The U.S. Geological Survey opened its Earth Resources Observation and Science (EROS) Data Center in 1973, northeast of Sioux Falls, S.D., surrounded by farmland. The then high-tech center included rooms for photo processing, computer storage, and microfilming. “Remote sensing” was still a novel term, and researchers couldn’t begin to imagine all the ways it would be used.
But those working at the center understood that much of the data gathered by the Earth Resources Technology Satellite, later renamed Landsat, would be useful to the U.S. Department of Agriculture (USDA) for crop forecasting. An EROS news release from 1966 said photographs taken from space indicated that lands “can be examined, evaluated, and mapped, and the type and vigor of plants can be determined.”
USDA became EROS’s largest customer, according to Tom Holm, chief of the EROS policy and communications office. Holm served as chief of the data services branch in the ’80s, when images were converted to digital tapes and shipped to customers by mail.
Watch this video about Planet's agricultural monitoring program and its use of the RapidEye satellite constellation.
“USDA’s Foreign Ag Service was the largest user of Landsat data, using thousands of images every year for global crop yield predictions,” Holm said. “The other user, almost equally large, was the Intelligence Community.”
Precision agriculture, the use of technology to improve crop yields and reduce costs, is considered the next big thing among venture capitalists, app designers, and tech-savvy farmers. But long before the first unmanned aerial vehicle (UAV) flew over a cornfield or the first app helped manage irrigation, the U.S. government was using satellite images to gather information about agriculture, which remains the foundation of economies throughout the world.
During the Cold War, Holm said, intelligence analysts looked at satellite images of Soviet fields. Analysts could determine the total wheat crop yield before harvest and before the USSR announced it would have a production shortfall—valuable intelligence on this side of the Iron Curtain.
“Sometimes people think of national security as having a well-trained army,” said Torreon Creekmore, program manager for the Intelligence Advanced Research Projects Activity’s HFGeo program, which develops high-frequency communication technologies. “We don’t think about how food security plays into that.”
When crops fail and prices rise, people don’t have the money to purchase food, which can lead to stealing, then riots, social unrest, and mass migrations. National security depends on agriculture, Creekmore continued. “In a global economy, it’s an integrated system.”
Industry experts and those working in humanitarian aid agree higher-yield farming can’t come fast enough as population growth and climate change increase demands on the world’s food supply chain. The United Nations predicts the world population of 7.3 billion will reach 9.7 billion in 2050.
“In that time, we need [to achieve] 50 to 70 percent more output per acre,” said Alex Thomasson, an agricultural engineering professor at Texas A&M University. “I’m very confident we can. Precision agriculture is one of the biggest ways we have to improve the situation.”
Venture Capital Turns to the Farm
Farmers aren’t generally known to be risk takers. But every season—whether they’re farming soybeans in the Midwest or sugar cane in Melanesia—they take chances and make decisions about how to produce a higher yield, minimize waste, and reduce environmental impact. With every tweak and modification, they determine when, where, and how to apply inputs: water, seed, fertilizer, pesticide, herbicide, fuel, and labor. Some decisions are based on experience and history, others on science. More than ever, those who work the land are turning to geospatial imagery and analytics to fine-tune practices and improve their bottom line.
According to Thomasson, who is working on multiple UAV projects, variability in weeds, soil type, and plant health, for example, calls for inputs on an as-needed basis—ideally on a plant-by-plant basis. “We’re far from that now,” Thomasson said, “but I’ll be working on it until I retire.”
The combined rise of small satellite constellations, cloud technology, and predictive analytics has led to a boom in precision agriculture. But the real value is in the analysis, and Silicon Valley is leading the way. According to a report from AgFunder, which funds startups in food and agriculture, investment in precision agriculture has nearly doubled since 2014, reaching $4.6 billion in 2015.
Precision agriculture includes revolutionary advances such as: self-driving tractors; tractors communicating with each other in the field; farmers making data-informed decisions about when to plant; consultants referencing UAV images to determine when specific plants are under stress; and growers receiving automated alerts when there’s a problem in the field. Such advances are significant. For example, in a farm of 1,000 acres precision agriculture can enable a tractor to self-drive within an inch of accuracy or show a farmer where he or she could add an extra row of soybeans.
“The holy grail for farmers right now is, ‘When will my crop mature?’” said Rob Laudati, director of Harris Geospatial Solutions. A strawberry grower in Florida, for example, needs to estimate his crop months in advance so a retailer can plan pricing and advertising. Laudati said machine learning is one of the key analytic advancements in precision agriculture—and a significant disruptor to the industry. Training a neural network could help determine not only whether a strawberry plant is under stress, but what is causing the stress.
Among Farmers, Cautious Optimism
In the United States, about one-third of commercial farmers fall into the category of those who are “completely consumed with wanting to know exactly what’s going on in their fields,” said Chris Clayton, Ag Policy editor for DTN/The Progressive Farmer. As the global population grows and the amount of arable land declines, the margin for error during a short growing season gets smaller.
“Famers turn to technology to boost yields per acre, and everyone’s looking for an angle,” Clayton said, adding that farmers—generally not GIS experts—are feeling overwhelmed by all the new vendors, products, and services. “It’s kind of like the beginning of the Blue Ray, HD, and DVR. Which technology is going to win out? There are so many startups in this field—which is pretty cool—but you don’t know which will survive and catch fire.”
Steve Hoffman, president of InDepth Agronomy, an independent crop consultant in Wisconsin, said farmers love new technology, but it needs to make sense for their bottom line. Most of them, he said, aren’t there yet. “There’s a lot of excitement among farmers,” he said. “It would be easy to adopt the technology and march ahead before we knew what the data meant.”
Among the precision agriculture startups are Descartes Labs and Orbital Insight, which both use satellite imagery and monitor conditions in real time. Descartes, for example, releases a weekly forecast for every corn and soy producing state and county in the U.S., well ahead of the monthly forecasts created by USDA’s National Agricultural Statistics Service. Customers ranging from humanitarian organizations to commodity traders use these highly accurate forecasts and similar products from Orbital.
Other leading startups include FarmLogs, which develops mobile apps to help farmers track metrics for their fields, and Farmers Edge, which offers field-centric data management and analysis. Both rely on satellite imagery from Planet. In October, Planet signed a $20 million contract with the National Geospatial-Intelligence Agency to provide imagery of at least 85 percent of the Earth every 15 days.
“If you were looking at images twice a year, that frequency of observation is only good for looking in the rearview mirror,” said Andrew Zolli, Planet’s vice president of global impact. “It doesn’t tell you anything that’s happening now and doesn’t get into the human decision cycle. Being able to look all the time not only tells you what happened and what is happening, but what may happen in the future.”
Planet operates the largest constellation of Earth imaging small satellites, which hit the sweet spot of low cost, high resolution, and high frequency. The company expects to have 100 satellites in orbit in 2017, collectively imaging the Earth daily at 3- to 5-meter resolutions.
“When you have that kind of daily imagery you will see patterns that not only tell you with much greater sophistication what’s happening on the ground, but a hint of what might be coming,” Zolli said.
IBM’s The Weather Company, which works with Farmers Edge to install advanced weather stations, uses predictive modeling to provide customers with real-time, hyper-local weather information.
In one case, IBM worked with Gallo Wines, which owns 15 wineries and had discovered—through satellite imagery—vines under varying degrees of water stress. Gallo’s new variable rate irrigation prototype allows the company to optimize the amount of water applied to each vine based on its specific need, resulting in a 23 percent larger grape harvest and the use of 20 percent less water. Carrie Gillespie, agriculture lead for The Weather Company, said in the future the company would deploy more than 400 personal weather stations in underdeveloped countries to help growers improve their agriculture and protect people from severe weather.
Large seed companies such as Monsanto—which bought The Climate Corporation in 2013—are getting into the game, providing variable-rate seeding prescriptions and recommendations to farmers about specific zones on their farm—executed through a planter control system. And John Deere offers web and mobile apps to its farmers to help analyze all the data collected by the company’s high-tech machines.
Lane Arthur, director of digital solutions for John Deere Intelligent Solutions Group, said the data is controlled by the growers, who can choose whether to share it with John Deere as well as their seed and chemical companies. Arthur said John Deere strives for transparency in its data use (to improve tractor performance, for example), but across the industry tension exists regarding how much data should be shared. Younger farmers are more comfortable in a sharing economy, but older generations are more private and often consider their data proprietary. This complicates things, as massive amounts of data are needed to make forecasts, predict trends, and support machine learning.
Startups aim to improve the lives of farmers, but selling aggregated information to third parties (i.e. chemical and seed companies) is also part of their business plans, said an executive who asked not to be named. The margin for selling data to farmers is so slim companies need additional income to remain profitable. He said while data sharing will benefit many, it could hurt agricultural retailers (which have historically made it difficult for farmers to compare prices) and agronomists (whose recommendations likely can’t touch the accuracy of those made by machines).
Turning images into action
In many ways, it seems as though UAVs were designed for farm work. They can image frequently, fly low, and capture detail about individual plants that satellites cannot. Analysts expect precision agriculture to be the largest U.S. end market for commercial UAVs, with a potential of $1.4 billion in sales before achieving saturation.
One of the market leaders in UAV-based precision agriculture is Kansas-based AgEagle, which offers a UAV that can fly in the wind, cover hundreds of acres in one flight, and land in the fields. AgEagle’s UAV creates stitched, geo-tagged images while airborne, and its multispectral sensors capture near-infrared bands that identify for farmers and agronomists which plants are stressed—days before the anomalies could be seen with the naked eye. Healthy plants show up green; stressed, yellow; dead, black.
USDA Senior Advisor for Agricultural Systems Seth Murray said the department may not be as engaged in precision agriculture as it is in some other fields such as plant breeding, but added it is yet to be determined exactly how UAV technology will be used.
“Many farmers have bought a drone thinking they can do something with it,” Murray said. “The gap between getting images and getting actionable information is a gulf a mile wide.”
He added USDA is working with partners to design and test UAV technologies and educate both farmers and the public; the department’s National Institute of Food and Agriculture is currently funding hundreds of UAV-related projects. He noted that research has been difficult, partly because of strict FAA regulations that USDA and their researchers must follow (while private companies may choose not to).
One company bridging the gap between image-gathering and data-driven decisions is Agribotix, which processes and analyzes UAV-gathered data to help farmers increase yield while reducing the environmental footprint. The company uses a cloud-based system called FarmLens to create useful tools such as prescription maps for fertilization and weed reports for geo-locating resistant patches.
“Our focus is how to make it easy after you’ve flown a drone to get the data,” said Agribotix Chief Operating Officer Paul Hoff. He said he expects UAV use for agricultural purposes to skyrocket in the next two years, more than any other precision technology.
What’s the next big thing? Creekmore suggests “nowcasting”—which tells a farmer, for example, the best time in the next hour to water a field or transport goods. Creekmore also predicts a move from multispectral to hyperspectral sensors will provide growers a more detailed fingerprint of their crops. For example, a multispectral sensor can be used to show forested areas versus crop areas, while hyperspectral sensors, which are more sensitive to subtle variations in reflected solar energy, can detect an oak, maple, or spruce tree within the forest.
Work is underway at a number of universities to further development of precision agriculture tools. At Texas A&M, Thomasson is exploring various remote and proximal forms of sensing—from robots in the field to the possibility of UAVs talking to tractors. One project looks at identifying a disease in cotton using UAV images, and another focuses on assessing the water and nutrition need of individual plants.
Precision Ag for Peacekeeping
While farmers both large and small use precision agriculture to improve efficiency, humanitarian leaders and the Intelligence Community are interested in how the technologies can be used globally to address food insecurity and the potential unrest it can create.
Global Harvest Initiative (GHI) is an organization that advocates for ways to help low-income countries boost agriculture through better information, training, irrigation, and seeds. High-tech farming tools are critical, said GHI Executive Director Margaret Zeigler, but equally important is making sure broadband is available for precision agriculture in rural areas. GHI’s recent Global Agricultural Productivity Report describes innovations such as the Soil Cares Scanner, being piloted in Kenya, which uses the infrared spectrum to provide affordable, real-time analysis about variability in soil and fertilizer recommendations through an app. Zeigler also pointed to the success of the International Center for Tropical Agriculture, which worked with Colombian rice farmers to avoid millions of dollars in losses after using weather and crop data to understand how climatic variation impacts rice yields.
Nonprofit PeaceTech Lab’s Noel Dickover leads the Open Situation Room eXchange project, which works to make big data more readily available in peacekeeping. He said big agriculture can provide an early warning sign about food insecurity and regional conflict, and can be a key decision support tool for local peace advocates. PeaceTech Lab is integrating weather data and agronomic models with existing indicators of conflict, working with Colorado-based aWhere, which monitors the weather at a nano level and communicates with farmers via SMS text messages.
“The military spends a lot of money and time trying to figure out all kinds of threats and risk assessments,” Zeigler said, referencing the Office of the Director of National Intelligence’s Global Food Security report, which concluded declining food security will “almost certainly contribute” to social disruptions and political instability. “So understanding where food crises might be coming is a way they can prepare.” As an example, Zeigler points to the failed wheat crops across continents, and a series of droughts in Syria that triggered global wheat shortages, caused a spike in wheat prices, and led to the Arab Spring.
“Because of that, the Intelligence Community has always focused on trying to get a handle on food prices and crop failure. There are now many more tools to do that.”