NASA Technology

When global food prices spiked dramatically in late 2007 and into 2008, with the costs of many basic dietary staples doubling or even tripling, protests and riots upset much of the developing world. Panicked governments stopped exporting food, aggravating the crisis. Almost as troubling as the widespread unrest and hunger, though, was the fact that it had taken the world by surprise.

Rice fields in the Mekong Delta
These two satellite images depict the An Giang Province in Vietnam’s Mekong Delta, a major rice-producing region, at different times of year. Dark blue and black areas are inundated and have low biomass, while white and gray areas are other crops like row crops and trees. The differences in color indicate a change in the ratio between soil moisture and biomass.

Following the price spike, world leaders gathered to figure out how to foresee and avert such market instabilities in the future, and a major product of those meetings was the Group on Earth Observations’ Global Agricultural Monitoring (GEOGLAM) initiative, carried out by a partnership of governments and international organizations in the G20 nations. The initiative relies primarily on satellite Earth-imaging data to improve projections of crop production and weather forecasting.

Making projections based on raw satellite data, however, is no simple task. “You have to be an expert to transform that data into useful information,” says Nathan Torbick, a director at Applied GeoSolutions, based in Newmarket, New Hampshire, which has been researching various applications for Earth-imaging satellite data for more than a decade. As far as crop projections, the company decided it would leverage its ongoing work on rice, which is the world’s most difficult major dietary staple to predict.

“There really is no rice futures market, relative to corn, soy, wheat, and wbeef,” Torbick says, noting that this is because traders don’t have reliable rice information or production forecasts on which to base such trades. This makes the rice market volatile, putting investors, producers, and, ultimately, consumers at risk. “If there’s a big drought in Thailand or other parts of Southeast Asia, hundreds of thousands of people worldwide might starve,” Torbick says.

Meanwhile, in the United States, parts of the country’s two biggest rice-producing regions are running out of water. In the Sacramento Valley area of California, which is home to about half a million acres of rice fields, drought has become the new normal, Torbick says. And in eastern Arkansas, where about a million and a half acres are dedicated to rice production, areas like the Grand Prairie region sit on aquifers that have been pumped nearly dry. One solution is to use satellite data to plan water use more efficiently, Torbick says.

Among the satellites most popularly used for Earth-imaging data are the Landsat series, in orbit since 1972, and the Terra and Aqua satellites, equipped with the Moderate-Resolution Imaging Spectrometer (MODIS) instrument. Although NASA built the satellites, the U.S. Geological Survey operates each Landsat once it’s in orbit.

Landsat 8, launched in 2013, covers Earth’s surface every 16 days, capturing images in nine visible, near-infrared, and shortwave infrared bands, as well as two thermal infrared bandwidths. The thermal information is especially important for detecting crop stress and supporting crop predictions, as it reveals moisture and temperature on the land surface, in plants, and in the lower atmosphere. MODIS, meanwhile, maps Earth every one to two days in 36 visible and infrared bandwidths, providing the frequent visits needed to monitor changes in crop stages, plants’ responses to weather, and farm activities such as irrigation or tillage.

Rice field in the Philippines

Rice is one of the most heavily produced food crops in the world, but it is one of the most difficult to predict, with yields often depending on weather. The Philippines’ famed Banaue Rice Terraces, pictured here, dried up completely in 2010. Applied Geosolutions’ Rice Decision Support System, which it developed with NASA funding, will use imagery from various satellites to make more accurate predictions, helping to stabilize the world’s rice markets.

Image courtesy of Adi Simionov, CC BY-SA 3.0

In addition to moisture and temperature observations, the imagers can detect the greenness, biomass vigor, and leaf moisture of rice plants, and the presence and depth of surface water, all useful for assessing crop health and predicting crop outcomes. By using the satellite imagery, Applied GeoSolutions can assess large areas at far less cost than traditional field visits.

Applied Geosolutions approached NASA with a plan to design Web-based software that would use current and archived data from Landsat, MODIS, and other satellites, incorporating measures of rice fields, yield modeling, and weather forecasts to generate information in real time about rice coverage, growth stages, deviation from normal, and expected yield around the globe, as well as calculate the statistical certainty of that information to address risks.

Technology Transfer

In 2012, Stennis Space Center granted the company Phase I and II Small Business Innovation Research (SBIR) contracts to create a Rice Decision Support System (RDSS), one that would inform rice-related decisions ranging from how to invest in rice futures to how to irrigate fields, where and when production risk exists, how to validate conservation practices, and when to buy, sell, or hold.

The system went into operation around the start of 2014.

In the United States, Applied Geosolutions predominantly uses Landsat and MODIS data, while in Southeast Asia it relies primarily on synthetic aperture radar (SAR) satellite data made available by governments and especially the Japan Aerospace Exploration Agency. In 2015 the company grew its use of the European Space Agency’s (ESA) Earth-imaging Sentinel platform, in addition to continuing its work with the ESA’s Phased Array Type L-Band SAR satellite and Canada’s Radarsat-2.

The various imagers have different resolutions and operate in different parts of the electromagnetic spectrum, with the RDSS overlaying all that information. “You combine all the satellites to give you a complete picture of what’s going on on the ground,” Torbick says.

Children eating rice

By making rice harvests easier to predict, Applied Geosolutions’ Rice Decision Support System will bring stability to the rice market, blunting food shortages that tend to hit developing countries especially hard. The system could also help inform programs like the U.S. Agency for International Development’s Famine Early Warning System, which can use the data to predict regional food crises and decide how much food aid will be needed and when and where to buy it.

Image courtesy of Feed My Starving Children, CC BY 2.0

Abroad, the system is focused on pilot sites in Java, Indonesia, and the Red River Delta in Vietnam, as well as in Brazil, which has recently ramped up rice production and has begun to offer incentives for sustainable crop management styles. RDSS monitors which farms are using these practices. “Tools to help monitor and validate rice-management practices will be central to developing ecosystem-services markets that will improve transparency and increase efficiency while rewarding stewardship,” Torbick says.

In Arkansas and California, Applied Geosolutions has paid partnerships with farmers and agencies to help them plan their growing season and manage resources, especially with regard to irrigation. As in Brazil, imagery gathered in the United States also proves which farms should receive incentives for using alternative irrigation methods.

The imagery from Southeast Asia is used primarily for supporting food security programs and commodity markets. In that part of the world, where rice is a major contributor to gross domestic product and supports the livelihood and diet of hundreds of millions of people, Torbick says, agriculture faces major challenges in the coming decade due to increasing resource pressures, weather and climate change, population growth, shifting diets, and economic development.

Historically, global rice production has been difficult to predict because this region, where the overwhelming majority of the world’s rice supply is produced, is too cloudy for conventional satellite imaging. SAR imaging technology helps overcome that problem, as it is able to detect vegetation structure and moisture content through cloud cover.