Editor's Note: Scientific American's George Musser will be chronicling his experiences installing solar panels in 60-Second Solar. Read his introduction here and see all posts here.

After months of planning and paperwork, our solar project has finally moved from talk to action. Last week, a roofing contractor began installing a new roof. For this, I have to thank a respondent to one of my earlier posts, who saw a picture of my roof and suggested it was “in rough shape,” an intuition that roofing contractors confirmed.

They told me that although the roof was less than 10 years old and showed no signs of leaking, the asphalt shingles were badly installed and wouldn’t last the 30-year lifetime of the solar system.

I can’t say I was entirely happy to learn I’d need to shell out $8,000 for an improvement that I couldn’t show off to house guests. The estimates struck me as steep, but we were told our Italianate-style house has a funny geometry and that the built-in gutters needed replacing, too.

Then it dawned on me that I could take this as an opportunity to improve our home’s energy efficiency. The first question was whether it’s better to have a black roof that absorbs sunlight, reducing heating costs in winter but increasing air-conditioning costs in summer, or a white roof that reflects sunlight, saving on cooling at the expense of heating. Here in New Jersey, most of our energy bill goes to heating, so at first glance you might think black would be greener.

But energy experts prefer white. EnergyStar roofs, which can qualify for tax credits, are reflective. According to roofing-comparison calculators at EPA and the Oak Ridge National Laboratory, you do take a hit on your heating bill, but even at our latitude, the electric savings are two to three times the heating penalty, for a net savings. Why is that? People have put forward a number of reasons of varying plausibility:

  • Hot air rises, so even if you heat the top of the house, you still need to run the furnace to heat the lower floors. That sounds reasonable, but on the other hand, wouldn’t reducing the temperature differential between the attic and top floor help to retain heat?
  • You don’t want to heat the roof, because snow will melt and refreeze, creating ice dams. The thermal barrier of the house shouldn't be the roof but insulation in the attic floor; the attic temperature should stay the same as the outdoor temperature. That sounds reasonable, but on the other hand, the roofing calculators don’t mention anything about attic-floor insulation, so do they really take this effect into account?
  • High reflectivity often means low thermal emissivity, so a roof that reflects sunlight in summer will also have lower radiative losses in winter, helping to keep heat in the house. That doesn’t sound so reasonable to me. Reflectivity and emissivity depend on the roof material, and the roofing calculators say nothing about emissivity, anyway.
  • In winter, daylight hours are shorter and the sun is lower in the sky, so the roof absorbs less heat than it does in summer. Thus any changes to roof reflectivity have less effect in winter than in summer. To test this explanation, I downloaded data for heating and cooling degree-days from DegreeDays.net and for incident solar radiation from PV Watts. For each month of the year, I weighted the effects of a change in roof reflectivity by the heating and cooling requirements. At my latitude, solar heating is three and a half times stronger in June than in December, so a reduction in the cooling load has a proportionately greater effect than a reduction in the heating requirements does. Yet the number of heating degree-days is so overwhelming greater than the number of cooling degree-days that this effect can't be the answer, either.
  • Joule for joule, natural gas is cheaper than electricity, so it’s cheaper to heat up a degree than cool down a degree. Maybe, but it clearly depends on the temperature range and type of heating system.

Does anyone have a convincing explanation?

Although I don’t quite understand the conventional wisdom, I’ve accepted it and sought to maximize the reflectivity of our roof. It’s ironic: we started off trying to collect energy from the sun, but here’s a case where we don’t, in fact, want that.

White asphalt shingles are about twice as reflective as the kind we have now, and the EPA calculator estimates the net energy savings would be a paltry $30 a year. But when I read that metal roofs have the highest reflectivity of all, a bell went off in my head. Long ago, my wife and I came across an 1870s survey of our neighborhood that showed that our house originally had a tin roof. Metal roofs are supposed to last almost forever. Was ours still there, buried under the shingles? When I first posed this question, local contractors thought I was nuts. Even my wife thought I was dreaming.

But a tin-roof restorer in suburban Philadelphia, Miriam Cunningham of RoofMenders, agreed to drive over and take a look. By peeking through the wooden roof slats in the attic, she showed us that the tin is still there. It’s punctured with nail holes, but nothing they can’t fix. Their restoration process includes a Acrymax coating with 85-percent reflectance, which the calculators estimate will save $150 a year in energy costs. That alone doesn’t justify the extra expense, but the tin roof has other benefits, such as moderating the sweltering conditions in our attic and enhancing the historical quality of our house. So instead of getting a fancy new roof, we’re getting a fancy old one.

Raymond and Lester of RoofMenders on top of George's roof