phases, will result in the addition of some trace elements and particularly salts, to pit wall runoff in all pit lake water balance types. The water quality of terminal sink lakes is expected to show increased metal(loid) and salt concentrations over time as solutes introduced through groundwater inflow and pit wall runoff are concentrated by evaporation [42]. However, both through-low or flowthrough water balances can lead to solute mass loss from the pit.

Groundwater inflows are the major contributor to the water quality of Pilbara pit lakes. Mass loading associated with groundwater solutes directly affects pit lake water quality [43], and groundwater flow to the pit lake will comprise the largest water volume inflow. Groundwater inflow volumes will also affect the residence time of the pit lake, which in turn can affect water quality. Sulphidic material in the cone of depression created by dewatering the orebody, and the hydraulic conductivity of this material can also influence groundwater quality as the water table recovers [44]. As groundwater levels rebound, oxidation products will dissolve in the inundating groundwater [45]. Seasonal fluctuations in lake levels will affect weathering process due to the continual re-exposure and oxidation of exposed lithologies [46]. Concomitantly, groundwater with high alkalinity can contribute to increasing the pH of a pit lake or buffering potential acidic drainage flowing from the lake.

Highly elevated nitrate concentrations can also occur in some regional aquifers, e.g., deep groundwaters in Australian Pilbara mining region [31,33] and are common to mine waters through blasting residues [47,48]. Nitrate toxicity is increasingly recognised as a threat to aquatic ecosystems, [49] and particularly those receiving mine waters [50], although local differences in ecological tolerances may exist in these areas also [51]. As such, nitrate is an emerging risk for even neutral pit lake water quality, especially during early years of development [52].

The generation of relatively dense, saline water at depth and periodic addition of fresh rainwater to pit lakes can result in a salinity stratified (monomictic) lake [53,54,55] or with sufficient flows can form a mixed (holomictic) freshwater pit lake [56]. If the pit lake has not mixed and remains stratified, it is likely that the water quality discharging from overtopping of the pit will be similar to site run off (Figure 3). The natural flooding of mine voids from intense rainfall events demonstrates the viability, on an opportunistic basis, of final voids as potential surface-water reservoirs [31].

Mixing of the pit lake during turnover can contribute to processes that may affect lake chemistry, such as the vertical homogenisation of dissolved chemical parameters, the vertical distribution of dissolved oxygen stored in the epilimnion, sediment re-suspension and vertical distribution of carbon dioxide stored in the hypolimnion followed by ex-solution at the lake surface [43]. Annual seasonal mixing of the entire water column is likely in even steep-sided pit lakes of Australia’s tropical north [57,58].

2.5.2. AMD

Modelling of sulphide oxidation has shown that the relative contributions to sulphate production may vary by about five orders of magnitude, with oxidation continuing until all sulphides are consumed or until reactive geologies are inundated. The indicative relative magnitudes are as follows [45]:

• Talus (1);
• Blast damage fractures adjacent to the pit wall (0.01);
• Pit wall surface (smooth) (0.0001);
• Pit wall surface (rough surface contribution) (0.0001), and;
• Non-blast damage fractures behind the MCS (0.00001).

Some geologies, although low in sulphur, may contain relatively geologically elevated concentrations of metals and metalloids, such as As, Se, Sn, and Mn [59]. Natural waters that contact these enriched geological materials may accumulate higher concentrations of solutes when compared with default water quality assessment criteria [60]. The mineralogical hosts for these elements will determine their long-term leachability [59] as many trace elements are partly or entirely bound within relatively unreactive minerals, such as crystalline silicates or iron oxides. However, some trace elements associate with more soluble mineral components, such as the following:

• Soluble sulphates and carbonates, as sources of transition metals, alkaline earths, uranium, and selenium;
• Manganese oxyhydroxides as sources of barium, transition metals, and thallium.

Oxidisable components are likely to be the principal contributors to long-term leaching from low sulphur wastes, e.g., trace sulphides, carbonaceous material, and redox-sensitive elements adsorbed on mineral surfaces. Nevertheless, only a minor proportion of the total trace element mass is usually soluble, and these leachable masses are unlikely to be of environmental significance compared to high sulphur materials. Low sulphur wastes, therefore, generally pose a negligible risk of AMD.

Mineral precipitation of trace elements may also be important as pit lake salinities increase. Dissolved mineral phases may precipitate from mixed pit lake waters, which can reduce the concentrations of some chemical parameters and provide surfaces for metal adsorption. Moreover, metal adsorption of metals to hydroxide precipitate surfaces, clay minerals, or organic material with reactive surfaces in solution can decrease concentrations of metals and metalloids. These processes might be very important where pit void shells and backfill materials present significant surface areas of iron hydroxides in particular. Speciation and redox processes of anions and metals in solution can be influenced by geochemical processes that occur within the pit lake. However, due to the pit lakes being largely mixed throughout time, the regular transformation of one species to another through oxidation or reduction reactions, precipitation or dissolution of mineral precipitates, or desorption/adsorption onto surfaces is unlikely.

Prior to the development by a new mining project, many sites may not have water quality conditions defined as default by ANZG [60] water quality guidelines (95% freshwater aquatic ecosystem biodiversity protection). An increase in contaminant concentrations is reasonably expected where the geochemistry of a site is already elevated for certain elements and characteristics that make this region of interest for mineral extraction. Defining water quality completion for closure is, therefore, more than just using generic national guidelines, but an explicit consideration of the baseline regional bio-physico-chemical context [61].

2.5.3. Salinisation

Although not a direct COPC in itself, elevated concentrations of conservative solutes, such as Ca, Na, and Cl, can present as salinity. At any time after the cessation of mining, the total solute load in the pit void is the cumulative sum of (i) the solute introduced by in-flowing groundwater, and (ii) reaction products mobilised and introduced to the pit void. Elevated major ion concentrations are recognised as being a key contributor to the toxicity of many mine waste waters. However, specific major ions do not clearly drive the toxicity of saline seepage waters, and the effects are probably due to total salinity [62].

The groundwater beneath the Fortescue Marsh is hypersaline (approximately 3–4 times the salinity concentration of sea water), but fresh to brackish in mining areas [63]. Generally, the EC measurements of in situ geological samples indicate they are not naturally saline, with the majority of EC measurements below 0.4 dS/cm [64].

However, the climate in the Pilbara is characterised by hot summers with periodic heavy rain and mild winters with occasional rain, which means that the periods between dissolution events may be long, and the volumes of water involved may be low; contact waters are, therefore, often saline [65], especially for secondary oxidation products [66].

The rate of salinisation of the pit lake is dependent upon the surface area exposed to evaporation, such that larger lake area will have greater evaporation rates [67]. Given the clear skies and high temperatures experienced in the Pilbara, the solar radiation input (and resultant evaporation rates) are very high throughout the year [68].

In increasing order of importance, short-term sources of salinity to Pilbara pit lakes are expected from the following (Figure 4):

• Direct rainfall;
• Catchment inflows;
• Pit shell exposures and talus;
• Groundwater inflows;
• Waste materials within catchment (including backfill).

Long-term, groundwater solutes are expected to dominate.

A major challenge for closure planning at BWT mining operations is predicting and managing evaporation from final void pit lakes, especially where pit lakes can impact on downstream groundwater and surface water flows and quality [69]. Long-term degradation of regional groundwater quality can be caused by seepage of poor pit lake water quality that can extend as plumes for many kilometres down-gradient [70], including for salinity [71]. The time taken for water in large voids to become highly saline is estimated in the order of centuries. Consequently, seepage rates are likely to be from metres to tens of metres per year, such that environmental impacts many kilometres away from the pit lake may not occur for more than one thousand years.

A closure approach that is increasingly being adopted is to infill the final voids with waste rock to above the pre-mining water table to prevent the development of a permanent pit lake and subsequent evaporative losses. However, with a stripping ratio of less than one for iron ore mines, there is typically insufficient waste rock to infill all mine areas to above the pre-mining water table. Instead, integration of hydrogeological modelling with mine planning can advise the selective infilling of pits, largely using run-of-mine waste dumping, and some diversion of surface water flows [69]. Diversion of some minor creek tributary flows into specific locations within the mine path may also induce hydraulic gradients that maintain aquifer throughflow and water quality.

2.5.4. Long-Term Water Quality

The long-term water quality of pit lakes plays a key role in mine closure considerations [72], and Western Australian pit lake closure planning and management needs to predict changes in water quality of any planned pit lake over time [73]. Potential deleterious impacts to water quality that require consideration include acidity, metalliferous, and saline effects, while the potential for pit lake stratification also plays an important role.

Factors that control the long-term evolution of water quality include local geology, hydrogeological and climatic conditions [74]. Although initially influenced by pit shell solutes dissolution, the water quality of Pilbara pit lakes will be predominantly controlled first by the following factors (Figure 5):

• Groundwater inflow quality;
• Evapoconcentration of this groundwater contribution to the total pit lake volume.

Direct precipitation on the lake surface can result in the dilution of dissolved constituents. Conversely, evaporation continually increases concentrations of salinity and other solutes. However, evaporation rates exceed rainfall and, therefore, rainfall has a minor dilution effect unless it is an extreme event. Consequently, water quality trends are primarily due to processes of groundwater inflow solutes being evapoconcentrated and the seepage of pit lake water to regional aquifers, flushing out solutes. The water quality of terminal sink lakes is, therefore, expected to deteriorate over time due to evaporation [75,76]. Since Pilbara pit lakes have little or no attenuation of salinity, this poor water quality is unlikely to be resolved naturally, even over long time scales (McCullough, 2008).

3. Risks

BWT lakes can present risks to humans, livestock, and wildlife from mine closure into perpetuity. The key risks and their causes are described below.

3.1. Creek Capture

The heavy, intense rainfall and subsequent flood events experienced in the Pilbara makes climatic factors a key influence on the stability of closure landforms. Iron ore mines in the Pilbara region operate in an environment with ephemeral creeks that are subject to flash flooding events following rainfall, predominantly during the wet season. Streamflow in the smaller flow channels is typically short in duration and ceases soon after rainfall passes [77]. However, runoff can persist for weeks in the larger river channels, which drain the larger catchment following major rainfall events resulting from tropical cyclones.

The closure implications of pits, waste dumps, and land bridges in relation to the creeks, and surface water flows and subsequent flood events, are better understood today than when many old Pilbara BWT mines were begun in the late 20th century when, globally, smaller watercourses were historically allowed to flow into mining pits. This practice has been recognised as a risk to surface water management, and diversion and complete reinstatement are now used to protect even smaller river channels [78,79].

However, river diversion (or river relocation) remains an issue for mine closure and rehabilitation [79] and historical mine footprints can often present significant closure challenges, with pits, waste dumps and land bridges often located adjacent to, or even within, creeklines and their floodplains. Creek capture to pit voids can occur during or following extreme rainfall events, especially where the track of these cyclones is in close enough proximity to cause heavy rains within the catchment and significant creek flows through the mine site. Without engineered surface water management structures, a natural watercourse can then scour a path into a mine void, effectively terminating downstream flow in a pit lake.

A risk-based approach can help determine which flood magnitude different pit lake and associated catchment designs meet [22]. Analysis of flood depth data from an inundation map can show predicted flood depths across the post-mining landscape and which pit lakes are predicted to intercept creek flows in the absence of further mitigation measures.

Although complete operational backfill is rarely available for low waste to ore strip ratio for iron ore mines [80], backfilling selected pits adjacent to creek lines during mine operations can manage the flood closure risk [22]. Creeks can also sometimes be reinstated over the pit and backfilled with detrital mine waste, adopting a similar length and grade to the original where hydraulic modelling shows catchment flows are of short duration with minor infiltration losses to backfill [77]. In these cases, drainage features can be re-establish over fully backfilled pits, partial pit backfill with land bridges to convey sediment and flows, and total or partial hydrologic disconnection between partially-filled open pits and the channel system [81]. Operational works can be used to reduce the associated earthmoving costs at closure and the overall disturbance footprint can be reduced with the placement of a waste in-pit and some waste landforms not being required.

Although well-engineered diversions can require less maintenance and demonstrate functioning similar to a natural systems, [82], the poor condition of river diversion channels can prevent mining companies from relinquishing their mine to the government after mining has ceased. Many regions lack guidance for designing river diversions from appropriate geomorphic and hydraulic analogue conditions. However, establishing baseline geomorphic reference criteria for unmodified catchments can guide restoration efforts to allow recovery and ensure the stability of the fluvial system [79].

3.2. Impact of Groundwater Loss

Pilbara surface water is scarce and shows high seasonality. As such, groundwater has become a strategic resource in the Pilbara region for ecological communities [39], and groundwater-dependent ecosystems (GDEs) are often highly reliant upon high-quality groundwater systems.

A significant risk from mine pit lakes occurs when there is a hydrogeological connection between the mine voids and important wetlands, waterways, or groundwater resources. Local groundwater level depression can occur where groundwater flow occurs in perpetuity with net evaporative pit lake water balances [69,83,84,85]. Projects involving extensive groundwater drawdown (sometimes through the interaction of dewatering operations at adjacent projects) have the potential to affect a large proportion of the population of a restricted species or to threaten the persistence of species with particularly small ranges [86].

Flora recorded from around the pit lakes generally comprise drought-tolerant plant taxa with no reliance on groundwater for survival, e.g., hummock grasslands, low tree steppe over spinifex, and low woodland. However, the dominant tree species of the surrounding creek line riparian zones are Eucalyptus camaldulensis ssp. refulgens (River Red Gum), Eucalyptus victrix (Coolibah), and Melaleuca argentea (silver cadjeput). A fourth, Sesbania formosa, while less widespread, is also a relevant species in Pilbara riparian systems. Although there have been few studies, the presence of these species is often used to infer the presence of a groundwater-dependent ecosystem (GDE), granted that dependency within species has been shown to vary both spatially and temporally [87]. Interactions between salinity and water level are also largely unknown [88].

Obligate phreatophytes are those species for which access to groundwater is critically important, and they can only inhabit areas where they have access to groundwater for at least some proportion of their environmental water requirements. Conversely, facultative phreatophytes are plant species for which access to groundwater is not necessarily important to their presence. E. camaldulensis and M. argentea are large, deep-rooted trees which occur along waterways where the water table is typically <5 m below the surface in tropical north Australia [89,90]. Although it does not form the only source of plant water use [91], groundwater is a critically important component of these species’ environmental water requirement. In particular, M. argentea is an obligate phreatophyte associated with shallow groundwater and/or permanent pools of surface water at points along the creek. While not a true phreatophyte, E. victrix is not dependent upon groundwater and is relatively drought tolerant, but it is likely to exhibit signs of stress with decreased access to groundwater [89,92].

Pilbara fractured rock and coarse alluvial aquifers also support diverse faunas comprising obligate groundwater inhabitants, largely crustaceans but also including insects, worms, gastropods, mites, and fish of typically short-range endemics, with the richest known groundwater and cave-dwelling faunal diversity in Australia having over 1000 species [93]. Although this subterranean fauna remains poorly understood, it is apparent that these numbers are globally significant. Taxonomic resolution among some groups of stygofauna is poor, and species richness is likely to have been substantially underestimated [86].

Although mining development has often occurred in areas that have a high value for subterranean fauna diversity, impact assessments suggest relatively modest impacts on long-term persistence for this component of biodiversity [94]. However, as few mines in the Pilbara have closed, little information is available on actual impacts [17]. In addition to localities where groundwater depression may threaten riparian and other GDEs, the Pilbara contains mining regions with concentrations of mine voids which pose a future cumulative risk to the environment [17].

3.3. Direct Contact and Drinking

The Pilbara region, whilst arid and relatively isolated, hosts a diversity of ecosystems that may be susceptible to COPC toxicity [24]. There are numerous priority ecological communities (PECs) and threatened ecological communities (TECs) which constitute environmental receptors; the relative paucity of both permanent surface freshwater sources and groundwater of useable quality makes the ecological systems based on these water resources especially sensitive to contamination. Much emphasis is placed on protecting GDEs, which include riparian vegetation communities and subterranean fauna.

Environmental receptors can make direct contact with pit lake waters, either through active movement toward the lake, or through passively receiving pit lake waters into their receiving environments (surface or groundwater). Pit lakes and other smaller water bodies created by mining are frequently used by wildlife [95]. These biota include birds [9] (such as long-distance migratory species [96]) and also bats [97]. Contact can lead to toxicity that is either lethal to individuals (acute) or sub-lethal chronic (has an impact on populations). Prolonged contact with contaminated pit lake water may cause skin disorders, like dermatitis [98]. Direct contact may also cause include impacts on fauna or human health through ingestion by drinking if void water becomes unpotable and is consumed [9].

Some Australian animals minimise water loss to such an extent that they meet all their water needs from the water content of their food and from metabolic water alone [99]. In particular, Australian marsupials demonstrate much lower water requirements than introduced mammal pest species [100]. Even native mammals, such as dingo (Canis lupus dingo), are adapted to infrequent drinking [101]. As a result, water requirements from Australian native mammals in the region are not expected to be high, and native terrestrial mammals are not expected to drink frequently from pit lakes. However, saline water provides limited food and no drinking water resources, and does not likely function as a significant habitat for many receptors [102,103]. Nevertheless, there are expected to still be many pathways from pit waters to environmental receptors under freshwater conditions, as presented in other papers [104,105,106].

3.4. Discharge to Waterways

In general, fluvial processes in the Pilbara, located in the arid subtropics, tend to be driven by infrequent, high intensity, and short duration hydrologic events that are related to the occurrence of tropical cyclones [107]. Regional waterways are ephemeral and, therefore, only support occasional and limited ecological values when flowing. Recharge to the groundwater system is thought to be dominated by flows in these ephemeral creeks [67]. However, where perennial river pools are present, they are connected to and interact with the underlying alluvial aquifer [108]. Aquatic invertebrates show high diversity for an arid zone, with about one-fifth of all species encountered currently believed to be endemic to the region [109], reflecting the abundance of consistently fresh, permanent water maintained by freshwater aquifers. Seedling recruitment and subsequent growth also mainly depend on heavy rainfall flooding events [110], which are likely to coincide with pit lake discharge [37,111].

Understanding pit lake discharge risk requires a good understanding of what frequency and conditions discharge might occur under. Stochastic water balance modelling will generally effectively advise risk likelihood. If discharge does occur, pit lake water quality, and the nature of the receiving systems downstream, will provide information on the consequences of decant risk. A freeboard should be maintained such that the discharge is not likely under predefined rainfall events, in-turn determined as a consequence of discharge [37].

Model performance has been shown to be dependent on the representativeness of meteorological observations; observations for modelling applications can be undertaken using either regional parameters from case studies [112] or, preferably, on site at the pit lake level [68]. A greater frequency of intense rainfall events is expected by both modelling for the region [113] and empirical evidence [114]. This short-term greater net precipitation can increase the risk of water outflows from pit lakes to both surface and groundwaters [115], even in arid regions [36,116]. Climate change can also increase the rates of pit lake hyper-salinisation and the associated risks through higher mean temperatures. Pit lake water balance and water quality modelling should consider expected changes to storm events under climate change [117].

3.5. Seepage to Groundwater

Backfill to prevent evaporative losses of groundwater through pit lakes may include mineralised and non-mineralised waste rock (Figure 6). This solution can result in the maintenance of adequate groundwater and surface water flows to support downstream water-dependent ecosystems and maintain beneficial use of water.

However, where regional groundwater flow is present and pits are backfilled, throughflow conditions are likely to develop following groundwater rebound located within the saturated zone [116]. Throughflow may release solutes accumulated from previous oxidation, which may have an impact on the receiving groundwater quality as AMD. Consequently, it is equally important to understand the possible impacts that the backfill could have on post-closure groundwater quality. Quantification of the potential for solute release under these conditions may require solute leaching under anoxic conditions expected within inundated backfill [41,118].

Comparatively little is known of the functioning and especially concerning the water requirements of many of the key riparian species of the Pilbara and the ecosystems in which they occur [88,119]. Salinity in creek surface water and groundwater is low and, therefore, salinity–flooding interactions are rarely studied [120]. When undertaken, studies have generally only evaluated salinity associated with rising water tables, and not flooding associated with large, episodic, and short-lived flows. However, riparian plant diversity in arid regions is sensitive to changes in groundwater depth and salinity, as well as changes in soil moisture and salinity [121].

Stygofauna are likely to be present in all aquifers of the Pilbara [86], and water chemistry appears to have limited influence on stygofauna occurrence in the Pilbara. However, the distribution of taxa within an aquifer is primarily governed by alkalinity, salinity, and pH, with many taxa being restricted to single aquifers [122]. Few ecotoxicological data exist with which to assess the risks of trace metal mobility and uptake by subterranean fauna in the Pilbara. As a result, there are currently no groundwater ecosystem-specific water quality criteria available to identify metal toxicity risks to stygofauna. In the absence of these, Australian freshwater quality criteria [60] are conventionally applied [123]. However, these currently applied water quality targets in Australia may not be entirely protective of groundwater ecosystems.

Density-Driven Saline Seepage

Where pit lakes contain their waters and are terminal, they can increase solute concentrations through evapoconcentration and become increasingly saline over very long periods. This solute increase can lead to increased water density, which may cause density-driven flow into surrounding groundwater under certain hydrogeological conditions [124] (Figure 7).

As the pit lake water becomes denser over time due to evapoconcentration, the potential for migration into the aquifer beneath the pit floor is more probable [125]. If such leakage occurs, then the concentration rise within the void water is reduced as seepage causes mass loss of solutes from the system. Mineral precipitation may also occur when some solutes reach saturation indices. Although it is extremely difficult to predict the rates of leakage in a fractured hard rock environment [126,127], density-driven seepage represents a potential contaminant transport pathway though seepage of very saline waters into local groundwaters [36].

3.6. Pests and Diseases

Introduced predators, such as feral cats, foxes, and wild dogs, are common, and research has shifted from pastoral concerns (Thomson et al. 1992a, 1992b) to understanding the best practices to control invasive mammals for the purposes of biodiversity conservation [128]. Early settlers also introduced many grazing animals, now feral, including cattle (Bos primigenius), pigs (Sus scrofa), rabbits (Oryctolagus cuniculus), camels (Camelus dromedarius), donkeys (Equus asinus), and horses (Equus ferus caballus) [129].

North-western Australia lacks native crayfish, but the introduced red claw crayfish (Cherax quadricarinatus) is now known from three of the region’s five drainage basins, with limited management options in this remote area [130]. The species has been established through deliberate releases to provide fishing opportunities in areas where public access is permitted [131]. C. quadricarinatus has typically established itself in reservoirs before subsequently colonising natural systems, highlighting the potential for pit lakes to act as sources for further spread of freshwater crayfish [132].

There are also numerous existing and potential pathways for incursions from other vertebrate, invertebrate, and plant species from the aquarium trade into freshwater ecosystems in tropical Australia [133], such as the Pilbara.

Pilbara pit lakes also present potential for harbouring water-borne diseases [9,23]. Water-borne disease can be significant in northern parts of Australia, with the Ross River virus, Barmah Forest virus, and Australian encephalitis of particular concern [134]. The availability of quiescent surface water bodies at abandoned mines may provide a permanent breeding habitat for mosquitoes, some of which could be vectors for these human diseases [31,104]. Native animals tend to congregate around artificial or natural water bodies and are natural hosts of viruses causing disease, and these animals may facilitate transmission, e.g., to mine company personnel and recreational users. Some mosquito species breed in saline water, so the long-term salinisation expected in some pits would not decrease the disease risk from this source [135].

While fibrous minerals, e.g., crystalline silicas, such as crocidolite, can be found in many parts of WA [136], they are particularly prominent in the BIF of the Pilbara [137]. Within weathered or mineralised areas, crocidolite is replaced by goethite, which makes the mineral brittle and, thus, means that it does not pose a health risk. Hazardous fibrous minerals are not found within iron ore, as they have been replaced by iron oxide as part of the mineralisation process. Tuffs of fibrous materials can remain around pit perimeters at the closure of BWT voids, whether they form lakes or not, and materials characterisation should, therefore, include the identification of fibrous and asbestiform materials [138]. Although there is little guidance provided for the management of fibrous materials at closure, they typically require covering with non-fibrous materials.

3.7. Safety

Due to the nature of their formation from previous mining operations, pit lakes tend to have steep sides, and access can present a significant hazard in itself. Increasing the risk pit edges may be formed from unstable and erosive materials, and below-water gradients be steep and/or with sudden depth increases [139]. Pit lakes present a number of hazards for humans, including highwall failures and falls and risk of drowning [9]. Some pit lakes may also be accessible by the general public and, as such, may present a safety risk to people [140].

The rock of a pit highwall has been fractured by blasting and can be highly unstable [141]. Pit lake highwalls may, therefore, be unstable, particularly following rebounding groundwater pore pressures [142] and decades of wave action. Steep pit walls are not well-stabilised by water pressure until the lake water level equilibrates [143], a process that may take hundreds of years.

The Pilbara is in the top three regions of Western Australia for drowning cases, with lakes and dams being the second most likely places to drown, and remoteness being the key risk factor. Almost 1/4 of people who drowned were Aboriginal, although 2/3 were also people visiting the area [144]. Drowning was identified as the major cause of death in accidents that occurred in inactive or abandoned US mine sites [141]. Steep drop-offs, deep water, sharp rocks, flooded equipment, submerged wire, and industrial waste make swimming risky. With 13 pit lakes established [145], drownings in pit lakes have occurred in the Collie Pit lake District in Western Australia, with three in the last five years alone [146,147,148,149]. Traditional owners have repeatedly requested access to future pit lakes for recreational use, and there will be risks associated with these activities.

In 2018, a man died after two jet skis collided at the abandoned Atlas Iron Pardoo mine site in the Pilbara [150]. The mine site had been in care and maintenance for many years and was closed to the public, with physical restrictions on access and signs advising the public not to enter the site. However, vehicle access was still readily possible given the site’s location on the Great Northern Highway.

4. Opportunities

Pit lakes can present a range of potential opportunities at closure [9,151]. Pit lakes have been shown to present a multitude of opportunities for human recreation and regional economies, including for wildlife habitats, recreation, and primary production.

The fundamental alteration of the landscape and the creation of new landforms based on mining features, especially pit lakes, requires consideration of an alternative land use. A variety of industrial end uses have been proposed for pit lakes, especially pumped hydropower [152,153,154]. The combined storage potential from existing pit voids around Newman, Tom Price, and the Paraburdoo or Pannawonica townships in the Pilbara is estimated at about 200 GWh [155].

However, most post-closure opportunities are principally limited by pit lake water quality [156]. Although water quality may initially be good in Pilbara BWT voids, ongoing evapoconcentration and resultant salinisation is expected to lead to high TDS waters with fewer opportunities available long-term. The Pilbara’s remoteness to established communities and accessibility [157] will also play important roles in determining what, and if, opportunities can be realised in these pit lakes [158]. These two factors will be key to limiting PMLU in Pilbara pit lakes.

Pilot studies assessed the practicalities of options, such as agri-business, tourism, and cultural experiences, to build social value at a closed BHP mine site in the Pilbara [159]. Studies have also identified that there may be alternative economies for the beneficial use of Pilbara pit lake water for economic diversification and recreational waters and amenity [160]. However, the opportunity for mine voids, pit lakes, and infrastructure associated with mine closure to support irrigated agriculture has not yet been assessed [161]. This is largely due to the lack of data on potential water yields and water quality constraints available post-mining [157]. Most industrial or commercial uses of pit lakes will probably make more use of the regional industry needs for water resources to support current and near-future regional mining activities. Instead, projects in the Pilbara are largely proposed to involve low-intensity grazing and pastoral use [162] or to return to self-sustaining native vegetation where the underlying use is vacant or allocated Crown land. Nevertheless, there are a host of PMLUs that have been realised for pit lakes globally and already informally undertaken in existent Pilbara closed or abandoned sites that could apply to Pilbara BWT voids that form pit lakes [10,151,156,163,164,165,166,167,168]. These include swimming, water sports, such as kayaking and waterskiing, hunting and fishing, and wildlife habitats and watching, as well as aesthetic appeal and traditional/spiritual uses, primarily dependent upon water quality, especially salinity (Figure 8).

Aquatic Ecosystems

The reinstatement of natural systems (i.e., restoration) is possibly an unrealistic expectation for most larger metalliferous mines within the Pilbara [161]. However, provided that water quality risks are acceptable for receptors, a wildlife habitat remains the most likely opportunity [5] for remote Pilbara pit lakes. This will likely be as freshwater habitats for potentially centuries, then of reduced habitat value, as saline to hypersaline conditions develop thereafter for most terminal evaporative lakes.

Very little is known about the aquatic ecology in Australian pit lakes, and the few studies are typically focused upon macroinvertebrate communities [169,170,171]. Tropical Australian pit lake aquatic biota data is particularly unavailable [172], let alone for the Pilbara region. Nevertheless, studies of pit lake aquatic biota have collectively shown macroinvertebrate communities of limited diversity dominated by cosmopolitan and pollution tolerant taxa [171,173]. These findings suggest that, regardless of water quality, macroinvertebrate communities in pit lakes do not appear to be representative of natural waterbodies, possibly due to reduced habitat diversity of low quality even under good water quality [174].

Where there is no connectivity, there will be a significant limitation to most aquatic wildlife becoming established in Pilbara pit lakes due to an absence of regional permanent water bodies that could supply aquatic biota, e.g., propagules. Sources of nutrients include loadings in creek inflows of available fractions of sediment and dissolved and particulate organic matter. Nutrients may also be lost through uptake, adsorption, or settling. Interaction with primary producers includes nutrient consumption through photosynthesis (primarily phytoplankton in most pit lakes) and released through respiration.

Although terminal pit lakes will have limited catchment for biodiversity connectivity, flowthrough pit lakes will receive propagules from upstream during storm events. However, these freshwater aquatic biota are unlikely to be successful in increasingly saline pit lake environments. Nevertheless, even saline lakes can have high diversity and ecosystem value worldwide [175,176], and especially in the Pilbara region where saline-adapted flora and fauna are widespread [109].

The food web expected for a saline pit lake is very simple with low diversity and a short chain, in particular, through high rates of primary production being captured by saline-tolerant zooplankters, which form the basis of the hypersaline food web [177]. Regional lentic saline–water systems contain algae, brine shrimp, aquatic insects, and insect larvae, which form a food source for various bird species. However, macroinvertebrates, aside from zooplankters, are unlikely to be abundant, with almost all insects orders absent from these saline lakes [178].

PMLU opportunities may also present an increased risk through increased utilisation of pit lakes by humans, livestock, and wildlife [9,98,104]. For instance, trophic transfer of COPC through pit lake food chains has the potential to affect higher-order (e.g., bird, mammal) populations through impacts on developing embryos and eggs [105,179,180]. Western Australian pit lakes have also been recognised by some traditional owners (TOs) as an undesirable deviation from previous terrestrial land uses [181]. The absence of potable water due to salinisation has been recognised as a limitation of Pilbara pit lakes by TOs [182]. Also of concern is the protection of public drinking water source areas (PDWSAs) including regional water supply borefields.

5. Conclusions

Closure planning of Pilbara BWT voids should consider key potential hazards (

Table 1. Potential hazards for consideration in closure planning of BWT voids in the Pilbara.

Hazard	Description
Creek capture	Seasonal flow events are often of great magnitude and, with many pits located within or near floodplains, creeks may flow into pit voids, leading to pit wall destabilisation and erosion, as well as reduced or even intercepted downstream flows.
Impact of groundwater loss	Pit lakes are likely to evaporate annually far more water than is accumulated from seasonal rainfall events, leading to constant abstraction of regional groundwaters as a terminal evaporative sink.
Contact and drinking	Very poor, e.g., acidic, pit lake water quality can have acute effects on humans and wildlife in contact with pit lake waters, including on individuals drinking from the lake.
Discharge to creek	Overtopping of lowest pit walls may occur during storm events where pit lakes waters are then discharged to lower waterway reaches, transporting geochemical contaminants to downstream receiving environments.
Seepage to groundwater	Pit lake water may seep to groundwater where the pit intercepts regional flow paths and/or becomes surcharged through storm event precipitation or creek capture. Seepage water quality may be poor through geochemical interactions with pit void/backfill materials and also evapoconcentration. Longer term, density-driven seepage may result if pit lake salinity leads to water densities that exceed groundwater head differences to the lake, potentially reversing flow away from the lake.
Pests and diseases	Exotic aquatic species may use pit lakes as habitat. Feral terrestrial species and birds may drink from pit lakes.
Safety	Steep sided and unstable pit highwalls can present fall risks. Steep beach angles can lead to drownings.

). Many BWT open cut mining projects will result in pit lakes in the Pilbara, many of which will be very large, and many of which will also degrade in terms of water quality with increasing salinisation over time. These lakes can present direct risks to receptors, with the pit lake representing a transport mechanism for geochemical COPC. However, as an arid region, BWT mining risks in the region are also dominated by alterations to the hydrology and hydrogeology of largely unmodified natural waterways and freshwater aquifers. Therefore, Pilbara hazard assessment should particularly consider the availability of water present in pit lakes as a novel regional landform [5] and the risk of discharge to local waterways in addition to the more typical reduced water availability considered when mining in semi-arid/arid regions, e.g., South Africa, Nevada USA, Chile, etc. This latter point is important given the proximity of many current and proposed mining operations to significant regional waterways that have expansive floodplains (Figure 1). Thorough and transparent risk-assessment processes, again focusing on long-term water balance and water quality processes and outcomes, should underpin all closure planning strategies from initial mining approvals and permitting, through to final and definitive closure plan acceptance and rehabilitation works.

Although remote, social risks may also present a threat to local communities, especially in terms of impacts on safety and groundwater values. Furthermore, this remoteness also decreases the potential for realising practicable development of post-mining land uses for pit lakes.

Closure planning for pit lakes requires the consideration of a technically diverse list of topics. Consequently, a wide range of disciplines should be involved with closure planning of pit lake landforms. Nevertheless, what sets pit lake closure planning apart from planning of other mine closure landforms in the Pilbara is the need to also ensure interaction between these disciplines, especially with water as a unifying theme [183]. Key inputs to BWT void closure planning should also include the following:

• Early and ongoing stakeholder engagement;
• Multi-disciplinary inputs from subject matter experts especially related to water;
• Sharing planning and benchmarking with other operations;
• Both publication for peer-review and incorporating learnings from case studies,
• Ad hoc research and trials;
• Expert third-party advice.

Explicitly considered risk-based decisions should determine closure outcomes for BWT voids and when pit backfill to prevent pit lake formation will be warranted. PMLU goals should also focus on realistic and practicable outcomes, from stakeholders (including regulators) that are well informed regarding both pit lake closure science and practice during the planning process.

However, maintaining an open pit lake or backfilling a void should also be considered against the balance of potential risks and opportunities.