Land Use Research Goes High-Tech

NASA researchers are mining satellite data to show the effects of land use change on soil carbon pools, the amount of land dedicated to biofuel feedstocks, and the overall productivity of croplands. When complete, the research may prove useful in determining where high-yield crop production can be maintained for long periods of time and provide an accurate estimation of greenhouse gas (GHG) emissions caused by various crop management decisions.
By Erin Voegele | June 03, 2009
An ongoing research project being completed at NASA's Ames Research Center utilizes current and historical satellite data to track how much land is being dedicated to producing crops that can be used to manufacture biofuels. The research also seeks to determine how land use changes affect soil carbon pools and greenhouse gas emissions (GHG) emissions.

"One of our main targets is to look at how soil carbon is changing," says Christopher Potter, a research scientist at the Ames Research Center, located in California's Silicon Valley. Potter says the techniques have two fundamental aims that his team is trying to develop. "One is how much is being produced on an acre-by-acre basis in any given crop," he says. "The other is how many new acres of a crop are being planted, presumably for biofuel generation."

The project currently focuses on two main regions in the United States - the Midwest and California. To date, research conducted in the Midwest has centered on Iowa and its surrounding areas and has focused on cropland used to grow corn, soybeans and switchgrass. Research centered in California has focused more on the aspects of soil carbon and GHG emissions. "In the case of California, it's more of generally looking and supporting our state's own greenhouse gas inventory," Potter says.

The Midwest Research
To complete the research, Potter and his team are compiling satellite data on Midwestern croplands spanning from 2000 to the present. The researchers are using data sourced from NASA's Moderate Resolution Imaging Spectroradiometer, which collects data on what Potter refers to as the "green-up and green-down curve" of agricultural lands.

Satellite data collected by Potter's team is used to create images displaying a region's "green cover" conditions. By combining data over an entire year, the team can estimate carbon gains and crop production cycles.
SOURCE: NASA Ames Research Center

"Wherever people are growing soybeans and corn, the shape of that green-up and green-down curveseems to have a very characteristic shape to it," he says. In other areas, such as fallow land or pasture, the curve has a different shape. The shape of this green-up and green-down curve allows the researchers to determine how the land is being utilized. "Right now we are in the process of characterizing what those shapes are for different crops and different mixtures of crops," Potter says.

Potter compares the data sourced from the satellites to the scrolling ticker that shows the rise and fall of stock prices on financial news websites. Although the satellites gather data on the Midwest region nearly every day, not all of that data is useful. This is because cloud cover and other conditions may prevent a satellite from collecting optimal data. "We take the best day over a two-week period, and it becomes part of our scrolling ticker for that area that year," Potter says. "We wait until the whole growing cycle is completed, and then we look at the shape of that whole crop growing profile, and things seems to separate out pretty well."

A work group in the University of Minnesota's department of computer science and engineering is assisting Potter's team in analyzing the raw satellite data that is collected for each year. "Basically, we are analyzing the data sets from satellites, and using data mining algorithms to find patterns of interest," says Vipin Kumar, department head. One such pattern, says Kumar, is how the use of cropland changes over time.

Data mining is a technology that is used to analyze large amounts of information and find useful patterns. "The job of data mining researchers is to develop algorithms that can produce meaningful patterns - meaningful results - that are useful," Kumar says. In other words, Kumar's team uses computer algorithms to convert large amounts of raw, low-quality satellite data into something useful. "Our job is to look at the data and see if we can find something actionable, something useful, something from which you can draw some conclusions," Kumar continues.

Putting the Research to Use
Researchers normally deal with crop statistics that are collected by county agencies. This data, however, usually provides only a rough mapping of where the best croplands are located. Potter's research should provide a clearer picture. "With the satellite imagery we can get down to a large farm plot and see individual plots and how they are producing," he says. "That gives us a big advantage in understanding within a county where the best areas seem to be for generating high yields and sustaining those high yields through several different kinds of years."

Besides finding how much farm land is dedicated to each crop, this research should also be able to determine how much land is transformed from a low production system, such as fallow grassland, to a high production corn system each year. It is expected to offer some insight into yield potential as well.

The researchers can also plug data mined from the satellites into computer models to simulate how soils may be affected; taking into account the removal of biomass sources - such as corn stover - that may be removed from the field for biofuel production. "If the soils become progressively depleted over several years of cropping because more and more of the plant materialis being taken away for biofuel, the thinking is that will detrimentally affect the soil carbon," Potter says.

According to Potter, his team is using the CASA model to determine these effects. "What it does is takes that satellite data and simulates how much carbon is being fixed and transferred into the roots below ground," he says. "In that case, we get the full carbon cycle - not just what you see above ground - but also below ground. [The model] also keeps track of the nitrogen used to grow a crop and all the greenhouse gases that are emitted during the production of that crop." The CASA model, which has been used for more than 15 years, can also take into account farming practices and can be used to analyze emissions of nitrous oxide and methane.

Potter's research may also be used to measure the productivity of idle land by building on data published yearly by the USDA using the historical data that shows there has been a major change in land use. Potter's team can then analyze the available satellite data on that piece of land to see how production was maintained in years following the conversion.

Potter says the strength of this kind of research is that it goes beyond simply mapping conversions by supplying data on the productivity of converted lands as well as carbon soil levels.

Ultimately the research will be able to produce a very accurate estimate of how much land is brought into production each year, as well as what kinds of crops are being grown on that land. The team will also be able to estimate the soil quality of idle land before it is brought into production, and determine how the conversion to active cropland will affect that soil.

Potter says there are three main ways in which this research will be useful to farmers, biofuel producers and policy makers. First, it will provide a map of where the most productive lands are and where production levels can be maintained for a significant period of time. In this way, the research will compliment other sources of information, such as inventories, that are available on the county level. Second, the research will provide an accounting of all the GHG that are or are not being emitted - into the atmosphere under different crop management decisions. Third, there is potential to extend the monitoring of converted lands into other areas of the globe, where a need exists to verify what land use changes are being reported by international organizations.

"We are trying to use this data in a way that would benefit the environmental community," Potter says. "We are pretty confident we can use these satellite records to help improve public policy and help improve industry decisions." The next step in the research project is to extend the study into other areas of the Midwest, such as Illinois and the Dakotas.

There may be opportunities to do the same sort of work in Brazil as well. "There are people in Brazil who would like to do the same thing, and we want to be responsive to their needs," Potter says.

It is also possible that future satellite data research could be used to support policy decisions regarding indirect land use change or to verify the accuracy of land use change models used in public policy, such as the controversial models recently proposed by the California Air Resources Board and the U.S. EPA. While Potter says he believes that current research may not be adequate for that exact purpose, he adds that after more data has been amassed and researchers become more experienced in interpreting the data, it could be used effectively for that purpose. The federal government has taken an interest in using data for biofuels policy purposes and might assist in speeding up that process. In May, the EPA announced it plans to look to satellite data in the future to project land use changes, GHG emissions and overall life-cycle GHG estimates. Some of that data could be used as evidence in the on-going review of the agency's proposed GHG emissions reduction rule which includes calculations for indirect land use change.

Potter's team expects to submit results of the research for scientific review in a technical journal by the end of this year.

Erin Voegele is an Ethanol Producer Magazine associate editor. Reach her at or (701) 373-8040.