Arizona Researchers to Compile First Ever High Resolution Global Cropland Map

Researchers hope that having a detailed picture of what’s happening with croplands around the world will help to inform the net effect of regional demographic and geological changes. Piecing together that accurate of a map will likely take five years, $3.5 million (funded by NASA), computation of thousands of satellite images, and collaboration with crop experts all over the globe.

The project is the brainchild of Prasad Thenkabail a research geographer at the United States Geological Survey with expertise in both remote sensing and cropland mapping. While global agricultural production has kept pace with population growth since the 1960’s, “that is coming to an end,” Thenkabail says. “Now the lands are becoming limited, more and more are being taken up by urbanization, and still more is being taken away for ethanol. Water, too is becoming scarce. In order to figure out how to produce more food with less land and less water, we must be able to examine how we are currently using land and water, Thenkabail said. He expects that his global cropland map will be able to provide that information.

The satellite images come from the LANDSAT Project, a program managed jointly by NASA and the USGS. LANDSAT is a satellite sensor that has been collecting images of the exact same resolution for 40 years. The shots upon first launch of the satellite in 1972 have the exact same scale of 30 meters by 30 meters per pixel as those taken during the most recent circling in 2013. As one of the first American satellites, LANDSAT has compiled the world’s largest compilation of consistent images. When in orbit, the satellite continuously circles the earth from north to south. “Every time it orbits the earth, it takes a skinny little strip of photos. It takes 104 minutes to make one circle. Then it goes around another time taking another skinny little strip of pictures next to [the first]. It takes 16 days to get the entire Earth,” explained Temuulen “Teki” Sankey, a plant ecologist and remote sensing specialist collaborating on the project from the University of Northern Arizona in Flagstaff.

The historical image record adds an additional temporal dimension to the map, Sankey says. Images taken over several decades show not only what we have today, but also how cropland quantity and productivity has changed in comparison to historic population growth. That visual timeline might provide some clues for how to accommodate future population growth. The researchers have selected 2010 as a baseline year. For just that one year, 1000s of images will have to be processed using algorithms created by researchers at NASA Goddard Space Flight Center in Maryland and processed by a supercomputer at NASA Ames Research Center in Calif.

Once the maps have been compiled, Thenkabail will reach out to local cropland experts around the globe to corroborate that the information is accurate. Many of these researchers have already taken land images of their region that can be compared to the satellite images. An expert in accuracy assessment from the University of New Hampshire will help to oversee the entire process. In the end, Thenkabail and Sankey hope to post the completed map online and make it freely accessible to the public. They are considering adding an interactive component that would allow the public to comment on how well the map reflects their current location.

Successful completion of such a detailed global cropland map could help inform other areas of crop research as well. Sankey expects that the satellite data will show not just where cropland lies, but also whether or not it has been irrigated, how many times farmers harvested crops in a season, and what yields looked like. “We’re trying to figure out, how much are we irrigating? In areas where we are irrigating, are yields higher? Could we be using water more efficiently?” she said. The long view of how croplands have changed over time could offer a window into the effects of climate change, urbanization, and crop intensification. Having accurate estimates of global vegetation cover can help inform carbon sequestration estimates.

Thenkabail and Sankey hope that this information will help reveal regions where productivity and resource efficiency are equally high.