You sit down to look at yield maps on your computer and notice some spots yielded more or less than you expected. Or perhaps there’s a big increase or decrease, and then the yield returns to typical levels. Are these changes real, or is there another explanation?
Bob Nielsen, Purdue University Exten-sion corn specialist, works with yield maps from dozens of fields each season. Along with agronomist Jim Camberato and plant pathologist Kiersten Wise, also both with Purdue Extension, Nielsen conducts field-size trials on farms each year. It’s his job to ensure the farmer’s combine is calibrated correctly. After harvest, Nielsen looks at yield maps to see if anything seems out of the ordinary.
Clean up maps
The specialists have no intention of altering true data. The data is yield information. However, there are ways to identify results in a yield map that aren’t due to real-world causes, and to “clean them up” so that maps more closely represent what truly happened.
“We use software programs to remove things in the data that are anomalies,” Nielsen says, “especially if they are affecting treatment strips unequally across the field, and leaving them in the map interferes with analyzing results.”
Some software programs are more complicated than others. Since he’s using data for research comparison purposes, Nielsen uses software that can make precise changes where he knows the map doesn’t reflect reality.
Here are a few abnormalities that can occur, along with how Nielsen deals with them:
End rows. One of the first things Nielsen often does is remove end rows and the first 50 to 75 feet of row length. Why? End rows often are more compacted or vary in other ways, and may differ in yield.
More importantly, he wants to remove the first 50 to 75 feet of row because it takes that long for the combine to refill with grain, and for the yield sensors to begin recording normally again.
“It’s not instantaneous,” he adds. “The combine needs to travel at least 50 to 75 feet before full grain flow occurs and yield estimates are accurate again.”
Water spots. If there’s a low swag running through the field and there’s no yield because the crop drowned out, the best thing is removing it from the map if your goal is comparing experimental treatments in an on-farm trial, and if water damage has nothing to do with treatment effects. This feature may be especially important this year in on-farm trials where excess water created problems earlier in the season.
Weed areas. Foxtail may be growing instead of corn in areas where the crop drowned out, or in areas where weeds got ahead of the corn. If they’re small areas that cut across a trial, Nielsen uses software to remove them. It’s especially important when the wet swag or weedy patch cuts across the trial unevenly.
Gullies. What happens when you reach a gully during harvest? You slow down, right? That affects flow of grain through the machine, which can cause a blip in yield results, Nielsen says.
He typically tries to remove yield within several feet on either side of a gully. Otherwise, the map reflects slowing down for the gully rather than actual yield.
Other cases. Remember that there’s a delay between corn reaching the time plants are snagged by the gathering chains and the time grain from those plants reaches the yield sensor.
This so-called “flow delay” can be 10 to 12 seconds of travel down the field. Not accounting for this time delay in mapping often results in zigzag yield patterns, he says.
Also, be sure to correctly input GPS offset values in your yield monitor or mapping software to account for the GPS antenna being 6 to 12 feet behind the header. If you’re trying to be precise, you may need to take these types of factors into account.
What you can't change after the fact - >>>
What you can’t change after the fact
Bob Nielsen was going to look at yield maps from a Purdue University and Indiana Prairie Farmer experiment at the Throckmorton Research Center near Romney, Ind., and tell which hybrid did better on two distinct soil types in the field-size trial. The goal was testing the multi-hybrid concept theory. Can you prescribe hybrids that do better on one soil type or another?
If you’re looking for that answer here, don’t get your hopes up. You won’t find it. Nielsen, Purdue University Extension corn specialist, ran into a roadblock.
Dave Nanda, a consultant for Seed Consultants Inc., Washington Courthouse, Ohio, selected two hybrids that were distinctly different. He wouldn’t say which should do better where. That’s part of the scientific process. To be a good researcher, he says you should assume no difference and see what results tell you.
Nielsen cleaned up data and removed the end rows and drowned-out spots. But when it came to matching hybrids and soil types, he couldn’t do it. Why? There was a major calibration error in setting up the yield monitor.
The monitor was calibrated for one hybrid, and that hybrid was run first. When the combine operator was asked to calibrate for the second hybrid, he tried his best. Due to changes in the yield monitoring program he wasn’t aware of, he couldn’t do it. Sure enough, weights began varying from actual grain cart weight on the second hybrid. Error went from 0.5% on the calibrated hybrid to as high as 2.5% on the second hybrid. The two hybrids produced very different types of kernels.
“That’s one thing I can’t fix after the fact,” Nielsen says. “There’s no way to go back and adjust yields for one hybrid vs. the other.”
So the data wasn’t available. “But it’s a real good example of why yield monitor calibration is so important,” Nielsen says. “You just needed to calibrate both hybrids.”
Decision Time: Production is independently produced by Penton Agriculture and brought to you through the support of Case IH. For more information, visit beready.caseih.com.