Seed Shadows

Puff on a dandelion gone to seed. How far can each seed fly? No one knows.

And suddenly, it matters.

drawing of dandelion in black and white

Soybeans in Delaware are being overwhelmed by a variety of a weed, commonly called horseweed or marestail, that is resistant to the herbicide glyphosate, or Roundup. “It has a six- to ten-fold level of resistance,” says Dave Mortensen, a weed ecologist at Penn State. “That means if a quart of herbicide used to kill it, now it takes six to ten quarts.”

How does such a weed spread from one field to the next? Can we keep it out of Pennsylvania?

“Marestail is like a miniature dandelion,” Mortensen explains. “Each seed has a fringed pappus that acts like a parachute. It floats in the air.” The question is, how far—and can we design our fields in such a way as to contain its spread? Will a road or a fence row or a field of corn knock a weed seed out of the sky, or simply send it farther? “What we call the seed shadow or seed cloud is determined by the landscape,” explains Mortensen, “both how far it moves and the likelihood the seeds will remain there once they land.”

Using a wind tunnel 15 meters long, Mortensen, post-doctoral scholar Robert Humston, and a team of graduate and undergraduate students simulated a variety of different breezes, changing the wind speed, direction, and turbulence to see how it affected the seeds' flight. They tested their setup with the larger dandelion seeds, then moved on to marestail, releasing 500 to 1,500 seeds for each run. (How did they “release” them? “We experimented with some elaborate fancy things,” laughs Mortensen. “We built a little stand and clipped the seed to it, a nice little platform of copper tubes.” In the end, the simplest way was best: They held the seed head before their lips and puffed, sending the seeds shooting down a short plastic tube into the wind tunnel.)

At different points along the tunnel, both across its breadth and on the floor, they trapped the seed cloud with sticky sheets of Tanglefoot, usually used to catch bugs for entomology studies. The result, once the data were entered into a computer, was a three-dimensional image of the seed cloud.

“In the literature, seeds move like something called a Gaussian tilted plume.” Mortensen draws a picture on his whiteboard of a slender egg, the smaller curve being the leading edge. “We're able to test what the plume looks like aerially as it moves down the tunnel. It's interesting mathematically. But we're also interested in interpolating beyond the tunnel.

“As the wind speed increased, especially with marestail, a lot of them went to the end of the tube, still in cloud formation. There was virtually no sedimentation on the floor. And these were just breezy-day winds, not hurricanes—which is bad news. Sixteen kilometers per hour was our highest wind speed. That's ten miles per hour. If you were flying a kite, you'd be trying hard to keep it up. A really breezy day, say 15 to 20 miles per hour, would not be unusual at all. But if we ran the wind tunnel at that speed, we'd blow every seed to the back.”

To expand to landscape scale, the researchers moved their study to Delaware, to fields so infested with marestail that the soybeans could hardly be seen. There they took “a seed's-eye view of the world,” Mortensen said. “In this mosaic, we asked, what is the diversity of land uses and how would they affect how well the seed will do when it lands? The way crops are managed is important to whether the seed has a chance of growing and being a problem to the farmer and maturing to blow seeds off to another place.

“Based on what we know about how air moves,” he adds, “we think edges—like a forest edge—will act like a filter. The wind speed will drop and it will trap the seeds—you should see a lot of seeds at an edge. But to my knowledge none of this has been measured in the field. And how important are less-sticky corridors, like a road or a creek? They could be good or bad. The complex interactions are fascinating, the interaction of open fields, edges, and corridors.” If you rely only on wind-tunnel data, he warns, “You can really mislead yourself about what's going on in the real world.”

Yet only recently, with the advent of supercomputers and satellite maps and Geographic Information System (GIS) tools, have weed ecologists had the capacity to work on a landscape
scale. Compared to former studies using potted weed plants at a research farm or greenhouse, where the experiment can be tightly controlled, an actual soybean field “is a very noisy environment. We have no control over which direction the wind is blowing, but we have a huge source of seed.” Mortensen's team ran transects 1,000 to 2,000 feet from the source, set up sticky cards of Tanglefoot on posts, and counted (with a magnifying glass) the seeds that stuck.

Onto this data, using GIS, they will superimpose a map of land use and a grid showing wind speed and direction. Finally they are devising a computer model that will predict what the field might look like, under different management practices, in three or five years: information a farmer needs in order to control the spread of the weed.

“My challenge is to constantly come back to the farmer's field,” says Mortensen. “It's always fascinated me that you can have a practical problem that requires basic ecological understanding to address. I like complicated questions that lie at the interface between science and the social side of things. To me, it's more rewarding to do something that might make a difference.”

David Mortensen, Ph.D., is an associate professor in the department of crop and soil sciences, College of Agricultural Sciences, 215 ASI Bldg., University Park, PA 16802; 814-865-1906; Robert Humston, Ph.D., is a postdoctoral scholar in the department. Other collaborators include Penn State master's degree student Joseph Dauer; Nora Peskin, an undergraduate exchange student from Argentina; Jose Gonzalazar, Ph.D., from Spain; and Mark Van Gessel, Ph.D., of the University of Delaware. Their work was funded by the University, the U.S.D.A., and The Nature Conservancy.

Last Updated May 01, 2003