Another popular means of evaluating a field's nitrogen needs is to eyeball it. "In those cases, the farmer is the sensor," Sudduth says. "He's looking at the corn or where he is in the field and thinking, 'The crop here is shorter and less green, so I'll add 10 percent more fertilizer.' Obviously, that's a pretty subjective thing."

Using sensors to take subjectivity out of the process is a key component of reducing nitrogen runoff and, by extension, stemming the spread of hypoxic waters in the Gulf's dead zone. But the technology, still in the experimental stage, will need some tweaking.

Last year, for instance, Sudduth and his colleagues relied upon their own customized equipment to gather data. This year, they have also installed the sensors on seven farmer-owned tractors. The switch has made their experiments more relevant to typical producers, who don't own special equipment. But it has also made applications less precise.

"The control rate on this equipment is fairly slow," says Scott Drummond, an information technology specialist with the USDA-ARS. "When you command a different rate, you're going to be three to five seconds down the field before that command takes effect. It's not an ideal situation, but it's better than nothing."

Another challenge involves the length of a tractor's boom, that long piece of hardware upon which the fertilizer application valves are mounted. The boom used in Nuelle's field sprays 16 rows of corn simultaneously, while the sensors, one mounted to each side of the tractor, see one row each. The computer averages the information from those two rows and then treats all 16 identically.

"This will probably do a good job on the big scale," Drummond says. "On a small scale, maybe you've got something weird going on in a 20-foot circle. We're not going to do as good a job on that. If you want to control those really small areas effectively, you need a system where the sensors control smaller sections of boom and the control happens close to instantaneously. It wouldn't have to be instantaneous, but it would have to be within the length of time it takes for the sensors to see it and the boom to travel to that point."

And then there are the algorithms, those mathematical formulas that allow farmers' computers to perform their magic. Thus far, the MU researchers have written these based on trial and error. They conducted experiments to determine how much nitrogen went into corn of varying shades of green, then used the data to write algorithms.

Investigators at other universities have conducted experiments of their own, and have arrived at different algorithms. Some focus more on plant size, others concentrate on color. These researchers, along with other interested parties from as far away as Argentina, periodically gather at workshops where they compare notes and discuss ways to test each other's findings. Such collaboration is crucial, the MU researchers say, because experience has taught them that nearly everyone's algorithms could benefit from fine-tuning.

In 2004, for example, Sudduth, Drummond and their colleagues conducted experiments in seven test fields using customized equipment. At two of the locations, the sensor-controlled nitrogen application recommended higher doses than the conventional rate -- a recommendation the researchers deemed valid. At four other locations, however, the sensor management resulted in greatly reduced application rates. It turns out that this was the correct call for one of those four fields, the researchers say, but the other three fields experienced significant yield and profit loss.

Averaged over all seven locations, the sensor approach applied 33 pounds of nitrogen per acre less than the recommended average of the conventional rate. That saved the producers about $10 per acre on fertilizer costs. But insufficient nitrogen application cost the producers about $6 per acre in lost yield. "So we're only making about $4 per acre," Scharf says.

Published by the Office of Research.

©2006 Curators of the University of Missouri. Click here to contact the editor.