“While we were confident in our original analysis of the PolyMet tailings dam, we have carefully examined the requests for reconsideration and related information about recent dam failures in other parts of the world,” Commissioner Sarah Strommen said in a news release. “We understand people’s concerns with these dam failures and whether those events indicate a fundamental design issue with PolyMet’s dam. Our analysis demonstrates that there are significant differences in site conditions, engineering design, and operating requirements and we remain confident in the safety of the PolyMet tailings dam as permitted."
The DNR said it understood concerns raised about dam failures in other countries, but that it subjected PolyMet's tailings dam proposal to "years of rigorous modeling and independent review" by its own engineers and outside dam safety experts.
The DNR determined that Fond du Lac and others did not raise any new issues that materially affect the DNR’s decision to issue the tailings dam permits. The DNR said it remained confident that the permitted PolyMet tailings basin dam, if properly constructed, operated, and maintained, will be safe and protective of human health and the environment.
The requests for reconsideration raised several concerns about the approved PolyMet tailings dam permits, the DNR said:
- the failure of the Brumadinho Dam in Brazil is evidence of new stability concerns regarding the upstream construction design of tailings basins;
- the use of the “Olson Method” to analyze the strength, liquefaction, and stability of the Brumadinho dam inherently means that its use to analyze the PolyMet tailings dam was flawed; and
- recent inspections of the existing LTV tailings basin at the PolyMet site call into question assumptions about tailings drainage and materials strength in the basin.
The DNR’s dam safety experts evaluated each of the claims in the requests for reconsideration, it said. The DNR found that while both the PolyMet and the Brumadinho dams include the use of “upstream construction” methods, there are "critical differences that must be understood and evaluated to draw technically valid comparisons and conclusions."
"There are multiple factors that go into the construction of a safe dam, and PolyMet’s permitted dam design is significantly different and safer than the Brumadinho dam" in the following ways, the DNR outlined:
- Safety factors for the PolyMet dam were established using conservative assumptions to assess the basin’s stability under extreme conditions. These assumptions include that the entire basin had liquefied, been subjected to extreme rainfall, and been subjected to an earthquake. The analysis of the Brumadinho dam failed to include anything close to this level of assessment.
- PolyMet’s dam will be built on flat topography, far from any community, and using ring-dike construction. The Brumadinho dam was constructed on a hillside, directly above a community, with higher-risk valley construction.
- PolyMet’s dam will have very gradual side slopes (7:1 overall) that are inherently more stable than the Brumadinho dam which had an overall slope of 4:1.
- PolyMet’s dam has virtually no inflow of surface waters into the basin. The Brumadinho dam had significant inflow from the adjacent watershed that necessitated the diversion of runoff from surrounding hillsides away from the tailings basin and dam. It appears this diversion system failed, resulting in heavy flows into the basin for weeks immediately prior to the dam’s failure.
- The PolyMet tailings dam is located in an area of little or no seismic activity. The Brumadinho dam is located in an area of moderate seismic activity.
- PolyMet’s dam is approximately eight miles from its mine site, which minimizes any risk of blasting impacts to the dam. Reports indicate that there had been mine blasting in close proximity to the Brumadinho dam on the morning of the failure.
The Olson Method, named after dam engineer Scott Olson, is an established and peer-reviewed method that was incorrectly applied at Brumadinho, resulting in inaccurate safety factors for the site and failure to recognize that the tailings at the site were highly liquefiable. According to the DNR, the analysis PolyMet supplied to the DNR correctly applied the Olson Method, with Olson’s oversight; assumed complete liquefaction; and resulted in higher factors of safety. The Olson Method analysis was subject to extensive review by DNR’s dam safety experts and independent experts under contract to the DNR, it said.
The modeling used to analyze the undrained strength of the tailings in the LTV basin already accounted for the conditions observed during recent inspections. Whether or not these portions of the basin are currently drained has no bearing on the undrained strength analysis of the dam. All of the material properties in the basin were obtained for the existing conditions and used in the undrained strength analysis.