FORECASTING THE WEEDS

'WeedCast' software fine-tunes weed control by weather patterns

By E. M. Morrison

A new software program, developed with funds from AURI's PRO program, uses local weather data to predict emergence and growth of annual weeds in corn and soybeans. "WeedCast" helps farmers apply weed ecology to improve weed control and cut herbicide use.

Herbicides are considered by many a "miracle" of modern agriculture. "They are easy to use and they work," says WeedCast developer Frank Forcella, an agronomist at the North Central Soil Conservation Research Laboratory in Morris. American farmers now apply more than 200,000 tons of herbicide a year, most of it to fields in the Corn Belt. In Minnesota, growers spend about $15 to $35 an acre on chemical weed control -- more than $350 million each season.

Too much of a miracle?

Because these chemicals are good at preventing yield losses, farmers routinely apply them more often than they need to, just to be on the safe side, say researchers. Forcella estimates that a quarter of prophylactic weed treatments are applied to fields "that don't have enough weeds to justify treatment."

At the same time, farmers worry about the effects of herbicides on the land and water. Jeff Arnold, manager of Trico Ag Service, a crop consulting company in Hancock, notes that in recent surveys of farmers, "environmental safety is now among the top five concerns."

Adding to environmental fears is the threat of herbicide-resistant weeds. "In some cases, just three or four years of the same chemical on the same field will build up a resistant population of weeds -- particularly with the newer herbicides," Forcella says. The trend towards two-crop rotations could increase that risk, says Paul Groneberg, a crop consultant for Centrol in Morris.

Forcella and others believe that a clearer understanding of the ecology of annual weeds could reduce the use of herbicides and these corresponding economic and environmental costs.

Wanted:
user-friendly science

For nearly two decades, Frank Forcella has been studying "the rebels of the plant world" -- redoubtable outcasts with vivid names like green foxtail, redroot pigweed, lambsquarters, black nightshade, velvetleaf. Scientists know a lot about these weeds, but it's tough to transfer that knowledge to farmers in useful ways, Forcella says.

Weed management models try to do just that. These computer programs sort through complex data about plant biology, economic conditions and weather to recommend weed control measures for specific growing conditions.

In 1994, AURI helped pilot test one weed management model, a sophisticated computer decision aid called WeedSim. Integrating weed seed counts from soil samples with a host of other variables, WeedSim gave specific advice on both mechanical and chemical weed control. On field tests in central Minnesota, WeedSim performed well, providing better weed control and higher profit margins than standard farming practices.

Nevertheless, WeedSim "did not include all the biological information that could be useful for management decisions. (Producers) needed a user-friendly weed ecology model," Forcella says. "That was the stimulus for WeedCast."

The new program is a snap to run, he says. WeedCast requires the user to enter only local daily rainfall, high and low temperatures, plus some key facts about cropping conditions: soil type, tillage system, residue, and weed species. From these few variables, WeedCast predicts the germination potential, time of seedling emergence, and daily growth of 17 common weeds. It charts the results on easy-to-read graphs.

Forcella says WeedCast is a good guide to what is happening out in the field, and a useful forecasting tool for fine-tuning weed control.

A good year for weeds?

For generations, farmers have observed that certain weeds are plentiful in some years, rare in others. The phenomenon, called secondary dormancy, occurs when unfavorable weather causes a weed seed to "shut down" until the next season.

Only recently have scientists identified some of the conditions that induce secondary dormancy in weeds. For example, a single 85-degree day is enough to induce secondary dormancy in giant foxtail. WeedCast incorporates this aspect of weed phenology, predicting how much of the weed seedbank will actually germinate under current weather conditions.

That knowledge can reduce the number of wasteful prophylactic treatments, Forcella says. "If you know ahead of time how many weeds there will be, you can make a more objective decision about whether you need to take any weed control measures at all, and if you do, whether they should be chemical or mechanical."

What's up and when

A vexing question for farmers each spring is: When will all the weeds be up? In good weather, emergence proceeds rapidly. But when the weather is cool or dry, emergence may slow or even stop.

Based on weather inputs from the farmer, WeedCast calculates the total percentage of weed seedlings that have emerged to date. Knowing which weeds are up can help farmers decide if there's time for a preemergent herbicide. It can also help them choose the best time to rotary hoe or apply a burn-down, says Groneberg, who used WeedCast to plan herbicide treatments on no-till fields. It can even tell the experts when it's safe to stop scouting for new weed flushes.

Timing it right

Once weeds are up, farmers face the question of when to knock them down. Timing can be tricky.

For example, Forcella remembers when farmers began complaining that they weren't getting consistent grass control with Accent, a leading postemergent herbicide. Like all postemergent herbicides, Accent is formulated for weeds of a specific height. "Apply too early or too late and it doesn't work well."

WeedCast helps farmers strike when the time is right. The model predicts the daily growth and height of weed seedlings using both observed and predicted weather data. By plugging in the five or seven-day weather forecast, farmers can estimate when a field is ready for spraying.

Weed forecasting can be especially useful for targeting fields under severe weed pressure, Groneberg says. "It fits in with the new philosophy of managing a farm field by field."

Nevertheless, weed management models have not yet gained wide acceptance by farmers, notes Groneberg. And Forcella admits that WeedCast, like other biological management models, probably won't bring dramatic improvements in the bottom line.

Still, he emphasizes, efficient and environmentally friendly weed control depends on a detailed understanding of weed ecology: "Many decisions may be greatly influenced by knowledge of when weeds emerge in spring, and how fast they grow subsequently."

Want to check out WeedCast? It's free software. The fully self-contained and executable program includes instructions for downloading plus a users manual. To download go to the home page of the USDA-ARS North Central Soil Conservation Research Laboratory at www.mrsars.usda.gov. Click on User Products, select Weed Ecology and Management, then follow the WeedCast instructions.