How districts actually decide to close school
There is no national standard and there is no formula. What there is, in every district that handles this well, is a set of written thresholds, a timeline that works backwards from the first bus, and a named person who owns the call. This is what those look like.
| Hazard | What you actually measure | Common action band | Source of the number | Decision by | Owns the call |
|---|---|---|---|---|---|
| Extreme cold | Wind chill at bus-stop height, at stop time | Delay at −15°F to −20°F wind chill; close below −25°F to −30°F | NWS wind chill advisory and warning criteria, which vary by region | 04:30 | Superintendent |
| Snow, falling rate | Inches per hour during the route window | Delay above 1"/hr; close above 2"/hr through the run | Plough cycle time — can crews clear faster than it falls | 04:45 | Supt. on transport advice |
| Snow, accumulated | Depth at rural turnarounds, not at the office | Close above 6–8" untreated; higher where the fleet is equipped | Road crew capacity and drift exposure | 04:45 | Transport director |
| Ice and freezing rain | Accretion in inches on treated surfaces | Close at any measurable accretion on bus routes | NWS ice storm warning at 0.25"; districts act far below it | 04:30 | Superintendent |
| Road surface temperature | Pavement temp vs air temp, on a route drive | Below 15°F, salt stops working; treat as untreated | Chloride effectiveness curve | 04:30 | Road crew report |
| Wind | Sustained speed and gust, on exposed routes | Review at 40 mph gusts; high-profile buses at real risk above 50–55 mph | NWS high wind warning; fleet guidance | 05:00 | Transport director |
| Extreme heat | Heat index, plus classroom temperature | Early release or closure at heat index 100–105°F without cooling | NHTSA/OSHA heat guidance; state cooling rules | Prior afternoon | Superintendent |
| Classroom temperature | Actual room temp at 10:00, worst room | Close above 85–88°F sustained; below 60–65°F in winter | State facility codes where they exist | Prior afternoon or 05:30 | Facilities |
| Air quality | Hourly PM2.5 AQI at a named monitor | Outdoor activity stops at 151; closure considered above 200 | EPA AQI bands; some state education agencies publish school rules | 06:00 | Superintendent |
| Flooding | Water moving over a roadway on any route | Route suspended at any flowing water; close on multiple routes | Route drive, county road closures | 04:30 | Transport director |
| Fog | Visibility in feet on the worst route segment | Delay below about 1/4 mile at route time | Driver judgement; NWS dense fog advisory | 05:00 | Transport director |
| No heat in a building | Overnight building temp trend | Close that building below about 60°F with no recovery path | State minimum temperature codes | 05:00 | Facilities and principal |
| No water or boil notice | Utility notice in writing | Bottled-water day if food service can adapt; close if not | Utility and county health department | 05:30 | Supt. and nutrition services |
| Staffing — drivers or subs | Unfilled routes or unfilled classrooms at 05:00 | Route-level action; district closure only in extremis | Internal fill-rate threshold | 05:00 | HR and transport |
Every band above is a convention, not a rule, and the variance between districts is enormous and mostly rational: a Minnesota district with a block-heater fleet and a −30°F culture closes at a number that would be unremarkable in Georgia, where the same reading would shut everything for two days. The point of writing your own numbers down is not that they are objectively correct. It is that when you announce a decision, you can say what governed it, and when you make the opposite decision to the district next door you can explain why in one sentence.
The 5am timeline, worked backwards
The decision looks like a judgement call and is mostly a schedule. Here is the version that works, for a district with a 06:10 first bus roll.
- The evening before, 18:00–20:00. Look at the forecast and decide whether tonight is a night-before call, a conditional, or a normal night. If it is a conditional, send it — families would rather be told at 8pm that you will decide at 5am than find out at 5am that there was a decision to make.
- 02:30–03:00. Somebody checks the forecast and radar. Usually the transport director. If nothing has changed, they go back to bed and the road crew report at 04:00 stands as the trigger.
- 03:30–04:00. Public works and county road crews report what they have cleared and what they will not reach. This is the single most informative input and it is the one most often skipped by districts that decide off a weather app.
- 04:00–04:40. Drive the routes. Not the main roads — the rural turnarounds, the hills, the shaded stretch by the creek, the neighbourhood the plough does last. Two or three staff in different parts of the district beats one person on a loop. This is where the decision is actually made; everything after it is confirmation.
- 04:40–04:55. Call the neighbouring superintendents. Not to copy them — to know what they are doing, because your families will ask, and because shared bus contractors, shared vocational centres and shared athletic schedules mean their decision has operational consequences for you.
- 05:00. Decide. One person, out loud, on a recorded call or in a group chat with a timestamp.
- 05:05–05:15. Send. All channels at once. Website, text, voice, email, social, and the local broadcasters if you still feed them.
- 05:35. First bus rolls, or does not. After this point a reversal costs you children standing at stops.
Two structural notes. First, the person who decides should not also be the person operating the messaging platform at 05:05, because the composing takes ten minutes and the decision window is fifteen. Draft the message before the decision, with both outcomes in the file. Second, whoever is going to press send should have sent a test to themselves the week before. The morning of a closure is not when you want to discover your account is locked.
Who decides, and the people they are deciding for
In almost every district the superintendent owns the call formally. In practice it is a recommendation from the transportation director, endorsed by the superintendent, occasionally overridden. That structure works. What does not work is a committee, and what really does not work is a committee at 04:50.
Write down, before the season: who recommends, who decides, who decides when that person is unreachable, and who presses send. Three names deep on each, with mobile numbers, on one page. The second name matters more than people expect — superintendents travel, and the deputy who has never made the call before will make a worse one at 05:00 with no written thresholds in front of them.
The harder question is who you are deciding for. A closure decision affects five populations with different interests, and being explicit about the trade-offs makes the decision defensible rather than arbitrary.
- Bussed rural students, whose safety is genuinely the binding constraint and whose conditions are worse than anything visible from the district office.
- Walkers, who are invisible in transport-led decisions. A district with high walker numbers has a different cold threshold than one that busses everybody, because a child walking twenty minutes in −10°F wind chill is exposed in a way a child at a bus stop for four minutes is not.
- Teenage drivers, the population with the worst outcomes in ice and the one nobody models. Some districts close secondary and hold elementary for exactly this reason.
- Families who cannot miss work, for whom a closure is a childcare emergency and, in a meaningful number of cases, a lost day's pay.
- Students who rely on the school for meals, warmth, medication or supervision, for whom a closure is materially worse than a marginal commute. This population is the strongest argument against closing on a forecast that has not yet materialised.
None of that produces a formula. It produces a better sentence in the message, which is what you are actually optimising: "we closed because the rural turnarounds were not passable at 4:30" is a decision. "Out of an abundance of caution" is not.
Why the night before beats the 5am call, and when it does not
The night-before decision is better for families in every way that matters, and districts under-use it. A closure announced at 8pm lets a parent arrange cover, swap a shift, tell a grandparent, or decide to work from home. The same closure announced at 05:10 gives them forty minutes, and the people with the least flexible jobs are hit hardest by the difference.
The argument against is accuracy, and it is a real argument. Forecasts at 8pm for a 5am condition are good on temperature, decent on accumulation totals, poor on timing, and unreliable on the rain-snow line. A district that calls it early on a marginal forecast and then wakes to clear roads has spent something it will need in February.
The practical resolution is to sort your situations into three and treat them differently:
- Call it the night before when the mechanism is temperature rather than precipitation — extreme cold is forecastable twelve hours out with real confidence — or when the storm has already arrived and is forecast to continue, or when a warning is in force through the morning window with high confidence.
- Send the conditional when the forecast is genuinely uncertain. State the exact time you will decide, commit to messaging either way including if the answer is that school is open, and tell people not to wait up. This is the message almost nobody publishes and it converts an anxious evening into a settled one.
- Wait for the morning when the condition depends on overnight timing, on the rain-snow line, or on whether crews get ahead of it — and accept that you have chosen accuracy over notice, which is a defensible choice you should be willing to explain.
Whatever you choose, the failure mode to avoid is the hedge. "We may need to close tomorrow" sent at 9pm produces the phone traffic of a closure and the planning value of nothing. Either decide, or say precisely when you will.
Communicating a decision you are not sure about
Most guidance on closure communication assumes the district is confident. The interesting cases are the ones where it is not, and there are three of them.
The marginal call you are making anyway. You have decided to open on conditions that are close to your threshold. Say so, name the measure, and give families explicit permission to disagree with you. "Our routes were driven at 4:30 and are passable, so schools are open. Conditions vary across the district and you know your road better than we do. If travel from your home is unsafe, keep your child home and it will be excused with no note." That sentence costs you a handful of absences and buys you the ability to open on marginal days for the rest of the winter without an argument.
The call that might reverse. Be explicit that it might, and attach a time. "Schools are on a two-hour delay. If the road crews do not get ahead of this by 07:00 we will move to a closure, and we will message by 07:00 either way." Families can plan around a stated risk. They cannot plan around a surprise, and a reversal that was flagged in advance costs a fraction of one that was not.
The call you got wrong. Districts almost never send this message and it is worth more than any other message in this guide. One paragraph, the morning after: what you decided, what actually happened, what you will do differently. "We opened yesterday and the roads in the north of the district were worse than our 4:30 check suggested. Twelve buses ran more than thirty minutes late and some students waited in conditions we would not have chosen. We are adding a second route check on [road] and moving our decision time to 04:45. We are sorry." Nothing else you send buys as much latitude for the next marginal decision.
What to write down before the season, on one page. Your thresholds by hazard, with a number against each and a note on what makes your district different. The decision timeline with clock times. Three names for recommend, decide and send, with mobile numbers. The draft closure, delay and conditional messages, already written, in English and Spanish, with blanks for date and time. The list of who else gets told and in what order — broadcasters, child nutrition, after-school providers, athletics, facility renters, the vocational centre, the neighbouring district that shares your buses. And the account credentials for your messaging platform, tested in September by the person who will actually use them in January. Districts that have this page make the same decisions as districts that do not, about fifteen minutes faster and with far less argument.
The mechanics of the 05:05 send
Three things go wrong on the send itself, and none of them are weather.
Quiet hours block it. Most platforms suppress overnight sends by default; Kastr's default window is 21:00–07:00. A 05:05 closure notice has to be flagged as an emergency to leave at all. That is deliberate — it makes waking the district a conscious act rather than an accident, and in Kastr the emergency flag sits behind a separate capability and is written to a tamper-evident audit log. But it is exactly the kind of setting you want to have exercised in October rather than discovered in January.
It sends twice. The page appears to hang, somebody presses send again, and 8,000 families get two texts at 05:06. Kastr keys each delivery on broadcast, person and channel, so the second attempt is a no-op at the database level rather than a second message. Ask any vendor how they handle this specifically; the answer tells you a lot.
It goes nowhere. Unregistered A2P traffic from a 10-digit number is filtered by carriers silently, with no error returned. See A2P 10DLC — registration takes weeks, and the time to do it is August. Relatedly, the number to ask your platform for is true reach, distinct families reached on at least one channel, rather than messages sent, which counts the ones that vanished.
Two features earn their keep specifically on this decision. A cost-and-reach preview before you send, because a four-character overrun turns one SMS segment into two across the whole district. And schedule-later with the audience resolved at send time rather than at schedule time, which is what makes a 20:00 night-before send safe. Kastr does both. What Kastr does not do is resolve audiences by school, grade or bus route — only named individuals and everyone — which matters if partial closures are routine for you. That limitation is stated on every page in this cluster because it is the one most likely to matter.
Questions people actually ask
At what temperature do schools close for cold?
There is no national threshold and the variance is genuinely large. Districts in the upper Midwest commonly delay somewhere around a −15°F to −20°F wind chill and close below −25°F to −30°F; districts in the South act at readings that would be an ordinary morning in Minnesota. What matters is measuring wind chill at bus-stop height at the time students are actually outside, adjusting for how many of your students walk, and writing the number down in advance so you are not negotiating it at 04:30.
How much snow does it take to close school?
Less than people think, and the depth matters less than the rate and the timing. Two inches falling per hour during the route window closes more districts than eight inches that fell and stopped overnight, because the question is whether ploughs can clear faster than it accumulates. Accumulated depth becomes decisive at rural turnarounds and unploughed side streets, not on the main roads. Measure where the buses have the most trouble, not at the district office.
Who makes the decision to close school?
Formally the superintendent, in almost every district. Operationally it is a recommendation from the transportation director based on a road check, endorsed or occasionally overridden. Write down three names for each role — recommend, decide, send — before the season, because the deputy making their first call at 05:00 without written thresholds in front of them makes a worse decision than the process deserves.
Is it better to announce a school closure the night before or at 5am?
The night before, whenever the forecast supports it, because it lets families arrange childcare rather than scramble. It supports it when the mechanism is temperature rather than precipitation timing, when the storm has already arrived, or when a warning is in force with high confidence through the morning. When the forecast is genuinely uncertain, send a conditional instead: name the exact time you will decide and commit to messaging either way, including if the answer is that school is open.
What should a district say when it makes a closure decision it is unsure about?
Name the measure you acted on, and give families explicit permission to make a different call for their own household. "Routes were driven at 4:30 and are passable, so schools are open. Conditions vary; if travel from your home is unsafe, keep your child home and it is excused with no note." If a reversal is possible, say so and attach a time by which you will confirm. And if you got it wrong, send the short message the next morning saying what you will change — it buys more credibility than anything else in this guide.
Should a district close because the neighbouring district closed?
No, but you should call them, and you should expect to explain any difference. Neighbouring decisions matter operationally — shared bus contractors, shared vocational centres, staff who live in the other district — and they matter reputationally, because your families will compare. The answer to "why are we open when they are closed" is a factual sentence about your own routes, not a defence of your judgement.
One price. Every feature. Locked for three years.
$3.50 per student per year under 5,000 students. No tiers, no add-on modules, no per-message fees. Published on the site because you should not have to book a call to learn a price.