Heat rises. Everyone knows that, right? It's absolutely true. Heat does rise. The problem is that sometimes people say this as if the flow of heat is driven by its wanting to rise. It's not. Heat can move up, down, or sideways, depending on the situation. What the laws of thermodynamics tell us is that heat moves from areas of higher temperature to areas of lower temperature. Put a torch to the top of a steel pole, and heat will travel downward by conduction. So, temperature difference is really what drives heat to move in any given direction.
When you're dealing with fluids, you have to account for density and buoyancy as well. Air is the fluid we live in, and this time of year we spend a lot of money pumping heat into it in our homes and workplaces. When we heat air, the molecules jiggle and zip around faster, which causes them to spread out. When a mass of air takes up more space, it has a lower density. When you have a lower density fluid immersed in a higher density fluid, the lower density fluid rises and the higher density fluid falls.
Think of air bubbles in water, as shown in the photo above. Think of a helium balloon. Think of a hot air balloon. Now, imagine an object with higher density immersed in a fluid. Put Wile E. Coyote's anvil in the air above his head, and it turns him into a pancake.
The point here is that it's easy to get confused by heat in the building science of air movement. Warm air rises when it's surrounded by cold air because of its lower density. Yes, that's due to heat, but density is the main factor causing the movement here. The name for this phenomenon is stack effect. Two factors affect how much stack effect a building experiences:
- Temperature difference between inside and out (because density depends on temperature)
- Height of the building
The problem with stack effect in buildings is that buildings aren't vacuum chambers. They leak. Obviously, a house isn't going to start floating up into the air like a balloon (although I recall with great fondness the Disney movies of my childhood that showed such magical events). But the low density air inside the house will move up and out into the cold, dense winter air when given the chance.
Try this experiment if you don't believe me. Open your pull-down stairs or scuttle hole to the attic on a cold day when your home is warm. Climb up into the attic and then put your face over the hole. You'll feel the stack effect pushing lots of warm air into the attic.
So, in winter, the warm, low density air inside your house wants to rise...if it can. If your house has no leaks, the warm air can't escape and do its thing. There's still a pressure difference across the building envelope, but that's OK if the air barrier's good. Positive pressure inside the house with nowhere to go because there are no pathways.
What happens in reality is that homes leak. Your nice warm air finds way to leak out (exfiltration) and cold air leaks in (infiltration). Because of its lower density, the warm air will leak out the top of the house if there are leaks there. When a cubic foot leaks out, however, it has to be made up by a cubic foot leaking in. As the warm air leaks out at the top, cold air leaks in at the bottom. The leakier your house is, the more temperature difference you'll notice between the top and bottom of the house.
All this happens because the warm air inside your home in winter is less dense than the cold air outside. In summer, the dense air is inside your home because that's where the temperature is lower, especially if you're air conditioning your home. What that means is that leaks in your house bring warm air in at the top and allow cool air to fall out at the bottom.
Ah, warm air falls! Heat sinks. That old expression, "Heat rises," is not a basic truth after all. As with many aspects of building science, you have to look at the full context to understand what's going on.
Rats to You, Daniel Bernoulli! - Understanding Air Pressure (with a cool video!)
Photo of water bubbles by Christian Haugen from flickr.com, used under a Creative Commons license.