Cooling energy savings potential of light-colored roofs for residential and commercial buildings in 11 US metropolitan areas

The U.S. Environmental Protection Agency (EPA) sponsored this project to estimate potential energy and monetary savings resulting from the implementation of light-colored roofs on residential and commercial buildings in major U.S. metropolitan areas. Light-colored roofs reflect more sunlight than dark roofs, so they keep buildings cooler and reduce air-conditioning demand. Typically, rooftops in the United States are dark, and thus there is a potential for saving energy and money by changing to reflective roofs. Naturally, the expected savings are higher in southern, sunny, and cloudless climates. In this study, we make quantitative estimates of reduction in peak power demand and annual cooling electricity use that would result from increasing the reflectivity of the roofs. Since light-colored roofs also reflect heat in the winter, the estimates of annual electricity savings are a net value corrected for the increased wintertime energy use. Savings estimates only include direct reduction in building energy use and do not account for the indirect benefit that would also occur from the reduction in ambient temperature, i.e. a reduction in the heat island effect.

This analysis is based on simulations of building energy use, using the DOE-2 building energy simulation program. Our methodology starts with specifying 11 prototypical buildings: single-family residential (old and new), office (old and new), retail store (old and new), school (primary and secondary), health (hospital and nursing home), and grocery store. Most prototypes are simulated with two heating systems: gas furnace and heat pumps. We then perform DOE-2 simulations of the prototypical buildings, with light and dark roofs, in a variety of climates and obtain estimates of the energy use for air conditioning and heating.

We proceed from the estimates of savings in individual buildings to the entire MSA, by calculating how much energy and money could be saved if the current building stock had its roofs changed from dark to light. This is done by scaling the simulated energy savings of the prototype buildings by the amount of air-conditioned space immediately beneath roofs in an entire MSA. For this, we use data in each MSA on the stock of commercia1 and residential buildings, the saturation of heating and cooling systems, the current roof reflectivities, and the local costs of electricity and gas.

The estimates of the direct savings are shown in Table EX-1. The largest potential for net annual dollar savings was found in Phoenix, $37 million ($37M), followed by Los Angeles ($35M), Houston ($27M), Miami/Fort Lauderdale ($20M), DallasFort Worth ($20M), New York City ($16M), Chicago ($ 10M), New Orleans ($9M), Atlanta ($9M), Washington, DC/Baltimore ($8M), and Philadelphia ($3M). The same quantities per 1000ft² of roof area of air-conditioned buildings for each MSA are shown in Table EX-2. To illustrate the climate effect, the results are plotted in Figures EX-1 to EX-4, superimposed on a map of the United States with contours of annual cooling hours for a typical residential building. The data per 1000ft² of roof area reflect the effects of climate, whereas the MSA savings are strongly affected by the sizes of the populations.

The sum total for all 11 MSAs are: electricity savings, 2.6 terawatt hours (TWh) (200 kilowatt hours per 1000ft² of roof area of air-conditioned buildings); natural gas deficit, 6.9 TBtu (5 therms per 1000ft²); net savings in energy bills, $194M ($15 per 1000ft²); and savings in peak demand 1.7 gigawatt (GW) (135 W per 1000ft²). Six building types account for over 90% of the annual electricity and net dollar savings: old residences more than 5596, new residences about 15%, and four other building types (old/new offices and old/new retail stores) together about 25%.

The results for the 11 MSAs are extrapolated to estimate the savings in the entire United States. This extrapolation is done first by scaling to the national population, and then by a method that accounts for the climatic variations of the savings. We find that the national savings are about four times the savings for the 11 MSAs: a decrease in annual direct electricity use by 9.3 to 11TWh (about 3.0% of the national cooling electricity use in residential and commercial buildings), an increase in natural gas use by 25 to 28 GBtu (1.6%), decrease peak electrical demand by 6.2 to 7.2 GW (2.5%) (equivalent to 12 to 14 power plants each with a capacity of 0.5 GW), and a decrease in net energy bills for the rate-payers by $680M to $850M.