The decision
How schools decide to close.
The call is made by a person, in the dark, on evidence collected a few hours earlier. Here is who that person is, what they look at, and why the answer differs one town over.
One person, several phone calls
In a US school district the decision belongs to the superintendent. In Canada it is made at the school board, usually by the director of education or a designated senior official, and in some regions it is coordinated with the transportation consortium that serves several boards at once. Either way it is one named person's call, and they own it publicly by around 5 or 6 AM.
They do not make it alone. The transportation director has already been out on the roads or has crews reporting in. The facilities lead knows whether the parking lots and walkways can be cleared in time and whether the buildings have heat. Somebody has spoken to the county or municipal road authority about what has been plowed and salted and what has not. In many areas the neighboring superintendents are on a group call comparing notes, because nobody wants to be the only district open.
What arrives on the desk is not a forecast. It is a set of observations about roads that already exist, gathered in the previous ninety minutes, plus the forecast for the next four hours. That distinction explains most of the gap between what a weather app says at 10 PM and what gets announced at 5:30 AM.
In the US
The superintendent of the school district decides, and announces to families directly. Districts vary enormously in size, from a few hundred students to several hundred thousand.
In Canada
The school board decides, and in much of Ontario and the prairies a shared transportation consortium cancels buses across several boards at once, which can close schools in one board because of road conditions in another.
Somebody drives the routes at 4 AM
The hours the call actually depends on, and what happens in them.
The decision window across most of the US and Canada is roughly 4 to 7 AM local time. It is not an arbitrary bracket. It is the span between the earliest a road can be usefully assessed and the moment the first bus has to leave the depot to make its first stop on time.
Around 4 AM, transportation staff drive sample routes: typically the worst ones, the hilly sections, the rural stretches that get plowed last, the bridges that ice first. They report conditions back. Somewhere between 4:30 and 5:30 the superintendent or director weighs those reports against the forecast for the next few hours, because a road that is passable now and freezing over by 6:30 is worse than one that is already bad and improving.
The announcement usually goes out between 5 and 6 AM. Districts aim to give families enough notice to arrange childcare before the working day starts, and the pressure to announce earlier is real. But announcing at 9 PM the night before means committing to conditions nobody has seen yet, and getting that wrong in either direction is expensive.
This is why a brief that looked settled at 9 PM can still change. It is also why the honest thing for us to do is show the hours the call depends on rather than a bare total, so you can compare our reasoning to what you know about your own roads.
It is almost always the buses
The question behind the question
Not “is it snowing” but “can a forty-foot vehicle stop safely on the worst hill on route 12, and can a child stand at the end of an unplowed driveway waiting for it in the dark?”
School buildings can nearly always open. They have heat, staff can usually get there, and the parking lot can be cleared. The constraint that actually binds is transportation, and it binds in two places at once: whether the vehicle can complete the route safely, and whether the children waiting for it are safe standing outside in the conditions and the darkness.
That second half is the part people underestimate. In midwinter, elementary students are waiting at rural stops before sunrise. Wind chill matters there in a way it does not matter to a car commute, which is why some districts publish an explicit temperature or wind-chill threshold that closes schools with no snow falling at all.
It also explains a pattern that looks strange from the outside: districts closing on a clear, dry morning after an overnight freeze. The snow finished at 8 PM, the roads were wet, they froze solid by 4 AM, and now every side street is a sheet of ice. The forecast total for that day is zero.
Why the district next door stayed open
Same storm, same forecast grid, different answer.
Geography inside the boundary. District lines were not drawn around terrain. One district can be almost entirely walkable neighborhoods; the one bordering it can include forty miles of county road, a river valley, and a hill that ices before anything else in the region. The forecast gridpoint they share tells you nothing about that difference.
Who plows what. Roads inside a district are cleared by whichever municipality, county, or province owns them, on that authority's own priority list. A district straddling two road authorities can have half its routes cleared by 5 AM and the other half untouched.
How the students travel. A district where most children walk has a completely different calculation from one busing them across a county. Colleges and campuses that house their students close on different thresholds again, because there is no commute to make safe.
Precedent and temperament. Decisions are made by people who remember the last time. A district that stayed open through a bad morning two years ago and took criticism for it will close earlier now. One that closed for a storm that never arrived will hold out longer. This is genuine institutional memory, it is not written down anywhere public, and it is the factor a national forecast map can never see.
It is also the honest limit of what we can model. We read the conditions the decision-maker is looking at. We do not know the judgment they will apply to them, or the phone call they took at 4:30 AM, which is why we publish a probability rather than a verdict.
Check tomorrow
Tomorrow's chance of closure for your location, with the decision window and the factors behind it.
Get my briefHow we calculate this
The four signals, the decision window, and an honest list of where the method runs out.
Read the methodPublished by Snow Brief · August 2026