Past present future tailings dams behaviour
Jun 1st, 2016
Three papers we have published in the past three years deal with Past present future tailings dams behaviour. Here we propose a summary and our conclusions.
Past present future tailings dams behaviour summary
Past tailings dams behaviour
At TMW 2013 (Oboni, Oboni, 2013) we attempted an estimate of the rate of failure of major tailings dams failures and to compare their risks to human life to well known social tolerance.
We started by clearly stating the limitations of the available data and the lack of clear definition of what constituted a major failure in commonly available statistics. We then found values varying between 10-3 (decade around ’79) and 2*10-4 (decade around ’99).
In the same paper we showed quantitatively how, over time, multiple hazards hit would significantly increase the probability of failure of a dam and lead to intolerable risks. The paper concluded “… Especially in the case of TDs located in areas where demographic pressure leads to settlements in the downstream areas, social and legal consequences of a failure will dramatically increase. This will particularly be the case if the methodologies used to perform the risk assessments prove to be in disconnect with the needs of our modern society.”
Mount Polley Tailings Dam Breach
A new study of tailings dams historic failures
The at TMW2015 we presented the results of a new study of tailings dams historic failures (Bowker, Chambers, 2015) which used detailed data and actuarial techniques to define historic rate of failures of tailings dams after attempting to define what constitutes “serious” and “very serious” TD failures.
We stated that “… The common practice approach of using oversimplified consequence functions (with “and/or” clauses as just defined above) is often used in research papers because of scope/budget limitations, but should not be accepted for a rational world-wide approach to decision making and tailings risks management for an industry that has significant societal impacts like mining. Tailings accidents generate multiple direct and indirect consequences on the environmental, human, H&S, operational and reputational areas and we believe it is time for the mining industry as a whole to adopt a uniform consequence function”.
Our paper concluded that
“… It is comforting that the results of the “quick and dirty” 2013 (Oboni & Oboni) study reached globally comparable results to the 2015 (Bowker & Chambers) very deep and solid analytical approach. We note that the selection of the time frame has a large influence on the conclusions of the 2015 study. Therefore we recommend to perform these comparative studies with constant duration. For example decade by decade, to avoid the hazard of drawing misleading conclusions. “Averaging” over 70 years, during which so many conditions have changed, may indeed mask decennial spikes. To prove this it is enough to look at the Table above which shows that the accident rates have actually decreased by 15%-24% from the 1990-1999 to the 2000- 2009 decades using the 2015 study’s own data.” The paper concluded “. We reiterate that the goal of zero tailings failures is impossible to achieve.
There will continue to be tailings failures. In fact, in the long term all tailings facilities will spiral toward significant increases of their pf and when they fail the tailings will go to downstream rivers, lakes, and the ocean as they did at every failure to date.
Present tailings dams behaviour
At TMW2015 (Caldwell, Oboni, Oboni, 2015) we reviewed a number of very recent (2014) tailings dams failures. We noted their common traits:
- Existing facilities that had been used for many years, reportedly following the “as it is impossible to anticipate everything, why bother” principle described in the first half of past century (Merton, 1936).
- Antiquated design and management practices, meaning poor initial conditions and overall structure category, as shown above for the cases where some data were available leading to substandard pf.
No risk assessment
- Absence of risk assessment or even of common health and safety programs. Let’s note that typical, common practice, risk matrices can only correctly and unambiguously compare a small fraction, reportedly less than 10%, of randomly selected pairs of hazards. Furthermore, they can assign identical ratings to quantitatively very different risks, a phenomena often referred to as “range compression” and can mistakenly assign higher qualitative ratings to quantitatively smaller risks and vice versa. These inaccuracies can lead to mistaken resource allocation. (Oboni, Oboni, 2012).
No peer review
- No peer review. Independent peer review of water dams is a long-standing practice. It is disgraceful that all regulators do not insist on it. Until they do, and peer reviews become very serious endeavors, we see no hope of reducing the incidence of tailings facility failure (Morgenstern, 2010, Caldwell, 2011)
- Limited engineer involvement that appears to have been aware of potential problems but not heard or acted on. Alarming disconnect comes from the poor definition of potential consequences of mishaps and their societal ripple effects. This aspect is indeed mostly ignored in codes, leaving professionals ample room to biases and censoring applied to potential losses (Oboni at Al., 2013, CDA, 2014).
- Overconfident mining companies that did not act when prudence may have so dictated. Assessors often censor and bias risk assessments towards “credible events”. However history, even recent, has shown that major failures occur when “incredible events” occur. Another cause can be a long chain of apparently benign events. The public has now got that clearly in mind, generating widespread controversy and projects’ opposition (Oboni at Al., 2013).
- Absence of significant regulatory oversight or involvement. Risks assessments are “at risk” if plagued by conflict of interest or overly optimistic cognitive biases, or censure. (Oboni, Oboni, 2014).
We concluded: “While there are similarities in root causes, there is an amazing difference in the response of the societies affected by the failures. From swift and dramatic, with new regulations and large fines for the responsible parties, to the usual committee meetings, review reports soon ignored, calls for action yet to be undertaken, finally a sort of void: no reports, no action, and the mere hope of something in the unspecified future.”
Future tailings dams behaviour
The theme of the long term survivability of TD was further detailed at TMW 2014 (Oboni, Oboni, Caldwell 2014) where attention was focused on modelling the aging process of a geo-structure as a series of discrete hits by hazardous conditions (these could be anything, from an earthquake to flooding, to icing, etc.). In that paper we drafted an attempt at multidimensional estimate of future consequences. The paper stated that “. Should the value of consequences increase, then even an “excellent dam” would soon pose a societally unacceptable risk. Any dam starting with a small initial FoS or reduced standards of care will see its risk evolve towards intolerable societal risks faster. Even if its consequences of failure remain constant. Thus it turns out that it is not necessary, for this discussion, to delve into conjectures related to future consequences.”
Over the years, we have demonstrated that consequences are not necessarily correlated with dam height or pond volume. As in many industries the “scary stuff” is not necessarily the riskier one. Our practice and research demonstrate that pf is often way higher in smaller structures than in major ones. That is simply because of higher standards for larger structures than for “insignificant ones”.
And examples like Stava or Bafokeng are there to show that “extreme” consequences can actually occur. We also demonstrated that the mining industry fatality rate lies way above the generally accepted “safe” thresholds for hazardous industries. The number of existing, operational, and closed tailings storage facilities around the world requires to prioritize the mitigation tasks. That is if we want to achieve a higher quality, be it at corporate or at national levels.”
Measuring and giving a sense to a complex problem
The methodology developed to date enables us to “measure” and give a sense to a complex problem. It allows to:
- transparently compare alternatives,
- discuss rationally and openly the survival conditions, or to
- evaluate the premature failure of a structure.
To slow down the increase of the probability of failure one can repair the damage occurring as a result of each hazard hit. Another way is to entirely avoid the damage. The second is generally “not feasible” for economic and constructional reasons. One can never reduce risks, especially long term ones, to nil.”
At TMW2016 we will give a course on this subject, introducing the latest developments of our research. We have sought to develop a simplified approach to frame the probability of failure of a dams. It applies to single dams or portfolios of dams. It is based on easy to retrieve archival data.
Come and follow our course. Also make sure you will get a copy of our paper: “A systemic look at tailings dams failure process”.
Tagged with: future tailings dams, Past tailings dams
Category: Consequences, Probabilities, Risk analysis, Risk management
interesting discussion. However recent failures (Mt.Polley and Bento Rodrigues) lead us to re evaluate the correct procedures and design criteria of tailings dams design e re consider different techiniques of tailings management.
Indeed one of most interesting is to consider the mechanical forced filtration of fines in order to eliminate the most important factor that is at origin of the failures: the water