Issue No. 22, Article 7/September 2, 2005
Nitrogen for Corn in 2006
Recent reports suggesting that nitrogen rate recommendations for corn in Illinois are inaccurate have created a great deal of confusion. This unfortunately is coming at a time when N prices are record-high, yields in some Illinois corn fields were reduced by drought in 2005, and corn prices have been held down by reports of higher supplies and some shipping problems.
We recognize that N rate recommendations have always tended to be a moving target, and that changes in hybrids and management have probably improved the efficiency of fertilizer N use. We applaud efforts to try to find ways to predict how much N the soil will supply to the crop each year; such an approach could help us to know better how much fertilizer N we need. We are not yet there, however. Thus statements like "Forget what you know about N", without a carefully-drawn and clear alternative, are without any merit whatsoever.
We fully recognize the limitations of the proven-yield method (N rate equals 1.2 times proven yield minus credits for soybean or other legume or manure), and that this method generally results in N rates that tend to be high for many fields. Having N at "almost always adequate, so sometimes excessive" rates has been the goal for most producers, given that N price has generally been low relative to the corn price. This method was based directly on data from field studies, which have tended to show that corn following corn responds to N up to the point when the N rate to yield ratio is about 1.1 to 1.2, depending on prices of N and corn. It was not simply pulled from the air, nor was it directly based on the amount of N found in a corn crop at a certain yield level.
Adjustments for soybean and other previous practices were based on direct comparison trials, in which we have found that corn following soybean responds less to N than corn following corn. Thus was born the "soybean N credit", which is an adjustment in the N rate to reflect the lower response to N fertilizer following soybean. Our most recent research confirms that the best N rate for corn following soybean is in fact about 40 lb of N per acre less than when corn follows corn. This difference might change when we have enough data to allow us to "decouple" N rates following soybean from those following corn, with separate recommendations for each. That will mean that adjustments for N "credits" won’t be needed, even though we know that corn following corn needs more N than if corn were following soybean in the same field.
Our newest approach has been to use the large amount of N response data generated by a number of recent and current research projects in Illinois to narrow in on what N rate is likely to produce the best return to N fertilizer in the field. To do this, we take the yield increase from N (above the yield without N) times the price of corn, and then subtract the cost of N (rate times price) at each N rate, ranging from zero to 200 lb or more. That produces a "return to N" (RTN) number, in dollars per acre, for each rate of N.
We have calculated the RTN for some 250 research trials, and then averaged these values over appropriate groups of data, for example, by geography or by general soil type. Using this approach allows us to calculate the range of N rates over which return to N is maximized. It is dynamic, in that we can adjust N rate guidelines based on prices of corn and N fertilizer. This approach depends heavily on having a lot of good data on corn response to N, and such data can only come from careful studies over a wide range of conditions. We will continue to carry out such studies in order to improve our confidence in the results. In the meantime, we think that current data provide reasonable guidelines for using N fertilizer on corn.
Using data from some 170 trials corn-following soybean conducted over the past decade in Illinois, we find that, with corn priced at $2.00 per bushel and N priced at 25 cents per lb, a rate of about 150 lb of N produces the maximum return to N (MRTN), with return being close to the maximum (within 1 dollar per acre) over the range of 130 to 175 lb of N. The average yield was 173 bu per acre over all trials at the MRTN (150 lb N per acre), which after the "soybean adjustment" makes for a ratio of about 1.1 lb N per bushel.
Among all of these trials, though, the best rate of N was NOT higher every time corn yield was high; there was no relationship between yield and the N rate required to reach that yield. How can this be? It can only be because the soil supplies widely variable amounts of N to the crop, due to factors such as soil organic matter content, soil depth, drainage, rooting depth, health of root systems, etc.. It is not unusual to see optimum N rates vary by 50 to 100 lb per acre over years when research trials are run annually in the same field. So far, it has been impossible to predict what the best N rate will be at the beginning of the season. Our approach thus acknowledges such variability, but at this point all we can do is to take an "average" approach to predict the N rate at which N will provide the greatest return. If and when we learn how to predict N fertilizer need before the season, we’ll be able to take a more exact approach.
If the corn price stays at $2.00 and N cost goes to 30, 35, or 40 cents per lb, the N rate providing the maximum return to N (MRTN) decreases, to 140, 130, and 122 lb N per acre, respectively, or about 2 lb less N for each 1-cent increase in the cost of a lb of N. The range over which return to N is relatively flat also shrinks at high N costs, to only about 35 lb (105 to 140 lb N per acre) when N costs 40 cents per lb. That means that the penalty for applying "safe" amounts--high enough to assure sufficient N under all conditions--increases as N gets more expensive.
For corn following corn, with a somewhat smaller database (some 80 site-years), our approach calculates maximum return to N at about 165 lb of N per acre, with a range of about 145-185 lb N, when corn is at $2.00 and N costs 25 cents per lb. As N cost rises, the rate at which N provides maximum return drops, again by about 2 lb N per acre for each 1-cent rise in N cost. When N costs 40 cents per lb, the rate expected to produce the highest return is only 137 lb per acre, and the suggested N rate range is 120 to 155 lb per acre.
Unlike the case when corn follows soybean, though, there is some tendency for corn following corn to need more N when yields are higher. This is likely because corn residue ties up some nitrogen, increasing the need for fertilizer N. It thus might be appropriate to adjust N rate within the suggested range, using higher amounts if there is reason to anticipate higher yields. We don't yet know what form this should take, but one possibility would be to take the N rate providing the maximum return to N and adjusting it up or down based on whether corn yield is expected to be higher or lower than average.
While we recognize the need to continue to improve nitrogen management, we think that N rates used by most producers are probably not too far from where they should be. One exception to this is in fields where corn follows soybean, and where proven yields, after the high yields of recent years, might be as high as 200 bu per acre. Using the proven-yield method, such yields call for N rates close to 200 lb per acre. Compared to the 140-lb rate that produces the maximum return with 30-cent N, applying 200 lb of N decreases profit by more than $6 per acre. When N costs 40 cents per lb, the loss in profit from applying 200 lb of N is more than $12 per acre. This doesn't count the cost of sending leftover N into our rivers. Such rates are simply higher than they should be.
So, while the proven-yield method has been mischaracterized as something it is not, it is clear that changes in hybrids and management in recent years have brought on the need to think again about N rates. Our best data tell us that N rates should be decreased in cases where rates are in the upper reaches of the current recommendations. For many, especially those who have moved to cut rates in recent years, the new approach is confirming that the economic consequences of such rate reductions may well be positive.
From our data, we know that there can occasionally be years when the N rate needed to maximize yield in a particular field can be as high as, or higher than, the highest rate recommended under the proven-yield method. There can also be years, even in the same field, when the N rate needed to maximize return to N can be much lower than even conservative methods might estimate. It helps to have an idea of how often to expect either of these extremes to occur, but we are far from being able to predict, and thus prevent, such occurrences. Until we can do so, we need to take an approach that averages a lot of response data together, in a way that allows up to understand how much N it takes to maximize return, and how much penalty there is for having rates too high or too low.
One final point: The recent attempt to discredit current N recommendations has included a contention, based only on the casual observations in very different fields, that high corn plant population require higher N rates. This was an attempt to explain why "higher than expected" N rates apparently were needed in some fields. The only way to test such a notion is to grow a corn hybrid at different populations and different N rates, all in the same field. We did this, and averaged over nine site-years, there was no difference in the optimum N rate among populations ranging from 25 to 40 thousand plants per acre. Others have shown the same results, and we know of no published information supporting the contention that high plant populations require higher N rates. Because the use of "high" plant populations has become known as a good management practice, the false belief that "high" populations need more N only encourages over-application of N.
When we acknowledge that optimum N rates are not related to yield level, we are also accepting the fact that other factors--soil, plant, or management--that might affect yield level may not necessarily affect N fertilizer need at the same time. Such connections may seem obvious, but they just don't operate consistently in the complex world of crops growing in fields with the weather we get. It is dangerous to use "common sense" to try to make such connections when the evidence shows clearly that the connection does not exist. This false contention lends support to the clearly erroneous notion that, because yield levels define N need, then surely N rate must limit yield levels, especially in fields that don't yield very well. It’s just not the way the world works, and to suggest otherwise is irresponsible.
For this fall, it should be business as usual for most Illinois producers in terms of N application, though some downward adjustments in rates are in order for some, based on our new approach. This includes waiting to apply until soil temperatures are 50 degrees or less (this usually means late October or early November), counting all forms of N, including that in DAP or MAP, and stabilizing fall-applied N to reduce N loss potential. For some, there may be difficult decisions due to price or form differences between fall and spring application. While we do not have the most convincing data to suggest that spring rates can be lowered from fall rates, much of our N response data were generated using spring applications, so they should be conservative if spring-applied N is occasionally used more efficiently than fall-applied N. Where corn in 2006 will follow corn that yielded less in 2005, some credit for unused N should be given.--Emerson Nafziger and Bob Hoeft