Buried inside the Montreal summit announcements was a deal that sounds technical and could matter more than the headlines: Canada and the European Union signed a cooperation arrangement on Destination Earth, the European programme building a digital twin of the planet. Signed on October 30 by the European Commission and Environment and Climate Change Canada, the arrangement names three priorities: wildfire prediction, Arctic change and climate risk in vulnerable regions including the Caribbean. The question now is whether Canadian researchers and emergency planners will actually get usable data, or only a signed paper.
The distinction matters because the European model is already running, while access to its data remains tightly controlled. The Commission published Data Lake terms draw a hard line: a user needs a validated personal account, with partial or upgraded access, the account cannot be transferred, underlying data remain European Union property, and passing data to a third party requires the Commission’s prior written consent. Those terms describe the existing service, not a new Canadian grant, and public files identify no Canadian account grant, no integrated wildfire run and no release date for a joint warning product. A project approval, a published model result or an operational service notice would change that record. None has yet appeared.
What the arrangement does commit to is direction. The cooperation will focus on artificial intelligence algorithms to improve predictions of wildfire ignition and spread, on simulations of Arctic sea ice decline, permafrost thaw and atmospheric shifts, and on high resolution forecasting for regions exposed to extreme weather. The European Centre for Medium Range Weather Forecasts says the climate and extreme event digital twins and their engine matured and became operational in phase two, with phase three running to June 2028. The engine runs workflows on EuroHPC systems, and Canada participation in the European High Performance Computing Joint Undertaking, agreed at Montreal, gives Canadian researchers a route into that computing power.
Why wildfire modelling is the test
Wildfire is the sharpest test because it is where a model either helps a warning or does not. A wildfire does not wait for a summit photograph, and a priority announced in Montreal only becomes useful when a specific team obtains access, runs a model and decides whether the output is fit for a warning. Canada wildfire seasons have repeatedly strained provincial resources, and better prediction of ignition and spread would have direct value for evacuation planning and firefighting deployment. But that value depends on data access decisions that the October 30 announcement does not itself make.
The climate modelling arrangement therefore sits in the same position as the wider Montreal package: signed, directional and unproven in operation. Its supporters can point to a real platform, real priorities and a real computing partnership. Its sceptics can point to the absence, so far, of a Canadian run, a Canadian warning product or a published Canadian result. The March review of the wider Canada Europe relationship will be an early moment to ask which side of that ledger has grown.
What to look for next
Three markers will show whether the arrangement is working: a named Canadian team with validated access, a published model result on wildfire or Arctic change, and a service notice describing how model output feeds an operational warning. Until those appear, the honest description is a cooperation arrangement with strong priorities and controlled data, not yet a climate service. For a country that loses forest, homes and firefighting seasons to wildfire, the difference between those two descriptions is the whole story.
