The problem is simple and ubiquitous. If you've ever grown a garden, it's highly probable that your good plants shared precious soil, water, and nutrients with weeds. Preventing or controlling weeds isn't easy. You can treat them chemically or get down in the dirt and pull them out by hand. And chances are they will grow back in a couple of weeks. Now imagine if your garden covered 1,000 acres or more. That's the dilemma facing today's crop producers.
Right after World War II, American farmers took advantage of new chemicals like 2,4-D to control weeds in corn. It was cheap and effective, though it has largely been abandoned because of environmental concerns.
Today's agricultural chemical prices have skyrocketed, partly because of increased oil prices (many chemicals are petroleum-based) and partly because of high research and development costs. Mounting concerns over water polluted by agricultural chemicals further complicate the picture.
But farmers can't afford to ignore the issue. Losses from weeds for 46 major commodities in the United States were estimated at $4.1 billion in 1991. Without herbicides, it is estimated, losses would have been more than $19 billion.
So agriculture's challenge is formidable: sustain food production by controlling weeds, but do it with fewer chemicals to avoid contaminating soil and water. To meet this challenge, crop producers must be armed with as much information as possible about weeds, control options, and consequences.
Colorado Agricultural Experiment Station researchers Donald Lybecker, a professor of agriculture and resource economics, and Phillip Westra, professor of bioagricultural sciences and pest management, joined forces with United States Department of Agriculture researcher Edward Schweizer to meet the challenge head- on. They combined their crop and weed science expertise with the number-crunching power of computers to develop bio-economic models that give farmers realistic choices for weed control schemes.
"What's neat about the bio-economic weed computer model," Lybecker says, "is that it integrates scientific information and data with specific information that producers provide about their crops."
A model named Weedcam focuses on irrigated corn production. After corn plants emerge, the producer counts weeds and weed species, measures distribution of the weeds, and records the maturity of the corn and weeds. The computer factors all this with current prices of herbicides available for weed control, costs of custom or self-application of herbicides, efficiency or effectiveness of treatments using various herbicides, and expected price of the crop at harvest. The computer program then delivers weed control choices to the producer with associated costs and estimated profitability.
But that's not the end of the story. The information is archived for future use in a computer database called the Weed Bank. The number of weeds counted in the field, for example, is an important factor regarding application of pre-emergence herbicides in succeeding crops.
Results show that growers who use computer models like Weedcam are about 80 percent more successful in reducing weeds, cutting herbicide applications and costs, and improving profitability than growers who make decisions without their help. Though that's a significant improvement, Lybecker and Westra are constantly tinkering with the models to make them better. "Like any process, we need to fine tune it," Lybecker says. They are considering ways to factor in mechanical weed control, effects of weeds left in the field, and information about risks associated with various levels of control. And finally, the models could include an environmental quality index that rates expected impact on the environment by different weed control schemes.
"The models are doing the job of helping us become smarter about weed control," says Lybecker. "They're useful to crop consultants, Cooperative Extension professionals, and farmers in that never-ending battle against those pesky weeds."