White roofs, London summers and a temperature claim

In London, the temperature claim attached to a white roof can refer to several different places: the roof surface, the room beneath it or the air across a modeled city. Moving between those scales is where an appealing idea can turn into an overstated promise. Reflectivity changes the energy reaching a surface. Insulation, weather, building use and the wider urban fabric help decide what that change delivers elsewhere.

Think of an imagined roof through a hot afternoon, a cold season and several years of weathering. Then step back far enough to see the neighborhood around it. The same material participates in all these scenes, but the question being asked changes.

Afternoon: the roof is not the room

Solar reflectance describes the fraction of incoming sunlight reflected by a surface. Thermal emittance concerns its ability to release heat through radiation. Both matter in understanding how a roof behaves. A visible color can be informative, but it is not a complete measurement of these properties.

The Cool Roof Rating Council’s explanation distinguishes those two characteristics and describes testing at initial condition and after three years of outdoor exposure. This is more specific than treating every pale surface as the same technology. It also leaves room for roofing products whose radiative properties cannot be reliably inferred from a casual glance.

Now follow the heat toward the occupied space. The insulation and construction between surface and room affect the transfer. The building’s other gains and its ventilation or cooling affect the room’s conditions. A top-floor room directly below the roof presents a different situation from a lower floor separated by several occupied levels.

For this reason, a large surface-temperature difference can coexist with a much smaller indoor-air change. That is not evidence that the surface measurement is false. It is evidence that the building is a system, with other paths and controls determining the indoor result. The comparison needs to measure the outcome being claimed.

An air-conditioned building introduces another distinction. A roof that reduces cooling demand may allow equipment to work less while the indoor setpoint remains unchanged. The benefit might appear in energy use rather than a lower room temperature. In a building without active cooling, the assessment may instead focus on indoor conditions. Advertising should not borrow the best outcome from both cases without identifying which one applies.

The EPA’s cool-roof account describes potential cooling and heat-island benefits alongside climatic limitations. It is guidance about an intervention, not a certificate that a particular building will obtain a stated saving. The building, baseline roof and local conditions still have to do the work in the assessment.

Winter: an annual claim needs an annual account

Sunlight that is unwanted during a cooling period can contribute heat during another part of the year. A more reflective roof can therefore reduce useful solar gain in a heating season. Whether this produces a significant annual penalty depends on the climate, roof assembly, operating schedule and heating system, among other factors.

This does not justify a simple reversal of the sales pitch—“white roofs are bad in cold places”—any more than summer benefits justify specifying them everywhere. It calls for an appropriate annual comparison. A hot-afternoon experiment answers a hot-afternoon question. Annual energy, cost and emissions results require the rest of the year and the relevant equipment and energy sources.

The baseline should be realistic too. If the existing roof needs replacement, the decision may be between feasible new assemblies rather than between a new reflective product and an imaginary roof that lasts forever at no cost. If only a coating is proposed, compatibility and the condition of the existing roof become important. These are different projects sharing an attractive color.

An owner also has to distinguish a product’s price from the full work. Preparation, access, installation, drainage and other necessary repairs can influence the actual cost. A program that reports only the coating price may understate what a household or building manager must arrange to obtain a dependable roof.

Timing can change the practical case without changing the physics. Combining a suitable reflective finish with scheduled roof work may be easier than arranging a separate intervention. Conversely, a building experiencing severe summer discomfort may need other actions sooner than its next roof replacement. The roof should fit the building’s adaptation plan rather than become a reason to delay every other measure.

Year three: the label needs a history

A new roof is not the same surface after exposure to weather, dirt and use. Performance should therefore be evaluated beyond the clean sample. That is why aged properties matter in a specification and why a bright installation photograph is an incomplete long-term record.

The CRRC Roof Rating Program publishes radiative-performance ratings; it explicitly says that being listed does not mean a product is “cool” under every code or program. Its process includes three-year weathering and aged testing, with interim ratings for qualifying products. A rating supplies data. The applicable standard and building assessment determine how those data are used.

Its product directory also distinguishes interim laboratory-aged values from measured three-year values and notes that actual building effects vary. This means a purchaser or analyst needs to know which value is being compared. Two neat numbers in a table may have different status, and a later measured value can replace an earlier estimate.

Maintenance belongs in the comparison as well. The Department of Energy’s Guidelines for Selecting Cool Roofs discusses soiling, cleaning and moisture considerations. It recommends using weathered reflectance when estimating savings and warns that cleaning methods can damage a roof if inappropriate. This is guidance on maintaining a roof system, not an instruction to perform unassessed work on an occupied building.

A sensible program would therefore record the installed product, its relevant rating, the condition of the assembly and the maintenance arrangement. It would avoid treating reduced reflectance as an inexplicable surprise after assuming the clean value for every year. The service being purchased is a functioning roof with a useful thermal characteristic, not a permanently immaculate test sample.

Moisture also needs a site-specific design. A surface that changes temperature can change drying conditions, but that does not make every reflective roof a moisture failure. The complete assembly, climate and detailing matter. Broad warnings and broad reassurance are equally poor substitutes for a competent assessment of the roof actually being proposed.

Step back: London is a modeled city, not a painted experiment

City-scale cooling is a separate line of evidence. A 2024 study in Geophysical Research Letters compared roof and vegetation interventions using urban-climate modeling for London during two hot days in 2018. Its citywide scenarios investigate what extensive deployment could do under specified conditions. They do not report an observed temperature drop after every London roof was coated.

A related Nature Cities study, available through UCL’s research repository, extended analysis across the summer of 2018 and estimated an average near-surface air-temperature change of about −0.8°C under its cool-roof scenario. This is a modeled scenario result for that city, period and deployment assumption, not a guaranteed reduction for each street or an indoor-temperature forecast for an individual home.

The two results become easier to read side by side:

Study scenario	Period modeled	Reported outdoor-air result
GRL extensive cool-roof deployment	26–27 July 2018	About 1.2°C average cooling over two days; about 2°C in some locations
Nature Cities all-roof cool-roof scenario	June–August 2018	About 0.8°C lower mean near-surface temperature

These are modeled scenarios with different study designs and averaging windows. Selecting the larger number without its time and spatial basis would change the claim. Neither row measures a particular bedroom or reports a completed citywide coating program.

The Nature Cities work also estimated about 0.3°C mean cooling in its all-roof photovoltaic scenario, alongside electricity generation. Its cooler-roof result therefore does not settle the entire rooftop decision: the photovoltaic option supplies an additional service. A public program comparing the two needs to state whether it is optimizing outdoor heat, electricity, building operation or a feasible combination.

A city average also leaves a distribution. Different neighborhoods and times can experience different outcomes. A program aimed at heat-exposed residents should examine whether the proposed roof stock, implementation route and local conditions align with that aim. A positive average is not the same as evidence that the most exposed households receive the largest benefit.

The comparison with trees or solar panels needs similar care. A model can compare particular scenarios under particular metrics. It does not establish that one intervention replaces all the functions of the others. A roof, a shaded walking route, habitat and electricity generation provide overlapping but different services. Choosing among them requires a statement of the objective, not simply a winner’s label from one model.

The roof has neighbors

Reflecting sunlight changes where energy goes. In a dense built setting, surrounding geometry and materials can matter, including potential reflections toward other surfaces or occupied places. The Department of Energy’s building-envelope guidance explicitly includes nearby surface reflectance and the possibility of unwanted reflections and glare in daylighting design.

That is another reason to review a proposed intervention in place. A roof out of view above a low-rise building is a different spatial problem from a surface near taller neighboring windows. The design should establish the relevant interactions rather than assume every reflected ray immediately leaves the urban environment.

There are other roof uses to consider. Equipment, access, drainage, vegetation and solar installations can all shape what is practical. A proposal should compare feasible combinations and their purposes instead of allocating the same area independently to several attractive plans. A rooftop is finite infrastructure, not a blank layer that every strategy can claim in full.

For public programs, this means selecting buildings with an explicit rationale. A climate-appropriate intervention, a sound roof and a clear route for installation and care can produce a credible demonstration. Counting treated square meters alone tells less about whether the program reached its intended residents or improved their conditions.

Specify the outcome before choosing the color

The imagined roof now has three assessments. Its surface performance is supported by relevant product data and condition. Its building effect depends on the assembly, use and annual climate. Its urban effect belongs to a wider deployment and modeling question. Keeping the assessments separate makes the combined case stronger where the evidence supports it.

It also makes a poor fit easier to recognize. A favorable surface rating cannot repair a failing assembly, a difficult ventilation strategy or an inaccessible maintenance plan. A city-scale cooling estimate cannot be placed directly into a household’s savings forecast. The roof may still be useful, but the reason needs to be the right one.

For evaluation after installation, record the relevant baseline and follow the outcome that justified the work. If the aim was cooling-energy reduction, account for weather and operating changes before attributing the whole bill difference to the roof. If the aim was better indoor conditions, measure the rooms and periods that motivated it.

White is a clear visual signal. It should not become a substitute for specifying the service: less unwanted heat, an appropriate annual energy result, a durable roof and a contribution to a locally justified plan. A reflective roof earns its place through those outcomes, one scale at a time.