Thinking About Nitrogen Recommendations

April 21, 2000
A news article in the April 14, 2000, issue of Illinois Agrinews about the 1999 National Corn Growers Association corn yield contest winners stated that "The average commercial nitrogen use per bushel of yield was . . . 0.99 for winning entrants, significantly less than the agronomic requirement of 1.1 to 1.2 pounds per bushel." Elsewhere in the story it was reported that the average yield of winning fields was 206 bushels per acre, and that 66 percent of the contest fields had corn following soybean, and about 20 percent had corn following corn.

While these are good yields that undoubtedly represent efficient use of inputs, this statement reflects a common misunderstanding of how nitrogen recommendations are developed and what we should expect them to do for us. First, the recommended fertilizer nitrogen rate for corn following soybean is not "1.1 to 1.2 pounds per acre." That is the suggested nitrogen fertilizer rate for corn following corn, but corn following soybean (as many of the contest fields did) must be given nitrogen credit for the soybean crop. In Illinois we suggest 40 pounds less of fertilizer nitrogen when corn follows soybean.

It is a technical point, but the credit given to corn following soybean is more likely a lower penalty than when corn follows corn. The penalty comes from the fact that as crop residue breaks down in the soil, nitrogen is absorbed from the soil, making less available for the crop. Because soybean leaves much less crop residue, and the residue has higher nitrogen content than that of corn, its breakdown ties up less nitrogen. In fact, when we grow corn following corn after removing cornstalks from the field, the response to nitrogen is similar to that for corn following soybean. And removing soybean residue before the following corn crop does not affect the nitrogen response of the corn crop.

The contestants in the 1999 NCGA corn yield contest apparently used an average of 206 bushels times 0.99 pound per bushel, or about 204 pounds of nitrogen fertilizer per acre. We'll assume that fields where corn followed soybean and those where corn followed corn yielded about the same. We would normally expect corn after soybean to yield more, but we know that the new record yield of 394 bushels per acre in Iowa was corn following corn, and that certainly brought up the average. That field, like some others in the contest, seems to have characteristics that make it uniquely suitable for corn, and in some cases corn is grown continuously on such fields. If these were expected yields, then the recommended rate for 206 bushels per acre would have been 206 x 1.2 = 247 pounds per acre for corn after corn, and 247 minus 40 pounds per acre, or 207 pounds per acre for corn following soybean.

The recommended N rate would have been even less for corn following a legume such as alfalfa, where as much as 100 pounds of nitrogen are credited per acre, and for corn following application of manure, where one-half of the nitrogen in that manure is credited to that year's crop. Given that in 80 percent of these fields corn followed a crop other than corn, and some fields likely had manure applied, the average rate of N fertilizer used was probably close to the recommended rate. Basing the recommended rate on a yield expectation above 200 bushels per acre is something that perhaps only contest participants ought to do routinely, but in any event the actual N rates were about what would have been recommended.

How were the nitrogen recommendations formulated in the first place? They were generated by running field experiments over many years, often with corn following corn and with corn following soybean or some other rotational crop. In every study like this done in Illinois, the response to nitrogen varies greatly with the year, depending on factors such as planting date and conditions, and especially on weather. We expect this: weather, soil moisture, and tillage affect root growth patterns and N uptake; dry weather that affects yield affects how corn uses the nitrogen it takes up; and weather has a very large influence on how much nitrogen soil microorganisms release from soil organic matter.

With such differences over years, the only way to formulate useful predictions that tell us how much nitrogen to use this year is to average previous results over years and then base a recommendation on that average response. Because of year-to-year variability, the recommended rate based on the average response over years is almost never the rate that we would have chosen had we known what the current year would be like. But it is our best guess at what we "on average" expect to happen this year. Some people may have their own reasons for expecting this year to differ from the average year in a predictable way, but it is difficult to have enough confidence in such predictions to act on them.

We recently analyzed 16 years of data from a nitrogen rate study at Monmouth, Illinois. The observed "best" nitrogen rate for corn following corn ranged from 53 to 240 pounds of nitrogen per acre, depending on the year. The predicted best rate of nitrogen based on all 16 years of data was 147 pounds per acre, with a yield of 144 bushels per acre, or slightly more than 1 pound of N per bushel. But when we "looked back" on each year to see how well the actual best rate that year agreed with the predicted best rate, we found that in only 3 years of the 16 did the actual best rate come within 20 pounds per acre of the predicted best rate, and in 7 of the 16 years it was more than 40 pounds per acre different from the recommended rate. The pounds of nitrogen per bushel of yield ranged from 0.6 to 2.1, and in only 2 years of the 16 did it fall within the range of 1.1 to 1.2 pounds N per bushel. The variability of these results did not surprise us; we've seen such swings in virtually every study we've conducted.

There's a tendency to look at such results and conclude that the nitrogen recommendation is certainly useless, if not simply wrong. That's understandable if we expect the recommended rate to be "right" every time we choose to test it. The recommendation is based on long-term, well-run rate studies, but we know that year-to-year variability in weather and in the response of corn to nitrogen is so great that the recommendation looks wrong almost every time we look back on a particular season to see how it worked. General recommendations simply cannot adjust for specific conditions.

This failure of the recommendation to adjust to each season's weather pattern does not mean that the recommendation failed or that the technique is flawed. But unless we can learn to predict the weather, and the response to nitrogen as affected by the weather, we will never do better than to recommend nitrogen rates based on sound studies conducted over a lot of variable years and/or variable locations with similar soils. Our expectations need to be fixed, not the method of producing N recommendations.

The tendency to look back on how well the recommendation "worked" rather than to have research-based confidence in the recommendation does nothing to improve management. For now, using 1.1 or 1.2 pounds of nitrogen per bushel of expected yield and then subtracting all of the appropriate credits for nitrogen are what the research results tell us to do. Let's use these recommendations with confidence, and recognize that second-guessing them is no more useful than second-guessing the weather. Hindsight may be 20/20, but we tend to run into walls and do damage when we're exercising it.--Emerson D. Nafziger and Robert G. Hoeft

Author: Emerson Nafziger Robert Hoeft