Whether operations or closure, downstream surface erosion of tailings embankments can be a perpetual challenge. During operations, the management of erosion and sediment transport is typically handled in due course of normal operating procedures. However, the challenges of erosion mount during reclamation and closure when owners face long-term environmental obligations and aspire to passive-care or walk-away solutions. As the recently published Global Tailings Standard highlights: “Few tailings facilities have a (design basis memorandum) that addresses long-term reclamation performance. Lack of clear agreement on design objectives and future performance creates a gap between what is planned by the mine and what is expected by regulators and local stakeholders.” Decisions on tailings reclamation may be further perplexed when the design and operation of tailing is at odds with the ideals of long-term closure (i.e. capital cost, storage, footprint, etc.). So, if we are required to design for and demonstrate the feasibility of safe closure, how do we quantify the long-term erosion performance of a closure landform to meet the expectations of industry guidelines?
There are tools available to designers to predict erosion such as RUSLE and WEPP as well as more sophisticated tools that combine erosion and long-term landform evolution such as SIBERIA and CAESAR-Lisflood. Recent work led by the University of Arizona, in collaboration with SRK and a soil science consultancy, has endeavored to study the erosion performance of a reclaimed upstream tailings embankment in the arid southwestern USA. The desert landscape receives modest, but intense, annual precipitation resulting in high-intensity erosive events. Portions of the facility cover experience rill erosion nearly every year. Using a decade of post-reclamation LiDAR data and two years of field data collected from various scales of rainfall-runoff-erosion test plots, the study team evaluated the performance and utility of the most common erosion models. Furthermore, the University of Arizona developed a new code, RILLGEN2D to model rill initiation in a rock-armor cover, which presents a novel approach to rill prediction based on material properties and landform geometry.
The study team was able to compare the different model codes against observed rilling and assess the strengths and weaknesses of each. Using RILLGEN2D, the team successfully retrodicted where rills form on an existing tailings embankment cover. Results to date show promise using SIBERIA to evaluate the long-term performance of landform designs and flag high-risk areas. Work is ongoing to link rill prediction using RILLGEN2D with sediment generation and transport in WEPP.
Some findings from this on-going study and considerations for erosion performance design:
- Know your Peak: A clear understanding of rainfall intensity, watershed scale, and runoff/infiltration are key to accurately predicting the erosive force that will be applied to a cover. For example, to predict rill formation the models needed sub-hourly rainfall intensity, contrary to many surface water designs that use longer duration storm events (i.e. 6-hr, 12-hr, 24-hr, etc.).
- Erosion Controls: For a given unit discharge at a point on a slope, erosion rates increase non-linearly with slope steepness. Understanding the impacts of slope geometry and material properties is key to predicting where erosion will occur. In the case of rock-armor covers, bigger isn’t always better.
- Un-natural Analog: Current trends towards geomorphic landforms and “natural analog” design approaches are desirable, but not always possible with legacy sites or sites with boundary limitations. In this case, designers must be prepared to tackle more “structural” approaches with analytical tools to determine the maximum slope lengths, gradients, and required cover gradations to resist erosion. Utilizing divergent topography and convex-concave slopes can improve designs where space is limited.
- Plan for Imperfection: A constructed closure landform can have unintended microtopography and subtle variations from the design that will increase the unit contributing area for a given slope. In this study micro-depressions resulted in unit discharges up to 5 times higher than adjacent areas, exceeding the intended design, and resulting in failure. Construction complexity and field tolerances should be closely considered when evaluating designs, even ones that appear simple.
