Princeton researchers use satellite images to track disease

Satellite imaging of nighttime lights helps track the migration and clustering of people, giving researchers the data they need to calculate the risk of outbreaks of infectious diseases such as measles and meningitis.

Tracking where humans migrate and cluster in any given country from season to season is, in some places, a tall order. Which makes tracking the risk of infectious disease outbreaks that thrive in dense populations tricky as well.

This 3D rendering shows the amount of brightness for urban areas in Niger over the course of an average year, with the height of each spike representing total brightness. The tallest spikes, then, indicate the most densely populated areas. Science/AAAS

Satellite images of nighttime lights could be the answer, according to researchers at Princeton, who report on their findings today in the journal Science.

Using nighttime images taken of Niger's three largest cities between 2000 and 2004 by a U.S. Department of Defense satellite, and checking those images against public health records compiled by Niger's Ministry of Health, they saw that new measles cases clearly occurred in the brightest areas.

"Temporary and seasonal migrants are very hard to measure," says Deborah Balk, a professor at the City University of New York who's been following what she calls in the school news release "pathbreaking" work. "The night lights are an important source of data for Africa and Asia, especially, where data is sometimes absent or poor."

The use of these images to track population changes may seem obvious in retrospect, but lead author and postdoctoral researcher Nita Bharti says the approach is quite unconventional given that these nighttime sat images are typically used to observe stable, not migratory, populations. "This is the first use, so we had to develop the method ourselves," she says.

The team began formulating the idea when Bharti and colleagues published a paper in 2010 finding that measles epidemics most often occur during Niger's dry season, when many farmers migrate to urban centers. The severity of outbreak, they concluded, must be linked to shifts in population density, not lower rainfall or other environmental factors. Without being able to measure these population changes, though, they were stuck with an untested hypothesis.

After using spacial analysis software ArcGIS and ERDAS Imagine, the team was not only able to confirm their suspicions but is also now looking into joining the night light technique with other population trackers--such as cell phone usage--in an attempt to get the most accurate data possible.

Moving forward, these images--which can typically be analyzed within 48 hours of being captured--could help epidemiologists not only study the patterns of human movement in specific regions over time but, by extension, also the patterns of a wide range of epidemics, which could then inform intervention strategies.


Join the discussion

Conversation powered by Livefyre

Show Comments Hide Comments