106. Read the following passage, then answer the question below.

Cities are usually warmer than the countryside around them, a difference known as the urban heat island effect. What the phrase hides is that there are two different measurements. One is surface temperature, which satellites record by sensing the heat radiated from roofs, roads, and treetops. The other is air temperature, which is measured by thermometers shielded from direct sunlight. The two do not always agree. On a summer afternoon, a dark parking lot can be thirty degrees hotter than a nearby lawn on the surface, yet the air above the two differs by only a degree or two, because wind constantly mixes it. Surface heat islands are strongest during the day, while air-temperature differences tend to peak at night, when buildings slowly release the heat they have stored. For public health, the nighttime figure matters more, since people who cannot cool off after dark are the most vulnerable in heat waves.

Planners have tried several remedies. Reflective, or 'cool', roofs send sunlight back into the sky, and they can reduce a building's energy use for air conditioning. Green roofs and street trees shade surfaces and release water vapor. Yet each has limits. A cool roof can lower the temperature of a single building, but its effect on street-level air is small unless many roofs are treated, and in winter a reflective roof may increase heating costs. Trees work well but take years to mature, and they need water, which is scarce in some of the hottest cities.

Some researchers argue that cities should focus less on average temperatures and more on the places where heat hurts the most. In many cities, the hottest neighborhoods are also those with the fewest trees and the lowest incomes, a pattern that traces back to decisions about where parks and roads were built. Targeting those areas, they say, would save more lives per dollar than a uniform citywide program, though it would raise hard questions about who decides where to invest first.

The author mentions all of the following as limits of heat-reduction measures EXCEPT that