A group of people who work on nitrogen fertilizer met in 2004 to talk about an alternative to the yield-goal-based N recommendation system that had been in widespread use for some three decades. The main concern with the yield-goal-based system was that, as corn yields increased over time, N rate trials were showing that the amount of fertilizer N needed to maximize yields was not going up as fast as yields. On the other hand, the amount of fertilizer N needed on lighter soils, including those in southern Illinois, was consistently higher than the amount suggested using the yield goal.
The result of this meeting was the development of an approach to turn recent N rate trial data into N rate guidelines. For each of several hundred trials in Illinois, we fit a curve to the data then calculated the “(net) return to N” (RTN) – yield produced by N x corn price minus N rate x N price – for each N rate over a range of rates from low to high. For each region – Northern, Central, and Southern Illinois – and for corn following corn and corn following soybean, we averaged the RTN curves from all the trials. The N rate that produced the high point of that average curve was dubbed the “maximum return to N” (MRTN) rate. Because the RTN curves were fairly flat on top, we also decided to include a range of rates over which the RTN was close (the default is within one dollar per acre) to the maximum. So along with the MRTN value is a range, typically about 15 pounds of N on either side of the MRTN, over which the data predicts about the same return to N.
The current version of the calculator that produces the MRTN values based on the latest N response data is located at http://cnrc.agron.iastate.edu/. Seven states – Minnesota, Iowa, Wisconsin, Illinois, Indiana, Michigan, and Ohio – currently use this approach. The calculator uses the trial data from each state or each region within a state, and for trials with corn following corn and corn following soybean to produce MRTN vales and ranges. For Illinois, the calculator runs on data updated through 2016; our plan is to update it every year, adding new data and selectively “retiring” data from older studies. This approach is an economic one, so requires the user to enter prices for corn and for N.
As part of the development of this approach, we looked to see if practices like tillage or if factors like soil type would enable us to provide separate guidelines based on such practices or characteristics. We failed to find anything that would produce different MRTN rates. This doesn’t mean that, for example, different soil types don’t need different amounts of N fertilizer. It simply means that the set of data from, say, trials run in different soil types did not produce differences in N rates that were large enough to justify having separate categories for soil type. The main reason we think this is the case is that N responses vary so much due to things such as weather, soil, and crop conditions.
Over time, with enough additional data, we might be able to make separate MRTN predictions for, say, lighter-textured versus heavier-textured soils. But we’ve also found that year-to-year variations in N response are often very large even within the same field or the same part of a field. This means that it won’t be easy to find factors that show a large enough or consistent enough difference in N response to stand out over time. While it might be tempting to select out and use data that “looks similar” in order to show “differences,” we can’t do that and still claim that the process is unbiased. It would, instead, mean that the results we chose don’t accurately predict future responses.
Here are some questions that have been raised about the N rate calculator, along with answers:
Q: Although the N rate calculator has been around for more than a decade and the Illinois Agronomy Handbook describes it as the best current method for making N rate decisions, it still doesn’t make sense that higher yields wouldn’t need more fertilizer N. But there’s no place to enter expected yields, and actual yields from trials aren’t visible. Are the yields in these trial just “average” yields, and doesn’t that mean that if we shoot for really high yields this won’t tell us to use enough N?
A: To better understand this, go to the calculator site and choose “Illinois” then “Central” (region) and “Corn following soybean” as the rotation. Leave the price of N and corn as the default values or put in different values, then hit “Calculate.” A figure with some numbers at the top will appear; the numbers include the MRTN value, which is 172 pounds N per acre at the default prices, and the range of N rates (158 to 186 pounds of N per acre) over which the RTN is close to the maximum. There is also a number called “Percent of Maximum Yield at MRTN Rate” which in this case is 98%. That means that 172 pounds of N produced 98% of the maximum yield, averaged over all of the trials. This is less than the maximum yield because adding enough N to maximize yield costs more money than those last few added bushels of yield are worth. On the left under “Display Charts,” click on “EONR vs. Yield” and a figure (Figure 1) will appear that has actual yields and the N rate required to meet that yield for each trial in the database.
Over the 245 trials in this database, yield at the optimum N rate (the MRTN for an individual trial) ranged from 100 to nearly 300 bushels per acre, and the optimum N rate ranged from less than 50 to more than 250 pounds N per acre. But there is no correlation between N rate and yield; high yields did not require more N and low yields did not require less N. In about a dozen trials, less than 100 pounds of N produced more than 200 bushels per acre, and in about the same number of trials, it took more than 200 pounds of N to produce less than 200 bushels per acre. So knowing beforehand exactly what the yield would turn out to be would have been of no help at all in knowing how much N to use.
The overall best (MRTN) rate of 172 pounds of N is more than the amount needed in nearly two-thirds of the trials and less than the amount needed in about one-third of the trials. We have no way to guess beforehand on which side of this rate the amount needed for an individual field would fall, but we do know that needing less than this amount is more likely than needing more than this amount. Using more N than the MRTN means, on average, spending more on adding N than we get back in added yield. So the MRTN is the best guess at the rate that maximizes return to N while minimizing the amount of N left over after the season.
Q: Since high-yielding corn obviously takes up more N than lower-yielding corn, how can it possibly not need more N fertilizer?
A: It is absolutely true that higher-yielding corn takes up more N: at its maximum N content during the season, which is usually during the kernel dent stage, the crop typically contains 0.8 to 1.0 pound of N per bushel of grain yield. So a 220-bushel crop might have 220 pounds of N in the plants, even if, say, only 160 pounds of fertilizer N was applied. If we estimate that 75 percent of the fertilizer N (120 pounds) got taken up (that’s a higher percentage than is often measured), that means that 220 minus 120, or 100 pounds of the N in the plants had to have come from a source other than fertilizer. Other than the small amount of N (usually less than 10 pounds per acre) fixed by lightning during storms, this only place this N could have come from is the soil.
Q: Can a soil really supply 100 pounds of N to a crop?
A: An acre of soil 6 inches deep weighs about 2 million pounds, so one percentage point of soil organic matter is about 20,000 pounds of SOM in that acre. Soil organic matter is about 5 percent N, so one percent SOM in the top 6 inches of soil would mean about 1,000 pounds of N per acre. This N is in organic form, which is not available to plants. A soil with 12 inches of topsoil and with 3 percent OM would contain about 6,000 pounds of N. On average under Illinois conditions, an estimated 2 percent of soil organic N is mineralized – freed from organic matter and available for crop uptake – each year. Two percent of 6,000 pounds would be 120 pounds of N per acre. So, while such a soil could provide 120 pounds of N to a crop, the actual amount provided depends on soil conditions – mineralization is carried out by soil microbes, so is slower when soils are cool, dry, or wet. Growing conditions can also affect the ability of roots to take up N, and too much when soils are warm or in light-textured soils can result in loss of mineralized N. This variability in soil N supply is what makes it so difficult to guess at how much N the crop will get from mineralized SOM, and so how much it needs from fertilizer.
Q: If more soil organic matter means more N for the crop, shouldn’t we decrease fertilizer N rates in high-organic-matter soils, including in higher-OM parts of a field?
A: If yields were uniform across the field that might make sense. But fields or parts of fields with higher OM tend to yield more, usually because organic matter can store more water, and sometimes nutrients, for the crop. Higher yields mean more plant N uptake, and so in higher-OM parts of a field, increased N uptake is matched with higher amounts of N supplied by the soil through mineralization. This doesn’t always happen – for example lower-lying parts of a field might have higher OM but also be more prone to damage to stands or roots from standing water, which can limit yield, N uptake, and the amount of soil N available to the crop. The soil is usually not capable of providing all of the N the crop needs, however; even when corn (following soybean) yields, say, 180 bushels per acre without N fertilizer, adding N fertilizer almost always adds yield. Some states include (or once included) SOM in their N rate calculations, though, and the amount of soil organic matter might make a reasonable basis for site-specific N management (that is, to apply less N where SOM is higher) within fields. Still, higher yields in such areas can increase the need for N more than higher OM increases the amount of N supplied by the soil.
Q: While I see that selecting rotation (previous crop) is required to run the calculator, why is there no place to enter the adjustment for the soybean N credit?
A: The data used to produce MRTN values for corn following corn is from trials in which corn followed corn, and for corn following soybean, is from trials where corn followed soybean. Most of these trials were in separate fields, and this was done on purpose so we could make separate calculations for the two rotations. In central Illinois, for example, there are 245 corn following soybean and 152 corn following corn trials currently in the database. So the soybean “N credit” – or more accurately, the corn-following-corn “N penalty” – is already built in. The difference between soy-corn and corn-corn MRTN rates is a little more than 40 pounds of N (which was the value of the soybean N credit in Illinois under the yield-based approach) in northern Illinois; a little less than 30 pounds of N in central Illinois; and only 10-12 pounds in southern Illinois. These differences are consistent with the idea that as we move south, more corn residue breaks down by spring and soils are warmer at planting, so residue effects on soil temperature and moisture – and so on mineralization and on N tieup – are less noticeable. It is important not to subtract any “credit” from the MRTN rates for corn following soybeans; doing so will mean not applying enough N.
Q: What if I’m applying manure?
A: To find the MRTN, run the calculator using the price of commercial fertilizer N, since fertilizer N is usually used to “top up” the N rate (if any more N is needed) in addition to the N from manure. The amount of N from manure is based on rate, type, and first-year availability estimates that can be found in a number of sources, including the Illinois Agronomy Handbook at http://extension.cropsciences.illinois.edu/handbook/pdfs/chapter09.pdf. Failure to credit manure N adequately is a common reason why corn in some fields has access to a total N supply that exceeds, sometimes by a considerable amount, the needs of the crop.
Q: Should I include the N from MAP or DAP in the total fertilizer rate?
A: All forms of N that we apply and that will be available to the crop should be counted in the N rate we apply. MAP and DAP contain N in the ammonium form, so it’s safe from loss at the time of application, though it needs some rainfall to move it into the soil to keep it in place. If we apply these materials in early October, higher temperatures (with some soil moisture) will usually mean that some portion of the N will convert to nitrate before winter. So waiting until soil temperatures are below 50 degrees (but before the soil freezes) or waiting until spring to apply will usually increase the availability to the next crop. We’re doing a study now to try to measure how much of the N from DAP is available following application in the fall or spring. Until we have better answers, I would suggest taking half credit for the N in MAP or DAP if these are applied before mid-October, three-fourths credit if applied in the fall when soils are cooler than 50 degrees and there isn’t heavy rainfall within the six weeks after application, and full credit if application is in the spring, within a few weeks of planting.
Q: Shouldn’t fertilizer N rates differ for different forms of N and different application times – for example, don’t we need less N if we apply in the spring rather than in the fall, and even less if we sidedress most of the N?
A: Most N fertilizers contain most of their N in the form of ammonia or urea, both of which convert to ammonium in a moderately moist soil. Ammonium is held tightly on soil exchange sites, and so is safe from loss. But the process by which bacteria convert ammonium to nitrate begins soon after N fertilizer is applied, and the conversion rate increases as soil temperature increases. Once N is in the nitrate form, it can move in the soil, and can move with water down to tile lines or to beneath the rooting zone. Nitrate can also be denitrified – converted to a gaseous form and lost to the air – if water stands for a few days when soils are warm. The reason we might consider adjusting rate based on N form or application timing is to counter the risk of loss. It is, though, better to manage N to decrease loss than to apply more N in order to counter loss the risk of loss from applying fertilizer N in ways that more often lead to loss.
We can lower the risk of N loss either by slowing the conversion of ammonium to nitrate, or by delaying application to shorten the time between application and crop uptake. We can slow the conversion to nitrate by using inhibitors – for example by adding N-Serve to fall-applied ammonia and by waiting until soils are cold to apply ammonia. Delaying ammonia application to spring is even more effective than applying later in the fall; we have found in recent comparisons that waiting to apply ammonia in the spring (without N-Serve) lowers the rate needed by about 20 pounds per acre, compared to fall ammonia with N-Serve. But compared to applying all of the N at planting, we have not found that we can consistently use less N if we wait to apply some or most of the N as sidedress. That’s because our sampling shows that we typically lose less N than we think between planting and early June when N uptake rates increase, and also because the crop seems to benefit sometimes from having more (or all) of the N available at planting in order to minimize the chances that the crop will develop deficiency. Loss of (nitrate) N only occurs when soils are wet, so little N will be lost if the soil stays relatively dry, even if all of the N is present in the nitrate form.
Q: Why can’t I compare different forms and prices of N on the calculator to see which are most cost-effective?
A: We do not have enough N source comparisons to allow us to calculate different MRTN values with use of different forms or times, though direct comparisons are showing that we might be able to make some adjustments, as discussed above for fall- versus spring-applied N. Most of the calculator database consists of trials with N applied in the spring; some with all of the N applied in early spring, some with all of it at planting, and others with some at planting and the rest as sidedressed N. There are also a number of trials with fall-applied N. But grouping trials by N time and form does not produce responses that are different enough to justify breaking down the database this way. The calculator allows the comparison of up to four different N prices at a time, and those can be from different forms of N. The MRTN rate calculation depends on the ratio of N price to corn price; the higher this ratio (high N price or low corn price), the more yield increase it takes to pay for the last amount of N, and the lower the N rate. So N forms that cost more per pound of N will produce lower MRTN rates. Either price per ton of fertilizer or price per pound of actual N can be put into the calculator, which converts one to the other.
Q: Can I calculate MRTN for fields where I grow cover crops?
A: Not with confidence, at least not yet. Allowing cereal rye or annual ryegrass ahead of corn to grow well into April almost always increases the amount of N needed by the crop, and even with more N the crop may not yield as well. Grass cover crop residue is similar to corn residue in that the breakdown of residue in the spring ties up some N, decreasing availability to the crop. Some of the tied-up N will become available to the corn crop, but the amount and timing of its release is unpredictable, and in most cases some of it will be released too late to be taken up by the crop that season. So it’s not a good idea to take a credit for the N in a grass cover crop to lower the fertilizer N rate. A cover crop legume that grew in the previous season (say, red clover planted into wheat the previous summer), that grew enough in the spring to have fixed some N, and that is killed early enough to break down and to supply some of that N to the corn crop would justify lowering the fertilizer N rate for corn, by perhaps 50 pounds per acre. Growing that amount of dry matter and fixing N would likely require leaving the cover crop grow into late April or, in northern Illinois, into May, so might delay corn planting. If a grass cover crop that overwinters is killed early enough (in March) so that there’s little intact residue at the time corn is planted, N rates and corn yields will probably be little affected by the cover crop. As a general observation, though, cover crop residue is not a particularly reliable source of N for the corn crop that follows it.