Predicting famine with satellite imagery: How one teen is combating global food insecurity
Our editorial transparency tool uses blockchain technology to permanently log all changes made to official releases after publication. However, this post is not an official release and therefore not tracked. Visit our learn more for more information.
Seventeen-year-old Lillian Kay Petersen can predict a famine four months before it occurs, giving aid organizations a vital head start on targeted hunger relief.
At a time when 26% of the global population—more than 2 billion people—suffers from food insecurity, Lillian is using satellite imagery to remotely monitor crop health throughout Africa, where a lack of mobile connectivity makes traditional monitoring methods ineffective, leaving millions at risk of starvation.
In 2015, Ethiopia experienced its worst drought in decades. Hundreds of thousands of farmers were left with failed crops and dead livestock. Aid organizations experienced more than double the usual demand for relief.
More than 8,000 miles away in Los Alamos, New Mexico, Lillian felt a personal connection. Before her three adopted siblings joined Lillian’s family, they all experienced food insecurity. Lillian saw them struggle with the long-term effects of malnutrition, like hindered brain development and delayed physical development.
Lillian wanted to help, so she turned to what she knows best—computer science.
Now, in 2020, Lillian’s remote monitoring model has won one of America’s most prestigious youth science award, providing valuable data to international aid organizations and saving lives.
A global crisis, close to home
The daughter of a former Navy computer scientist and a Ph.D. in applied mathematics, Lillian is no stranger to seeing code transform a complicated problem into a simple solution.
Her father, who programs models of Earth to predict ocean health over the next century, taught Lillian to solve problems using computer programming in the fifth grade.
“He would make little challenges for me,” says Lillian. “Like, ‘Try to build a Christmas tree where the user can change its shape.’ I thought it was so much fun—to figure out how to get the computer to do things like that.”
Lillian has shared her findings with organizations across the U.S. and globally to help prevent food insecurity. Photo credit: Lillian Petersen
Her fascination with computer programming grew, and when Lillian’s family adopted her three younger siblings, she saw a challenge that programming just might be able to solve.
“Even though they grew up in the U.S.,” Lillian says, “all of my younger siblings experienced food insecurity in their former homes.” Malnutrition delayed their physical and mental development.
“I wanted to find a way to help others like them, to make a difference,” she says. “I spent a year researching how heat extremes affected crop yields in the United States and then how that was supposed to change in the future.”
She saw that a bad crop yield in the U.S.—or even crop failure—was less catastrophic to the food supply than in developing countries. “I knew from my experiences with my younger siblings that malnutrition doesn’t just go away with the next season or an influx of food aid—the impacts are lifelong.”
After learning about it in the news, Lillian couldn’t stop thinking about the Ethiopian drought and the impact it would have on kids.
From Illinois to Ethiopia: A crop prediction model for the world
In 2016, approximately 24.7% of the population in Africa suffered from food insecurity, compared with just 1.2% in North America and Europe. Variations in African climates and terrain create difficult growing conditions for farmers.
During her research, Lillian used satellite imagery from farms in Illinois to find a correlation between image color and the outcome of the crops—the greener the satellite image, the healthier the crops.
Field plots in Africa are much smaller than those in the U.S., making crop predictions more difficult.
“Illinois has a very good reporting of crop yields,” she says. “The fields are really big, and they grow a small number of cereal crops. So, the models can be highly tuned to an individual region.”
By gathering enormous amounts of satellite data every day, Lillian was able to accurately predict the crop yield in Illinois.
Emboldened, the high schooler set out to better connect food aid to African droughts like the one in Ethiopia.
Supplementing mobile with remote monitoring
But, when it was time to apply her model to African countries, she had to make some adjustments. “In Ethiopia and similar countries,” she says, “the plots are relatively tiny, and the crops are more varied than they are in Illinois. Plus, there’s not nearly as much monitoring on the ground throughout the growing season.”
But, because a lot of farmers in Africa don’t have mobile phones, and fewer people are travelling to observe crop health during the pandemic, many aid organizations are currently unaware of conditions on the ground. “So, for now it’s important that we have an affordable and simple remote monitoring system,” says Lillian. “It’s a challenge in Africa to have a model that can be applied to any location, crop and climate, but my method does that successfully with high accuracy.”
One of the key crops Lillian monitors is Sorghum in Ethiopia, a staple food, second only to corn in production.
Looking at comparative satellite images of Ethiopia, Lillian saw that 2013 was a very wet year, with green images and, thus, a high crop yield. Then the drought hit in 2015. The pixels in the satellite images became more brown than green.
When Lillian sees something like this—brown pixels that indicate unhealthy crops—she often discusses her findings with aid organizations. “We see what other models are saying, see what correlates and what doesn’t. We try to figure out exactly what’s going on and the best way to respond,” she says. “They’ll have a three-month head start on bringing relief to that region.”
Connecting to prevent famine
“There is a whole community of researchers and aid organizations working together to find the best way to monitor crop yields in developing countries,” says Lillian. “The goal is to make sure that we are prepared when a food crisis is coming so we can get food there to prevent a famine.”
Right now, during the COVID-19 pandemic, a lot of aid organizations’ models, which combine remote sensing and ground data, may be breaking down because observers on the ground can’t travel, she says, making Lillian’s model even more useful.
Using tech to solve our biggest challenges
Lillian was recognized for her breakthrough method with the top prize in the prestigious Regeneron Science Talent Search competition for the Society for Science & the Public, in which high school seniors submit their research in critically important scientific fields.
Now a freshman at Harvard, Lillian continues to investigate the intersection of nature and technology. “I think molecular biology, in companion with computer science and big data analysis can have the largest impact for lifesaving drugs and therapies. It’s the fastest advancing field, and there’s so much to be gained,” she says.
With technologies like CRISPR gene editing and genetically modified organisms, there could be a “perfect solution for solving hunger and stopping deforestation at the same time. And these same tools can be used for cancer treatments and for improving human health and well-being,” says Lillian.
Recently, Verizon committed $1 million to help other businesses use technology to become carbon neutral. The initiative underscores Citizen Verizon, the company’s responsible business plan for economic, environmental, and social advancement.
“If a single user with a personal computer can produce reasonable crop forecasts for a whole continent,” she says, “then perhaps anything is possible.”
Know someone who's creating a positive impact using technology and connectivity? Send us a message at email@example.com. They could be next in our story series.