Write the critical review of the article below:

Contents lists available at ScienceDirect

Resources Policy

journal homepage: www.elsevier.com/locate/resourpol

Environmentally sustainable mining: The case of tailings storage facilities

Erica Schoenberger

Department of Geography and Environmental Engineering, The Johns Hopkins University, Baltimore, MD 21210, USA

article info

Article history:

Received 13 November 2015 Received in revised form 24 April 2016 Accepted 25 April 2016 Available online 18 May 2016

Keywords:

Sustainability Mining Mine tailings Environment and society Environmental regulation

1. Introduction

Mining is unavoidably environmentally disruptive. Huge quantities of earth and rock are moved, some of it processed to recover valuable minerals, the rest discarded as waste. The mate- rials that are left over after processing, known as tailings, are es- timated to be produced at a rate of anywhere from five to fourteen billion tons per year. They may include sulfide minerals that can induce the formation of acid drainage, other processing chemicals, and process water. Tailings can be disposed of in a variety of ways. In the worst of the cases, they are dumped into adjacent water- bodies, whether rivers, lakes or the sea. They may be backfilled into pits left over from underground mining. Much of the time however, tailings are stored behind dams constructed of mine wastes (Edraki et al., 2014; Adiansyah et al., 2015).

Environmental disruption related to mining is inevitable. En- vironmental disaster, on the other hand, should not be, the more so as environmental disasters often trigger social disasters. The most critical arena for reducing the likelihood of mining-related environmental disasters lies in the handling of tailings.

Tailings dam failures account for about three-fourths of major mining-related environmental disasters (MMSD, 2002a). A tailings storage facility (TSF) can occupy several square kilometers of land with dams that can reach in the tens of meters. Tailings dams are not like water retention dams. They are built in stages as mining and waste production progresses and they are built usually of

E-mail address: e..s@jhu.edu

http://dx.doi.org/10.1016/j.resourpol.2016.04.009

0301-4207/& 2016 Elsevier Ltd. All rights reserved.

abstract

This paper addresses the question of whether mining can be done in a way that contains and remediates environmental impacts and thereby safeguards the livelihoods of local populations. It focuses on tailings storage facilities (TSF) as the source of most mining-related disasters. It compares outcomes at three mines - two which ended in disaster and one notable success - to try to get at what factors are critical in producing these outcomes. Although the design and construction of TSFs is technically challenging,the paper concludes that the basic causes of TSF failure are political, not technical. A second purpose of this paper is to suggest that a social scientific analysis of engineered projects needs to pay attention to the engineering.

& 2016 Elsevier Ltd. All rights reserved.

mine wastes rather than concrete. Water management is the cri- tical problem. An adequate amount of freeboard must be main- tained, calibrated on maximum likely storm activity. If water is adjacent to the dam itself, erosional or seepage processes may lead to breaching. The foundational geology is also a critical issue bearing on the stability of the embankments. TSFs in seismically active or unusually high rainfall areas are especially vulnerable (Vick, 1990; McLeod and Murray, 2003).

The technical challenges of storing mine wastes are significant. Nevertheless, I will argue here that the principal causes of TSF failures are political rather than technical. Much is known within the mine engineering community about how to manage tailings in an environmentally sustainable way (Vick, 1990). This generally involves different techniques for removing the water. These tech- niques are costly, however. Some companies may adopt them voluntarily. It seems reasonable to suppose, however, that until the companies generally are held to higher standards of best practice in managing tailings, we will continue to see catastrophic TSF failures.

Best practice bears on two issues in particular for the purposes of this paper. The first concerns when and how environmental considerations - in particular, the design of TSFs - are built into the mine development process. The second concerns the actual tech- niques involved.

I will show that when mining companies are held to the highest standards, they can and do meet them. Whether or not they are held to those standards depends in significant measure on the regulatory environment. How exigent are the regulations, how

120 E. Schoenberger / Resources Policy 49 (2016) 119-128

comprehensive are they, and how well are they enforced? The answers to these questions, I will suggest, have in part to do with the influence of the industry in particular jurisdictions compared with other land-intensive uses, especially as this bears on reg- ulatory capacity and competence. Second, the social composition of the surrounding population also matters. Local populations with political and financial resources will have a much greater chance of escaping environmental disasters than those without such resources.

In this paper, I will explore the histories of three mines. Two of them suffered major TSF dam collapses with widespread and on- going environmental damage: the Ok Tedi mine in Papua New Guinea (PNG), and the Mount Polley mine in British Columbia. The third mine - the McLaughlin mine in Northern California - is a rare success story in which all of the environmental dislocations ne- cessarily associated with mining were confined on site and, to a significant degree, remediated after active mining ceased. The TSF has retained its integrity. I have explored the Ok Tedi and McLaughlin mine histories elsewhere and will summarize them briefly here (Schoenberger, 2015). The third case is more recent, dating to August 2014. I will focus on the construction and maintenance of tailings dams.

What I want to work through in this paper is why the failures failed and why the McLaughlin mine succeeded at mining in an environmentally sound and responsible way. Because the en- vironmental damages of mining are closely linked to social harms (through impacts on livelihoods, exposure to environmental toxins and the like), it is particularly worthwhile getting at the causes of both success and failure in an effort to determine whether mining can increasingly be done in a way that contains and remediates environmental harms.

A second purpose of this paper is to suggest that a social sci- entific analysis of engineered projects needs to pay attention to the engineering. Because of the complex interplay among the environmental, the social and the engineered, we risk missing important information if we treat the engineered as a kind of black box. The reverse is probably also true. A quick search through recent journal publications on the topic of tailings storage facilities shows that they are all in technical journals unlikely to reach a social science or policy audience.

An important and promising exception to this is the 2011 paper by Franks et al. in the journal Resources Policy. It provides an as- sessment of the advantages and disadvantages of a range of waste disposal methods and proposes a set of principles that could be used to guide industry practice (Franks et al., 2011). I think we need to press further in three ways.

First, it is clear that best practice under these principles will be more expensive than many of the approaches that are in use today. The industry as a whole has expressed its commitment to more socially and environmentally responsible methods and, all other things equal, many operations can afford the additional costs and may well implement them voluntarily (ICMM, 2008). But marginal operations may be hard-pressed or simply unwilling to adopt them. Declining ore grades and declining commodity prices se- parately and together are no doubt putting considerable pressure on mining companies at the margin (Mudd, 2007). So we need to consider the degree to which voluntary adherence to the princi- ples proposed by Franks et al. can be relied upon.

Second, I will try to show that the way the design of TSFs is integrated into the overall development plan of the mine matters. In brief, it needs to be an integral part of the process of designing the mine itself rather than being viewed as a separate problem.

Third, there is a question of who is able to comment author- itatively on the design and operation of TSFs. The industry as a whole is increasingly committed to meaningful participation by local communities which is all to the good. Here, though, I want to

argue in favor of binding independent peer review of both the design and operation of TSFs in additional to local stakeholder participation.

Section 2 of this paper describes the research method. Sub- sequent sections (3 through 5) describe and analyze the perfor- mance of the three mines in question. Section 4 considers the problems of TSFs more generally, focusing on what is considered best practice by the engineering community and what conditions might foster the wider implementation of this knowledge in the design, construction, maintenance and closure of TSFs. Section 5 offers some concluding thoughts. An epilogue brings some aspects of the story up to date.

2. Research method

This research is qualitative and, in a sense, forensic. It is based on a review of published and unpublished documents related to the specific cases and to the engineering of TSFs in general. These documents include technical post mortems of the two failed TSFs. Other information was gathered from correspondence with and conference presentations of practicing engineers with many dec- ades of experience in the construction and maintenance of TSFs. Information was also gathered from company websites, govern- ment websites and newspaper accounts.

I have only been able to make one site visit. This was to the McLaughlin mine where I was guided by the former environ- mental manager and the current manager of the TSF. One very experienced field engineer was kind enough to review this manuscript for technical accuracy. Some of my correspondents have preferred to remain anonymous and I am obliged to respect that request.

Case studies do not allow for statistical validation or general- ization. They can, however, shed light on highly complex situations and possibly provide the grounds for developing testable hy- potheses (Schoenberger, 1991).

3. Tailings storage facilities: lessons from three mines

3.1. Ok Tedi

The Ok Tedi is an open pit copper and gold mine in Papua New Guinea (PNG) developed from the early1980s by a consortium headed by the Australian firm, BHP Billiton. It cost about US$1.4 billion to develop the mine which sits near the headwaters of the Ok Tedi River in the highlands of Western PNG. The Ok Tedi's waters flow into the Fly River, and thence into the Bay of Papua (see Map 1). The Fly is notable for its extraordinary biological di- versity (Townsend and Townsend, 2004).

According to the terms of a 1976 PNG law, the mine developers were required to prepare an Environmental Impact Statement (EIS). However, the company was only required to spend a max- imum of US$220,000 on this study whose scope was, accordingly, quite limited. A second, government-commissioned EIS was more thorough, funded at US$1 million (Hyndman, 1988; Townsend and Townsend, 2004).

The critical element here for our purposes is that the second EIS was delivered in 1982, a year after construction had started. No alternatives to the Consortium's original design were considered (Townsend and Townsend, 2004). BHP Billiton had, however, promised the government of PNG that 100% of the tailings would be contained (MMSD, 2002b). Since the operation generated about 30 million tons of ore and 55 million tons of waste each year, management of the tailings and the waste rock were critical issues (MMSD, 2002b).

E. Schoenberger / Resources Policy 49 (2016) 119-128 121

Map 1. The Ok Tedi Mine, Papua New Guinea. Cartography: Chris Kelley.

A series of disasters hit the mine in 1984, including a large sodium cyanide spill and a smaller untreated tailings spill. In ad- dition, the still-unfinished tailings containment dam was de- stroyed by a landslide. Since then, all the tailings and waste rock have been discharged into the Ok Tedi River and thence into the Fly, creating a large and still growing dead zone downriver that is likely to reach 2-3000 km2 (Day et al., 1993; Hettler et al., 1997; Harper and Israel, 1999; MMSD, 2002b; Kirsch, 2003:117-118).

A legal agreement reached in 1996 required BHP Billiton to pay compensation amounting to about US$500 million and to develop a new tailings containment plan. The company eventually con- cluded, however, that there was no feasible way to construct a permanent TSF or mitigate the on-going environmental damage, even if mining were stopped immediately. Indeed, closing the mine would plausibly have increased the damage. The Ok Tedi and the Fly were naturally alkaline, and the geology around the mine was a mix of limestone and sulfides. So long as the limestone continued to be discharged with the sulfides, the acid-generating potential of the sulfides would be mitigated. The potential for mobilizing heavy metals into the water was, thereby, lessened (Chapman et al., 1999; Kirsch, 2003; MMSD, 2002b).

In the end, BHP Billiton transferred majority ownership of the mine to the Papua New Guinea Sustainable Development Program, Ltd. The government of Papua New Guinea also holds a separate interest directly (BHP Billiton, 2002).

How did such a disaster happen? The Ok Tedi mine is located in a seismically unstable region in steep terrain in one of the wettest places on earth - it receives some 10,000 mm of rain a year, often very intense rain. For this reason, landslides may be expected to occur on almost a daily basis (MMSD, 2002b:H-4). In that sense, the eventual disaster was not entirely unpredictable.

An analysis by two mining engineers with a combined fifty years experience in PNG, including with the Ok Tedi mine, is in- structive (Murray and Thompson, n.d.). Key elements of this ana- lysis are as follows. The difficulties of developing on-land disposal facilities in PNG mean that until now riverine or marine disposal of tailings have been the norm. The steepness of the terrain and the narrowness of valleys severely limit the available space for TSFs. Rapidly downcutting streams and rivers produce weak and

unstable colluvium deposits; landslides can be set off by high rainfall, earthquakes or with no obvious trigger at all. PNG is one of the most seismically active zones on earth, with an average of two earthquakes of Richter magnitude 7.0 or greater each year. The landscape is characterized by extensive development of Karstic features due to the combination of limestone, high runoff water and the acids produced by abundant rotting vegetation. Murray and Thompson caution that Karst conditions create exceptional difficulties for ensuring the integrity of TSF impoundments. Long- term maintenance presents particular problems. They offer a comparison of the problems of managing tailings in several mining regions including Canada, Australia, Peru and Chile. PNG ranks the highest in difficulty by a wide margin.

Mining activity in PNG, which has been going on for about a century, continues to be widespread. Almost none of the existing and historical industrial-scale mines in that country have em- ployed even conventional engineered mine waste and tailings disposal strategies. Ok Tedi, with the planned TSF, would have been an outlier in this respect. Mines in PNG have generally relied on riverine and marine waste disposal or simple dumping subject to erosion and other impacts, although a more recent mine in Morobe Province did implement a TSF with somewhat mixed re- sults (Mudd and Roche, 2014).

Ok Tedi and many other sites in PNG, it seems reasonable to conclude, could not be mined in an environmentally and socially acceptable way given the terrain, the geology and the climate. On the other hand, Papua New Guinea is very poor and it needs the revenues. On the model of carbon offsets or paying American farmers not to grow corn, one could imagine putting the land in trust once an ore body was identified, indemnifying local groups that might have gained something from the mining operation (e.g., land rent, employment, social and physical infrastructure), possi- bly indemnifying the mine developer, and providing twenty years of virtual royalties to the government to persuade everyone that they did not actually have to dig in order to benefit. The goal would be to buy out the whole operation before it starts.

Such an approach could run to billions of dollars so it would not be a trivial effort. Yet the costs of disasters are also very high. BHP Billiton, after all, walked away from a $1.4 billion investment in the

122 E. Schoenberger / Resources Policy 49 (2016) 119-128

Map 2. McLaughlin Mine, California, USA. Cartography: Chris Kelley.

mine and paid on the order of $500 million in compensation to local residents and other remediation steps as part of the legal settlement (Kirsch, 2003). In a more recent case, BHP Billiton and its partner Vale may be liable for as much as US$5 billion following the disaster at their jointly owned iron ore project, Samarco, in Brazil (see Section 6).

In any case, a decision against permitting mines in en- vironmentally vulnerable sites is not unheard-of. In post- apartheid South Africa, permission to mine a rich titanium source in an ecologically fragile area was denied and the area was designated a World Heritage Site (MMSD, 2002b). IRMA - the Initiative for Responsible Mining Assurance, a multi-sector body that is developing best practice standards for all aspects of mining - also recognizes that in some cases the best option will be no mining and acknowledges that additional steps need to be taken to ensure that less responsible operators do not step in subsequently (IRMA, 2015).1

3.2. McLaughlin

The McLaughlin Mine in California was contemporaneous with Ok Tedi, with active mining running from 1985 to 2004.2 McLaughlin was developed by the Homestake Mining Company (now owned by Barrick), headquartered in San Francisco. The mine is located a few hours north of San Francisco, east of the Coast Ranges, and its operations touch on three counties: Napa, Lake and Yolo (see Map 2).

In 1985, Napa County hosted an immensely valuable wine

1 On the home page of the IRMA website, where they state their basic prin- ciples, number 7 reads: "We recognize that in certain cases, whether or not there is governmental approval, due to potential impacts or other values or benefits, no mining could be the best option. We seek to advance methodologies that allow such decisions to be made within a sustainable development context. We also re- cognize that we must pursue solutions that avoid simply leaving the mining of such sites to less responsible operators." www.responsiblemining.net accessed most recently on 1 May, 2015.

2 I am relying in this section on the invaluable set of oral histories conducted by Eleanor Swent, transcripts of which are housed at the Regional Oral History Office of the Bancroft Library at the University of California at Berkeley.

industry. Along with neighboring Sonoma County, Napa is the center of premium wine production in the US, with output worth some US$5 billion in the mid-2000s. It was just in 1983 that the industry had persuaded the Federal Bureau of Alcohol, Tobacco and Firearms to allow geographic denominations of origin fol- lowing the example of French regional appellations (Walker, 2009). Land prices in Napa increased from US$1000 per acre in 1960 to US$40,000 per acre in the mid-1980s and US$120,000 per acre by 2000. This is the most valuable agricultural land in North America (Walker, 2009: 185).

The profits of the industry and the value of the land are directly related to the fact and the perception of environmental quality. An accidental release of contaminated tailings - even if they never touched any grapevines - could be expected to inflict severe da- mage on the industry. Moreover, the San Francisco Bay Area has been on the leading edge of the environmental movement in the US. People in the region are acutely attuned to issues of environ- mental quality (Walker, 2009). A mine processing roughly 7500 tpd - a moderate scale for a gold mine but hardly invisible - would not be an obviously welcome land use. Nevertheless, after exhaustive review, the mine project was permitted with the sup- port of local environmental groups.

Today, most of the land once deeply scarred by the McLaughlin mine's operations is part of a nature reserve and environmental research station managed by the University of California at Davis. The discharge of toxic substances into local streams is zero and was throughout the life of the mine. The excavated earth was stockpiled for the life of the mine and has been used to cap and immobilize contaminated tailings and to resurface and revegetate the surrounding landscape. As described below, the waste dumps were constructed to isolate and entrap acid producing waste so they are not leaching acids into the adjacent settlement ponds and creeks. The TSF is fenced off and is still managed by the company.3

3 I was able to visit the mine site and vicinity courtesy of the Donald and Sylvia McLaughlin Natural Reserve and Homestake. My invaluable and very generous guide was Ray Krauss, the Environmental Manager of the mine and the person who devised the idea of converting the old mine site into a nature reserve. Scotty Moore provided a detailed and informative tour of the TSF.

Map 3. The Mount Polley Mine, British Columbia, Canada. Source: Google Earth.

How did the McLaughlin mine succeed in containing and mi- tigating its environmental presence? The answer involves a mix of regulatory exigency, management style, and the serendipity of the coming together of a like-minded group of people from a variety of professional backgrounds - not always mining related - who re- thought what a mine could look like and how it might relate to the land and the community around it. The geology - highly im- permeable - was also helpful (see Schoenberger, 2015 for more detail).

Mining in California is big business, but the state's economy is enormous (if the state were a separate country its economy would rank eighth largest in the world) and is dominated by manu- facturing, services, finance and - not at all incidentally - agri- culture. Mining companies lack the political clout they can wield in other places and are less able to influence the terms of regulation and enforcement.

Mines in many jurisdictions must present an environmental management and restoration plan before they are allowed to go ahead. Several things distinguished the McLaughlin mine in this respect. The first was that the environmental compliance and re- clamation plan was built into the design of the mine from the outset. The developers did not design the ideal mine from an economic point of view and then add environmental mitigations to it. It is worth emphasizing the contrast with the Ok Tedi mine where the main environmental impact study was only produced a year after construction on the mine started.

The company went to unusual lengths to demonstrate its commitment to environmental health to its prospective neighbors. It financed a 30,000 acre inventory of plants prior to starting up and the original footprint of the waste dump was shifted to avoid a patch of a rare native plant. A colony of bats that lived in a cave in the path of the mine was successfully relocated. There was a plan from the outset to segregate and stockpile the topsoil excavated from the mine and from a reservoir built for water management purposes. This would be used for capping the TSF and restoring the surrounding landscape. In addition, before mining commenced, the company cleaned up an old mercury mine nearby.

continually modified to deal with unexpected environmental is- sues. For example, early testing had indicated that a very small proportion of the waste rock was likely to generate acid drainage and that this would be neutralized by the carbonates in the waste. Once mining was underway, it became clear they faced a much bigger potential acid drainage problem. To deal with this, they devised a system in which acid-producing waste was identified at the mine face and mined separately from non-acid producing waste. The different categories of waste were then arranged in the waste dumps in such a way that the acid-producing waste was encapsulated on all sides with low-permeability clays and buffered by non-acid producing materials (Krauss, 2002). Among other things, this suggests that a rigid regulatory approach that specifies how to develop the mine is not the best option. The regulations arguably should bear on outcomes, giving the developer the flex- ibility to adapt as conditions change.

Third, the original plan of the mine in some important ways exceeded the regulatory requirements. The regulations, for ex- ample, required that the TSF be able to withstand a 100-year, 24-h storm without breaching. At the same time, other regulations re- quired zero discharge to adjacent land and watercourses. The environmental manager concluded that these regulations were incompatible: guaranteeing that the TSF could withstand the 100- year storm would not guarantee zero discharge. The TSF in fact was constructed to withstand a 1000 year, 72-h storm (Krauss, 2002 and Personal communication, 2009).

The TSF is the last site with active reclamation still ongoing and remains under the management of the company. It is surrounded by berms, flanked by diversion channels to redirect runoff water away from the tailings. The winter rains create seasonal wetlands around the basin's periphery which have been recolonized by wind-blown volunteers - cattails and tules - and plantings of willow and cottonwood. In the winter, these wetlands are crowded with waterfowl. The mine has lived up to its commitment of zero discharge to the surrounding area.

Second, as the mining progressed, operating plans were

E. Schoenberger / Resources Policy 49 (2016) 119-128 123

124 E. Schoenberger / Resources Policy 49 (2016) 119-128

Map 4. Mount Polley Mine, Area of Damage. Source: Cariboo Regional District Emergency Operations.

3.3. Mount Polley

Mount Polley is an open-pit copper-gold mine located in South Central British Columbia, near Lake Quesnel (see Map 3). It was developed, beginning in 1997, by British Columbia-based Imperial Metals Corporation. The mine processed about 20,000 t per day (tpd). Run-of-mine ore was crushed on-site and put through a closed-circuit flotation separation process. After dewatering, the concentrate was shipped overseas - mainly to Asia - for smelting.

The tailings dam at Mount Polley breached on August 4th, 2014, flushing some 25 million cubic meters of water, tailings solids and construction material into Polley Lake, Hazeltine Creek and thence into Lake Quesnel (see Map 4).4 This is the largest tailings dam collapse in Canadian history and among the largest in the world during the last fifty years. Mining ceased immediately (but see Epilogue).

3.3.1. The regulatory environment in British Columbia

Sub-surface mineral rights in British Columbia are owned by the provincial government (Stano and Lehrer, 2013). The govern- ment awards rights to develop mines according to the terms of a law dating to 1859 and largely unchanged since then. Eighty seven percent of the land surface of the province - including privately owned land - is classified as 'mineral lands' where mining es- sentially trumps all other uses and interests. The "free entry" mi- neral tenure system requires little to no prior consultation with any other stakeholders including First Nations, landowners and local governments. So long as the proper steps are followed, the government has no authority to deny a lease application, no matter the competing land uses, the technical or financial capacity of the applicant or the applicant's past environmental record

4 The mine's history and operations data are taken from Imperial Metals' website: http://www.imperialmetals.com/s/MountPolleyMine.asp?Re portID14584863. The figures are taken from Imperial Metals' website: http://www. imperialmetals.com/s/Mt_Polley_Update.asp?ReportID 14 671041. Both were ac- cessed most recently on 16 March, 2015. Further information can be found on the BC Ministry of the Environment website: http://www.env.gov.bc.ca/eemp/in cidents/2014/mount-polley/ also most recently accessed on 16 March, 2015.

(Stano and Lehrer, 2013). The development and operation of a mine do have to accord

with environmental and safety regulations with fines potentially amounting to C$100,000. However, prior to Mt. Polley, no penal- ties had been imposed on anyone by the Ministry of Energy and Mines (MEM) since 1989. The Environment Ministry had issued about a dozen citations in the previous eight years. The current Liberal provincial government had sharply reduced its oversight: the annual number of geotechnical mine inspections since 2001 had fallen to less than half the number of inspections carried out during the 1990s. In 2001, MEM employed five geotechnical en- gineers. Between 2004 and 2011, that number fell to one, although it is now back up to three (Hoekstra, 2014).

The socio-economic context is also dramatically different from the California case. The Cariboo Region, in which the Mount Polley mine is located, is sparsely populated, relatively poor, and highly dependent on natural resource extraction industries. First Nations account for a relatively high share of the region's population.

First Nations in principle have a right to exercise informed consent about projects that may affect their land. Under B.C. law, they have a 30-day period to review proposed mining projects. Given how long the planning and development process is for mines, one might conclude that a 30-day response period is un- necessarily restrictive. In any case, the right to block proposals is only relevant to the small share of First Nations' lands that are covered by formal treaties with the Crown (Stano and Lehrer, 2013). It is also the case that First Nations have few good prospects for economic development and have looked to mining as a critical source of revenue. In general, the guiding principle has been non- opposition to proposed developments, provided that appropriate conditions are upheld (Stano and Lehrer, 2013).

Thus, although mining companies are encouraged to negotiate an "Impacts and Benefits" agreement with affected First Nations groups, this is not mandatory and does not have to be achieved before the mine is developed. Imperial Metals has been developing a second mine in the northwest part of B.C. - the Red Chris mine. At the time of the Mount Polley disaster, Red Chris was reportedly "in the final stages of construction and one of the last items

Imperial needed to check off was a benefits agreement" with the First Nations (Penner, 2014).

3.3.2. Why did the breach happen?

The government of British Columbia commissioned a report from a panel of experts to investigate why the disaster occurred. The Report on the Mount Polley Tailings Storage Facilities Breach (referred to hereinafter as The Report) was published on January 30, 2015 (IEEIRP, 2015). Here is a brief summary of its findings.

1. Inadequate geotechnical analysis: the design engineering firm, Knight Piesold, failed to detect critical characteristics of the soils underlying the proposed tailings dam. It appears the testing did not probe sufficiently deeply and the instrumentation built into the dam to provide continuing status information was inadequate.

2. Inadequate design:

The water balance model was based on average climatic

conditions which could not accommodate exceptionally wet

years.

The design criteria for the tailings beach - the solid surface

above the waterline and against the embankment - were inadequate at 10m width and, in any case, were never achieved.

The design proceeded on an ad hoc basis, a year at a time. The tailings dam was contingent on the weather, mine operations and the water balance. The TSF had already experienced a near-overtopping episode in May of 2014.

3. The as-built dam differed in critical respects from the design criteria that had been approved by the Ministry of Energy and Mines (MEM):

Stage 2 of the dam was constructed in an entirely 'upstream'

configuration rather than the planned 'centerline' configura- tion. An upstream dam is built towards the interior of the storage facility so that the embankment crest moves steadily inwards. A downstream dam is built up on the outside face with the embankment crest moving outwards. A centerline dam is built along the middle. Upstream dams are cheaper to build and, of the three designs, considered the least reliable, especially in high rainfall and high seismicity areas. They are banned in countries such as Chile and Peru where there is a lot of seismic activity. Downstream dams are most expensive - Vick estimates that they require roughly three times the amount of fill material compared with an upstream dam for the same cross-sectional area - and inherently the most re- liable (see Vick, 1990; USEPA, 1994; State of Victoria, 2006).

The steepness of the embankment at the breach site was substantially greater than planned: 1.3 H:1 V rather than 2.0 H:1 V. Planned buttresses were not in place. The reason for this is that the embankments were being built with mine waste which was not being produced rapidly enough to construct them to design.

The design called for a 'tailings beach' of at least 10 m width. This was never achieved. The result was twofold. The dam was retaining a lot of water as well as tailings. As a water- retention facility, a different design for the dam would have been called for altogether. The second consequence was that water was in direct contact with the embankments, weak- ening them.

It can be noted that the Engineer of Record (EOR) resigned in 2011 and responsibility was transferred to a new EOR which undoubtedly added to the complexities of the situation. In its letter of resignation, Knight Piesold noted that "The em- bankments and the overall tailings impoundment are getting large and it is extremely important that they be monitored, constructed and operated properly to prevent problems in the

future".5 It seems reasonable to suppose that the low-key language masks a sense of considerable urgency and that this would have been apparent to the recipient and to the Chief Inspector of Mines who was copied on the letter.

4. Inadequate regulation and regulatory supervision a. MEM was chronically short of staff during much of this per- iod. The Manager of Geotechnical Engineering position was vacant from 2009 to 2011, during which time there were no

inspections of the mine. b. MEM by statute has limited influence over the design. The

analysis and design are the work of the EOR, not the Ministry; inspections by MEM personnel will not unearth problems the EOR has not detected

c. MEM issued a "Departure from Approval" notice concerning the tailings beach, but the problem was not rectified

d. MEM repeatedly urged raising the factor of safety (FOS) from 1.3 to 1.5, yet the mine never exceeded an FOS of 1.3 and for some periods of time ran below even that figure.

The expert panel calculated that, based on historical experience and the number of active tailings dams in British Columbia, the province could expect on average two failures every 10 years.

The Report provided some general observations about mining practice which are worth considering. It recommends integrating tailings planning into mine planning: "This has not been common practice in the industry to date..." (p. 121). It notes that "Tailings management is often not a core skill in many mining organiza- tions" (p. 125) and that tailings storage facility design studies submitted for regulatory approval "are often lacking in detail re- garding the factors that need to be considered in assuring safety of the facility." (p. 126). As late as 2005, an account of the 1994 Merriespruit tailings dam collapse in South Africa was able to conclude that "The appropriate construction and safe operations of tailings dams is now seen as an important management respon- sibility by the mining industry" (Van Niekerk and Viljoen, 2005, emphasis added; see also Gowan, 2006).

All of this implies that one of the most environmentally salient aspects of mining has long been routinely handled in a worrisomely ad hoc manner. It will be recalled that the McLaughlin mine in- tegrated environmental management planning into the design and development process and that the design of the mine was altered to accommodate unexpected environmental conditions.

Given all these factors, it is hard to avoid the conclusion that the Mount Polley tailings dam failure was an accident waiting to happen. Further, without serious modifications to the design and construction process and the regulatory and institutional struc- ture, we can anticipate additional catastrophic failures in the fu- ture in Canada and, plainly, elsewhere (see Section 6).

4. What to do with tailings

The Report on Mount Polley suggests that best standards of practice for tailings dams now require independent technical re- views by a panel of experienced engineers that meets at least annually. One would wish also to require an annual report from the EOR on what had happened during the year in terms of design, construction, instrumentation and the like, and that this report be backed up by a report from an independent engineering firm on the dam's status. The Mineral, Mining and Sustainable Develop- ment Project in fact recommends independent peer review of TSF

5 Letter from Ken Brower, Managing Director of Knight Piesold to Brian Kynoch of the Mt. Polley Mine Corporation dated 10 February, 2011. The letter was coun- tersigned by Jeremy Haile, President of Knight Piesold and copied to Al Hoffman, Chief Inspector of Mines.

E. Schoenberger / Resources Policy 49 (2016) 119-128 125

126 E. Schoenberger / Resources Policy 49 (2016) 119-128

designs and regular independent audits of operations. Further, it recommends a process of international certification of the de- signers of TSFs (MMSD, 2002a; see also State of Victoria, 2006).6

The main geotechnical cause of TSF failures is water in the wrong place, whether causing overtopping of the embankment or foundation failure (Gowan, 2012). Within the mine engineering community, a great deal is known about how to manage tailings safely. Aside from the methods already discussed, backfilling into underground mines and disposal in mined-out open pit mines are increasingly employed. Both methods mostly avoid the problem of surface storage, but issues of water management and the surface containment of certain kinds of tailings are not entirely absent. Significantly, underground disposal methods require that tailings storage be considered in the design of the mine itself. Thickened or paste discharge involves converting a very liquid slurry into a highly viscous substance, such that the tailings deposit retains its position and shape without large retention dams. It is probably most relevant to fairly flat topographies; water management and vulnerability to seismic activity remain important concerns (Vick (1990) is the classic text).

I'd like to look more closely at just one approach that is growing in importance: filter-pressed "dry stack" tailings. This technique involves vacuum or pressure filtered elimination of water to a level of roughly 70-80% solids compared to 30-50% in conventional storage. Thus, although the tailings are solid and relatively easily handled, they are not exactly dry (Vick, 1990; Murphy and Caldwell, 2012; Gowan, 2006; AMEC Earth and En- vironmental, Inc., 2008; MMSD, 2002a; Mining Watch Canada, 2009).

The benefits of this approach are considerable, especially where water conservation is important. Drier tailings lead to less loss of water to the environment in general (e.g., through eva- poration) and are not likely to be mobilized even in the face of seismic activity. In short, physical failure is not an issue. The dry stack footprint is typically much smaller than conventional sto- rage. Removing the water also reduces the propensity for chemical reactions, especially important in the case of sulfidic minerals, prone to acid production. Dry stacks do not require impoundment dams and long term management of tailings ponds. This reduces potential liability costs. Further, reclamation can begin rather early in the life of the mine. In sum, the risks to the environment are substantially lowered and, as a side benefit, the costs of re- habilitation may also be lower.

The principal technical issues for dry stack storage include surface and groundwater management, dust and erosion control. The principal drawback is the cost. Conventional storage is still considerably cheaper.

All of these methods may be more or less suitable depending on the exact character of the tailings - whether sandy or clayey, for example - so the choice of disposal technique requires very close analysis. What remains common to all is that water management is a critical variable.

Wet storage of tailings requires a dam that will last literally forever. During the life of a mine and for some plausibly lengthy period after closure, one can imagine that regular inspection and maintenance would maintain the dam's integrity. If these safe- guards cease, the dam will eventually fail. The threat to people, property and the environment is not on a scale of radioactive waste storage, but the length of time proper management is re- quired is roughly the same. The impacts of climate change,

6 State of Victoria, Dept. of Primary Industries, 2006. "Management of Tailings Storage Facilities." Melbourne: author; Mining, Minerals and Sustainable Develop- ment Project, 2002. "Mining, Minerals and the Environment, Chapter 10." Breaking New Ground: Mining, Minerals and Sustainable Development, Final Report. London: International Institute for Environment and Development.

especially as these affect the hydrologic cycle in general and the frequency and intensity of severe weather events in particular, will arguably greatly augment these already substantial risks (Azam and Li, 2010).

5. Conclusion

The technical demands of safely sequestering and storing tail- ings are many and varied. It takes a high degree of engineering training and insight to get it right. But it is possible to do.

Many in the mine engineering community are working hard to improve their knowledge and skill in this arena. Nevertheless, whether as consultants or employees, engineers may find them- selves under irresistible pressure from companies to speed up their work, cut costs and cut corners. For those who are unable to talk back or walk away for whatever reasons, it seems plausible that strong, well-conceived, consistent regulatory environments would help them do what they know is right. They would be able to insist on the practice not because it is good, but because it is necessary.

In at least some quarters within the mine engineering com- munity, there is a clear sense that environmental issues need to be addressed in the original feasibility and design studies of any mining project rather than added on to a design that is solely concerned with economics. This is especially true for managing wastes (Hart and Boger, 2005; IEEIRP, 2015). Comments from a keynote address at a mine waste management conference in 2005 are especially pertinent and worth quoting:

The technology is here now to make dry [tailings] disposal a reality. Yet the industry continues to pursue unsustainable disposal practices such as riverine disposal and tailings dams. Tailings dams fail on occasions but the argument is always made that it is too expensive to deal with a dry disposal technology up front in most operations.... (Hart and Boger, 2005, p. 6).

They add that the mining companies are persisting in un- sustainable practices in part because the gap between the tech- nical people and the people who ultimately make the decisions is so great. For this reason, best practice technologies are not being widely implemented even though they are well known within the technical community (Hart and Boger, 2005).

Will more stringent regulatory environments in some areas tend to push mining investment to laxer regions? It would not appear so. The Fraser Institute is a libertarian-leaning think tank based in Vancouver that conducts an annual survey of senior managers and executives in the mining industry (Jackson and Green, 2015). Their goal is to develop a picture of the pure mineral potential of 122 jurisdictions around the world (in its latest sur- vey) and the mining community's perception of the policy en- vironment in each of these jurisdictions. This latter concerns the degree to which the policy environment encourages or dis- courages investment in particular places. Survey questions place particular emphasis on the uncertainty and the onerousness of regulations concerning the environment, responsibility to local stakeholders, taxation and trade barriers, all thought to be a dis- incentive to investment. It also assesses the legal environment, political stability, infrastructure and labor supplies. A high ranking means that the regulatory environment does not discourage investment.

The policy rankings do not tilt toward developing countries with lax regulatory regimes as one might have expected. Finland is first. Botswana is the highest ranking developing country at #13 and it is widely regarded as having its act together in terms of

governance. British Columbia is ranked 42 and California is 48. Papua New Guinea is ranked 100 and Honduras is last.7

It is difficult to untangle the precise meaning of these rankings, especially since they are based on an index that amalgamates such divergent ratings as tax burden and labor supply. In addition, states or provinces have equal status with entire countries. But the results are of interest if only because the Institute is so demon- strably in favor of the unfettered, unregulated free market. Despite what one might imagine are its fondest hopes, experienced mining people weight the quality of the mineral deposit at 60% and the policy environment at 40% in terms of influence on investment decisions. Here it appears evident that developed-country reg- ulatory environments - albeit not as effective as one might like - are not a systematic disincentive for investment. If the resource is there, that is where the companies will go. In fact, it may be that institutional strength, good infrastructure and political stability are much more important attractions for investment than many suppose.

If we set the question of TSFs in this context, several issues can be drawn out. One is that an exacting regulatory environment is not necessarily a barrier to investment in the mining industry. Second, an exacting regulatory environment can produce good environmental outcomes, as the case of McLaughlin illustrates. In short, well-designed regulatory structures and effective enforce- ment plausibly will cause the mining industry to adopt improved practices and so internalize its environmental externalities. This means that marginal ore bodies will not be developed. It also means that mining will not take place in areas that are en- vironmentally fragile or stressed and where the cost of protecting the environment makes the venture insufficiently profitable.

A further point is that mandated peer review of TSF design and operation could play an important role in moving the industry in the direction of environmental sustainability. Peer review, it seems to me, is unlike external audits of garment factories in such places as Bangladesh or Samoa which have well-known weaknesses.8 It is done by experts who really understand the implications of tech- nical decisions. If it is done before the mine plan is permitted, it will help us avoid more environmental catastrophes. If it is con- tinued periodically while the mine is in development, in operation and during closure, it will also help us avoid more environmental catastrophes. If the designers and developers know that they will be peer reviewed, it seems plausible that they will strive to do their best work so as not to be embarrassed. If they do build a TSF that fails catastrophically, it will presumably be easier to de- termine liability.

Further, peer review is something that can be documented. The campaigns to encourage consumers to boycott resources mined in socially and environmentally unsustainable ways do not give the consumer that much to go on. If entire countries adopt regulations requiring peer review, then we have a much better chance of

7 Jackson, T., Green, K.P., 2015. Fraser Institute Annual Survey of Mining Com- panies, 2014. Fraser Institute, Figure 2, p. 8, available at http://www.fraserinstitute. org/uploadedFiles/fraser-ca/Content/research-news/research/publications/survey- of-mining-companies-2014.pdf, accessed most recently on 11 April, 2015. The survey is based on 425 responses.

8 The IRMA standards, currently in draft form, must be met for formal certi- fication by IRMA. They specify independent audits by IRMA-trained auditors on a number of counts. Whether this will mandate peer review on technical issues is not clear to me. Their website says: "Certificates will be issued by third-party certifi- cation bodies whose auditors are trained by IRMA. To determine whether or not a mine site has met the IRMA requirements, certification bodies will carry out audits of the mining operation. The audit will involve on-site visits, as well as consulta- tions with the mining company employees, workers, union representatives, af- fected community members and other stakeholders. The certification body will publish a summary of its findings."

http://www.responsiblemining.net/the-irma-process/faqs/#HOWCERT accessed most recently on 1 May, 2015.

distinguishing sustainably produced from unsustainably produced minerals.

Getting to a point where adequate regulatory capacity and peer review are a normal feature of mining is a political problem, not a technical one. That may be more promising than it sounds. It gives people concerned about the social and environmental damages associated with modern mining something concrete to promote and to monitor. It gives the companies something concrete to offer when they talk about the 'social license' under which they operate. The main limiting factor might well be the number of potential peer reviewers that exist at present.

Three case studies cannot provide a comprehensive picture of industry practices. Nevertheless, the lessons that emerge from them are illuminating. The technology is advancing in ways that promise considerable improvement in the management of mine wastes. But this will not happen in the absence of significant po- litical change.

I would like to make two points in conclusion. First, the pro- blem of tailings storage facilities at any particular mine is a chal- lenging technical one. The problem of tailings storage facilities in general is a political and social one.

Second, the ability of social scientists and policy analysts to shed light on the complex social and environmental issues sur- rounding mining will be enhanced if we learn more about the engineering involved. The kind of interdisciplinary problems posed by mining need a more far-reaching interdisciplinary en- gagement than, I think, has been the case.

6. Epilogue

The Mount Polley mine restarted operations at half capacity almost exactly a year to the day from the breach, in August 2015. Tailings will be deposited in a nearby pit until a long term plan for operations - which may include a water treatment facility - can be developed. The government of British Columbia declined to bring charges against the company. No fines have been imposed since the province's mining laws do not permit administrative penalties, although changes in the law have since been proposed that would permit such penalties. The company has, so far, spent about C$70 million on clean-up (Hoekstra, 2015, 2016).

More recently, a huge new TSF breach occurred in Minas Gerais, Brazil at the Samarco iron ore mine, jointly owned by Vale and BHP Billiton. This is the worst mining accident in Brazilian history. Two tailings dams collapsed, sending an estimated 60 m3 of water and wastes 600 km down the Rio Doce to the Atlantic Ocean. Nineteen people have been confirmed dead. The state police have recommended charges of "qualified homicide" against seven cor- porate employees. The parent corporations are in negotiations with the federal government over a settlement which some esti- mate at US$5 billion (Jamasmie, 2016).