Donlin Mine: An independent dam breach analysis

The Donlin Mine, also known as the Donlin Gold Project, is a proposed gold mine located in the Yukon-Kuskowkim Delta of southwestern Alaska, approximately 145 miles northeast of Bethel and 10 miles north of the village of Crooked Creek. 

Donlin Gold, LLC has proposed to develop what would become one of the world's largest open-pit gold mines, with an initial projected lifespan of 27 years, although there are additional mine claims of various adjacent Federal, Satte, and private lands. The Mine infrastructure will span across multiple tributaries to the Kuskokwim River, approximately over a 25-square-mile footprint. 

The Donlin Mine tailings storage facility (TSF) would sit at the headwaters of Anaconda Creek approximately 10 miles upstream of the village of Crooked Creek near the confluence of the Kuskokwim River. The TSF is expected to store waste material during the mine’s life and in perpetuity. Although total failure of the Donlin Mine TSF is very unlikely, any release from the TSF would pose a significant threat to downstream communities with the potential for devastating and long-term ecological impacts along the Crooked Creek and Kuskokwim River watersheds as a whole.

In order to ensure adequate public planning and safety measures if the dam were to fail in the future, Mother Kuskokwim Tribal Coalition has commissioned Lynker Intel to conduct an independent dam breach analysis of the proposed Donlin Mine TSF.

Exploration of the Donlin Creek gold deposit has been underway since the mid-1990s. In 2012, Barrick and NovaGold announced their intention to begin the permitting process, and in August of that year,  initial permitting applications were filed. The final Environmental Impact Statement (EIS) was issued in April 2018, concluding the environmental review process, though several outstanding state permits remain, including the State of Alaska’s Certificate of Approval for Dam Safety. As part of the final EIS, Donlin Gold was required to evaluate a set of spill scenarios for the various hazardous materials necessary for the mining operations, including a release of tailings and water into Crooked Creek and the Kuskokwim River. The scenario presented in the final EIS evaluates a small partial tailings dam failure in which just 0.5 percent of the stored tailings are released (US EPA, 2018). At no point in the EIS do regulatory agencies evaluate a complete and catastrophic failure scenario consistent with recent tailings dam failures, such as Mount Polley in British Columbia, Canada, or Brumadinho in Minas Gerais, Brazil.

Here, we present an independent dam breach analysis of the proposed Donlin Mine TSF evaluating a range of failure scenarios using a numerical model to simulate the outflow and inundation of the downstream areas.

The Donlin Mine has an estimated resource base of 34 million ounces of gold. To support mining operations and store waste byproducts, the TSF design reports the construction of a 471 feet high and 5,800 feet long earthen tailings dam with a designed maximum storage capacity of 15,538 million cubic feet of tailings and water.

Using a published, peer-reviewed dataset of global tailings dam failures (Piciullo et al, 2022), we can study the historical relationship between pre-breach TSF capacity and the observed post-breach release volumes (such as Mount Polley, green circle, below) to understand how a hypothetical breach of the Donlin Mine TSF might compare to past TSF failures. While this empirical analysis does not incorporate mine-specific characteristics (e.g., tailings composition and local topography), it does provide a first-order approximation of the possible scale of downstream impacts based on the storage capacity of the TSF. Our results highlight several important conclusions:

1. Larger TSFs tend to fail more catastrophically, releasing a larger amount of tailings if and when they do fail
2. The Donlin Mine TSF is larger than any facility within this tailings dam failure dataset
3. Because the Donlin Mine TSF is so large, estimating possible release volumes requires extrapolation of this mathematical model. If we do so, we find that the empirical estimate of a failure of the Donlin Mine TSF (black triangle) is more than 10 times larger than the partial tailings dam failure evaluated within the final EIS (black square).

The above conclusions suggest that the partial breach scenario evaluated within the final EIS is, at best, an evaluation of a lower-end breach scenario, and at worse, inconsistent with published, peer-reviewed data from historical TSF failures, which estimate a release volume more than 11.5 times greater than estimates conducted by BGC Engineering on behalf of Donlin Gold. These findings imply that a failure of the Donlin Mine TSF, though very unlikely, may pose a far greater threat to downstream communities than is captured within the final EIS. This independent study is motivated by this concerning finding.

Beyond looking at past tailings dam failures, one technique we can use to better understand the risks of a specific breach event is to run a flow-routing simulation. A key challenge, however, is that the circumstances of a TSF failure are inherently uncertain. For example, a TSF failure might happen under sunny-day conditions due to structural failures of the dam itself, as was the case during the Mount Polley breach, or due to undetected piping within the dam. A TSF failure could also be caused by the overtopping and erosion of the dam under a pump failure scenario during intense flooding conditions. In more seismically active regions of the world, it is even possible for an earthquake to lead to complete liquefaction of the tailings and the earth dam itself, causing rapid and catastrophic impacts downstream. In this study, we represent these uncertainties by defining a large set (i.e., several dozen) of credible TSF breach scenarios which we then run through a numerical model designed to simulate the types of hyperconcentrated mudflows characteristic of a TSF failure. Our sensitivity analysis, which covers breach volumes of 5%, 50%, and 90%, can be reviewed in the full report.

For this study, we use the FLO-2D model to simulate the potential impacts of a catastrophic failure of the Donlin Mine TSF using an ensemble of breach scenarios. This numerical model generates a debris flow based on the estimated composition and volume of the stored tailings, and routes this flow downstream based on the physical characteristics of the river valley, including topography and surface roughness. The results presented here are from a simulated failure releasing 5%, 50%, and 90% of the final maximum TSF capacity.

The following animations show the possible downstream impacts based on a 90% breach numerical model simulation.

The first animation shows the timing and extent of the tailings flow as it moves downstream the Crooked Creek and into the Kuskokwim River

The second animation offers a static look at the maximum simulated flow depths, where brighter colors indicate higher maximum flow depths.

A potential failure of the Donlin Mine TSF has the potential to cause significant and widespread impacts to the Crooked Creek and Kuskokwim River watersheds. The final EIS notes that: “…the overall anticipated direct and indirect water quality impacts on fish and aquatic habitat in Crooked Creek may be measurable or noticeable.” Though the nature of these impacts is not explicitly simulated by the numerical routing model used in this study, this section discusses possible impacts based on existing habitat ranges in the Kuskokwim River watershed and known impacts from gold mining operations.

Crooked Creek is a productive salmon stream (Forest Service, n.d.), and as specified by the Anadromous Fish Act, Bell Creek and Anaconda Creek are also important for the “spawning, rearing, or migration of anadromous fishes.” According to a 10-year aerial survey from 2004 to 2014 (Donlin Gold, 2020), salmon populations in these areas mostly consist of Chinook (King), Coho, and Chum salmon, the former of which remains a significant subsistence fishery in the Kuskokwim River watershed. As for resident fish, the slimy sculpin, Dolly Varden, burbot, and Arctic grayling are stable populations in these waters (Fish Habitat Permit, 2022). In addition, these tributaries to the Kuskokwim River are rich in periphyton and phytoplankton communities (Donlin Gold, 2020). The surrounding wetlands are productive systems, driving critical ecological and hydrological processes that maintain Alaska’s biodiversity (ADFG, 2015).

Infrastructure to support the Donlin Mine poses risks from the increased usage of the canal barges and the construction of a 313 milenatural gas pipeline (USACE, 2013). It is expected this barge traffic will increase 200% along the Kuskokwim River, and this increase “would potentially impact salmon, broad and humpback whitefish, sheefish, and rainbow smelt, all subsistence species important to villages on the Kuskokwim River” (Donlin Gold, 2020).

In addition to physical disturbances from the proposed mining operations, there are also possible risks from gold mining byproducts such as mercury and cyanide and other toxic heavy metals like arsenic, cadmium, and lead. The effects of these chemicals in the environment have the potential to affect survival, growth, and functions in plants, animals, and humans. Of particular concern to the environment are mercury and cyanide, both of which present significant risks if not properly contained, such as in the event of an accidental spill or tailings dam breach.

Mercury is a persistent element in the environment which can transform into methylmercury and bioaccumulate through the food chain. Mercury is naturally occurring in the local sediments and has been historically mined near Red Devil, Alaska (DEC, 2023), but the mining operations may add to the contamination load. The bioaccumulation of methylmercury has been a public health concern, as elevated levels of mercury are present in the middle Kuskokwim River (Matz, 2012; USACE, 2013).

Cyanide is a chemical commonly used to leach gold from gold-bearing minerals in ore. Even at low concentrations, cyanide is very toxic chemical to all living organisms (FEIS appendix N; BLM, 2018). Although the mine plan includes converting cyanide to cyanate, a compound that is less toxic, before discharging into the TSF, the presence and transport of cyanide as part of the mining operations remains a risk.

There are a number of historical cyanide spills from gold mining operations that demonstrate the risk of an accidental spill to downstream waterways and communities, such as the Baia Mare Gold Mine dam breach in Romania. In January of 2000, an estimated 3.5 Mft ³  of cyanide contaminated waste and water was released, leading to one of the most significant mining-related environmental disasters in Europe. Practically all aquatic life was killed in the nearby rivers (Kovac, 2000) and drinking water systems had cyanide levels about 80 times above the permissible limit (UNEP & OCHA, 2000).

The release of hazardous materials from the Donlin Mine operations is not limited to a TSF failure. There are numerous spill reports from variable causes like from human error, equipment failure, transportation-related accidents, to name a few (Lubetkin, 2022). While these types of spills are likely much smaller than what could come from a TSF failure, the released materials are still contaminated and pose a public health threat to nearby communities. For example, the Fort Knox Gold Mine near Fairbanks, Alaska spilled 305,300 gallons of cyanide water solution in 2010 from a failure in the ore processing facility (DEC, n.d.b), and another 45,000 gallons of cyanide water solution in 2012 caused by a bulldozer disrupting a buried pipeline (DEC, n.d.a).

Tailings storage facility (TSF) failures are relatively rare events, but observational records confirm that they can and do happen. A near instantaneous, catastrophic failure, while very unlikely, does have precedence and should be sufficiently considered. Our analyses suggest that a catastrophic failure of the proposed Donlin Mine tailings storage facility poses significant risk to the community of Crooked Creek Village, with the debris wave arriving as quickly as 60 minutes after the breach initiation. In the absence of a rapid and total evacuation, such an event could lead to significant loss of life. Downstream of Crooked Creek, the continued flow and deposition of contaminated tailings and wastewater would very likely enter the Kuskokwim River, at which point the fate and transport of those materials would become difficult if not impossible to mitigate and/or contain, with impacts as far downstream as Bethel.

Based on the available historic data, a breach release volume between 5% and 50% of the TSF storage volume are most consistent with past failure events such as Mount Polley and thus should be considered plausible, though very unlikely, events. The higher 90% release scenarios in this study should be considered within the context of emergency planning as a theoretical upper bound, with an assumption of considerable uncertainty.

Alaska Department of Environmental Conservation (DEC) (2023). Red Devil Mine.  https://dec.alaska.gov/spar/csp/sites/red-devil/ 

Alaska Department of Fish and Game (ADFG) (2015). State Wildlife Action Plan.    https://www.adfg.alaska.gov/static/species/wildlife_action_plan/2015_alaska_wildlife_action_plan.pdf 

Alaska Department of Environmental Conservation (DEC) (n.d.a). Fairbanks Gold Mining, Inc. cyanide water spill. https://dec.alaska.gov/spar/ppr/spill-information/response/2012/14-gold/

Alaska Department of Environmental Conservation (DEC) (n.d.b). Fort Knox Gold Mine cyanide water spill. https://dec.alaska.gov/spar/ppr/spill-information/response/2010/08-fortknox/

Boin, A., & McConnell, A. (2007). Preparing for Critical Infrastructure Breakdowns: The Limits of Crisis Management and the Need for Resilience. Journal of Contingencies and Crisis Management, 15(1), 50–59. https://doi.org/10.1111/j.1468-5973.2007.00504.x

Bureau of Land Management (BLM) (2018). Donlin Gold Project Final Environmental Impact Statement, Appendix N.  https://www.doi.gov/sites/doi.gov/files/encl-ykd-appendix-n-section-810-analysis.pdf 

Donlin Gold, LLC. and Owl Ridge Natural Resource Consultants, Inc. (2020). Aquatic Resources Monitoring Plan. Plan of Operations – Volume VII C.  https://dnr.alaska.gov/mlw/mining/large-mines/donlin/pdf/poo-armp2020.pdf 

Fish Habitat Permit, FH22-III-0053 (2022).  https://dnr.alaska.gov/mlw/mining/large-mines/donlin/pdf/fh22-iii-0053.pdf 

Forest Service (n.d.). Region 10 - recreation: Habitat of Crooked Creek Information Site. Forest Service National Website. https://www.fs.usda.gov/detail/r10/recreation/?cid=fsbdev2_038702

Kovac C. (2000). Cyanide spill threatens health in Hungary. BMJ, 320(7234), 536A. https://www.ncbi.nlm.nih.gov/pmc/articles/PMC1117597/

Lubetkin, S. (2022). Alaska Mining Spills Retrospective Analysis. Brooks Range Council, Earthworks, National Parks Conservation Association, Norton Bay Intertribal Watershed Council, and Tanana Chiefs Conference. https://earthworks.org/wp-content/uploads/2022/06/Alaska-Mining-Spills-Retrospective-Analysis-4-2022-2.pdf

Lumbroso, D., Davison, M., Body, R., & Petkovšek, G. (2021). Modelling the Brumadinho tailings dam failure, the subsequent loss of life and how it could have been reduced. Natural Hazards and Earth System Sciences, 21(1), 21–37. https://doi.org/10.5194/nhess-21-21-2021

Matz, A. C. (2012). Mercury, arsenic, and antimony in aquatic biota from the middle Kuskokwim River region, Alaska, 2010–2011. Interim Technical Report prepared for the Bureau of Land Management, Alaska State Office, Anchorage, Alaska.

Piciullo, L., Storrøsten, E. B., Liu, Z., Nadim, F., & Lacasse, S. (2022). A new look at the statistics of tailings dam failures. Engineering Geology, 303, 106657. https://doi.org/10.1016/j.enggeo.2022.106657

United Nations Environment Programme (UNEP) and the Office for the Co-ordination of Humanitarian Affairs (OCHA) (2000). The Cyanide Spill at Baia Mare, Romania: Before, during, and after. The Regional Environmental Center for Central and Eastern Europe (REC). https://eecentre.org/wp-content/uploads/2019/06/Baia-Mare-cyanide.pdf

U.S. Army Corps of Engineers (USACE), Alaska District (2013). Donlin Gold Project Final Environmental Impact Statement: Final Scoping Report.  https://dnr.alaska.gov/mlw/mining/large-mines/donlin/pdf/dg-feis-scope-report.pdf