Days of heat, not gallons
A tank gauge tells you a fraction. It does not tell you a duration, and duration is the only thing that matters when a delivery is a week out and the forecast is bad. Converting one into the other takes three numbers: how much heat the house is actually taking, how efficiently the appliance turns fuel into that heat, and how much energy is in a unit of the fuel.
The chain runs like this. Average heat delivered, in BTU per hour, divided by the efficiency, gives the fuel energy the appliance has to consume. Multiply by 24 for a day. Divide by the energy content of one gallon, bag or cord and you have units per day. Divide the usable fuel by that and you have days. Everything else on this page is a consequence of that one calculation.
The input people get wrong is the load. A heat loss calculation gives the design load, which is what the house needs on the coldest hour of a design winter, and the seasonal average is far below it. If you have past delivery records, the most reliable way to get this number is backwards: divide the fuel used over a known period by the days in it, and adjust the load field until the consumption figure here matches what you actually burned. The heat loss calculator gives you the design figure if you want to start from the building instead.
Energy content, and why these are approximations
| Fuel | Nominal energy | What moves it |
|---|---|---|
| Propane | About 91,500 BTU per gallon | Fairly consistent. Tank temperature affects the vaporisation rate rather than the energy |
| Heating oil | About 138,500 BTU per gallon | Grade and blend |
| Kerosene | About 135,000 BTU per gallon | Grade |
| Wood pellets | About 320,000 BTU per 40 lb bag | Species, ash content and how dry the bag has stayed |
| Seasoned hardwood | About 20 million BTU per cord | Species and moisture, which between them swing this by half or more |
Every figure above is nominal and every one is editable through the custom option. Wood is the least reliable by a wide margin, because a cord is a stacked volume rather than a mass and because unseasoned wood spends a large part of its energy boiling off its own water before any heat reaches the room. If you burn wood, the number that matters is not what the supplier calls it but how long it has been stacked somewhere it can dry.
The refill trigger, which is the point of the page
A reserve is not measured in fuel. It is measured in whether the fuel outlasts the delivery. The trigger level here is the amount you will not draw below, plus the consumption over the delivery lead time, plus a margin. Reach that level and you order, regardless of how comfortable the gauge looks, because everything below it is committed.
Two things make lead times worse than the one you have in your head. The first is correlation: a regional cold spell makes every household in the area low at the same moment, and suppliers who normally deliver in three days start quoting ten. The second is access, since the same weather that empties tanks is the weather that keeps a truck off a rural road. Both argue for a larger margin than feels necessary in October.
The cold snap field exists to make that vivid. Set it to 50 or 100 percent and watch the days figure halve. That is the scenario the reserve is for, and a reserve sized against the seasonal average is a reserve sized against the wrong case.
Backup heat, and the one thing that must not be improvised
When the reserve does run short, or when the power fails and takes the controls and blower with it, households reach for backup heat. Fuel-burning heaters and generators produce carbon monoxide, and they must never be run indoors, in a basement, or in an attached garage. That is true of unvented portable heaters used for a night, of a generator by an open window, and of a grill under any circumstances at all. Carbon monoxide alarms belong in the house before any backup heat source does, and permanent heating equipment is installed and vented by somebody qualified to do it.
Beyond that, the cheapest heat in a cold event is the heat you do not lose. Closing off unused rooms, blocking draughts, and heating one room rather than a house stretches a reserve further than any adjustment to the burn rate, and the cold weather guide covers the household side of that. If the outage is the problem rather than the fuel, the power outage timeline shows what else in the house is on a clock at the same time.
Questions people ask
How long will my propane last?
Divide the usable gallons by your daily burn, and get the daily burn from your average heat load rather than the design load. At 25,000 BTU per hour delivered through an 80 percent efficient appliance, the consumption works out at roughly 8.2 gallons a day, so 225 usable gallons is about 27 days. Two things distort that: efficiency figures on old equipment are lower than the nameplate, and a hard cold spell can double consumption. If you have past delivery records, work the load backwards from them, because your own history beats any table.
What level should I reorder at?
At the level that covers the delivery, not at a fixed fraction of the tank. That is the amount you will not draw below plus the fuel your lead time consumes plus a margin. On the defaults here it works out around 82 gallons in a 500 gallon store, which is a gauge reading somewhere near 16 percent, and it is far higher than the quarter-tank habit most people use in ordinary weather because a quarter tank means nothing without a lead time attached to it. Ask your supplier what their current lead time actually is in cold weather rather than assuming the summer one.
Why is my fuel use so much higher than this says?
Usually the load figure. The average heat delivered over a season is much lower than the design load a heat loss calculation produces, so entering the design figure overstates consumption, while entering a mild-weather figure understates it in a cold month. Efficiency is the other common gap: an older appliance, a system with long uninsulated runs, or an open fire delivers far less of the fuel energy into the rooms than the rating suggests. Calibrate against a past delivery. Divide the fuel used by the days it covered and adjust the load until this page agrees with reality.
How much wood is a cord and how long does it last?
A cord is a stacked volume of 128 cubic feet, conventionally 4 by 4 by 8 feet, and it is a volume rather than a weight, which is why the energy in one varies so much. Seasoned hardwood runs somewhere around 20 million BTU per cord as a planning figure, softwoods considerably less, and unseasoned wood less again because a large part of the energy goes into boiling off its own water. The duration follows from the same arithmetic as any other fuel, but treat the answer as approximate and buy earlier than the calculation suggests, since drying takes months you cannot compress.
Can I use a portable heater or a generator as backup?
Both are used, and both have one non-negotiable condition: anything that burns fuel produces carbon monoxide and must never run indoors, in a basement or in an attached garage, including with a door or window open. That covers unvented portable heaters, generators, camp stoves and grills without exception. Working carbon monoxide alarms belong in the house before any backup appliance does. Beyond that, installation and venting of heating equipment is work for a qualified installer, and the operating instructions for a specific appliance come from its manual rather than from a calculator.