Content uploaded by lindsay newland Bowker
Author content
All content in this area was uploaded by lindsay newland Bowker on Nov 03, 2015
Content may be subject to copyright.
The RISK, PUBLIC LIABILITY, & ECONOMICS
of TAILINGS STORAGE FACILITY FAILURES
Lindsay Newland Bowker1 & David M. Chambers2
July 21, 2015
1. INTRODUCTION
Prior works interpreting the history of Tailings Storage Facility (TSF) failures, 1910-2010, have concluded that the
lower numbers of failures and incidents in the two most recent decades evidence the success of modern mining
regulation, improved industry practices and modern technology. When examined more closely the 100 years of TSF
failures shows an emerging and pronounced trend since 1960 toward a higher incidence of “Serious”3 and “Very
Serious”4 failures. That is, the consequence of loss is becoming increasingly greater.
In a keynote address at a 2011 tailings conference Dr. A. Mac G Robertson described this trend and its implications
going forward as elevating risk potential by a factor of 20 every 1/3 century. His address called a “red flag” on the
current “Mining Metric” which results in ever larger and higher TSFs (Robertson 2011).
The Mining Metric creating this exponentially increasing consequence in the event
of a tailings dam failure, is driven by continuously lower grades in identified
resources and continuously falling real prices of most metals. The costs to excavate
more material for a ton of end product at a lower price has been made possible
through technology improvements in milling and concentration processes, bulk
mining and economies of scale. There have been some new technologies e.g. dry
stack and paste tailings and the more
prevalent use of center line over upstream dam designs which offer the
potential for lower consequence in the event of failure, and perhaps a
lower overall risk of failure. However, many of the same features of
modern mining that create economic feasibility in lower grades of ore
also pose greater challenges for the management of mine waste and
waste water. One of the manifestations of these challenges overall is a
greater frequency of Very Serious tailings dam failures with significant
levels of social and economic consequence, sometimes non remediable.
49% (33/67) of all recorded Serious and Very Serious failures from 1940-
2010 have occurred since 1990. Of all 525 recorded incidents cited, 1990-
2010, 17 (33%) were Serious failures, i.e. large enough to cause
significant impacts or involved loss of life. Another 16 (31%), were Very
Serious failures, i.e. catastrophic dam failures that released more than 1
1 Bowker Associates Science & Research In The Public Interest, 15 Cove Meadow Rd Stonington, Maine, 04681, Email:
lindsaynewlandbowker@gmail.com, Tel: 207-367-5145
2 Corresponding Author, Center For Science In Public Participation, 224 N. Church Ave., Bozeman, Montana, 59715,
www.csp2.org, Email: dchambers@csp2.org, Tel: 406-585-9854
3 We defined Serious failures as having a release of greater than 100,000 cubic meters and/or loss of life.
4 We defined Very Serious failures as having a release of at least 1 million cubic meters, and/or a release that travelled 20 Km or
more, and/or multiple deaths (generally ≥ 20).
5 Our study included authoritatively documented TSF failures that were not in the WISE or ICOLD inventories. See Appendix
1, TSF Failure Data Table, for a complete list of TSF incidents & failures included in our study and the basis on which they
were classified.
Risk potential has
increased by a
factor of 20 every
1/3 century.
(Robertson 2011)
The modern “Mining
Metric” is well mapped:
higher mine production
necessitated by lower grades
of ore, a century of declining
prices offset by declining
costs per ton. The metric is
to continuously develop the
resource through economies
of scale, larger and deeper
footprints, more efficient
operations, bigger and better
bulk mining technology.
2 | Page
million cubic meters of tailings and in some instances resulted in multiple loss of life. 63% of all incidents and failures
since 1990 were Serious or Very Serious. The total costs for just 7 of these 16 large failures was $3.8 billion, at an
average cost of $543 million per failure (See Appendix 3). These losses, according to dam committee reports and
government accounts are almost all the result of failure to follow accepted practice. These failures are a direct result
of the increasing prevalence of TSF’s with greater than a 5 million cubic meter total capacity necessitated by lower
grades of ore and the higher volumes of ore production required to attain or expand a given tonnage of finished
product. We project 11 Very Serious failures 2010-2020 at total unfunded unfundable public cost of $6 billion. We
estimate an additional $1 billion for 12 Serious failures this decade. These losses are uninsurable. Very few miners
can simply absorb a loss at this scale without risking bankruptcy and permanent closure of a resource that has not
yet been “mined out”. There is no organized industry attempt to pool these losses in the context of a risk management
loss prevention program, and no political jurisdiction issuing permits is large enough to prefund a low frequency
high consequence loss of this scale. The inevitable result is either government pays or the damages go unremediated.
Much of our data on cost of large scale failures was sourced from court cases or proceedings where government
sought unsuccessfully to recover what had been spent on remediation, compensation for damages or assigned as
value for actual socio economic and natural resources loss. Shielded via wholly owned subsidiaries who can legally
declare bankruptcy when liabilities exceed assets of the subsidiary (not the parent), the parent companies paid little
or nothing toward most of these large losses. In countries founded on the common law tradition that all are
responsible for the consequence of their actions, this gap between outcome and expectation for the most serious local
impacts violates the terms and conditions of a “social license to operate” and fails to meet a standard of “polluter
pays”.
As we have seen with Mt. Polley, very large releases do not just occur at very large
mines. In comparison to the scale envisioned by mines like Pebble or KSM, the Mt.
Polley TSF was relatively small, only about 35 meters high at failure with a total
capacity of about 74 million cubic meters (Independent Panel 2015). In fact this is the
pattern we see on close examination of Very Serious and Serious failures; older TSFs
with smaller footprints are pushed to unplanned heights to accommodate additional
production that was not anticipated when the tailings dams were originally designed and the permits originally
issued.. Capital markets and investors don’t finance clean ups. They finance production that is profitable. Smaller
companies operate on tighter margins within the same overall metric affecting all miners but are less able to take
advantage of and finance optimizations or achieve economies of
scale that will keep production costs low enough to maintain a
specific mine site as economically feasible.
Our sense of the data, and the case histories we have looked to for a
deeper understanding of the data, is that “mining economics” plays
a significant role in TSF failures. It is important in permitting, and
in the checks and balances built into the regulatory process over the
life of a TSF, to look beyond “mechanisms of failure” to the
fundamental financials of the miner, the mine, and mega trends that
shape decisions and realities at the level of miner and individual
mine.
Taking our study of the relationship between “mining economics”
and TSF failures 1910-2010 into account, it is our expectation that
large failures in the near term (through 2020) will continue to come
from operating mines under ownership of smaller miners first
Miners Must Move
Forward or Perish
(Jones 2014)
Our sense of the data and the
case histories we have looked to
for a deeper understanding of
the data is that “mining
economics” plays a significant
role in TSF failures and that it
is important in permitting and
in the regulatory process to look
beyond the “mechanisms of
failure” to the fundamental
financials of the miner, the
mine, and mega trends that
shape decisions and realities at
the individual mine.
3 | Page
commissioned from the late 60’s to the early or late 80’s. These smaller older mines are producing within the Mining
Metric of lower grades and now steeply rising production costs against the continuous possibility of a sharp adverse
price swing but with much less capital, as compared with larger mines, to buffer contingencies or provide required
levels of stewardship for TSFs from design through closure. For a mega mine like the 100 year old Bingham Canyon
mine it was possible to respond to an identified threat of failure and the growing environmental problems of age. It
is not clear how smaller old mines will find the funds to identify or respond in a timely fashion to threats at their
facilities, or whether regulatory structures now in place will serve well enough to identify such “at risk” facilities.
If they are identified in time, it is not clear how smaller miners skating on thin balance sheets will finance the closure
or improvements at TSFs and carve out the funds for new TSFs where necessary. Larger mining companies, however,
are better positioned financially to manage and mitigate these threats.
This study anticipates the future trend of Serious and Very Serious TSF failures over the next decade, through 2020,
and estimates the total public economic consequence of those failures, which are presently unfunded and un-
fundable. We borrow the applicable elements of “loss development” in insurance rate making utilizing 100 years of
data on loss and consequence and on the production levels of the mining metric producing TSF waste volumes to
project an expected number of failures and an average expected loss per failure from which global estimates of
expected public loss can be reasonably estimated.
Having something more like “actuarial data” to refer to is important in understanding the potential magnitude of
loss from an individual dam or a
permitting districts portfolio of dams
and TSFs. With such low frequency
high severity losses we can never assign
risk to an individual TSF based on its
design and receiving environment
parameters. Unless it has an identified
flaw that puts it at near certain risk of
imminent failure, we can’t say whether
a given dam “will” fail. We can only say
what the consequence would be in
economic terms if it failed.
Satellite imagery has lead us to the realization that
tailings facilities are probably the largest man-made
structures on earth. Their safety, for the protection of life,
the environment and property is an essential need in
today’s mining operations. These factors, and the
relatively poor safety record revealed by the numbers of
failures in tailings dams have led to an increasing
awareness of the need for enhanced safety provisions in
the design and operation of tailings dams. (ICOLD 2001)
4 | Page
2. INCREASING CONSEQUENCE OF FAILURES
For this study we are interested primarily in the history and trend of Serious and Very Serious Failures rather than
all incidents in the International Commission on Large Dams (ICOLD) or the World Information Service on Energy
(WISE) compilations. These are the failures that cause consequential compromise of environmental security beyond
the mine site. Serious and Very Serious failures accounted for 31% (67) of the 214 TSF failures and accidents 1940-
2010, but comprise 63% (33/52) of the 52 total incidents, 1990-2010, with sufficient data for meaningful analysis.
We defined Serious failures as having a release of greater than 100,000 cubic meters and/or loss of life. 38 recorded
incidents out of the 214 failures and accidents in the period 1940 to 2010 (18%) that had sufficient data for analysis
met that criteria. 17 of those (45%) occurred in the last two decades.
We defined Very Serious failures as having a release of at least 1 million cubic meters, and/or a release that travelled
20 Km or more, and/or multiple deaths (generally ≥ 20). Very Serious failures comprised 14% of total historic events
(29/214), but 31% (16/52) of all incidents and events in the past two decades (1990-2010). The complete list and
criteria is presented in Appendix 1, TSF Failure Data Table.
This very clear trend to larger and more consequential losses is apparent in Figure 2.1 below. The clear aqua and
paler blue is the distribution of incidents other than failures, most of which are very small with little or no release or
consequential damage. Prior to 1980 Other Failures and Accidents (pale and aqua blue) were most prevalent. Post-
1990 Serious and Very Serious failures (deep and dark blue) dominate.
0
10
20
30
40
50
60
1940-49 1950-59 1960-69 1970-79 1980-89 1990-99 2000-09
Number of TSF Failures/Accidents
Decade
Other Accidents
Other Failures
Serious Failures
Very Serious Failures
Figure 2.1 Increasing Severity of TSF Failures Globally 1940-2010
Sources:
ICOLD(2001)
WISE (2015)
Wei et. al (2012)
Rico et. al (2007)
Other (see data
base)
5 | Page
3.0 RELATIONSHIP BETWEEN LARGE FAILURES & THE MINING METRIC
Our aim was to explore the relationship between economic factors not explicitly accounted for in the permitting
and regulatory oversight of mines and the observed trend toward failure incidents of greater consequence. Our
data base included a count by decade of failures (Serious failures, Very Serious failures, Other failures, and Other
Accidents) and a data set of variables describing the main economic trends driving mine production: price, costs to
produce and grade. The following chart for copper prepared by the Raw Materials Group for the World Bank
(World Bank 2006) describes the generic fundamental elements of the Mining Metric affecting all primary metals
and most precious metals.
The chart is highlighting the very dramatic change in the relationship between metals output (the red line) which
increased only 17% over the decade 1990-2000 and ore production6 which increased 63% as grades continued to
decline. The two key elements missing from this chart that explain how it was possible to “grow the resource” against
a long trend of falling prices and falling grades the economic viability of these trends are the market price of the red
line (the final refined product) and the costs to produce are highlighted by Richard Schodde, who noted that the
declining costs to produce more than offset a century of falling prices. (Schodde 2010)
This fuller context is shown in Figure 3.2 below. That production costs have offset price is apparent through 1990.
6 In our analysis we have used copper ore production data taken from the World Bank/Raw Materials Group graph because it
is the only available published data for copper ore production. We have also done a comparison by using average copper ore
grade and metal production to back-calculate to ore produced. For the back-calculation we used metal production data from
Kelly & Matos (USGS 2014a), Schmitz/ABARE (Mudd 2012), the International Copper Study Group (ICSG 2014), and copper
grade data from Mudd (2012). These data compared very favorably with the World Bank/Raw Materials Group data. We
made several attempts to contact the Raw Materials Group through their corporate parent, SNL Metals & Mining, in an
attempt to both verify the data (World Bank 2006) and the method(s) they used to develop it, but did not receive a response
to these inquiries.
Figure3.1.CopperProduction&OreGrade
6 | Page
In correlation analysis, Table 3.1, price had a lower correlation than production cost with all failure classes. The most
significant correlations with the four failure variables were with Cu Production Cost, Cu Grade and annual Cu Ore
Production volume and Cu Metal Production. The correlations were only notable with the two highest failure
severity categories. Cu Metal Production had higher correlations with both Very Serious failures (0.881) and Serious
failures (0.826) as compared with Cu Ore Production. Cu Ore Production is more closely related, however, to TSF
waste volume and also seems to distinguish between the two highest severity classes. This small difference also
occurs with Cu Grade (greater negative for Serious) and Cu Production Cost (greater negative for Very Serious).
Table3.1CorrelationBetweenFailureSeverityandMiningMetricIndicators
CuOre
Production
CuMetal
Production
Cu
GradeCuProdCostCuPrice
VerySeriousFailures0.8600.881 ‐0.794 ‐0.788‐0.427
SeriousFailures0.7200.826 ‐0.884 ‐0.682‐0.126
OtherFailures‐0.265 ‐0.0990.2980.3000.489
OtherAccidents‐0.216 ‐0.050 ‐0.3120.281 0.485
Abbreviations:
CuProdCost=Costtoproducecopperconcentratefromcopperore,includingwastedisposal
CuGrade=gradeofcopperintheore
CuProd=copperoreproduction
OtherFailures=tailingsdamfailuresandincidentsotherthanSeriousorVerySeriousFailures
SeriousFailures=Serioustailingsdamfailures
VerySeriousFailures=VerySerioustailingsdamfailures
Sources: USGS Metal Statistics (2014a), Schodde (2010), ICOLD (2001),
WISE (2015) & additional
0
5,000
10,000
15,000
20,000
25,000
30,000
35,000
1910 1920 1930 1940 1950 1960 1970 1980 1990 2000
Copper Production (Milliom Metric Tonnes)
Decade
Copper Grade
Copper Production Cost
Copper Price
CU Metals Production
Copper Ore Production
Linear (Copper Grade)
Figure 3.2 Mining Metric 1910-2010
Declining Prices Offset By Lower Production Costs
SOURCES
USGS Metal Statistics (2014a)
Schodde (2010)
Mudd (2012)
World Bank (2010)
VERTICAL SCALE
Cu Gade, Cu Price, and Cu
Production Cost are scaled to
fit on the vertical axis of this
graph, so the values
displayed are only
proportional to Cu Production
1910-19 1920-29 1930-39 1940-49 1950-59 1960-69 1970-79 1980-89 1990-99 2000-09
7 | Page
Therefore, we chose Cu Ore Production, Cu Grade and Cu Production Cost to produce for further analysis. We did
not include, or have a basis for deeper consideration, of copper price. These relationships are graphically presented
in Figure 3.3 below.
The key mining metric variable, Copper Production Cost to produce, dropped from $85/tonne in 1900 to only
$15/tonne in 2000. Over this same period price dropped from $7,723/tonne to $3,292 per tonne. The largest cluster
of Serious and Very Serious failures of TSFs, 88% (59/67), occurred in the long downward price trend from 1970 to
2000. 86% (25/29) of Very Serious failures and 89% (34/38) of Serious failures occurred during this period. 2000
marked the beginning of an upward trend in price but also a 33% increase in costs to produce, from $15/tonne in
2000 to $20/tonne by 2010 but with Serious and Very Serious failures still representing 71% (15/21) of all failures for
the decade 2000-2010.
The dramatic shift emphasized in the World Bank/Raw Metals charts (Figure 3.1) co- occurs with an upward swing
in costs to produce while grade continues to fall (Figure 3.3). This suggests a higher level of financial risk
beginning in 1990, which co-occurs with the emergence of Very Serious TSF failures.
Our data suggests that the many smaller mines and miners that became part of global production of all primary and
precious metals post-1950 were not as able to take full advantage of as many of the technologies and economies of
scale as larger miners, and therefore remained more sensitive to price changes than larger miners, with frequent
shutdowns in a small portfolio of investments as price changes made continued production unviable. Smaller miners
run on thinner balance sheets with more price vulnerability in comparison to the larger miners.
Another major factor affecting stewardship for TSFs and other mining environmental liabilities, which was not
mapped sufficiently for inclusion in our database, is access to capital markets. Smaller mines have always had access
only to more risk tolerant markets, such as the Toronto Stock Exchange, and sometimes, as in the case of Mt. Polley,
R² = 0.9984
R² = 0.8565
R² = 0.8474
R² = 0.7912
0
5,000
10,000
15,000
20,000
25,000
30,000
1940 1950 1960 1970 1980 1990 2000 2010
Copper Production (Milliom Metric Tonnes)
Decade
Cu Production Cost
Expon. (Copper Ore Production)
Linear (Copper Grade)
Linear (Very Serious Failures)
Linear (Serious Failures)
Figure 3.3 Relationship Between Mining Metric
and TSF Very Serious and Serious Failures 1940-2010
SOURCES
USGS Metal Statistics (2014)
Schodde (2010)
Mudd (2012)
ICOLD (2001)
WISE as of 2/15/15
VERTICAL SCALE
Cu Gade, Cu Price, and Cu
Production Cost are scaled to
fit on the vertical axis of this
graph, so the values
displayed are only
proportional to Cu Production
1940-49 1950-59 1960-69 1970-79 1980-89 1990-99 2000-09 2010-19
8 | Page
with one or two specific backers. The top miners are financed through markets with tight, well defined credit
standards and an increasing underwriting emphasis on full disclosure and accounting of environmental liabilities.
Smaller miners have almost no meaningful access to insurance for their environmental liabilities, whereas larger
miners have more integral relationships with insurance and reinsurance markets (even though the types of risks that
are insurable are no different between large and small insurers). These large market relationships create more
external accountability to environmental risk management and to financial risk management for larger miners than
exists for small miners, and a more rigorous ongoing process of review and reckoning. Regulatory structures don’t
include enough structure on assessment of financial capacity to balance that difference creating an “apparent norm”
of higher financial risk in smaller mines that translates into the higher losses we see in the historical data.
Two significant changes in financial risk also weigh more heavily for smaller mines than for larger mines: a radical
contraction of all capital markets for mining (Jones 2014); and, a 30% increase in costs to produce. The increase in
costs to produce is across the board and attributable, according to informed market analysts, to both an increase in
energy costs and also in foreign exchange rates. Chile, a major producer of copper globally, has had to commit to a
major capital program to improve its mining infrastructure to maintain grade and hold its place in world concentrate
markets.
While each principal base metal (iron, aluminum, copper, zinc, etc.) has its own version of the Mining Metric, the
basic “shape” and slope of trend lines for production and price for all base metals are the same. The basic bottom
line, vis-a-vis manifest environmental loss across all metals, is the same. All operate on close margins. Those with
larger budgets, better quality assets, lower production costs and uniform corporate policies on optimization and
efficiency at each site, and who can also achieve economies of scale, will generally fare better than smaller miners
with tighter budgets and less access to global capital markets. The global capital markets are able to provide external
checks and balances on financial/risk management relationships that hold miners to account on environmental
liability management, even when regulatory structures don’t – but only if the miner in question is working in the
global capital market.
Copper is widely recognized as a bellwether base metal for the mining industry. Most works on mining economics
use copper as the “index metal”. Beyond that, the greater quality and detail of regularly produced copper commodity
information over the entire last century led us to explore its use as the index metal for TSF failures, i.e. expressing
TSF failures per million tons of copper production. The USGS publishes metal statistics on two of Mining Metric
elements, price and mine production, but no historical data on costs to produce or grade. So copper is the only metal
for which it was possible to establish a full century long “actuarial” data base on the relationship between the
economics of mining and environmental loss attributable to TSF failures. Going forward it will be possible to build
the data base for other metals from current and data and short term projections. In the next section we present the
statistical correlation between mining economics and TSF failures.
9 | Page
4.0 THE STATISTICAL CORRELATION BETWEEN MINING ECONOMICS &
ENVIRONMENTAL LOSS FROM TSF FAILURES
We chose Canonical Correlation Analysis (CCA) as a way of further exploring the relationship between the failure
severity categories we created for this research and the main elements of the mining metric that affect all miners and
all mine sites. We were interested in knowing whether there is a significant relationship and if so, whether it warrants
greater attention in permitting standards and oversight of mine permits. We know from past study of TSF failures
that there are many physical attributes of a TSF that influence severity as well as other often noted but so far
unstudied factors such as the structure of the regulatory framework and the technical capacity available to oversight.
Canonical Correlation is a multivariate technique that aims at identifying the degree of influence of one data set with
another (rather than causality). We had no pre conceived notion of what the degree of influence might be, nor did
we have the data set we would like to have had. Nevertheless, the results of this exploration strongly suggest that
the influence of the mining metric on frequency and severity of TSF failure is unexpectedly strong.
The First Canonical variant F1 explained 95% of the variability between the two data sets (failures v mining metrics
elements). The correlations between F1 and both high severity variables are strong: Very Serious (-0.922); and,
Serious (-0.995). The Wilks Lambda on F1 was 0.046 indicating a high degree of certainty that the two data sets
(Failures and Mining Metric are not independent of one another). The Eigenvalue for F1, 0.903, suggests a very
strong linear relationship between the two data sets (See Appendix 2, Technical Documentation on Canonical
Correlation Analysis, for the data set and complete technical documentation on the Canonical Correlation).
Table4.1CanonicalCorrelationValues
F1
CanonicalCorrelation0.950
Eigenvalue0.903
Wilks'Lambda0.046
Correlationbetween:
VerySeriousfailures&F1‐0.922
Seriousfailures&F1‐0.995
Because no other research team that we could find had explored the dimensionality of this relationship, we began
with a larger set of mining metric variables beyond the 4 basic variables (Cu Production, Cu Production Cost, Cu
Price and Cu Grade), and also attempted to create variables indicating the characteristics of TSF’s so that the degree
of influence of the mining metric variables could be compared with dam characteristics. We integrated all
ICOLD/WISE recorded incidents from 1910 to 2010 into a single reconciled data set, and in the course of our research
on consequence of those incidents discovered several compilations that added to WISE/ICOLD, and which also filled
in gaps on our main indicators of consequence (total TSF release and release run out). We used both correlation
matrix analysis and canonical correlations to find the strongest set of mining metric variables, which turned out to
be tons of Cu Ore Production, Cu Production Cost, and Cu Grade. As there was only one recorded Serious failure
prior to 1940 and very little information on all incidents, our final data set and analysis focused on the period 1940-
2010.
Initially, none of our created synthetic variables for the Mining Metric were as strong the four main variables (copper
price, production cost, grade, and copper ore production). One variable, Risk Factor, which combined cost and
production volume into a single indicator actually had higher correlations with each of the two most Serious failure
10 | Page
categories and also in linear regressions on each of the two highest severity categories. It did not perform as well in
lieu of production and cost, though, in a canonical correlation. Further work is needed to evaluate Risk Factor so we
are not presenting it here. Within the 4 basic variables price and cost canceled each other out, and cost was the
stronger correlation, so the final data set for the Canonical Correlation was only cost, production and grade.
We were not able to develop a meaningful data set on dam characteristics for comparisons of degree of influence as
between the variables of the Mining Metric and various dam characteristics (dam height, volume, etc.).
Even though these results are not conclusive, because the number of observations is very small for a CCA, they are
persuasive evidence of a greater than expected and very significant influence of Mining Metric mega trends on the
frequency and severity of TSF failures. Further, it is important to note that these are not “individual measurements”
in the usual sense, but rather aggregations by decade of over 200 observations, and so should be afforded more
consideration and weight than would normally attend such a small set of observations. The data set and the full
CCA output are at Appendix 2, Technical Documentation on Canonical Correlation Analysis, along with additional
technical annotation.
Although further research would be useful to shed more light on how these mega trend variables interact to affect
failure, these results in our opinion support a conclusion that financial feasibility of the mine and financial capacity
of the miner require greater specific consideration on permit issuance and permit oversight.
Strength of Influence of Copper Ore Production
Among the variables in the Mining Metric data set we were especially interested in the relative degree of
influence/connection between copper ore volumes and the TSF failure categories especially whether it could be a
reliable denominator for TSF failure rates. The conventional one to one correlations, which are a standard output of
CCA in XLSTAT©, showed that both Very Serious and Serious failures were strongly correlated with copper ore
production, 0.860 and 0.720 respectively. We had both production and price data on all metals 1900-2010 from the
USGS metal statistics (USGS 2014a), but the correlations with aggregate all metals production and the failure
variables were not nearly as strong. So the CCA output also lent support to copper ore production as the most
reliable and meaningful denominator for TSF failure rates.
Although we did reasonably form an expectation that the mega trends would have a measureable and significant
effect on the failure categories established (i.e. that the mega trends contribute to severity), we also know from dam
committee reports and other research that many other dam specific elements have a known effect on severity of
failure. The final output of a canonical correlation is a set of synthetic variables which maximize the accounting for
mutual variability between the two sets of variables. Thus it is an approach which inherently recognizes that all of
the information needed to explain the output of interest, the severity of failures over time, are not contained in the
analysis, and further that the influence that may exist within in the expected determinant set (the mega trend
variables) may result from complex interactions among the determinant data set.
While Canonical Correlation Analysis, and its focus on dimensionality rather than causality, may be the perfect tool
for exploring the effect of mega trends of the Mining Metric on the trends in severity of TSF failures, many key
variables that would shed more light were not available. We would hope in the future to have a more rich and
complete data set, including standing TSFs that didn’t fail with the same geographic distribution as those that did.
11 | Page
At present there is no comprehensive compilation of recent or historic tailings dam failures. This is partly
understandable given the multi-national nature of the mining industry, but given the severity of the problem,
coupled with the fact that it is probably not realistic to think that the problem can be solved without a full analysis
of the nature of the problem, it is disappointing that someone has not stepped forward to perform this service.
5.0 FREQUENCIES & PROJECTIONS FROM COPPER PRODUCTION VOLUMES
The results of the correlation analyses give strong support that copper production volumes are a meaningful
denominator for TSF failures. Even if there were a centrally professionally maintained inventory of TSFs it would,
in our opinion, still be preferable to express TSF failures on the basis of mine production.
Copper metal production is the only reliably managed data element we have available globally that correlates
directly with TSF risk potential. The analysis shows us, however, that copper ore production distinguishes more
clearly between the two high severity failure categories and is a better descriptor of risk. While it is not routinely
and authoritatively compiled and reported as metal production is, the World Bank/Raw Materials Group data
(Figure 3.1) did give us an authoritative and reliable historical compilation. As ore production volume is more
directly related to TSF waste, in our opinion Cu Ore Production is the better predictor to use. We don’t have a global
census inventory of standing TSFs. To be meaningful any denominator must be available for all TSFs globally as it
is only through data on the global whole that meaningful expectations and comparisons can be made at the level of
a nation, province or state.
Secondly, we know there is a great deal of variation in the standing operating TSFs at any point in time. Size and
therefore possible maximum consequence of failure varies from small mines with a total capacity of less than 105
cubic meters to those over 107 cubic meters. Therefore, failure frequency per TSF isn’t meaningful without enough
attending globally available data to adjust for size and other known risk factors. Post failure it is possible to
reexamine the losses more closely, taking account of the specific characteristics of the particular TSF (and eventually
to recompile findings if enough new information is developed or if there is more systematic capture of these elements
in WISE or other data sources).
Thirdly, we know that the risk profile of TSFs is constantly changing based on production volumes, and how the
waste volumes generated from that production are managed. We know that 90% of all TSF failures in Europe (Rico
et. al. 2008), to 95% in China (Wei et. al. 2012), occur during operations, as opposed to being in standby or in closure.
Cu Ore Production provides an equalized basis for looking across an inventory of TSFs with highly varying size, and
it is more directly tied to the phase of active life for the TSFs in which most failures occur (Rico 2008).
Table 5.1, below, shows the failure incidents data for Very Serious failures, Serious failures and Other failures by
decade, expressed per million tons of copper ore production. For example, a 0.0020 rate for Other failures in 1940-
1949 on 2,545 million tons of ore production describes 1 event. A 0.0006 rate on 16,437 million tons (16.44 billion) of
ore production in 1980 describes 10 Other failure events.
12 | Page
Table5.1FailuresperMillionTonnesCopperMineProduction1940‐2011
Decade
CuOre
Prod
(MMt)
Very
Serious
failures
(#)
Very
Serious
failures
rate
Serious
failures
(#)
Serious
failures
rate
Other
Failures
(#)
Other
Failures
rate
Other
Accidents
(#)
Other
Accidents
rate
1940‐492,54510.000400.000050.002000.0000
1950‐593,68000.000000.000070.001900.0000
1960‐695,00430.000640.0008250.0050170.0034
1970‐797,44540.000580.0011230.0031150.0020
1980‐8910,57550.000590.0009220.0021140.0013
1990‐9916,43790.000590.0005100.000630.0002
2000‐0923,65870.000380.000350.000210.0000
Total/Ave69,344290.0004380.0005970.0021500.0010
Abbreviations:
CuProd=copperoreproductioninthedecadenotedinmillionsofmetrictonnes
VerySeriousfailure=multiplelossoflife(~20)and/orreleaseof≥1,000,000m3semi‐solidsdischarge,
and/orreleasetravelof20kmormore.
Seriousfailure=lossoflifeand/orreleaseof≥100,000m3semi‐solidsdischarge
Otherfailures=ICOLDCategory1failuresotherthanthoseclassifiedasVerySeriousorSerious
OtherAccidents=ICOLDCategory2accidentsotherthanthoseclassifiedasVerySeriousorSerious
FailureRate=numberoffailurespermillionmetrictonnes(MMt)CuOreProduced
The overall rate of Very Serious failures and Serious failures 1940-2010 were comparable, 0 00004 and 0.0005
respectively. As expected, the higher the severity the lower the frequency. The frequency rates for all the lower
severity loss categories were much lower; 0.021 Other Failures, 0.0010 for Other Accidents.
As shown in Figure 5.1 below the most dramatic change occurred with the shift from predominantly Other Failures
(less Serious failure events) to predominantly more Serious failures post 1970. Across the board for each failure
category, the rate of failure per ton of copper production has decreased. However, as noted in the introductory
section, the severity of failures has steadily increased. More of the failures that occur are Serious or Very Serious).
Our data is incomplete (we don’t have actual loss data for every Serious and Very Serious failure), however it is
certain that that the absolute consequence of all TSF failures has increased and is increasing substantially. This is
obvious in that 55% (16/29) of all catastrophic (Very Serious failures) over the past 100 years have occurred since
1990, and that 74% (17/23) of all failure events post-2000 are Serious or Very Serious.
0.0000
0.0010
0.0020
0.0030
0.0040
0.0050
0.0060
1940 1950 1960 1970 1980 1990 2000
FAILURE RATE
(per Million Metric Tonnes
Cu Ore Produced)
Figure 5.1 Failure Rates by TSF Failure Severity
Very Serious Failure Rate
Serious Failure Rate
Other Failuress
Other Accidents
1940-49 1950-59 1960-69 1970-79 1980-89 1990-99 2000-09
13 | Page
6.0 PROJECTIONS FROM COPPER MINE PRODUCTION V. FAILURE TRENDS
The heart of risk analysis is to reliably measure and forecast expected losses that are beyond control (and to hopefully
finance these losses via third party transfers, i.e. insurance or risk pool). We know that will not apply to TSF failure
losses, as almost without exception all losses were subject to control and prevention. The basic techniques for
forecasting future losses, based on past loss experience, are nevertheless applicable to anticipating the future
consequences of continuing the Mining Metric without some new forms of regulatory control and oversight which
takes more adequate account of the financial viability of the deposit and the miner.
The Copper ore production estimate for this decade (2010-2019) is advanced from the equation associated with the
trend line which had an extremely high R square, 0.9984. The result is 36,338 million metric tonnes, a projected
increase of 54%.
In insurance rate making the normal procedure for estimating future losses is to combine the last four years of loss
data. For this data, though, each cell represents 10 years of experience data not 1, and we can see from analysis of
the variables over 100 years that the events that shape loss and failure are unique to each decade, i.e. that each decade
has its own pattern of determinant/loss-affecting characteristics.
Table 6.1 below compares three estimates of next decade failures based on three approaches to uses of copper
production based frequencies: (1) average of last three decades; (2) last decade only; and, (3) “50-50” weighting
between most recent decade and last three decades. The trended values based on failure data alone are presented in
Table 6.1 in the last row of the table.
The chart values in Table 6.1 are computed from the trend line equations as they appear in Figure 6.2 (The trend lines
in Figure 6.2 are linear data projections, rounded to the nearest whole number).
Very Serious failures 2020 = 0.1393*2020-271.64 = 9.746
Serious failures = 0.1643*2020-3189.6 = 12.026
y = 1E-28e0.0373x
R² = 0.9984
0
5,000
10,000
15,000
20,000
25,000
30,000
35,000
40,000
1940 1950 1960 1970 1980 1990 2000 2010
Cu ORE PRODUCTION
(Million Metric Tonnes)
DECADE
Figure 6.1. Cu Ore Production
1940-49 1950-59 1960-69 1970-79 1980-89 1990-99 2000-09 2010-2020
14 | Page
Table6.1Predictions2010‐2020FromHistoricFailureRates
VerySerious
failuresSeriousfailuresOtherFailuresOtherAccidents
BasisRatePred.RatePred.RatePred.RatePred.
Last3DecadeAve 0.000415.90.000621.00.001035.10.000518.8
LastDecade 0.000310.80.000312.30.00027.70.00001.5
50‐50Weighting 0.000413.30.000516.70.000621.40.000310.1
Chart 9.5 12.0
Rate=numberoffailurespermillionmetrictonnes(MMt)oremined
Pred=numberofpredictedfailuresintheperiod2010‐2019
The high R-squared values on the trend lines for both Serious failures and Very Serious failures indicate a “goodness
of fit” that is apparent on visual inspection alone (i.e. the markers closely track the trend line). The calculated
predictions by chart trend line equation most closely matches the prediction based on the most recent decade failure
rates.
The canonical correlation demonstrates that the trends in the high severity failures are shaped by the entire metric
(as represented in grade, cost and production). Inspection of the data set shows that the main elements of the metric
as of 2009 were very different than those of either of the prior two decades. It is not likely costs will return to as low
as $15 or that prices will fall to as low as they were in either of the two most recent decades. Therefore we have
greater confidence in the most recent failure rate by class than we do in the either the average of the last three decades,
or a 50-50 weighting between the average of the last three decades and current decade. Still there are already clear
indications that this decade involves uncertainty about the direction of cost to produce, price, and perhaps even
production volumes. The previous two decades both had constant costs of production against failing prices, a very
different pattern with an expected higher rate of failure. Mid-decade 2010-2019 the overall environment seems to be
trending toward higher financial risk, and therefore higher potential environmental liability than the 2000-2009
decade.
We are though projecting 12 Serious failures and 11 Very Serious failures for the present decade (2010–2019) relying
on the failure rates of the most current decade (see Table 6.1).
15 | Page
Our dataset included 5 failures 1910-2010 that met our criteria for Very Serious that were not listed in WISE or ICOLD
data bases, from a compilation of Chinese major failures and a compilation of Philippine significant tailings incidents.
The frequency rate 2000-2009 was essentially the same with or without these five failures. We cannot say that
whatever undercount actually exists in WISE/ICOLD data would have no bearing, however, in our view this is a
conservative projection quite apart from the possible undercount issue. It makes no allowances for the possibly
higher risks of price jitters on many metals (e.g. molybdenum, iron, zinc, gold), of rising production costs mostly
from energy and foreign exchange rates, and the uncertainty about the roles China, Chile, and Peru (as producers,
and China and India (as consumers) will play, and how that could elevate financial risks for smaller mines and
smaller miners.
7.0 PROJECTED COST OF REMEDIATION AND NON REMEDIABLE
UNCONTROLLED RELEASES FROM TSFS
We searched the historic record for what local authorities had deemed the costs of public damages from the major
releases in our database, and found sufficient authoritative documentation on a total of 6 of the 14 post-1990 Very
Serious uncontrolled TSF originating release incidents. Our process was to translate from foreign currency to US in
the year of the incident and then to convert those $US to 2014-$US. The average cost of the 7 incidents for which we
found authoritative data was $543 million (Figure 7.1). That translates to a projected public liability for remediation
of 11 Very Serious releases from TSFs at cost of approximately $5 billion globally before the end of this decade (2020).
We did not attempt any estimates for the expected 12 Serious failures by 2020 but a guess of an additional $1 billion
is probably not unreasonable.
R² = 0.8474
R² = 0.7912
0
5
10
15
20
25
30
1940 1950 1960 1970 1980 1990 2000 2010
Number of Failures
Decade
Very Serious Failures
Serious Failures
Other Failures
Other Accidents
Linear (Very Serious Failures)
Linear (Serious Failures)
Figure 6.1 Failure Predictions By Trend Line
1940-49 1950-59 1960-69 1970-79 1980-89 1990-99 2000-09 2010-2020
16 | Page
Usually losses are forecast from a record of homogeneous data maintained by one source over time by the entity
which has actually incurred or paid out those losses (i.e. an insurer or a rating bureau like the Insurance Services
Office), or a company’s or agency’s risk manager. That is not true of our loss history data for TSF failures.
Although WISE has followed with some detail on a few cases involving litigation for recovery of outlays (e.g. for
Los Frailes), descriptions of consequence are brief and narrative. There are few links to more in-depth authoritative
analysis on consequence. Losses are not systematically or uniformly captured or developed as part of either the
WISE or ICOLD databases. The costs data we present here is all we could find for Very Serious post-1990 failures
which pertained to environmental losses, and which were cited or developed by authoritative or credible sources.
We aimed for as much homogeneity as possible in choosing amounts documented for inclusion in our loss history
(i.e. to include only natural resources/environmental losses whether or not cleanup was ordered or undertaken. In
one case, Omai, we used a token amount to acknowledge what farmers, fisherman, and NGOs attempted to
recover, and to acknowledge what is widely agreed was environmental damage notwithstanding the governments
judgments to the contrary. The token amount allocated to Omai actually lowers the overall average cost estimate
but, given all the litigation and controversy that has attended, simply admitting to the extent of environmental
damage we felt Omai could not simply be left off the list, even though we could not find documentation on what
part of $2 billion joint damage claim was attributable to documented environmental damages to lands and waters.
While sketchily sourced and documented, the few failures which are systematically and authoritatively developed
give us a high level of confidence that our average natural resource loss of $543 million for a catastrophic failure is
not overstated. For example, the estimated costs to clean up the Los Frailes spill was borne primarily by the
Andalusian Government as a non-remediable loss. We think that situations like this, where the actual costs are so
high or cleanup costs so astronomical that losses from Very Serious TSF failures will more and more be permanent
non-recoverable losses. Mt Polley is a possible example of a tailings spill into a creek and lake that will not be
retrieved. Such losses will, hopefully, still have a complete accounting of value whether or not remediation is
ordered, undertaken, or possible.
The data on the 7 failures forming the basis of our average loss amount of $543 million and its sources are presented
in Table 7.1, below. See Appendix 3 for more detail on this chart.
Apply this to our projections of the number of Very Serious failures, 11 results in a projected unfunded unfundable
public liability loss of $6.0 billion from Very Serious TSF failures for the decade 2010-2019.
Our sense of the data and case histories is that this decades’ TSF failures will continue to arise mostly from standing
operating TSFs, pushing older TSFs up to and past their original designs, or stretching the limits of TSFs that were
not built or managed to best practices in the first place. We expect most to arise from smaller mines and miners. We
see in the record an indication that in many instances releases and events suggesting fundamental problems with the
structure of the TSF preceded a final catastrophe by two to four years. In the cases of Golden Cross (New Zealand),
Bingham Canyon (Utah), and Mike Horse (Montana) long term issues with dam stability led to closures in time to
avert catastrophe at costs that were significantly lower than the remediation costs or assessed damages would have
been for a structural failure.
17 | Page
Table7.1DocumentedTSFVerySeriousNaturalResourceLosses1990–2010
TSFFailureYear
Original
Currency
(Millions)
Failure
Year
MUS$
2014
MUS$Ore
Release
(Mm3)
Run
Out
(km)
Deaths
KingstonFossilPlant,Harriman,
Tennessee,USA2008US1,200$1,200$1,300 5.44.1
Taoshi,LinfenCity,Xiangfen,
ShanxiProvince,China2008US1,300$1,300$1,429Fe0.192.5277
BaiaMare,Romania2000US179$179$246Au0.15.2
LosFrailes,Spain1998EU275$301$437Zn/Cu
/Pb4.65
MarinduqueIsland,Philippines1996P180+
US114$123$185Cu1.627
Omai,Guyana1995US100$100$156Au4.280
Merriespruit,SouthAfrica1994R100$29$46Au0.6217
======
AverageUS$2014:$543$3,799
Reviewing their own role in creating and perpetuating the environment in which we have allowed TSFs at risk of
consequential failure to proliferate, the International Bank for Reconstruction & Development and the International
Development Association put it well:
“Governance should be strengthened until it is able to withstand the risks of developing major extractions. Once that has
happened, the International Bank for Reconstruction and Development (IBRD) and the International Development
Association (IDA) can add support for the promotion of a well-governed extractive sector. Similarly, when the International
Finance Corporation and the Multilateral Investment Guarantee Agency (MIGA) consider investing in an oil, gas, or
mining project, they need to specifically assess the governance adequacy of the country as well as the anticipated impacts of
the project and then only support projects when a country’s government is prepared and able to withstand the inherent
social, environmental, and governance challenges.” (IFC 2003)
Our study has provided a very conservative estimate of future unfunded public liabilities for standing, already
operating, and permitted TSFs globally. We know globally that every one of those failures can be prevented for a
cost much less than $6.0 billion for just the 11 Very Serious failures we are predicting by 2020.
We know globally, and in Canada and the US, the regulatory structure is not presently in place to identify and correct
these at-risk TSFs before they fail, and we know many of them are operated by companies whose balance sheets are
too thin to fund repairs and closure where necessary.
We hope our work will begin a collaborative and highly focused multi-disciplinary dialogue to prevent the
materialization of these $6.0 billion in public losses by 2020.
18 | Page
8.0 SUMMARY & CONCLUSIONS
The advances in mining technology over the past 100 years which have made it economically feasible to mine lower
grades of ore against a century of declining prices have not been counterbalanced with advances in economically
efficient means of managing the exponentially expanding volume of associated environmental liabilities in waste
rock, tailings and waste waters. In fact those new technologies which do offer better management of mine wastes
usually add significant cost and are often detrimental to bottom line financial feasibility. This is evidenced in a post-
1990 trend toward un-fundable environmental losses of greater consequence. This interdisciplinary review of TSF
failures 1910-2010 establishes a clear and irrefutable relationship between the mega trends that squeeze cash flows
for all miners at all locations, and this indisputably clear trend toward failures of ever greater environmental
consequence.
The implication of our findings is that a continuation of the present Mining Metric is not environmentally or
economically sustainable, and that regulatory systems must begin to understand and address financial capacity of
the miner, and the financial feasibility of mining itself, both in permitting criteria and in oversight of mine water
management over the life of the mine.
Our findings point toward undocumented and unstudied risks of failure in the standing operating already permitted
mines of smaller miners globally where cash flow pressures have led to an avoidance of best practices in waste
management, and where political pressures have led to avoided close scrutiny of decades of neglect and shortfalls.
We have not identified an existing statutory or regulatory system anywhere that has the authority and capacity to
identify and prevent the $6 billion in losses we estimate the public globally will be liable for by the end of this decade.
#####
19 | Page
9.0 REFERENCES & BIBLIOGRAPHY
Advani 2011. Advani, Sameer “Real Historical Copper Price”, May 30,2011, Sameer Advani’s Instablog
http://seekingalpha.com/instablog/946411-sameer-advani/182708-real-historical-copper-price
Azam & Li 2010. Azam, Shahid Li, Quiran “Tailings Dam Failures: A Review of the Last 100 Years”, Geotechnical
News December 2010 http://www.infomine.com/library/publications/docs/Azam2010.pdf
Bertram 2010. Bertram, Geoff “Mining Economics & The Conservation Estate” Royal Forest & Bird Protection
Society of New Zealand Inc., September 2010 accessed November 2014
http://www.geoffbertram.com/fileadmin/Mining%20Economics%20and%20the%20Conservation%20Estate%20
main%20text.pdf
Boswell & Sobkowicz 2011. Boswell, Jeremy, Sobkowicz, John “Duty Of Care Applied To Tailings Operations”,
Proceeding Tailings & Mine Waster, Vancouver B.C. Nov 6-9 2011
http://www.infomine.com/library/publications/docs/Boswell2011.pdf
BCMM 2013. British Columbia Ministry of Mines “ Provincial Summary Exploration and Mining In British
Columbia 2013”
http://www.empr.gov.bc.ca/Mining/Geoscience/PublicationsCatalogue/InformationCirculars/Documents/IC
_2014-01.pdf
BCOAG 2011. British Columbia, Office of the Auditor General “An Audit of the Enviornmnetal Assessments Office
Oversight of Certified Projects, July 2011 http://www.bcauditor.com/pubs/2011/report4/audit-bc-
environmental-assessment-office-EAO
Caldwell 2006. Caldwell,Jack “British Columbia Tailings Failure at The HB Mine Near Salmo, British Columbia”,
TechnoMine July 6, 2006
http://technology.infomine.com/articles/1/8455/tailings.failure.canada/british.columbia.tailings.aspx
COAG 2014. Canada, Office of the Auditor General “Implementation of the Canadian Environmental Assessment
Act, 2012, 2014 Fall Report of the Commissioner of Environment and Sustainable Development http://www.oag-
bvg.gc.ca/internet/English/parl_cesd_201410_04_e_39851.html
Chambers & Higman 2011. Chambers, David M., Higman, Bretwood ”Long Term Risks Of Tailings Dam Failure,”
October 2001, accessed November 2014 at http://www.csp2.org/technical-reports
Davies et. al. 2002. Davies, Michael, Martin,Todd and Peter Lighthall “Mine Tailings Dams: When Things Go
Wrong,” AGRA Earth & Environmental Limited, Burnaby, BC 2002
http://www.infomine.com/library/publications/docs/Davies2002d.pdf
Davies & Martin 2009. Davies, Michael, Martin,Todd “Mining Market Cycles and Tailings Dam Incidents,”
AMEC Earth & Environmental, Vancouver, British Columbia, Proceeding of the Thirteenth International
Conference on Tailings and Mine Waste, November 1, 2009, Banff, Alberta, Canada accessed November 2014 at
http://www.infomine.com/library/publications/docs/Davies2009.pdf
EGN 2014. Ecology Global Network “Population Estimates: Year 1 to 2050” accessed November 2014 at
http://www.ecology.com/population-estimates-year-2050/
20 | Page
ICMM 2012. International Council on Mining & Metals, “Trends in the Mining & Metals Industry” Mining’s
Contribution To Sustainable Development October 2012. Accessed November 2014 at
http://www.icmm.com/document/4441 M
ICOLD 1995. International Commission on Large Dams, “Dam Failures Statistical Analysis,” Bulletin 99, 1995
ICOLD 2001. International Commission on Large Dams and the United Nations Environmental Programme
(UNEP) Division of Technology, Industry and Economics (DTIE), “Tailings Dams Risks of Dangerous
Occurrences Lessons Learned From Practical Experiences,” Bulletin 121, 2001
ICSG 2014. International Copper Study Group, Copper production statistics, November 13, 2014.
IFC 2003. Striking a Better Balance, the World Bank Group and Extractive Industries, the Final Report of the
Extractive Industries Review, International Finance Corporation, December 2003
Independent Panel 2015. Report on Mount Polley Tailings Storage Facility Breach, Independent Expert
Engineering Investigation and Review Panel, Province of British Columbia, January 30, 2015
Jakubekl et. al. 2013. Jakubekl, Jaret , Clayon, Russel, Guest Alan R., “Mudrush Risk Evaluation” CIM Journal 2013
accessed November 2014 at
http://www.srk.com/files/pdfs/mudrushriskevaluation_paper_cim_journal_fnl_20130322_jj.pdfailings
Jones 2014. Jones, Sam Jordan, Editor/Compiler “Global Mining Finance Guide 2014” Aspermont Media 2013
accessed November 2014 at http://www.mayerbrown.com/files/Publication/3e2dc5e3-878a-40a2-8b06-
cb75f8d632ba/Presentation/PublicationAttachment/86b233d1-beb2-4f47-9b85-
cd4793b90b92/global_mining_finance_guide_jan14.PDF
Kelly & Matos 2013. Kelly, T.D., and Matos, G.R., comps., 2013, Historical statistics for mineral and material
commodities in the United States (2013 version): U.S. Geological Survey Data Series 140, accessed November
2014, at http://minerals.usgs.gov/minerals/pubs/historical-statistics/.
Knight Piesold 2011. Knight Piesold Ltd., Mount Polley Mining Corporation Mt Polley Mine Storage Facility
Report of 2010 Annual Inspection”, Prepared for Mt. Polley Mining Corporation, January 25, 2011
http://www.scribd.com/doc/241108453/Mount-Polley-Mine-Tailings-Storage-Facility-Report-on-2010-annual-
inspection-by-Knight-Piesold
Kossoff et. al. 2014. Kossoff, D. Dubbin, W.E.Porter, Edwards, S.J. , Macklin, M.G., Hudson Edwards, K.A. “Mine
Tailings Dams: Characteristics, Failure, Environmental Impacts, and Remediation”, Applied Geochemistry V.51,
2014 PP 229-245 accessed November 2014 at http://dx.doi.org/10.1016/j.apgeochem.2014.09.010
Mudd et.al. 2012. Gavin M. Mudd, Zhehan Weng, and Simon M. Jowitt, A Detailed Assessment of Global Cu
Resource Trends and Endowments, Economic Geology, v. 108, pp. 1163–1183, August 30, 2012
Porter & Beliwas 2003. Porter,Kenneth E., Beliwas ,Donald I., Physical Aspects Of Waste Storage From a
Hypothetical Open Pit Porphyrry Operation, U.S. Department of the Interior, U.S. Geological Survey, 2003,
accessed November 2014 at http://pubs.usgs.gov/of/2003/of03-143/of03-143.pdf
21 | Page
Repetto 2014. Repetto, Robert “Silence is Golden, Leaden and Copper Disclosure of Material Environmental
Information in the Hardrock Mining Industry” Yale School Of Forestry & Environmental Studies, July 2004
accessed November 2014 at http://environment.research.yale.edu/documents/downloads/o-
u/repetto_report_execsum.pdf
Rico et al. 2007. Rico, M., Benito, G., Díez-Herrero, A. “Floods From Tailings Dam Failures” Geological Hazards
Unit, Spanish Geological Survey (IGME), Madrid, Spain, October 2, 2007
http://digital.csic.es/bitstream/10261/12706/3/MayteRico_10.pdf
Rico et al. 2008. Rico, M. Benito, G., Salgueriro, A. Diez-Herrero, A. & Pereira, H.G. Reported Tailings Dam
Failures: A review of the European Incidents in the Worldwide Context, Journal of Hazardous Materials 152
(2008), pp. 846–852, http://digital.csic.es/bitstream/10261/12716/3/MayteRico_08.pdf
Robertson 2011. “Mine Waste Management in the 21st Century Challenges & Solutions Beyond Incremental
Changes” Andrew Macgregor Robertson, Key Note Address Tailings & Mine Waste, Vancouver B.C., 2011,
http://www.infomine.com/library/publications/docs/Robertson2011c.pdf
Samis et. al. 2012. Samis, M., Martinez, L., Davis, G. A., and Whyte, J. B., “Using Dynamic Discounted Cash Flow
and Real Option Methods for Economic Analysis in NI 43-101 Technical Reports.” in the Valmin Seminar Series
2011-12 Proceedings, The Australian Institute of Mining and Metallurgy, Publication Series No 3/2012, 149-160
accessed December 2014 at http://inside.mines.edu/~gdavis/Papers/ValMin.pdf
Schodde 2010. Schodde, Richard “The Key Drivers Behind Resource Growth: An Analysis of the Copper Industry
For the Last 100 Years”, Mine EX Consulting, 2010 MEMS Conference Minerals and Metals Over the Long Term,
Phoenix Arizona, March 3, 2010 . http://www.slideshare.net/RichardSchodde/growth-factors-for-copper-
schodde-sme-mems-march-2010-final
Stano 2011. Stano, Maya “Rave Mine: A Regulatory & Fiscal Black Hole” Environmental Law Centre, Spring 2011
http://www.elc.uvic.ca/publications/documents/2010-02-03-RavenMine-RegulatoryandFiscalBlackhole_Stano-
REVISEDJul2011.pdf
Troncosco et. al. 1993. Troncosco, J.H. , Avendano, A. Vergara, A. “The Seismic Failure of Barahona Tailings Dam”
Procedings: 3rd International Conference on Case Histories in Geotechnical Engineering, St. Louis, Missouri, June
1-4, 1993, Paper # 2.56, Accessed November 2014 at
https://mospace.umsystem.edu/xmlui/bitstream/handle/10355/33917/P%201473-
%20The%20Seismic%20Failure%20of%20Barahona%20Tailings%20Dam.pdf?sequence=1
USGS 2014a. U.S. Geological Survey “Historical Statistics for Mineral and Material Commodities in the United
States” Thomas D. Kelly and Grecia R. Matos, with major contributions provided by David A. Buckingham, Carl
A. DiFrancesco, Kenneth E. Porter, and USGS mineral commodity specialists. Data Series 140, April 2014.
Accessed November 2014 at http://minerals.usgs.gov/minerals/pubs/historical-statistics/#copper
USGS 2014b. U.S. Geological Survey “Iron Ore Statistics” Compiled by T.D. Kelly (retired), W.S. Kirk (retired), J.D.
Jorgenson (retired), and C.A. Tuck. Data Series 140 2014 Accessed November 2014 at
http://minerals.usgs.gov/minerals/pubs/historical-statistics/#copper
USGS 2014c. U.S. Geological Survey “Zinc Statistics” April 2014 Compiled By DiFranceso, C.A. Plachy, J., Gabby
P.N., Bleiwas,D.I., Tolcin, A.C. Data Series 140 April 2014 http://minerals.usgs.gov/minerals/pubs/historical-
statistics/#copper
22 | Page
Utah 2011. State of Utah, “Groundwater Quality Discharge Permit UGW350011 Kennecott Utah Copper, Magna
Utah, January 2011, Accessed November 2014 at
http://www.waterquality.utah.gov/GroundWater/gwpermits/kennecott-tailings/kennecott-tailingsSOB.pdf
Valero et. al. 2011. Valero, Alicia, Valero, Antonio and Domiguez, Adriana “Trends of Energy Costs and Ore Grade
in Global Mining”, Centre Of Research Energy Resources & Consumption ( CIRCE), Spain, 2011, Accessed
November 2014 at http://teide.cps.unizar.es:8080/pub/publicir.nsf/codigospub/0751/$FILE/cp0751.pdf
Vick 1996. Vick, Steven G. Failure of the Omai Storage Dam Geotechnical News, September 1996, Accessed
November 2014 at https://ejatlas.s3.amazonaws.com/docs/omai-gold-mine-tailings-dam-
guyana/Failure_of_the_Omai_Tailings_Dam.pdf
Vogt 2013. Vogt, Craig, International Assessment of Marine & Riverine Disposal of MineTailings Final Report
Adopted By The International Maritime Organization, London Convention.Protocol, October 18, 2013
http://www.craigvogt.com/links/Mine_Tailings_Marine_and_Riverine_Disposal.pdf
Wei et. al. 2012. Wei, Zuan, Yin, Guangszhi, Wang J.G, Ling, Wan, Guangzhi, Li “Design Construction and
Management of Tailings Storage Facilities For Surface Disposal In China: Case Studies of Failures” Waste
Management An Research Vol 31 p 106-112 Sage Publications October 11,2012
http://wmr.sagepub.com/content/31/1/106.full.pdf+html
WISE 2015. World Information Service on Energy Uranium Project, Chronology of Major Tailings Dam Failures,
from 1960, updated 18 May 2015, http://www.wise-uranium.org/mdaf.html
Witt & Schonhardt 2004. Witt, K.J. Schonhardt M. Editors Tailings Management Facilities, Risk &Reliability,
TAILSAFE September 2004 http://www.tailsafe.bam.de/pdf-documents/TAILSAFE_Risk_and_Reliability.pdf
World Bank 2006. World Bank Group, Oil Gas Mining & Chemicals Group, Background Paper The Outlook For
Metals Markets Prepared For G-20 Deputies Meeting Sydney 2006, September 2006
http://siteresources.worldbank.org/INTOGMC/Resources/outlook_for_metals_market.pdf
APPENDIX 1
TSF Failure Data Table
Page1of12
TSF FAILURE DATA TABLE LEGEND
DAM TYPE
Key
DAM FILL
MATERIAL
Key INCIDENT
TYPE Key
INCIDENT
CAUSE
Key
US
Upstream
T
Tailings
1A
Failure Active Impoundment
SI
Slope instability
DS
Downstream
CST
Cycloned sand t
a
1B
Failure Inactive Impoundment
SE
Seepage
CL
Centerline
MW
Mine waste
2A
Accident Active Impoundment
FN
Foundation
WR
Water retention
E
Earthfill
2B
Accident Inactive Impoundment
OT
Overtopping
NR
Not reported
R
Rockfill
3
Groundwater
ST
Structural
EQ
Earthquake
MS
Mine subsidence
ER
Erosion
U
Unknown, or
NR
Not Reported
SOURCES:
(1)
ICOLD. International Committee on Large Dams, Bulletin 121 “Tailings Dams
Risks of Dangerous Occurrences Lessons Learned From Practical
Experiences”
(2)
WISE. World Information Service on Energy Uranium Project (http://www.wise-
uranium.org/mdaf.html) as of December 10, 2014
(3)
Rico, M., Benito, G., Díez-Herrero, A. “Floods From Tailings Dam Failures”
Geological Hazards Unit, Spanish Geological Survey (IGME), Madrid, Spain
http://digital.csic.es/bitstream/10261/12706/3/MayteRico_10.pdf
(4)
Wei, Zuoan, Yin, Guangszhi, Wang J.G, Ling, Wan, Guangzhi, Li “Design
Construction and Management of Tailings Storage Facilities For Surface
Disposal In China: Case Studies of Failures” Waste Management An Research
Vol 31 p 106-112 Sage Publications October 11,2012
http://wmr.sagepub.com/content/31/1/106.full.pdf+html
(5)
Repetto, Robert “Silence is Golden, Leaden and Copper Disclosure of Material
Environmental Information in the Hardrock Mining Industry” Yale School Of
Forestry & Environmental Studies, July 2004 accessed November 2014 at
http://environment.research.yale.edu/documents/downloads/o-
u/repetto_report_execsum.pdf
(6)
Piplinks. “Chronology of Tailings Dam Failures In The Philippines (1982-2007),”
accessed January 2015 at http://www.piplinks.org
(7)
Tailings.info. Tailings Related Accidents - Failures, Breaches and Mudflows,
http://www.tailings.info/knowledge/accidents.htm
(8)
United Nations, Department of Economic & Social Affairs, International Expert
Group Meeting on Indigenous Peoples And Protection of the Environment,
“Case Study of the Impact of Mining & Dams on the Environment and
Indigenous Peoples in Benguet, Cordillera, Philippines,” Aug 27-29, 2007
GENERAL NOTE
We found small variations source to source on total release, run out, deaths and other details, but we
found no ambiguities or inconsistencies that precluded a clear classification as "Serious" or "Very
Serious".
Overall we found much more detailed accounts of "consequence" in local compilations or regional or
national studies. WISE & ICOLD occasionally including details on consequence, or linked to sources
detailing consequence. Our bibliogtaphy includes a more extensive list of materials related to the
consequence of TSF failures
VerySerious30VerySerious=multiplelossoflife(~20)and/orreleaseof≥1,000,000m3
semi‐solidsdischarge,and/orreleasetravelof20kmormore
Serious38Serious=lossoflifeand/orreleaseof≥100,000m3semi‐solidsdischarge
OtherFailures98OtherFailures=ICOLDCategory1failuresotherthanthoseclassifiedas
VerySeriousorSerious
OtherAccidents50OtherAccidents=ICOLDCategory2accidentsotherthanthoseclassified
asVerySeriousorSerious
Non‐DamFailure10Non‐DamFailures=groundwater,wasterock,etc.
=======
Total226
COLORCODE
MINE/PROJECT&LOCATION
DAM
TYPE
DAMFILL
MATERIAL
DAM
HEIGHT
(meters)
STORAGE
VOLUME
(cu.meters)
ICOLD
TYPE
INCIDENT
DATE
RELEASE
VOLUME
(cu.meters)
RUNOUT
(km)
DEATHS
Source
Color
CodeSOURCESNOTES
Page2of12
Karamken,MagadanRegion,
Russia 1A29‐Aug‐09 1,200,0001 WISE,MACE 11houseslost,1death(KaramkenUpdate‐MACE2012‐02‐10)
HuayuanCounty,Xiangxi
AutonomousPrefecture,Hunan
Province,China
1A14‐May‐09 50,0003 WISE3killed,4injured
Kingstonfossilplant,Harriman,
Tennessee,USA 1A22‐Dec‐08 5,400,0004.1 WISE
5.4millioncubicyards(1.09billiongallons)offlyashwasreleased
(http://www.sourcewatch.org/index.php?title=TVA_Kingston_Fos
sil_Plant_coal_ash_spill#TVA_Reaction)
Taoshi,LinfenCity,Xiangfen
county,Shanxiprovince,ChinaUS 50.7290,0001A8‐Sep‐08190,0002.5277 WISEAtleast254deadand35injured.
GlebeMines,UK E 1B‐
OT22‐Jan‐0720,000 HSEReport
InitialReportoftheHSEinvestigationintotheGlebeMinesStony
Middletondamfailure2007,HSECentralDivision‐Nottingham,
UK,23Feb07
Miliang,Zhen'anCounty,
Shangluo,ShaanxiProvince,China 1A30‐Apr‐06 517 WISE17missing
PinchiLake,BC,CanadaWRE122A‐
ER30‐Nov‐04 6,000‐8,000 WISEMercurycontaminatedtailingsintoPinchiLake
Riverview,Florida,USA 1A5‐Sep‐04227,000 WISE
Partizansk,PrimorskiKrai,Russia 1A22‐May‐04 166,000 WISE
Malvési,Aude,France 1A20‐Mar‐04 30,000 WISEUraniumslurrieselevatednitrateinriver
CerroNegro,nearSantiago,Chile,
(5of5)UST 1A‐
ER3‐Oct‐0380,00020 WISE
ElCobre,Chile,2,3,4,5UST 1B‐
OT22‐Sep‐02 8,000 Villavicencio
(2014)
SanMarcelinoZambales,
Philippines,Bayarongdam
(9/11/02)
47,000,000 1B11‐Sep‐02 WISE,
Piplinks
Sep.11:lowlyingvillagesfloodedwithminewaste;250families
evacuated;
SanMarcelinoZambales,
Philippines,Camalcadam
(8/27/02)
1B27‐Aug‐02 WISE,
Piplinks
Aug.27:sometailingsspilledintoMapanuepeLakeandeventually
intotheSt.TomasRiver.
ElCobre,ChileUST 1B‐
OT11‐Aug‐02 4,500 Villavicencio
(2014)
SebastiãodasÁguasClaras,Nova
Limadistrict,MinasGerais,Brazil 1A22‐Jun‐0182 WISE2killed,3missing.Tailings8kmdownstreamtheCórrego
Taquarasstream,mudaffectedanareaof30hectares
NandanTinmine,Dachang,
Guangxi 1A18‐Oct‐00 28 WISE,WeiWISE:15killed,100missing,100housesdestroyed
COLORCODE
MINE/PROJECT&LOCATION
DAM
TYPE
DAMFILL
MATERIAL
DAM
HEIGHT
(meters)
STORAGE
VOLUME
(cu.meters)
ICOLD
TYPE
INCIDENT
DATE
RELEASE
VOLUME
(cu.meters)
RUNOUT
(km)
DEATHS
Source
Color
CodeSOURCESNOTES
Page3of12
Inez,MartinCounty,Kentucky,
USA 1A11‐Oct‐00 950,000120 Table1ICOLD,WISE
Aitikmine,nearGällivare,SwedenDSMW&E1515,000,000 1A‐
ER8‐Sep‐001,800,0005.2 Table1ICOLD,WISE
BaiaMare,RomaniaEsmerelda
Exploration
DS
then
US
TAfewm800,0001A‐ST 30‐Jan‐00100,000>100 221ICOLD,WISE,
Rico
Killedtonnesoffishandpoisoneddrinkingwaterofmorethan2
millionpeopleinHungary
Borsa,Romania 1A200022,000t Table1ICOLD,WISE Company:ReminSA
SurigaoDelNortePlacer,
Philippines(#3of3) 1A26‐Apr‐99 700,000t124Table1ICOLD,
Piplinks
ToledoCity(Philippines) 1B19995,700,000 PiplinksDrainagetunnelblowout
Huelva,Spain 1A31‐Dec‐98 Table1ICOLD,WISE Fertiberiaphosphatemine
LosFrailes,nearSeville,SpainWRR2715,000,000 1A‐
FN25‐Apr‐98 6,800,00041 209ICOLD,WISE,
Rico
ZamboangaDelNorte,Sibutad
GoldProject 1A‐
OT6‐Nov‐97 Piplinks
PintoValley,Arizona,USA 1B22‐Oct‐97 230,000 Table1ICOLD,WISE
Amatista,Nazca,Peru 1A‐
EQ12‐Nov‐96 300,000 WISEduetoM6.4earthquake
ElPorco,Bolivia 1A29‐Aug‐96 400,000300 Table1ICOLD,WISE 300kmofPilcomayorivercontaminated
Marcopper,MarinduqueIsland,
Philippines(3/24)(#2of2) 1A‐ST 24‐Mar‐96 1,600,00026 208ICOLD,WISE,
Piplinks
Drainagetunnelplugfailed.26kmoftheMakulaquitandBoac
riversystemsfilledwithtailingsrenderingthemunusable;US$80
millionindamage
Sgurigrad,BulgariaUST451,520,000 1A‐SI 1996220,0006 220ICOLD,Rico
NegrosOccidental,BulawanMine
SipalayRiver 1A8‐Dec‐95 Piplinks
GoldenCross,WaitekauriValley,
NewZealandR25‐303,000,000 1A‐
FNDec‐959,999 207ICOLD
SurigaodelNortePlacer,
Philippines(#2of3)WRE171B‐SI 2‐Sep‐9550,00012206ICOLD,WISE
OmaiMine,TailingsdamNo1,2,
GuyanaWRR445,250,000 1A‐
ER19‐Aug‐95 4,200,00080 205ICOLD,WISE,
Rico80kmofEssequiboRiverdeclaredenvironmentaldisasterzone
MiddleArm,Launceston,
TasmaniaCLE425,0001A‐
OT25‐Jun‐955,000 204ICOLD
Riltec,Mathinna,TasmaniaCLE7120,0002A‐SE Jun‐9540,000 203ICOLD
HopewellMine,Hillsborough
County,Florida,USA 1A19‐Nov‐94 1,900,000 WISE
COLORCODE
MINE/PROJECT&LOCATION
DAM
TYPE
DAMFILL
MATERIAL
DAM
HEIGHT
(meters)
STORAGE
VOLUME
(cu.meters)
ICOLD
TYPE
INCIDENT
DATE
RELEASE
VOLUME
(cu.meters)
RUNOUT
(km)
DEATHS
Source
Color
CodeSOURCESNOTES
Page4of12
PayneCreekMine,PolkCounty,
Florida,USA 1A2‐Oct‐946,800,000 WISE
Merriespruit,nearVirginia,South
Africa,Harmony2,3
US
paddockT317,040,000 1B‐
OT22‐Feb‐94 600,000417202ICOLD,WISE,
Rico
OlympicDam,RoxbyDowns,
SouthAustralia 314‐Feb‐94 5,000,000 WISEDesignedgroundwaterleakagefromunlinedtailings
impoundmentintogroundwater
MineraSeraGrande:Crixas,Goias,
Brazil
DS
then
US
CST412.25Mt2A‐SI Feb‐94None 214ICOLD
FortMeade,Florida,Cargill
phosphate(#3of3) 1A2‐Jan‐9476,000 WISE
Longjiaoshan,DayeIronOremine,
Hubei 1A199431 Wei
Marcopper,MarinduqueIsland,
MogpogPhilippines(12/6)(#1of
2)
1B6‐Dec‐932 PiplinksSiltationdamfailure.MogpogRiverandMogpogtownflooded.
TD7,Chingola,ZambiaUST&E51A‐
OTAug‐93100t 200ICOLD
Itogon‐Suyoc,Baguiogolddistrict,
Luzon,Philippines 1A‐
OT26‐Jun‐93 199ICOLD,
Piplinks
Marsa,Peru 1A‐
OTJan‐936 WISE
Kojkovac,MontenegroWRE 3,500,000 2B‐
ERNov‐92none 198ICOLD
Saaiplaas,SouthAfrica,2 CST 1A‐IS 19‐Mar‐92 Table1ICOLD3separateeventswithin4days
MaritsaIstok1,Bulgaria Ash1552,000,000 1A‐
ER1‐Mar‐92500,000 218ICOLD,WISE
Tubu,Benguet,No.2Tailings
Pond,Padcal,Luzon,Philippines 80,000,000 1A‐
FN2‐Jan‐9280,000,000 197Piplinks
IronDyke,SullivanMine,
Kimberley,BC,CanadaUS 211A‐SI 23‐Aug‐91 75,000 196ICOLD
SodaLake,California,USAUSE32A‐
EQ17‐Oct‐89 111ICOLD
SilverKing,Idaho,USADSE937,0002A‐
OT5‐Aug‐89Small 108ICOLD
BigFour,Florida,USACLE 2A‐
FN1989 14ICOLD
CyprusThompsonCreek,Idaho,
USACLCST14627,000,000 2A‐SE 1989 34ICOLD
COLORCODE
MINE/PROJECT&LOCATION
DAM
TYPE
DAMFILL
MATERIAL
DAM
HEIGHT
(meters)
STORAGE
VOLUME
(cu.meters)
ICOLD
TYPE
INCIDENT
DATE
RELEASE
VOLUME
(cu.meters)
RUNOUT
(km)
DEATHS
Source
Color
CodeSOURCESNOTES
Page5of12
SouthernClay,Tennessee,USAWRE51A‐SE 1989300 112ICOLD
Stancil,Maryland,USAUSE974,0001A‐SI 198938,0000.1 116ICOLD,Rico
Unidentified,Hernando,County,
Florida,USA#2USE123,300,000 1A‐
OTSep‐884,600 163ICOLD
Jinduicheng,ShaanxiProvince.,
ChinaUS 401A‐
OT30‐Apr‐88 700,000~20195ICOLD,WISE
ConsolidatedCoalNo.1,
Tennessee,USA,DSMW851,000,000 2A‐ST 19‐Jan‐88250,000 121ICOLD,WISE
RainStarterDam,Elko,Nevada,
USAWRER271,500,000 3‐ 1988 98ICOLD
Unidentified,Hernando,County,
Florida,USADSE122A‐
FN1988 164ICOLD
SurigaoDelNortePlacer,
Philippines(#1of3) 1A9‐Jul‐87 Piplinks
MontcoalNo.7,RaleighCounty,
WestVirginia,USA 1A8‐Apr‐8787,00080 WISEtailingsflow80kmdownstream
Bekovsky,WesternSiberiaUSArgillite,
aleurolite5352,000,000 1A25‐Mar‐87 None 212ICOLD
Xishimen,ChinaUST311A‐SI 21‐Mar‐87 2,230 194ICOLD
MontanaTunnels,MT,USADSMW33250,0003‐ 1987 87ICOLD
MariannaMine#58,PA,USE37300,0002A‐SI 19‐Nov‐86 77ICOLD
Mankayan,Luzon,Philippines,
No.3TailingsPondE 1A‐ST 17‐Oct‐86 193ICOLD,
PiplinksSiltationoftheAbraRiverwhichaffected9municipalities
Lepanto,Mankayan,Benguet,
Philippines 1A17‐Oct‐86 PiplinksSiltationoftheAbraRiverwhichaffected9municipalities
PicodeSaoLuis,Gerais,Brazil T201A‐
ER2‐Oct‐86 192ICOLD
Rossarden,TasmaniaWRE7.5200,0001B‐
OT16‐May‐86 190ICOLD
Story’sCreek,TasmaniaValley
side1730,0001B‐
OT16‐May‐86 Minimal 191ICOLD
Itabirito,MinasGerais,BrazilGravi
tyMasonry301A‐ST May‐86100,000127189ICOLD,WISE,
Rico
MineralKing,BC,CanadaCLCST6Small1B‐
OT20‐Mar‐86 188ICOLD
Huangmeishan,China 1A198619 WISE
COLORCODE
MINE/PROJECT&LOCATION
DAM
TYPE
DAMFILL
MATERIAL
DAM
HEIGHT
(meters)
STORAGE
VOLUME
(cu.meters)
ICOLD
TYPE
INCIDENT
DATE
RELEASE
VOLUME
(cu.meters)
RUNOUT
(km)
DEATHS
Source
Color
CodeSOURCESNOTES
Page6of12
SpringCreekPlant,Borger,Texas,
USA 530,0001A‐
OT1986 114ICOLD
Bonsal,NorthCarolina,USAWRE638,0001A‐
OT17‐Aug‐86 11,000 17ICOLD
Stava,NorthItaly,2,3USCST29.5300,0001A‐SI 19‐Jul‐85200,0008269117ICOLD,WISE,
Rico
LaBelle,Pennsylvania,USADSMW791,230,000 2A‐
FN17‐Jul‐85 68ICOLD
CerroNegroNo.(4of5)USCST402,000,000 1A‐
EQ3‐Mar‐85500,0008 30ICOLDWISE,
Rico
VetadeAguaUST24700,0001A‐
EQ3‐Mar‐85280,0005 178ICOLD,WISE,
Rico
ElCobreNo.4DSCST502A‐
EQ3‐Mar‐85 44ICOLD
Niujiaolong,ShizhuyuanNon‐
ferrousMetalsCo.,Hunan 1AJan‐85731,0004.249 Wei
Marga,Chile 1B‐
OT1985 76ICOLD
Ollinghouse,Nevada,USAWRE5120,0001A‐SE 198525,0001.5 91ICOLD,Rico
Texasgulf4BPond,Beaufort,Co.,
NorthCarolina,USAWRT812,300,000 2A‐SI Apr‐84 122ICOLD
Mirolubovka,SouthernUkraineUSE&T3280,000,000 2A‐SI 15‐Jan‐84‐ 210ICOLD
BattleMt.Gold,Nevada,DSE81,540,000 2A‐SI 1984 11ICOLD
VirginiaVermiculite,Louisa
County,Virginia,USAWRE91A‐ST 1984 179ICOLD
ClaytonMine,Idaho,USACLT24215,0002A‐ST 2‐Jun‐83 32ICOLD
GoldenSunlight,MT,USACLCST 3‐ 5‐Jan‐83 51ICOLD
GreyEagle,California,USADSE 3‐ 1983 53ICOLD
Vallenar1and2 1B‐
OT1983 175ICOLD
Sipalay,Philippines,No.3Tailings
PondWRMW 37,000,000 1A‐
FN8‐Nov‐8228,000,000 187ICOLD,WISE,
Piplinks
Damfailure,duetoslippageoffoundationsonclayeysoils.
Widespreadinundationofagriculturallandupto1.5mhigh
Royster,Florida,USAUST211A‐
FN1982 102ICOLD
Ages,HarlanCounty,Kentucky,
USA 1A18‐Dec‐81 96,0001631 WISE
DixieMine,Colorado,USA 1B‐UApr‐81 39ICOLD
COLORCODE
MINE/PROJECT&LOCATION
DAM
TYPE
DAMFILL
MATERIAL
DAM
HEIGHT
(meters)
STORAGE
VOLUME
(cu.meters)
ICOLD
TYPE
INCIDENT
DATE
RELEASE
VOLUME
(cu.meters)
RUNOUT
(km)
DEATHS
Source
Color
CodeSOURCESNOTES
Page7of12
BalkaChuficheva,RussiaUSCST2527,000,000 1A‐SI 20‐Jan‐813,500,0001.3 211ICOLD,WISE
TexasgulfNo.1Pond,Beaufort
Co.,NorthCarolina,USAWRE 24,700,000 2A‐SI 1981 123ICOLD
VetadeAquaA 1A‐
EQ1981 176ICOLD
VetadeAguaB 1A‐
EQ1981 177ICOLD
Tyrone,NewMexico,Phelps‐
DodgeUSCST662,500,000 1A‐SI 13‐Oct‐80 2,000,0008 94ICOLD,WISE,
Rico
SweeneyTailingsDam,Longmont,
Colorado,USA 71A‐SE May‐80 119ICOLD
KyaniteMining,Virginia,USA 11430,0002A‐
OT1980 67ICOLD
Churchrock,NewMexico,United
NuclearWRE11370,0001A‐
FN16‐Jul‐79370,000110 173
ICOLD,
Wikipedia,
Rico
UnionCarbide,Uravan,Colorado,
USAUST432A‐SI Mar‐79 172ICOLD
IncidentNo.1,Elliot,Ontario,
CanadaWRE93‐ 1979 35ICOLD
SuncorE‐WDike,Alberta,CanadaWRMW302A‐SI 1979 118ICOLD
Arcturus,ZimbabweUST251.7‐2.0Mt 1A‐
OT31‐Jan‐7839,0000.31185ICOLD,WISE,
Rico
MochikoshiNo.1,Japan(1of2)UST28480,0001A‐
EQ14‐Jan‐7880,0008184ICOLD,WISE,
RicoDamfailureduetoearthquake
Norosawa,JapanDS 24225,0002B‐
EQ14‐Jan‐78 90ICOLD
Hirayama,JapanDS 987,0002B‐
EQ1978 56ICOLD
MochikoshiNo.2,Japan(2of2)UST191A‐
EQ19783,0000.15 85ICOLD,Rico damfailureduetoaftershock
Syncrude,Alberta,CanadaCLT 2A‐
FN1978 120ICOLD
Madison,Missouri,USAWRE111A‐
OT28‐Feb‐77 74ICOLD
Homestake,N.Mexico,USAUST211A‐ST Feb‐7730,000 59ICOLD
PitNo.2,WesternUST91A‐SI 1977 96ICOLD
COLORCODE
MINE/PROJECT&LOCATION
DAM
TYPE
DAMFILL
MATERIAL
DAM
HEIGHT
(meters)
STORAGE
VOLUME
(cu.meters)
ICOLD
TYPE
INCIDENT
DATE
RELEASE
VOLUME
(cu.meters)
RUNOUT
(km)
DEATHS
Source
Color
CodeSOURCESNOTES
Page8of12
Unidentified,Hernando,County,
Florida,USACLE62A‐
FN1977 162ICOLD
WesternNuclear,JeffreyCity,
Wyoming,USA#2 1A‐SI 197740 180ICOLD
Kerr‐McGee,Churchrock,New
Mexico,USAWRE91A‐
FNApr‐76 64ICOLD
ZlevotoNo.4,YugoslaviaUST251,000,000 1A‐SI Mar‐76300,000 184ICOLD,WISE
Dashihe.ChinaUS 372A‐
EQ1976 36ICOLD
Unidentified,Idaho,USADSE342A‐SI 1976 149ICOLD
CadetNo.2,Montana,CLE212A‐SI Sep‐75 18ICOLD
Madjarevo,BulgariaUST403,000,000 1A‐ST 1‐Apr‐75250,000 219ICOLD
CarrFork,Utah,USA 101A‐ST Feb‐75 22ICOLD
DresserNo.4,Montana,CLE151A‐
FN1975 40ICOLD
KeystoneMine,CrestedButte,
Colorado,USA 1B‐U1975 65ICOLD
MikeHorse,Montana,USAUST18750,0001B‐
OT1975150,000 79ICOLD
PCSRocanville,Saskatchewan,
CanadaUST123‐ 1975 92ICOLD
Unidentified,GreenRiver,
Wyoming,USAWRE183‐ 1975 161ICOLD
HeathSteelemaindam,
Brunswick,CanadaWRR,E302A‐
FN1975 186ICOLD
Bafokeng,SouthAfricaUST2013,000,000 1A‐SE 11‐Nov‐74 3,000,00045127ICOLD,WISE,
Rico
GoldenGilpinMine,Colorado,
USA 121B‐UNov‐