The data and maps explained

Where our data came from
We used data from US Air Force weather satellites to create our new light pollution maps. These have light detecting instruments which are used to assess cloud cover at night, by picking up moonlight reflected from clouds. But these instruments — operational line scanners — can also detect light from towns and cities, fires, gas flares and heavily lit fishing boats when there is no cloud cover. Scientists from the US Federal Government's National Oceanographic and Atmospheric Administration have developed ways of using the stream of data from these scanners to measure the total brightness of artificial nighttime lights within small areas of the earth's surface.

The scanners make repeated measurements of light beaming upwards, capturing the rays from various angles as it passes overhead dozens of times. Abnormal, transient brightness, such as might be caused by a large fire, is filtered out in order to provide a reliable estimate of the amount of artificial light normally beaming upwards.

How we made the maps
We divided the land surface into small squares, or pixels, each representing a square kilometre. We gave each pixel a value, according to the accumulated brightness of the light within. The light is measured on a range from 0 to 255; 0 means the satellite is detecting no light in that pixel and 255 means the satellite's detector is saturated with light.

We divided the pixels into five bands, each with a different colour:

The red areas cover the largest towns and cities. Around and between them are yellow areas, covering suburbs, smaller towns and brightly lit highways. Light pollution diminishes the darker the shade of blue.

In the red and yellow areas, there is no chance of seeing the Milky Way — our own galaxy — on even the darkest, clearest night.

In our maps, the most spectacular growth appears to be in the blue and bright blue areas, where low to medium levels of light pollution now intrude.