Why an Air Mattress Sleeps Colder Than a Foam Mat

This page contains affiliate links. If you buy through one we may earn a commission, at no cost to you.

Sleeping on the ground is not better. What happened is that your air mattress was not insulation, and the premise worth correcting is the one underneath the question: that a thick layer of air between you and the ground must be warmer than a thin one. It is not, and the reason is specific. The dominant way heat leaves you through an open air chamber is radiation, straight across the gap from the warm upper skin to the cold lower one, and radiative exchange between two parallel surfaces does not depend on how far apart they are.

That is not a rough statement. Put a 2 inch chamber and an 8 inch chamber side by side with the same plastic skins and the radiation term is identical in both, at about 114 watts per square metre. Only the conduction through the still air changes, from 12.8 watts per square metre to 3.2. Totals of 127 and 117. Four times the thickness, an eight percent improvement.

Where the heat actually goes

Three paths cross the chamber, and the sizes are nothing like the intuition.

Radiation first, because it is the big one. Two surfaces facing each other across a gap exchange heat according to the Stefan Boltzmann law, scaled by an effective emissivity. Plastic films are not reflective: PVC, TPU and nylon all sit around 0.9 on the emissivity scale, which is close to the 1.0 of a perfect black body. Put two such surfaces facing each other and the effective emissivity is one over the quantity one over 0.9 plus one over 0.9 minus one, which is 0.818. Take a top skin under a sleeper at 30 degrees Celsius, 303 kelvin, and a bottom skin against ground at 5 degrees, 278 kelvin. Fourth powers: 303 to the fourth is 8.43 times ten to the ninth, 278 to the fourth is 5.97 times ten to the ninth, and the difference is 2.46 times ten to the ninth. Multiply by 0.818 and by the Stefan Boltzmann constant of 5.67 times ten to the minus eight, and you get 114 watts per square metre.

Conduction second. Still air has a thermal conductivity of about 0.026 watts per metre kelvin, so across 10 centimetres with 25 degrees of difference you get 0.026 times 25 divided by 0.10, which is 6.5 watts per square metre. One seventeenth of the radiation.

And convection third, which is where a correction is owed, including to the brief this article was written from. The usual story is that warm air rises off your body, cools against the ground and circulates back, so the chamber stirs itself. That cannot happen in this geometry. You are the warm surface and you are on top. The ground is the cold surface and it is underneath. Warm air above cold air is stably stratified, and buoyancy holds it there: the hot fluid is already where it wants to be. Rayleigh and Benard's instability, the one that makes a fluid layer overturn, requires heating from below, which is the arrangement in a saucepan and the opposite of a sleeping pad. So in the ideal case convection is suppressed rather than driven.

What does stir an air chamber in practice is the edges. Your body does not cover the whole top surface, so the uncovered parts of the top skin are cold while the part under you is warm, and a horizontal temperature difference along a boundary drives circulation whatever the stratification does. The side walls are cold too, so air sliding down them feeds inward along the floor. Those effects are real, they are why the edges of an air bed feel coldest, and they are still second order next to 114 watts per square metre of radiation crossing the middle.

The 8 inch air bed and the 1 centimetre mat

Take the thickest thing we carry. A coldpa double camping air mattress with a built in foot pump publishes 8 inches of thickness, a 84.6 by 51.5 inch surface, nylon with a TPU coating, coil beam construction and a flocked waterproof top. Run the arithmetic above at 20.3 centimetres and you get 114 plus 3.2, or 117 watts per square metre, which is a resistance of 0.213 in SI units and 1.21 in the ft2 F h per BTU units that camping R-values are published in.

Now a sheet of closed cell foam one centimetre thick, at a thermal conductivity of 0.035 watts per metre kelvin. Resistance equals thickness over conductivity, which is 0.286 SI, or 1.62. The foam sliver is 34 percent better than the 8 inch bed at one twentieth of the height, and 2 centimetres of it is 3.24, which is almost three times the air bed. That is the answer to the question in one line.

What closed cell foam has that an air chamber does not is interruption. Every sealed cell wall absorbs radiation and re-emits it, so instead of one clean 20 centimetre jump from skin to skin the energy makes thousands of hops of a fraction of a millimetre each, and the whole transfer collapses into an effective conduction with a very short mean free path. The gas inside the cells cannot circulate either, because the cells are sealed, which is what the words closed cell mean and why the distinction from open cell foam is a thermal one rather than a texture one.

Two things in that coldpa listing deserve fair treatment. The flocked top is genuinely worth something: a short pile fabric traps a thin layer of still air at the contact surface and raises the surface's effective emissivity path a little, so it is not marketing. It is worth millimetres of equivalent foam rather than inches. And the record's own cons are candid that inflatables can soften overnight and that durability claims are maker stated. What the record does not do, and what no air mattress listing does, is publish an R-value, which is the figure that would settle this. Details in our coldpa double air mattress review.

The one feature on an air pad that attacks the right term

If radiation is 94 percent of the loss, the only thing worth paying for is something that lowers an emissivity, and exactly one product category does it: the reflective internal barrier.

A CKG ultralight camping sleeping pad, 4 inches thick with a heat reflective thermal barrier is built around that idea. Our own record describes the feature as one that "reflects body heat and cuts the heat loss into cold ground that makes you cold", and that is correct in mechanism, which is more than most feature copy manages. Here is what it is worth in numbers. Replace one of the two 0.9 emissivity skins with a metallized film and the effective emissivity collapses, because the pair is dominated by the better surface. A film at 0.1 gives an effective 0.099 and drops the radiation from 114 to 13.8 watts per square metre. Add the 6.4 of conduction at 4 inches and the total is 20.2, which is a resistance of 7.04. A film at 0.05 would give 10.7.

And this is where the honesty has to come in, because that result is too good. Real reflective air pads are tested at R-values of roughly two to four, not seven to eleven. So the film in a real product is doing well under half what an ideal shield would. The reasons are all mundane and all worth knowing before you pay for the feature: a film laminated between plastic layers faces its own plastic rather than the chamber, a film in contact with the opposite surface conducts around the gap instead of blocking it, printed or patterned metallic layers are not continuous mirrors, the welds that hold the pad together bridge straight through, and the perimeter is a short circuit all the way round. None of those is visible from a listing.

Which is why one sentence in our own sleeping pads guide is the best advice on the site about this, and we should say so rather than improve it: "If a pad states no R-value and mentions only a thermal or reflective barrier, treat it as a helpful three-season pad, not a guaranteed cold-weather one." That is exactly right, and the arithmetic above is the reason. A barrier attacks the dominant term, so it helps a great deal, and nothing about naming it tells you how much of the ideal it achieves. The CKG listing publishes the barrier, the 4 inch loft, 50D polyester, 1.92 pounds and a 13.8 inch pack size, and no R-value. Our record's own cons say it is "not for deep winter without a high R-value", which is the correct place to leave it. See our CKG sleeping pad review, and read the R-value guide for what number to want for which night, because this page does not re-argue that.

Drop stitch does nothing thermal, and the arithmetic says how little

This is the second correction, and it goes against the idea the article was commissioned on, so it is worth doing in full.

The appealing story is that drop stitch and internal baffles subdivide one big chamber into many small ones and thereby cut the loss. Test it. Drop stitch is thousands of fine threads running from the top skin to the bottom one. Take a typical density of around ten threads per square centimetre at perhaps 0.3 millimetres diameter: that is a hundred thousand threads per square metre with a combined cross sectional area of 0.7 percent of the surface. Polyester conducts at about 0.14 watts per metre kelvin, so the heat those threads carry is 0.007 times 0.14 times 25 divided by 0.10, which is a quarter of a watt per square metre. Set that next to 114 and it is noise. The threads do not hurt.

They do not help either, and this is the part the subdivision story gets wrong. A radiation shield works because it is a surface that intercepts rays. A thread is a thin cylinder, and it blocks radiation only across its own narrow silhouette, which is the same 0.7 percent of the area. The air between the threads is continuous from top to bottom, so nothing is subdivided in the direction that matters.

Baffles fail the same test for a cleaner reason: they run the wrong way. A baffle or a coil beam is a vertical partition through the thickness of the pad. The radiation path that matters is vertical, from the top skin down to the bottom one. A vertical partition is parallel to that path and never crosses it. Baffles stop air moving sideways and stop the pad ballooning into a sausage, and that is their job.

So drop stitch is a structural technology, not a thermal one. It exists because threads let a flat panel hold high pressure without bulging into a cylinder, which is what gives a firm, level surface. A Lanryn drop stitch camping sleeping pad with an air pump publishes exactly the specification that confirms it: high elastic drop stitch construction, a load rating of up to 700 kilograms, PVC free fabric, a dual action valve setting firmness between 80 and 95 percent, and a stated 100 hours of air retention. Every one of those is a mechanical figure. Our own record then does something unusually honest for this category: its cons field states, in as many words, that "cold ground may still need extra insulation". That is the correct conclusion, and the arithmetic on this page is why. Details in our Lanryn drop stitch pad review.

Our guide to choosing a camping sleeping pad describes drop stitch as internal threads giving "firm, level support that resists bottoming out, which suits rough ground and heavier sleepers", and that is accurate. It is a structural claim and it stays a structural claim. Nothing there needs correcting; what was missing is anyone saying out loud that the thermal column of the ledger is blank.

The honest limit

Not one of the three pads above publishes an R-value, and the ASTM F3340 test that produces one is the only figure that would let you compare them. My 114 watts per square metre rests on an assumed emissivity of 0.9 for the skins and an assumed 25 degree difference, and both are reasonable rather than measured. The implied 1.21 for an 8 inch air bed happens to sit close to where real uninsulated air mattresses test, which is a reassurance about the method and not a measurement of that product.

And the free fix beats every purchase here, because the dominant term is a surface property and you already own surfaces. Put anything opaque and fibrous between you and the air bed, and you have inserted the interception that the chamber lacks. A folded blanket, a closed cell foam mat, a foil backed windscreen shade, even the clothes you are not wearing, laid flat on top of the mattress rather than over you. All of it works, and it works upward of what its thickness suggests, because you are not adding resistance in series with a good insulator, you are breaking the single long radiation jump that was carrying almost all of the loss.

Three practical consequences fall out of that. A blanket is worth far more under you than over you, because above you it competes with a sleeping bag that is already working and below you it replaces insulation your body weight has crushed flat. Thickness is the wrong thing to shop for on an air pad and emissivity is the right thing, which unfortunately nobody publishes. And if you are car camping and cold, the real upgrade is not a thicker air bed. It is a thin foam mat to put on top of the one you have.