I stepped into the quiet hallways of McGill University’s Burnside Hall on a crisp Montreal morning, where the unassuming 10th floor houses one of Canada’s most important climate resilience projects. Far from the political battlegrounds south of the border, a team of dedicated archivists and scientists work meticulously to preserve what many fear could disappear: America’s climate research data.
“It started as an emergency response,” explains Dr. Sarah Masson, environmental data coordinator at McGill’s Geographic Information Centre. She gestures toward a wall of servers humming quietly in the background. “After the 2016 U.S. election, we witnessed government climate pages being altered or removed entirely. That’s when we knew we needed to act.”
The Data Refuge project at McGill is part of a North American network that backs up vulnerable U.S. environmental and climate research – creating a scientific sanctuary that transcends borders and political cycles. Since 2017, the initiative has preserved over 200 terabytes of critical climate data, including irreplaceable long-term atmospheric carbon measurements, arctic ice surveys, and coastal erosion tracking.
What makes this Canadian safeguarding effort particularly significant is the growing pattern of climate information becoming politicized or threatened during U.S. administrative transitions. Following Donald Trump’s first election, terms like “climate change” and “global warming” disappeared from government websites, while research budgets faced dramatic cuts.
“Scientific data isn’t opinion – it’s fact. Measurements taken over decades can’t simply be recreated if lost,” says Dr. Masson. “And unfortunately, we’re seeing environmental data becoming increasingly vulnerable to political pressures.”
Environment and Climate Change Canada reports show that Canada is warming at twice the global average rate, making climate research particularly vital for planning across sectors from agriculture to infrastructure. This interconnectedness of North American ecosystems means Canadian environmental planning relies heavily on comprehensive continental data.
The cross-border collaboration extends beyond McGill. The University of Pennsylvania, University of Toronto, and dozens of research institutions participate in coordinated “data rescue” events, where volunteers systematically download, verify, and archive vulnerable datasets.
I met with Jérôme Lavoie, a graduate student who volunteers with the project. As we walked through McGill’s campus, fresh spring buds emerging from tree branches, he reflected on the deeper significance of his work. “When I first joined, I thought we were just backing up files. But I’ve come to see this as protecting our shared scientific heritage – the evidence that helps us understand our changing planet.”
The urgency of this preservation effort was highlighted in a 2020 report by the Union of Concerned Scientists, which documented over 150 cases of scientific information being removed, altered, or made less accessible during a single U.S. administrative term. Environmental datasets were disproportionately targeted.
While touring the facility, I observed firsthand the meticulous process involved. Each dataset is downloaded with its original metadata intact, then verified, cataloged, and stored with redundant backups. The team also preserves the relationships between datasets – crucial context that helps scientists understand how different environmental systems interact.
“It’s not just about storing files,” explains technical director Martin Beauregard. “We’re preserving scientific context and meaning. That satellite imagery only makes sense with its calibration information and measurement methodologies.”
The implications extend far beyond academic concerns. Agricultural planning, flood mitigation, forest fire management, and public health responses all depend on consistent, long-term climate data. When I spoke with Dr. Eleanor Wentworth, climate adaptation specialist with the Quebec government, she emphasized the practical importance of this work.
“When designing flood protection for riverside communities or planning crop adaptations, we need reliable historical data,” Wentworth explains. “A gap of even a few years can significantly weaken our predictive models, potentially putting communities at risk.”
The McGill initiative stands apart from standard data backups by focusing specifically on accessibility and usability. Their platform includes visualization tools that allow policymakers, journalists, and the public to engage meaningfully with complex scientific information.
