Thermodynamics of roof color

[Hari Seldon]

A "borough" in Montreal has decreed that all new roofs be white. They argue, obviously correctly, that this will reduce solar heating in the summer. Others argue that black roofs are more effective in increasing solar heat in the winter and Montreal has a lot more winter than summer. Let me leave aside the point that all reasonably well-insulated roofs are white for nearly all the time December to March and ask if this point is even correct. The way I see it, on mid-winter day, we get about 9 hours of day and 15 of night and a black roof would absorb more during the day but also radiate more during the night and therefore a white roof is still preferable. Is this argument correct?

[CalMeacham]

Is this argument correct?

I suspect not. I haven't the time to do calculations now, but during the day you're absorbing heat from the very hot sun, and it's being communicated by conduction to the air in the house and heat it up. That's a big thermal reservoir. During the night, the roof will radiate heat away more effectively than white, but I strongly suspect that conduction to the moving air is going to be a bigger loss mechanism, and that's pretty much independent of the roof color.



Assume solar heat is all you have (not likely, but assume it) With a dark roof, you absorb lotsa heat during the day, and can afford to lose some at night. But if you have a white roof, you don't absorb very much during the day. It's not much comfort to know that you'll be re-radiating very little of very little at night.

[Una Persson]

Some crackpot named Cecil wrote a bit about this, with some madwoman named Una doing some research for him...

[Machine Elf]

If you insulate the bajeezus out of your attic and ventilate it like you're supposed to, roof color shouldn't matter much at all. Most homes fall short on both counts.

[johnpost]

new construction has a cold attic (well ventilated, floor insulated) unless the attic is made for living or heated storage. a old style (more sealed, roof insulated) attic leaves a reservoir of heat above your house that keeps your house warm in summer, if you use air conditioning then that is additional load to cool.

[Dewey Finn]

This is a brochure (in PDF format) from the US Department of Energy that advocates for cool roofs. which aren't necessarily the light-colored ones.

[Hari Seldon]

Originally Posted by Una Persson

Some crackpot named Cecil wrote a bit about this, with some madwoman named Una doing some research for him...

Sarcasm I needed. But I don't have air conditioning (although I wouldn't at all mind a cooler second floor in summer) and I sure use plenty of heat all winter. On the other hand, as I mentioned I have a white roof anyway for at least four months in the winter. Still, my roof is only ten years old and will doubtless outlive me.

[scr4]

Originally Posted by Hari Seldon

...a black roof would absorb more during the day but also radiate more during the night ...

The second half of that sentence is necessarily true.

The radiation from the sun peaks in the visible light, so the color of a surface in visible light determines how much radiation it absorbs. That is, if it looks black to the human eye, it absorbs heat from the sun pretty well.

But the roof does not emit visible light. It emits infrared light. And the ability to radiate infrared (called "emissivity") is related to the color in infrared, which may be different from its color in visible light. A surface may look white in visible light but black in the infrared.

If we're talking about painted surfaces or ceramic roof tiles, I think that black vs. white surfaces both have similar, high emissivity. So the radiative heat loss is not very different. A black roof would therefore be preferable in winter.

I think for overall efficiency, a shiny metal roof (e.g. stainless steel) would be the best year-round. Metal surfaces have very small emissivity, at least when they're clean and not oxidized. So it would have the least radiative heat loss in winter (day and night), and minimal heat absorption during the day.

(Personally, I think the best solution is to plant a big deciduous tree on the south side of the house, but that's just me...)

[Napier]

Ditto scr4, except the bit about black and white surfaces both having high emissivity. I don't know if that's true. It's likelier true for black, which is usually made with carbon black, which I know does have high emissivity.

People are a good example. We all have emissivity of about 0.98 +/- 0.01, which is very black, regardless of skin color in the visible.

One other bit - shiny metals have low emissivity, but it is a surprise how shiny they are in the thermal infrared. Metals we think of as very dull in the visible often look quite shiny there. Similarly, things we think are quite shiny are often matte and dull in the ultraviolet. It's because waves reflect pretty well off of things having roughnesses smaller than the waves, which in the case of the thermal infrared includes many more metal surfaces.