Why the forecast temperature is the wrong input
Pipes do not freeze at an outdoor temperature. They freeze at their own temperature, and the gap between the two is where almost all the useful information lives. A pipe on the warm side of a wall cavity in a 68 degree house sits far above the outdoor air. The identical pipe stapled to the sheathing on the cold side of the same insulation sits close to the outdoor air, and it is that pipe that splits. Two runs a few inches apart in the same wall have entirely different winters.
This is why the location selector on this page carries more weight than the temperature field. It sets an estimated temperature at the pipe by placing it somewhere on the gradient between outdoors and indoors, and that estimate then drives everything else. Every other input adjusts how quickly heat leaves once you know how cold it is out there.
The two stages, and why the second is the long one
Freezing standing water is two separate jobs. First the water has to be cooled to 32 degrees, which takes one BTU per pound per degree. Then it has to change state, which takes about 143 BTU per pound at no change in temperature at all. That second number is the reason a pipe that reaches freezing point in twenty minutes may take hours to actually block.
For half inch copper, a foot of pipe holds about a tenth of a pound of water. Cooling it from 50 degrees to freezing is under 2 BTU. Turning it to ice is another 14.5. The latent stage is roughly eight times the sensible stage, and the calculator reports them separately because people consistently expect the opposite proportion.
A consequence worth noticing: the time barely depends on how long the cold run is. A longer pipe holds proportionally more water and presents proportionally more surface for heat to leave through, so the two scale together and cancel. What length does change is how much water you lose if it bursts.
What actually moves the number
| Change | Rough effect on time to freeze | Why |
|---|---|---|
| Stopping wind or draught on the pipe | Two to three times longer | Moving air strips heat off a surface far faster than still air |
| Sealed foam sleeve on a bare pipe | Around three times longer | Adds resistance the heat has to cross to reach the air |
| Raising the space temperature 10°F | Depends heavily on how cold it was | The driving force is the gap below freezing, so 10°F is huge at 25°F and small at minus 10 |
| Larger pipe | Somewhat longer | Volume grows faster than surface, so bigger pipes hold out longer |
| Running a drip | Changes the problem rather than the number | Replaces the coldest water and relieves pressure behind any plug |
The wind row is the one that surprises people, and it is why an air leak at a rim joist or a vent left open matters more than an inch of insulation somewhere else. Insulation slows heat leaving through a surface. A draught replaces the air at that surface continuously, which stops the surface from ever warming its own boundary layer.
Insulation does not add heat
A foam sleeve on a pipe in an unheated space buys time and nothing more. It slows the loss; it supplies nothing. Over a long enough cold snap the water reaches the temperature of the space regardless, and if that space is below freezing the pipe eventually freezes with the insulation on it. Insulation is a delay measure, useful precisely because most cold snaps are shorter than the delay it buys.
What is not a delay measure is getting the pipe onto the warm side. A run relocated inside the insulated envelope, or a cavity air-sealed and insulated so the pipe is in conditioned space, does not have a clock on it. That work is the same work as the rest of an envelope improvement, which the home insulation guide covers, and the heat it saves shows up in the heat loss calculator as well.
What this estimate is worth
It is rough and the page says so more than once, deliberately. The model is a bare pipe of standing water losing heat to air at a single steady temperature. Real pipes are connected to plumbing that is warmer at one end, sit in cavities that buffer them, are bounded by structure on some faces, and face an outdoor temperature that is falling through the night rather than holding. Any of those can shift the answer by a factor of several.
Read the comparisons rather than the clock. The useful outputs are that a bare pipe in wind fails several times sooner than the same pipe still and sleeved, that the temperature at the pipe matters far more than the temperature on the forecast, and that a drip changes the failure mode rather than the arithmetic. If the whole house is the question rather than one pipe, the cold wave guide is the wider checklist, and the heating fuel reserve calculator answers the other question a long cold snap raises.
Questions people ask
At what temperature do pipes freeze?
There is no single temperature, which is the honest and unsatisfying answer. Water freezes at 32°F, but a pipe reaches that only after it has lost enough heat, and how fast that happens depends on where the pipe runs, whether air is moving over it, what covers it, how big it is and how long the cold lasts. Figures around 20°F outdoors get repeated as a threshold, and they describe a common case rather than a rule: a pipe on the cold side of a wall in a draught can freeze well above that, and a pipe on the warm side may never freeze at any outdoor temperature the location sees.
Does letting a tap drip actually prevent frozen pipes?
It helps, and for a reason people usually get slightly wrong. The point is not that moving water cannot freeze. It is that a drip continuously replaces the coldest water in the run with water from further back in the warmer part of the system, and just as importantly it relieves pressure. Most burst pipes do not fail at the ice plug itself; they fail downstream of it, where water trapped between the growing plug and a closed tap has nowhere to go as the ice expands. An open path releases that pressure. A drip costs water, does nothing for a pipe that is already blocked, and is a measure for a specific cold night rather than a way to run a winter.
Is PEX safe from freezing?
PEX tolerates freezing better than rigid pipe because it can expand somewhat and is more likely to survive a freeze that would split copper, and that is a genuine advantage. It is not immunity. PEX still freezes at the same temperature and still blocks, and fittings, manifolds, valves and any rigid section in the run remain vulnerable to the pressure that builds behind an ice plug. A frozen PEX line also stops delivering water regardless of whether it survives. Treat the material as a difference in consequence rather than a reason to skip the precautions.
Why did my pipe burst somewhere other than where it froze?
Because the damage is done by pressure rather than by the ice directly. As a plug grows it pushes water ahead of it down the pipe, and if the far end is a closed tap or valve, that water has nowhere to go. Pressure in the trapped section climbs until the weakest point gives way, and the weakest point is frequently well away from the ice, at a fitting, a solder joint or a thinner section. This is also why opening a tap on the affected line is the standard advice: it gives the trapped water somewhere to go, which is a different thing from thawing anything.
How accurate is this time estimate?
Rough, and it is designed to be read as rough. It treats a bare length of pipe holding still water, losing heat to air at one steady temperature, which is a simplification of a real installation in several ways at once. Connected warm plumbing, an enclosing cavity, structure on some faces, a draught on another and an outdoor temperature that keeps falling can each move the result substantially. Use it comparatively: it is reliable about what insulation buys relative to what blocking a draught buys, and about the fact that the temperature at the pipe matters far more than the one on the forecast. Do not use it as a countdown.