The two researchers, Mark Z. Jacobson and John E. Ten Hoeve, developed a model that "meshed data on land use, vegetation, albedo (the reflective capacity of different land uses) and soil-type. Then, they ran two 20-year-long simulations to see how much heat islands contributed to "gross global warming" (warming before cooling factors) – and what impact a lot of white paint might have."
In the end, the study showed how the urban heat island effect may only contribute 2-4 percent of gross global warming.
White Roofs on the other hand could actually warm the Earth on a global scale. "Although white surfaces are cooler, the increased sunlight they reflect back into the atmosphere by can increase absorption of light by dark pollutants such as black carbon, which increases heating."