The data was recorded by the Defense Meteorological Satellite Program (DMSP) in the National Geophysical Data Center (NGDC), now part of NOAA National Centers for Environmental Information (NCEI). The data was collected using polar orbiting satellites that provide full cover of the globe twice a day. The satellites have an Operation Linescan system which allows them to detect low levels of visible-near infrared radiance at night.
With this data, it is possible to detect clouds illuminated by moonlight, lights from cities and towns, industrial sites, gas flares, fires, lightning, and aurora. This particular visualization shows only the lights generated from electricity on two different maps, one with electric lights from 1992, the second with electric lights from 2002. By comparing the two images, viewers can see areas of growth and decline. The Nighttime Lights of the World data set was complied from Defense Meteorological Satellite Program (DMSP) data spanning October 1994 - March 1995.
As with any of the Earth at Night datasets, most of the coast lines are well highlighted, as it seems people like to live by the water. The Nile River in Africa it outlined by the lights that border it. In the United States, it is visible that the eastern half of the country is more densely populated than most other areas. Major highways can be identified by the lights along them.
This is a color composite of three years - 1992, 2000, 2008 - of DMSP nighttime lights. Each year has been assigned to one of the primary colors: 1992 = blue, 2000 = green, 2008 = red. The contrast has been enhanced to show all the detected lights, including the dim lighting often detected in populated rural areas. Locations that had bright lighting in all three years are white. There are many rural areas in India and elsewhere that have a bronze color, indicating that they had no detected lighting in 1992 and dim lighting detected in both 2000 and 2008. Notice the areas of blue lighting in several parts of the Former Soviet Union, the result of collapse and loss of lighting in the mid-1990's. The dataset is available with and without country borders. For more information about the comparison dataset, visit here.