San Francisco

The US Environmental Protection Agency (EPA) is hatching a scheme that could let it monitor genetically modified crops from space. Experiments will begin next spring to determine whether subtle differences in the way leaves reflect the Sun's rays can distinguish transgenic from conventional maize.

If it works, the technology would allow the EPA to track farmers' compliance with planting guidelines, and might even spot the emergence of insecticide-resistant pests. This would be a remarkable feat of detection because, to the naked eye, transgenic plants appear identical to normal ones.

Satellite imagery is already a well-established agricultural tool. Stressed plants absorb less infrared light than healthy plants, so they appear 'warmer' in infrared photographs. Farmers already make use of images from satellites and aeroplanes to spot thirsty crops, pest infestations and plant diseases. Some large-scale growers use infrared data to classify weeds in their fields to help them decide which herbicides to apply.

All of these things can also be spotted by a farmer in the field — the main advantage of observations from the air is that they allow large areas to be examined quickly. But the EPA's scheme to monitor transgenic crops by satellite would involve discerning more information about a plant from space than an observer could do on the ground.

The idea rests on the possibility that subtle genetic differences between plants might influence the spectral qualities of the solar radiation reflected by the leaves, says project leader John Glaser of the EPA's Cincinnati office. Transgenic varieties would appear different, not because of their inserted genes, but because the starting strains from which they were generated were different from those of conventional crops, he says.

To achieve this sensitivity, computer models will be needed that can peel away background noise, such as the effects of rainfall, air temperature and pest attack. With this in mind, Glaser has recruited agricultural experts at Pennsylvania State University and the University of Nebraska in Lincoln.

The researchers admit that this may be easier said than done. “The big question is, can we refine it enough to see genetic differences?” says project collaborator Joe Russo, president of ZedX, a company in Bellefonte, Pennsylvania, that builds computer models for agriculture.

The project was born mainly out of concern that overuse of Bt maize (corn) — which is genetically modified to produce a natural insecticide — will result in the development of resistant insects, rendering the technology useless. US farmers who plant Bt maize are required to keep conventional maize on 20% of their acreage to minimize the risk, but up to one-fifth are thought to be flouting this rule, according to a study released in June (see Nature 424, 116; 200310.1038/424116a).

In theory, the project could also spot fields of Bt maize that are under attack by insects, helping regulators to nip resistance in the bud. The technology, which would not necessarily be limited to maize, could ultimately help to track the use of transgenic crops worldwide, Glaser says.