C. Mayhew & R. Simmon, NASA/GSFC, NOAA/NGDC, DMSP Digital Archive
Looking down from orbit hundreds or thousands of kilometers above the sunlit surface of the Earth, the signposts of civilization are clearly visible. The vast checkerboard patterns of farmland; the thin traceries of roads, railways, and airplane contrails; Egypt’s Great Pyramid and China’s Great Wall—all are easily seen with the naked eye. But passing into night, the view changes. The signs of humanity fade save for one: light from artificial sources like electric lamps, oil-refinery gas flares, and intentionally set fires.
In 1962, when John Glenn became the first American to orbit the Earth, he also clearly demonstrated that lights at night could trace human activity from space. Passing over western Australia, he reported seeing a very bright light on the ground; the citizens of Perth had coordinated to turn on as many lights as possible to signal to Glenn as he flew overhead.
Since that time, satellite imagery of nighttime lights has proved useful as a piecemeal method to measure development upon our planet. Static snapshots can reveal stark disparities, like the brightly glowing prosperity of South Korea beneath the darkness of the impoverished, totalitarian north. Gradual changes can be discerned, too, like when the intensity of Soviet-bloc countries’ nighttime lights slowly increased after the USSR’s fall. All such observations to date have been quite crude, constrained by the fact that most of the data comes from one low-resolution source, the Defense Meteorological Satellite Program (DMSP), an aging network of satellites formerly operated by the US Air Force but now controlled by the National Oceanic and Atmospheric Administration (NOAA).
But new analytical techniques and observational satellites may soon open a more rigorous frontier for measuring economic activity from space. Brown University economists J. Vernon Henderson and David Weil, along with their graduate student Adam Storeygard, recently released an analysis of a decade’s worth of global night-light data from DMSP. Their research shows a link between changes in a country’s gross domestic product and the intensity of its electric lighting: On average, as a country’s GDP increases, its nighttime light emission becomes more intense. The work is particularly promising for measuring growth in the developing world, where the quality of collected economic data is notoriously poor.
“A lot of activity in these developing countries is in the untaxed, off-the-books informal sector, but very little information is gathered about it,” Henderson says. “So when [statistical agencies] estimate total economic activity, they don’t really know the size of that sector even though it may account for a majority of the employment in the country. When you get another metric to compare the numbers to, you can be shocked by how much they are off.”
Henderson cites the Democratic Republic of Congo as an example of where data quality is poor. According to the Penn World Table, a standard source used to measure economic growth across countries, during the period from 1992 to 2003 the country had negative GDP growth. In that same period, the satellite data shows a marked increase in nighttime light intensity, suggestive of positive growth, likely in the informal sector. Henderson says Myanmar’s numbers, on the other hand, may show political manipulation: Nocturnal lights indicate significantly lower GDP growth than that stated by the ruling military junta.
Henderson and his colleagues also used the DMSP data to examine economic activity on sub-national scales, investigating the relationship of African cities to nearby agricultural regions. They found that years of crop-boosting high rainfall in a city’s hinterlands significantly correlated with increased growth and development in the urban center as measured by artificial lighting intensity.
“There’s a lot of literature assuming hinterlands aren’t relevant to urban growth and development, but that’s not really the case,” Henderson says. “Farmers demand a bunch of products, they demand services, they demand inputs into agricultural production—this makes a lot of difference to a nearby city. In some sense it’s an academic debate, but this really is a fundamental matter of development and policymaking that can change people’s lives.”
Of course, there are dangers in using nighttime lights as an indicator for economic growth. The DMSP satellites have limited capabilities to measure light intensity, and their image resolution is on the order of a kilometer-per-pixel. A NOAA team led by Chris Elvidge removes contaminating natural phenomena from the images like moonlight, cloud cover, lightning, polar auroras, and forest fires, but human choices of how buildings and streets are lit, the ways windows are shuttered, and even which variety of light bulbs are used all alter the patterns and intensity of light, adding uncertainty to any conclusions.
Still, even uncertain data seems better than none at all. Elvidge and his team have used nighttime DMSP data to monitor gas flares, the deliberate burning of excess natural gas leftover from petroleum production. Gas flares may account for as much as 1 or 2 percent of all annual anthropogenic carbon emissions. The DMSP satellite data proved in a way no other observations could that some countries, notably Russia, were grossly underreporting their gas flaring. Such monitoring will become increasingly important if global cap-and-trade schemes emerge for curtailing carbon emissions.
According to Paul Sutton, a geographer at the University of Denver and frequent collaborator with the NOAA team, measuring artificial lighting may in fact become one of the best ways for estimating how technology affects humanity’s entire ecological footprint on Earth. Along with Elvidge and several other prominent researchers, Sutton is lobbying NASA to develop and launch a Nightsat, a state-of-the-art satellite devoted solely to gathering high-resolution, full-color data on nocturnal lights across the entire globe.
“In my father’s lifetime, the world’s population went from 2 billion to 6 billion—so my father was alive for a tripling of the total human population. In my lifetime, we’re probably going to hit 7 billion and 8 billion and 9 billion,” Sutton says. “Our ability to know what all the human beings are doing on this planet is getting increasingly sketchy, and a Nightsat mission is one way to get objective global coverage of what’s going on. Its data could help quantify carbon emissions, energy usage, urban development, economic growth, and more.
“This all goes back to the idea of ‘carrying capacity’,” Sutton adds. “You can only put so many goldfish in a tank before they’re killed by their own wastes, and you can only raise so many livestock on a certain size of land.” Humans are in a unique position, he says, because Earth’s carrying capacity for us depends intimately on how we choose to live and use technology: Billions of bicycle-riding, solar-panel-building vegetarians will produce a different planet than billions of Hummer-driving, coal-burning carnivores. Thus, the influence of technology on Earth’s carrying capacity is difficult to predict and discern. “Some cultures will use technology to increase their environmental impact; some will use it to decrease it,” Sutton says. “But nighttime satellite imagery may offer a good proxy for its use now.”
Originally published September 21, 2009