Told to update climate models, N.W.T.'s Diavik mine says it's considering how and when to do so
A manager at one of the N.W.T.'s diamond mines says the company is considering when to update climate models it's been using to map out what may happen at the site in the decades after it's shut down.
The Diavik mine is expected to stop production in 2026 and to be fully closed down by 2029. It is one of three diamond mines approaching closure in the territory, and is striving to set itself apart from a long history of abandoned mines by successfully relinquishing the land back to the territory.
Rio Tinto, which owns the mine, submitted the first version of its final closure and reclamation plan to the Wek'èezhìi Land and Water Board (WLWB) in 2022. That plan contained data from climate models that forecast 100 years into the future.
Earlier this year, the WLWB told the company it needed to update its climate-change modelling in its final performance assessment report. It also told the mine to run models beyond 100 years – or to propose a time by which it would do so, in the next version of its closure plan.
The question to be answered now is when the company will do both those things.
Sean Sinclair, Diavik's closure manager, at the company's office in Yellowknife. (Liny Lamberink/CBC)
Sean Sinclair, Diavik's closure manager, told CBC News in an interview last week that it was talking with partners and stakeholders about "when and how it's appropriate and best to update" the models.
He said that performance assessment reports are something that happen after the mine is closed and post-closure monitoring has begun. They will show the board whether Diavik is meeting a variety of criteria and objectives of closure.
It has not yet been decided how often those reports will happen, or when a final one might be done.
Sinclair said Diavik's "final position" on how and when to update climate models would be in the next version of its final closure and reclamation plan, which is expected in April.
What's a climate model?
A climate model is a computer simulation that represents how energy and matter interact in different parts of the ocean, atmosphere and land.
They are tested by running them backwards into the past and seeing how they stack up to what really happened — and they can be run forward to forecast conditions under different scenarios of warming.
A pair of physical 3D models — not climate models — show what Diavik's operation looked like in 2017, left, and what it's expected to look like when it's closed. (Liny Lamberink/CBC)
Sinclair said climate models have long been built into Diavik's closure plans. They've been used to engineer a site that, once relinquished back to the territory, can manage flooding, big rainfall events and large amounts of snow that may happen in the decades to come.
Climate scenarios were also factored into the design of the mine's north country rock pile and the processed kimberlite containment facility — two sites that are making use of the cold temperature of the continuous permafrost that underlies East Island in Lac de Gras.
'It blows my mind'
A modelling expert at the University of Toronto is impressed that the WLWB is telling Rio Tinto it needs to look further into the future with its closure plan for the N.W.T. diamond mine.
"It blows my mind, in a good way, that the regulator is asking these questions," said Steve Easterbrook, a computer science professor and director of the university's school of environment.
But, he said, running climate models further into the future will be a challenge to carry out — and the results may not be all that reliable.
WSP, the company which prepared Diavik's climate projections, said it gleaned information from 24 models from climatedata.ca, an online portal that is supported in part by the federal government.
Easterbrook said it was the right data to be using — it's what he uses too — but that the organization likely doesn't run its models beyond 100 years. The problem, he said, is the further into the future you go, the more unreliable the predictions become as uncertainties add up over time.
The A21 pit at Diavik diamond mine. (Liny Lamberink/CBC)
"A lot of climate scientists that I've spoken to say they don't like to run their models more than 100 years into the future just because they know after that there's just too much guesswork."
Industry standard
Running climate models for about 75 years — until 2100 — appears to be an industry standard.
The Mining Association of Canada put out guidelines for climate change adaptation in 2021. It said longer-term models being run until 2100 should be used for closure and post-closure work.
The federal government also has a guide on assessing projects' climate change resilience. It says the majority of climate change simulations extend to 2100, and that simulations beyond that are available but there aren't as many and they come with more uncertainty.
The federal government is also running models to 2100 for its cleanup of the abandoned Giant Mine in Yellowknife — though the Giant Mine Oversight Board called that time frame "short-sighted" earlier this year.
Kirsten Zickfeld, a distinguished professor of climate science at Simon Fraser University in B.C., has worked on long-term climate simulations before. She, too, acknowledges they become less reliable the further into the future they go, but said they are very important for understanding components of the climate system that take longer to respond to change — like oceans, ice sheets and permafrost.
She said project planners — for example, a company winding down a diamond mine — could address uncertainties in long-term projections by pairing them with risk assessments.
"So maybe there's a 70 per cent chance under a given scenario … that permafrost might thaw in that specific area and then you sort of, as a decision maker, you have to decide, 'OK, what does that mean?'"
If Diavik uses climate models to understand what may happen at the mine site beyond 100 years from now, Easterbrook said it would need to look for other sources of data.
"It can be done, but yeah, they're going to have to go further than just using publicly available data on standard data sites," he said. "It will cut down on the number of experts available to actually do this kind of analysis."