2.1 Overview

In the past decade, there has been strong growth in annual environmental or sustainability reporting by numerous mining companies (7–10), including many South African and especially pgm companies. In general, sustainability reports cover the environmental, economic and social performance of a company alongside statutory financial reporting. The compilation and analysis of the reported data can provide critical insights into a given mining sector, as well as valuable data for broader analyses of other mineral commodities. This section briefly describes the sustainability context and challenges for mining, and outlines the methodology adopted in this study.

2.2 Sustainability and Mining

At first glance, applying the principles of sustainability to mining is seemingly a simple oxymoron – since mining means to extract a resource which is finite and ‘non-renewable’. The nature of mining is therefore widely considered to be unsustainable, since it is depleting a stock (or ‘natural capital’). The paradox, however, is that the global mining industry is now larger than ever in history, producing minerals and metals at a rate which dwarfs previous generations of mines (11).

While there is evidence to suggest that many mineral commodities have shown growth in known economic resources over recent decades in some countries (for example, Australia (11)) as growing demand has encouraged exploration and developments in extraction technology – it is increasingly clear that the historical patterns of discovery and development of mineral resources cannot simply be assumed to continue unaltered into the future. The primary constraints may vary from social and governance issues in one region, to water or energy resources in another, or GGEs globally.

The application of sustainability principles to mining is therefore complex. The global mining industry, as part of their contribution to the Johannesburg Earth Summit convened by the UN in 2002, released a major report on mining and sustainability. The ‘Mining, Minerals and Sustainable Development’ (MMSD) report (12) was a major shift from arguing the historical case of growing resources over time, to a position where mining can contribute to sustainable development – even if a mining project is only a relatively short to medium term endeavour compared to other industries which are more perpetual in nature (for example, agriculture or tourism).

Sustainability is commonly defined as ensuring the ability of current generations to meet their needs without compromising the ability of future generations to meet their needs (i.e. the Brundtland 1987 definition (13)). In the context of mining, this can be taken to include the availability of resources and a productive environment at former mining or milling sites. The context for sustainable development and mining can therefore be taken back to first principles as balancing the potential environmental, social and economic risks. Further discussion of sustainability and mining are given in (7, 11).

2.3 Sustainability Reporting

An increasingly popular way of demonstrating performance against sustainability objectives is through sustainability reporting. This involves reporting and discussing all aspects of performance for a given year, covering social, economic and environmental aspects. Some mining companies, such as WMC Resources Ltd (now part of BHP Billiton Ltd) and Placer Dome Inc (now part of Barrick Gold Corp), began releasing annual environmental reports in the mid-1990s and these evolved into broader sustainability reports by 2000. Since the 2002 Johannesburg Earth Summit and the release of the MMSD report, numerous mining companies now report sustainability alongside statutory financial performance.

Early methods for reporting used internal company schemes. Due to the need to ensure consistency across companies, industry sectors or other organisations, the Global Reporting Initiative (GRI) was established in 1997 to develop protocols and promote and enhance sustainability reporting. The third edition was released in 2006 (14), with a pilot mining sector supplement in 2005 and the final mining sector supplement released early in 2010 (15). The GRI protocol is now the most common sustainability reporting tool used by mining companies (9).

The GRI itself is voluntary, and can be applied in whole or in part. There are five main sections of reporting, including economic, environmental, labour practices, human rights and social aspects. The qualitative and quantitative indicators used for each area are either core or voluntary. There are 30 environmental indicators in total (16), with some of the most significant examples including:

Overall, the emergence of and continuing improvement in sustainability reporting is providing a valuable basis to assess the environmental sustainability aspects of mining. PGM companies in South Africa are certainly at the forefront in this regard, with Anglo American Platinum perhaps showing the best quality reporting of data and analysis (9).

2.4 Quantifying Sustainability and the Platinum Group Metals

The availability of growing data sets on wastes, energy, water and GGEs from sustainability reporting can be easily combined with normal production statistics from financial performance. In this way it is possible to link aspects such as energy, GGEs and water costs with ore grade, annual throughput or project configuration, providing some useful benchmarks to compare individual site operations and also enable the environmental implications to be understood as pgm production continues to grow to meet rising demand.

The ‘sustainability metrics’ used in this study are unit consumption of energy or water per unit pgm production (GJ kg⁻¹ pgm or m³ kg⁻¹ pgm) with respect to ore grade measured for the four elements platinum, palladium, rhodium and gold (4E g t⁻¹), unit consumption of water or energy per tonne of ore milled (m³ t⁻¹ ore or GJ t⁻¹ ore) with respect to mill throughput (Mt ore year⁻¹), and unit output of GGEs in carbon dioxide equivalents per unit pgm production (t CO₂e kg⁻¹ pgm).

3.1 Extent of Sustainability Reporting and GRI Indicators

There are fourteen primary pgm companies, with numerous more at the exploration stage (mostly concentrated in South Africa), plus four diversified mining companies operating Ni-Cu-pgm projects. Of these companies, six of the pgm and all of the Ni-Cu-pgm companies produce an annual sustainability report, with all companies using the GRI as their reporting basis. The extent of reporting against GRI indicators for these ten companies for 2009 is compiled in Table I. For the table, a score of 0 was noted where no information or data is provided, 1 was given for qualitative information or a brief discussion of this indicator, and 2 was given where the data and information is fully compliant with the GRI indicator, giving a maximum score of 60. Although there are many possible assessment approaches (for example, (17)), this system was considered effective to allow a comparison of reporting performance between companies as well as for critical indicators.

The total scores in Table I range from 10 (representing poor reporting performance) to 43 (representing fairly strong reporting performance). Only the energy indicators (EN3 and EN4) were assigned a perfect score for all ten companies. All other indicators varied widely. A major issue which is perhaps not immediately obvious is the lack of site data given by some companies, mainly the larger diversified companies. In order for any data to be materially useful, individual site data is best and should be reported. This is common practice for production.

The two top companies for sustainability reporting were Anglo American Platinum and Vale Inco. These are clearly the leaders in this area for the pgms sector. Anglo American Platinum’s first report was published in 2002 (to coincide with the Johannesburg 2002 Earth Summit), and the content and presentation has continued to evolve over subsequent years. In general, Anglo American Platinum released site data up until year 2007, with the 2008 and 2009 reports giving group totals only (the site data was provided upon request to the present author (18)). Their group energy consumption for 2009 is shown in Figure 2 (19), including the split between mining, milling, smelting and refining as well as sources of energy for each major stage and GGEs by energy source. This is an excellent approach, and allows a quick and substantive interpretation of energy consumption and sources, as well as links to GGEs. Furthermore, in their 2006 and 2007 reports, Anglo American Platinum presented mining, milling and smelting data by site, as well as including site-specific energy intensity targets, shown in Figure 3 (20). This level of detail allows detailed analyses of energy performance, although the 2008 and 2009 reports stopped reporting in this manner. In contrast, Vale Inco have always reported by group totals and, although this allows comparison for Vale’s performance year-on-year, this does not facilitate direct cross-sector evaluation of pgms production (as shown later in this paper). As a reporting model, therefore, Anglo American Platinum’s provides greater depth to facilitate analysis of key factors, trends and issues.

By comparison, Lonmin scored much lower in Table I. However an example of their data is given in Figure 4 (21), showing declining total energy consumption but increasing unit energy consumption.

3.2 Sustainability Metrics

The compiled sustainability metrics are given in Table II, divided by mines and companies where possible.

4.1 Sustainability Metrics: Historical Comparison

In 1955 the Rustenburg and Union mines processed ~1.6 Mt ore and consumed 306,000 GJ of electricity and 2,157,000 m³ of water (26). Assuming all South African production of 11.87 t pgm in 1955 was from these mines, this gives a yield of about 7.42 g t⁻¹ (an ore grade of ~9.3 g t⁻¹ assuming 80% recovery) and unit energy and water costs of 25.8 GJ kg⁻¹ pgm and 182 m³ kg⁻¹ pgm, respectively, plus 1.35 m³ t⁻¹ ore. Compared to recent values (Table II), this suggests that energy costs have increased over time but water costs have only marginally increased. The increasing energy costs are probably related to the gradually increasing depth of the mines and lower ore grades.

4.2 Sustainability Metrics: Energy

With respect to energy, the data show a moderate negative correlation between ore grade and unit energy costs – as ore grades decline the unit energy costs increase (Figure 6). Although average 4E ore grades in the Bushveld have not declined as dramatically as gold ores elsewhere in South Africa, the current average ore grade of ~3.9 g t⁻¹ (4E) is marginally below the 2009 average ore grade of Bushveld reserves and resources of ~4.3 g t⁻¹ (4E) (3). Average ore grade may decline as shallower but lower grade projects (for example, the Platreef or Great Dyke mines) are expanded or developed in preference to deeper Merensky/UG2 projects. Therefore there will be only slight pressure on energy costs from declining grades, and overall, this suggests that total energy consumption will largely be a function of pgm production. Declining grades, however, may still affect individual companies or major projects.

Unlike gold mining, energy in pgm mining and production is dominated by electricity consumption, related to the prevalence of underground mining in the Bushveld region and the more complex processing that is required for pgm extraction. The typical range for unit energy costs is presently 122–319 GJ kg⁻¹ pgm, with a production-weighted average of 222 GJ kg⁻¹ pgm (based on refined production only). Most projects show variable unit energy costs over time and no clear trend, although 2007 to 2009 are often the highest on record – presumably related to the South African electricity supply crisis affecting production. A recent study of gold mining (7) showed that the typical unit energy costs of gold ranged from 120 to 213 GJ kg⁻¹ Au and averaged 143 GJ kg⁻¹ Au. The unit energy costs for pgms are clearly higher than those for gold, but not as much as could be expected based on the differences in mining and processing.

Anglo American Platinum’s data show the split in energy costs between mining, milling, smelting and refining (Figures 2 and 3), showing the dominance of the first three of the four major stages in pgms production. In addition, the inclusion from 2006 to 2007 of energy costs split between mining and milling is a rare example of detailed reporting in the global mining industry, and allows improved understanding of the various stages of pgm mining and production. Unfortunately, as individual smelter pgm production statistics are not reported (for example, grade and tonnes of concentrate processed to matte produced), it is not possible to provide a breakdown to unit energy costs (GJ kg⁻¹ pgm) for each major stage which could then be expanded into a rigorous, process-based model of pgms production.

4.3 Sustainability Metrics: Greenhouse Gas Emissions

The unit GGEs also show a moderate negative correlation to ore grade, similarly to energy – as should be expected given the dominance of coal in South Africa’s electricity mix. The typical range for unit emissions is 25.8–78.3 t CO₂e kg⁻¹ pgm, with a production-weighted average of 51.2 t CO₂e kg⁻¹ pgm (based on refined production only). The unit GGEs for gold typically range from 10.3 to 16.4 t CO₂e kg⁻¹ Au and average 11.5 t CO₂e kg⁻¹ Au (7). The significantly higher unit GGEs for pgm production are influenced by the high proportion of coal-based electricity for Bushveld projects as well as the more intense processing associated with smelting and refining.

The energy and GGEs intensities of Bushveld pgm production give rise to perhaps a unique situation in global mining. Due to the high proportion of electricity used per unit production, it should be possible to examine future low GGE electricity sources, including renewable energy projects like wind power, baseload solar thermal and/or photovoltaics, to progressively replace existing coal-based electricity. This would still allow electricity needs to be met but provide for a significant reduction in emissions. In contrast, gold mining is often dominated by open cut techniques, for which sustainable alternatives to diesel as an energy source appear very limited at present. Although viable energy sources in the Bushveld are outside the scope of this study, it is clear that energy efficiency as well as the choice of energy sources will be critical in determining both the energy and emissions intensities of pgms production.

4.4 Sustainability Metrics: Water

The extent of water consumption for pgms is within typical ranges for various metals (8). In terms of water consumed in milling, the range found in this study for processing pgm ore is 0.18–3.39 m³ t⁻¹ ore with one project averaging some 15.6 m³ t⁻¹ ore. The average of 1.04 m³ t⁻¹ ore (excluding the single high value) is similar to gold (1.37), copper-gold (1.22), copper (1.27), lead-zinc-silver (2.67) and nickel (1.01) ore processing (all m³ t⁻¹ ore; (8)). Almost all of these ore types undergo grinding and flotation to produce concentrates in a similar manner to pgm ore processing.

A perhaps surprising outcome is the degree to which the Bokoni project has reduced water consumption from a high of 6.07 million m³ in 2004 to just 279,000 m³ in 2008 (in 2009, Bokoni became a non-managed joint venture (JV) and water reporting reflects this change in ownership as a full year’s data was not available in 2009 (18)). Most projects, however, have not been successful in this regard, with Anglo American Platinum’s Rustenburg group increasing total consumption of water from ~3.1 million m³ between 2002–2004 to about 6 million m³ in 2009 despite only a marginal increase in production. The opportunity to save water would appear to be very site and project specific but remains a critical area for future sustainability in the pgm sector.

Despite the increasing reporting of water consumption by pgm companies, it remains a challenging area for sustainability reporting. Under the GRI, total water consumption (EN8) and water discharges (EN21) are core reporting indicators while impacts on water resources (EN9) and water recycling (EN10) are voluntary. Very few pgm companies report on all of these indicators in detail (see Table I), with most simply reporting total water consumption. As such, it has not been possible to present an account of the extent of water recycling in pgm ore processing.

The typical range for unit water consumption per unit pgms produced of 241–2743 m³ kg⁻¹ pgm, with a production weighted average of 800 m³ kg⁻¹ pgm, compares similarly to gold mining where production requires an average of 691 m³ kg⁻¹ Au (typical range 224–1783 m³ kg⁻¹ Au) (7). The differences are probably related to ore grades, the degree of water recycling and the more intensive smelting and refining processes that are required for pgms.

A major weakness in the GRI’s approach to water aspects is that water quality is not considered, except for external water discharges to the environment (8). This is critical since water quality is the primary factor in determining its potential use, recyclability and impact on the environment. At present, not all pgm companies divulge data on the quality of consumed water, and there is limited information on the quality of water discharged to the environment. Anglo American Platinum’s statement on water resource impacts at Mogalakwena versus those of ActionAid, discussed earlier, are an exception in this regard.

4.5 Sustainability Metrics: Mine Wastes

The mining industry is the largest annual producer of solid wastes globally (12, 27). The relatively low grade of pgm ores, of the order of a few grams per tonne, means that > 99.99% of the ore becomes solid waste. However, despite numerous major tailings dam failures, riverine or marine disposal of tailings, or ongoing acid and metalliferous drainage at innumerable current and former mine sites around the world (12, 27–29), the GRI still does not mandate that large volume mine wastes be fully and accurately reported. Under the main GRI protocol (14), the core indicator for solid wastes (EN22) is typically taken to refer to putrescible and/or hazardous wastes, such as those sent to landfill, recycled or treated further (for example, chemicals, oils, metals, woods, etc.) (9). The GRI mining and metals sector supplement (15) proposes that ‘total amounts of overburden, rock, tailings, and sludges and their associated risks’ (p. 33, indicator MM3) should be a core indicator but fails to make it an explicit requirement to report complete data. It is expected that only ‘hazardous’ mine wastes should be reported, effectively leaving the reporting of large volume mine wastes at the discretion of a site ‘risk assessment’ to judge what is ‘hazardous’. This is unfortunate, since it allows an easy escape clause to avoid reporting such wastes. Based on a comprehensive survey of sustainability reporting by many major mining companies, full and accurate reporting of large volume mine wastes is a key strategic weakness across the global mining industry (9).

Within the pgms mining sector, Anglo American Platinum are certainly the leader with respect to reporting mine waste data, although all pgm companies can and need to improve their reporting in this regard. The recent controversy over the perceived water resources impacts at Mogalakwena highlights the importance of proactive management of large tonnage mine wastes, especially where these border large local communities which depend on water resources adjacent to mining projects.

4.6 Sustainability Metrics: Biodiversity

The reporting of biodiversity data and information was highly variable and often poor (indicators EN11 to EN15 and EN25). Given the public prominence of biodiversity issues, legislation and numerous international conventions or treaties and the potential impacts of mining on protected areas and threatened and endangered species, this is rather surprising. For some indicators, a review of International Union for Conservation of Nature (IUCN) information on ‘Red List’ species is all that is required, while other indicators simply require land tenure information and a discussion of protected or high conservation areas relative to land uses and management plans.

The best example of detailed biodiversity reporting was by Vale Inco, who scored perfectly for indicators EN11 to EN15. Vale Inco provide detailed biodiversity data, descriptions of their operations relevant to specific regions or biodiversity values, and detailed discussions of management plans. It is perhaps no coincidence that Vale Inco has been at the forefront of these challenges in their home country Brazil, home of the Amazon rainforest, compared to other companies which are in established agricultural and/or mining provinces where biodiversity issues have not historically been as prominent.