What Actually Closes a School for Snow
Published 11/3/2026 · 7 min read · Everyday calculators
Lena Hoffmann — Science & education writer at OneKitly
Mathematics · Physics
Checked against 3 sources
No school closes because of a depth of snow. It closes because the roads the buses use cannot be made safe by the time the buses need them, and that depends far more on how well a place is equipped than on how much falls. The heuristic behind the calculator says the same thing in numbers: with a 3 in forecast at 28 °F, an inch already down and snow-covered roads, it gives 86 % in the US South, 74 % in the Mid-Atlantic, 61 % in the Midwest, 58 % in the Northeast and 54 % in the Mountain West — a 32-point spread on identical weather. The other big lever is ice: freezing rain adds as much to the odds as 6 cm of snow, because ice defeats the ploughs entirely.
The same 8 cm forecast gives 86 % odds in a mild-winter area and 58 % in a snowy one. That gap is the whole answer: closures track the mismatch between the snow and the ploughs, not the depth on the ruler.
The decision is about buses, not about snow
A school does not decide whether snow is dangerous in general. It decides whether the specific roads its pupils arrive on will be passable at the specific hour they arrive, and whether they will still be passable at the hour everyone has to go home. That is why the ranking of factors in the model is not the ranking most people expect. Road condition carries a thirty-point spread, from minus ten for a clear road to plus twenty for an icy one, and it is doing the work that snow depth only appears to do.
The school-type term makes the same point from the other side. A public school scores plus five, a private one zero, a university minus ten — a fifteen-point gap that has nothing to do with the weather. Universities run no bus fleet, their students are adults who arrange their own transport, and their liability for a nineteen-year-old walking to a lecture is not the liability of putting a seven-year-old on a coach. Two institutions on the same street, in the same storm, reach opposite decisions for reasons no forecast contains.
Why a mild region beats a cold one on identical weather
The region term ranges from plus twenty for a mild-winter area down to minus twelve for the mountains, and it is not measuring how cold a place is. It is standing in for how many ploughs, gritters and trained drivers that place owns per kilometre of road, and for whether its households own winter tyres. A mountain district that sees two metres a season has the fleet, the routine and the drivers to absorb 8 cm before breakfast. A mild coastal town that sees snow twice a decade has none of that, and the same 8 cm is a genuine emergency.
This also explains the model's most counter-intuitive line. Cold helps a closure — below −1 °C adds eight points and below −8 °C adds eight more, because snow that stays frozen stays on the road — but above 2 °C the model subtracts fifteen. Snow falling onto warm ground melts as it lands, and a forecast of 10 cm at 4 °C closes nothing. The depth on the forecast and the depth on the road are different numbers, and only the second one matters.
Two places the model is deliberately crude, and one where it is simply wrong
The snowfall term saturates. Every centimetre adds four points up to a ceiling of sixty, which is reached at 15 cm, so a 15 cm forecast and a 30 cm forecast score identically. That is deliberate and defensible: past a certain depth the closure is already certain and more snow cannot make it more certain. The wind term only engages above 30 km/h, and it is modelling drifting rather than cold — wind that fills a cleared road back in, not wind that makes a bus stop unpleasant.
The place it is simply wrong is that the factors are added independently and never checked against each other. Ask it about a mild-winter area at −3 °C with snow-covered roads, a public school, no snow forecast and none on the ground, and it returns 51 % — a coin flip drawn from an impossible situation, since roads cannot be snow-covered when no snow has fallen or is falling. That is worth knowing before trusting a middling number: the region and road terms alone can carry the score past fifty with no weather behind them. The verdict labels are honest about the rest — anything under 25 % reads as a normal day.
Who actually makes the call, and when
In the United States the decision belongs to the school district superintendent and is usually taken between four and six in the morning, after road crews and transport managers have driven the routes. In France, Germany, Spain, Italy and Portugal the school itself rarely decides alone: a prefecture, municipality or Land authority issues the closure, and it often follows from the suspension of school transport rather than from a judgement about the buildings. The chain of causation runs through the bus in every one of them.
Which is the reason the tool carries a disclaimer rather than a claim. It is a fun heuristic with a defensible shape, not a forecast: the only authority on whether your school opens tomorrow is your school, and the only reliable signal is the one it publishes itself. Use the calculator to understand which factors are doing the work — that is a genuinely useful thing to know at nine in the evening — and use the district's own channel to find out what happened.
| Factor | Range of effect | Note |
|---|---|---|
| Snowfall forecast | 0 to +60 | Saturates at 6 in — 12 in scores the same as 6 |
| Freezing rain | 0 or +25 | A yes/no worth 2.4 in of snow |
| Region | −12 to +20 | The equipment proxy — the biggest single lever |
| Road condition | −10 to +20 | Icy beats snow-covered by 5 points |
| Temperature | −15 to +16 | Above 36 °F it subtracts 15 — snow that melts is not a closure |
| Snow already down | 0 to +20 | Crews already stretched before it starts |
| Wind | 0 to +15 | Only above 19 mph — drifting, not cold |
| School type | −10 to +5 | Universities almost never close: no buses, adult students |
Worked with our own calculator
Snow day calculator
Given
- Unit system
- Imperial (°F, in, mph)
- Temperature (°F or °C)
- 14
- Snowfall forecast (in or cm)
- 0
- Wind speed (mph or km/h)
- 0
- Snow already on ground (in or cm)
- 0
- School type
- Public school
- Region
- US South
- Road conditions
- Clear / dry
- Freezing rain expected?
- No
Result
- Snow-day probability
- 36%
These figures are produced by the calculator below, not typed in by hand — they are recomputed whenever the tool changes.
Run it on your own figures →Frequently asked questions
- How much snow does it take to close a school?
- There is no single depth, and that is the honest answer. In the model, a 3 in forecast at 28 °F with snow-covered roads reaches 86 % in the US South and 54 % in the Mountain West. The same weather, opposite verdicts. If you want a rule of thumb for your own area, look at what closed it last time rather than at any national figure.
- Why does freezing rain matter more than snow?
- Because a plough can move snow and cannot move ice. Freezing rain falls as liquid and freezes on contact, bonding to the road surface in a sheet that grit takes hours to break and that reforms as soon as traffic polishes it. The model gives it a flat +25, the same weight as 6 cm of snow, and that is if anything conservative: a serious ice storm closes places that shrug off half a metre of powder.
- The regions in the tool are American. Can I use it in Europe?
- Yes, through the two generic options. Seven of the nine regions name American zones, but the list also carries a mild-winter area and a snowy-winter area, and those are the ones to pick outside the United States: mild for Lisbon, Seville, Rome or the French Atlantic coast, snowy for Bavaria, the Alps, Scandinavia or northern Poland. The mountain option works anywhere mountainous. Switch the unit selector to metric first, or the tool will read your centimetres as inches.
Articles you may find interesting
All guides →Related tools
Sources
Spotted a mistake in this article?