Stove Fuel#

Estimate how much canister gas to carry from the water you need, how much comes from snow, your stove system, altitude range and reserve.

How it works#

Everything follows from the energy needed to get water to a boil, and how much of the fuel’s energy actually ends up in the pot.

Energy required#

Liquid water only needs heating. Snow additionally has to be melted first, which is where most winter fuel goes:

where m is the mass of water in kg (1 L ≈ 1 kg), c_w = 4.186 kJ/(kg·K) is the specific heat of liquid water, L_f = 334 kJ/kg is the latent heat of fusion of ice, T₀ = 5 °C is the assumed liquid-water starting temperature, and T_ref = 100 °C. Snow is still treated as 0→100 °C. The endpoint remains 100 °C at every altitude so the calculator never recommends less fuel merely because water boils sooner on a high mountain.

That latent-heat term explains much of the extra fuel needed in winter. Melting a kilogram of snow and heating it to the reference temperature takes about 1.8 times the energy of heating a kilogram of water that was already liquid at 0 °C, and over twice that of 15 °C stream water. Mountaineers have noticed this for a long time — it is sometimes called the Shipton rule, the observation that melting the ice takes about as long again as heating the resulting water.

Stove efficiency presets#

Efficiency is the one number in the model that cannot be derived from physics, so it has to come from measurement. Manufacturers do not publish thermal efficiency, but most publish how many litres a canister will boil, and that can be turned back into an efficiency. Assuming the usual 1 L from 20 °C to 100 °C at sea level:

Two of these are capped rather than used as published. Jetboil’s MiniMo figure implies 88% thermal efficiency, which is not a credible number for any real cooking system and almost certainly reflects a warmer start, a lower boil endpoint, or generous rounding. The Flash’s implied 73% sits at the very top of what independent testing supports. Both are pulled back to a conservative 60%.

The generic entries are guided by independent stove-efficiency testing rather than manufacturer claims. A published stove-efficiency testing protocol weighs fuel to 0.1 g against a measured temperature rise, and its upright-canister examples show why real-world stove performance varies widely. A separate heat-exchanger pot test shows why integrated systems can do better: heat-exchanger geometry can reduce the fuel penalty versus a plain pot.

Treat the single figures as nominal. A more honest statement of each is a band: Reactor 58–75%, WindBurner 50–70%, integrated Jetboil-type 50–75%, upright canister with a good windscreen 38–58%, bare upright burner 30–50%. The calculator shows one number to stay readable, but the real uncertainty is roughly ±10 percentage points.

Why only some stoves have a windscreen option#

Wind is not a global input in this calculator, but it is not ignored either — it is built into the presets for the two designs where it dominates.

An exposed upright burner like the PocketRocket 2 holds the pot on supports above an open flame, with nothing between the burner and a crosswind. A windscreen is by far the largest efficiency lever available to it. Manufacturer boil figures are still-air bench results, so the published value is used for the windscreened case and derated about 20% for the bare case.

An integrated system like the Reactor, WindBurner or a Jetboil encloses the burner inside the pot’s heat-exchanger skirt, so it is already shielded. Side-by-side wind testing found the WindBurner showed no discernible performance loss in gusts forecast around 56 km/h — conditions in which an exposed-burner Jetboil Sol repeatedly failed to boil at all. Offering a windscreen toggle for these would be inventing precision that does not exist, so the calculator does not.

Boiling point at altitude#

Air pressure falls with altitude, and water boils when its vapour pressure matches the ambient pressure — so it boils cooler the higher you go. Pressure comes from the International Standard Atmosphere barometric formula (the same one behind the Altitude & Oxygen calculator), and the boiling point from the Antoine equation:

with P in mmHg, T_b in °C, and Stull’s coefficients for water A = 8.07131, B = 1730.63, C = 233.426. That gives 100 °C at sea level, 90 °C at 3000 m and 80 °C at 6000 m.

The boiling point is shown as useful context, but it does not reduce the fuel estimate. In strict thermodynamic terms a lower boiling point makes a boil cheaper. In expedition practice that saving is overwhelmed by colder equipment and surroundings, poorer canister output, greater heat loss, wind exposure, and less oxygen per litre of air available to the burner. A pot reaching 80 °C at 6000 m is also not equivalent to a 100 °C sea-level boil for cooking.

High-altitude allowance#

There is no published universal efficiency curve for backpacking stoves at altitude, so the calculator uses a deliberately simple planning heuristic:

The dropdown uses broad altitude ranges rather than an exact-height input because this is only a planning heuristic, not a combustion-efficiency measurement. It adds 5% per 1000 m band from the 2000–2999 m range upward, capped at 35% for >8000 m. This is a combined operating allowance for the conditions that predictably accompany altitude: falling air and oxygen density, colder stove and pot hardware, reduced canister output as the fuel gets colder, and greater heat loss. The progression is intentionally modest at trekking altitudes and conservative on 6000–8000 m expeditions.

This treatment is consistent with practical fuel-planning guidance: start from expected stove consumption, then allow for elevation, cold and reserve rather than assuming bench-test performance. One published fuel-planning approach lays out that field method in detail. MSR’s explanation of canister fuels in cold and at altitude describes why canister pressure and fuel composition make the real result system- and temperature-dependent.

Fuel required#

The reference energy is divided by the useful fraction of the fuel’s heat, then multiplied by the altitude allowance and the chosen safety reserve:

where η is the thermal efficiency of the stove system, f_h is the high-altitude allowance, r is the safety margin, and LHV ≈ 45.8 MJ/kg is the estimated lower heating value for an 80/20 isobutane/propane blend — the composition MSR publishes for IsoPro. Lower heating value is the right choice here because a camping stove vents its exhaust to the air and never recovers the heat of condensation from the water vapour it produces.

References: water and ice properties from the NIST/IAPWS steam tables; fuel heating values from the US DOE Alternative Fuels Data Center and the IPCC default net calorific values; atmospheric model from the U.S. Standard Atmosphere, 1976.

Limitations#

• Liquid water is assumed to start at 5 °C; snow is assumed to start at 0 °C. Snow temperature below freezing is not modelled. Snow at −20 °C needs roughly 40 kJ/kg more to reach freezing point, about 5% on top of the melt-and-boil total.
• The default assumes 80% of water is melted from snow. Some liquid water is usually available on approaches, but this calculator is intended for alpinists using it in the mountains, where melting snow is often the main water source. Adjust the snow percentage to match the specific trip.
• The altitude allowance is a heuristic, not a measured efficiency curve. Its +5% per 1000 m above 2000 m combines several effects that cannot be separated reliably in the field: thinner air, cold equipment, weaker fuel delivery from a cold canister, and heat loss. A warm, sheltered Reactor may beat it; an exposed upright stove in a storm may use far more. Do not add another generic altitude percentage on top of it.
• Wind is only modelled through the windscreen presets — and in the field it may well be the largest single factor. An exposed burner in moderate wind can use 1.3–2.5 times the fuel, and in strong wind may never reach a boil at all, at which point no multiplier means anything.
• Cold attacks the canister, not just the water. Isobutane boils at −12 °C and n-butane at around −0.5 °C, so vapour pressure collapses as things get cold, and running the stove chills the canister further still. Propane boils off preferentially, so a half-used canister performs worse in the cold than a fresh one. This “canister fade” is a failure mode, not a quantity problem — carrying more fuel does not fix it. Keep canisters in your sleeping bag, and consider a stove with a pressure regulator or an invertible remote canister.
• Efficiency belongs to the system, not the burner. Pot diameter, lid, heat-exchanger geometry and how hard you run the flame can matter more than the badge on the stove. Running at full throttle is usually less efficient than a moderate setting.
• Manufacturer figures are marketing. “Litres boiled per canister” rarely states the starting temperature, elevation, flame setting, or what counted as a boil. They are used here because they are the only per-model data available, not because they are rigorous.
• Only boiling is modelled. Simmering, priming, relighting, spilled pots and long idle time all burn fuel and are covered solely by the safety margin.
• Snow density does not change the energy — a kilogram of water is a kilogram of water — but it changes everything about the experience. Dry powder means many more pot refills, more handling and more lost heat than dense spring snow.
• Weigh your own canister. The most useful thing you can do is weigh one before and after a representative trip. Your own measured grams per day beats any model on this page.
• This is a planning estimate, not a safety guarantee. On a serious expedition, follow route-specific guidance and carry a genuine reserve.