Issue No. 24, Article 7/November 11, 2005
New Nitrogen Rate Guidelines for Illinois
First, a summary of this lengthy article: University agronomists from a number of Corn Belt states have been collaborating on a new approach to nitrogen rate recommendations. The approach uses recent research data and applies economics to the decision on N rates. Over the last few months, N prices have risen sharply, raising questions about what rates should be for the 2006 corn crop. What will these N rate recommendations look like for Illinois?
- The Proven Yield method now used for N recommendations produces rates that are higher than needed to maximize return to N, especially for corn that follows soybean, where yield levels are high and when the N price increases.
- The guidelines will now be to apply N rates within a range, defined as the rates that provide returns close to the maximum economic return to N. Ranges will be based on N response data from some 250 Illinois trials.
- The recommended range will change when the price ratio between N and corn changes.
- There will be separate ranges for corn following soybean (SC) and corn following corn (CC). The "soybean N credit" will no longer be subtracted, though it will be included directly, since the SC range will be based on SC research and the CC range on CC research. (Adding the soybean N credit to the SC recommended range to produce "total N supply" might help with acceptability of this approach but should no longer be subtracted from the recommended CC range or otherwise be part of the recommendation.)
- Calculating separate ranges based on the data suggests that for southern and central Illinois they are very similar, but for northern Illinois the range is actually less than for the rest of the state. A single recommended range for all of Illinois, while conservative for northern Illinois, may be acceptable until more data become available.
- For CC there is some correlation between yield and the N rate it takes to produce that yield, indicating that some adjustment based on yield is justified. We suggest using the expected yield to move to higher or lower rates, while staying within the suggested range.
- Using a corn price of $2 per bushel and an N price of 30 cents per lb ($492 per ton of anhydrous ammonia), the recommended range for SC in Illinois is 122 to 162 lb N per acre. This range drops by about 10 lb if the N price increases by 6 cents per pound of N ($100 per ton of anhydrous ammonia), providing the corn price stays the same.
- For CC using the same prices, the recommended range is 137 to 174 lb N per acre. Within this range, increase the N rate by 0.4 lb of N for each bushel of yield expected above 148 bushels per acre. (If yield is expected to be 180 bushels per acre, add 32 x 0.4 = 13 lb N to the 156 lb N that is the midpoint, so use at least 169 lb N.)
- Producers and advisers will need to decide where to set N rates within a recommended range. This decision can involve risk management philosophy, previous experience with N rates, and possible considerations regarding N movement into the environment.
There has been a great deal of recent work in Illinois on the response of corn yield to fertilizer N rates. Much of this work has been stimulated by the need to provide a stronger basis for determining economically optimal rates of N application and to minimize environmental consequences of corn production. The large variability in N response we have seen has led us to look for a new approach to using rate data to formulate use guidelines.
In the 1950s and 1960s, the observed variability in corn's response to N was handled by making recommendations to use N rates within a wide range. For example, the 1971 Illinois Agronomy Handbook suggests N rates "between 125 and 250 lb per acre" for corn following corn. In the 1970s, this wide range was narrowed by using the results of trials averaged over years, then calculating a factor that related N rate to yield. This resulted in the "Yield Goal" or "Proven Yield" method, the guideline currently in use in Illinois. The current recommendation is to apply "no more than" 1.2 lb of N for each bushel of expected yield for corn following corn, with credits given when corn follows a legume or when manure has been applied to the field. This recommendation has worked quite well most of the time, especially when N costs have been relatively low compared to corn price.
Corn yield is directly related to the amount of N the crop needs to take up. Harvested grain has about 0.8 lb N per bushel, and the crop residue (stalks, leaves, husks, cobs) left in the field has about 0.45 lb of N per bushel, for a total of about 1.25 lb of N per bushel. That means that a 180-bushel corn crop has a total N content (or requirement) of about 225 lb. Some of this 225 lb of N comes from soil organic matter and some from the air. We don't have a good way to predict how much N will come from the soil in a season, but the amount typically ranges from very little to more than 100 lb, depending on soil organic matter content, soil moisture, temperature, and how the crop grows.
Let's assume here that the soil provides 80 lb of N to the crop. There is also some N from the air, fixed by lightning or emitted from industrial plants and brought down by rain. This might be about 25 lb of N per acre per year. In our example, the soil and air thus provide 125 of the 225 lb of N the crop needs. In our studies where corn follows corn, yields with no N fertilizer are about 60 to 80 bushels per acre, suggesting that the crop gets about 80 to 100 lb of N from the soil and air. For corn following soybean, yields are 80 to 120 without fertilizer N, indicating that the soil and air provide 100 to 150 lb of N. Our example is somewhere in the middle.
In our example, if 125 lb of N comes from the soil and air, the other 100 lb needs to come from N fertilizer. Because some fertilizer N is immobilized by soil microbes and some might be lost due to leaching or denitrification, not all applied fertilizer N is available to the crop in the year it is applied. The amount available varies depending on N form, N loss, and other soil and weather factors. For our example, we'll assume that two-thirds of applied fertilizer N is available to the plant. To provide the 100 lb of N the crop needs, we would need to apply 150 lb of N as fertilizer.
We can now start to understand why the actual N fertilizer requirement for a crop in a field is so unpredictable. Even though the proven-yield recommendation was based on data from field trials, we have always known that the actual amount of N fertilizer needed in a field in a given year is only as predictable as the corn yield (which sets the N need by the crop), the N supply from the soil for that year, and possible N losses. Yield and N supply from the soil are not at all predictable, so the actual amount of fertilizer N we need to apply is a best guess based on previous studies. The actual amount of fertilizer N needed is seldom very close to this guess. But it's the best we can do.
When the ratio of N cost to corn price is low, there is little economic penalty to setting the N rate relatively high, since the financial consequence of applying more N than the crop needs is relatively small. There are several difficulties with this approach, however. First, corn yields have been very high in many Illinois fields in recent years, resulting in a rapid increase in "proven yield" in many fields, which in turn calls for increased N rates. With an expected yield of 200 bushels per acre, applying the 1.2 factor and subtracting the 40 lb N credit for soybean results in a suggested application rate of 200 lb N per acre. In only five of 40 site-years in one of our studies has the economically optimum N rate (EONR) exceeded 200 lb N per acre, and the average EONR has been only 132 lb N per acre. If we use a 50% higher N cost to reflect recent price increases, then the EONR was at or above 200 lb N per acre in only two of these 40 site-years, and the average EONR was only 120 lb N per acre.
A second problem with the proven yield method when corn follows soybean is that our recent research shows very little correlation between yield and the N rate needed to reach that yield. In other words, yields do not dictate required N rates, and applied N rates do not dictate yields. So even if we somehow knew what a yield was going to be, our results tell us that we would have little idea what N rate would be needed to get it. As an example, in more than 40 separate trials in one study we are doing, yields above 200 bushels per acre for corn following soybean have required, depending on the year and location, anywhere from 65 to 202 lb of N per acre. In cases where yields were less than 130 bushels per acre, the N rate has ranged from 34 to 180 lb N per acre.
For corn following corn, the N rate to yield ratio we calculate using recent data is often in the same range of 1.1 to 1.2 lb N per bushel that was calculated decades ago. There is also some correlation between yield and the N rate in continuous corn than in corn following soybean, even though the relationship is not very strong. Still, the use of expected yield to suggest an appropriate N rate is more justifiable for corn following corn, and we think that at least some adjustment based on expected yield level makes sense. Based on our results, it seems appropriate to treat corn following corn and corn following soybean as different "systems" with regard to setting N rates rather than using a single approach with an adjustment for corn following soybean. The new guidelines reflect this.
Working with agronomists in other Corn Belt states, we have developed a new approach to making N rate guidelines. The approach uses data from most of the recent N response work conducted in Illinois, including both small-plot and on-farm results. The method involves calculating net return to N fertilizer (RTN) based on data from each trial, then averaging these returns (at each N rate increment) over appropriate trials.
To do this, we first "fit a function" to (find a mathematical relationship between) N rate and yield data for each trial. We use that function to calculate the yield increase from N (above the baseline yield without N fertilizer), the gross return (yield increase times the price of corn), and RTN, which is the gross return minus the N cost (N rate times N price) at each N rate. For the example here, we will use an N price of 30 cents per pound and a corn price of $2 per bushel.
The database we used to formulate new N rate guidelines for Illinois consists of 172 separate trials where corn followed soybean, with a mixture of on-farm strip trials, small-plot trials at research centers, and small-plot trials run in production fields. We first calculated RTN at each N rate for each N rate trial in the database and then averaged the RTN values over all trials. We then use this "Net RTN" curve as the basis for the N rate guidelines. There is a high point on this curve, which we call the maximum return to N, or MRTN. This is the N rate that, averaged over all trials, provided the maximum profit to the use of nitrogen.
Using the N price of 30 cents per pound and corn price of $2 per bushel, we find that the MRTN for Illinois is $85.74 per acre, and it comes at an N rate of 142 lb N per acre. The RTN curve is relatively flat on top, indicating that the return to the investment in N is about the same over a range of N rates. We arbitrarily chose to set this range as the range of N rates within which the RTN was within $1 of the MRTN. This N rate range for corn following soybean is 122 to 162 lb N per acre, which in this case is 20 lb of N on either side of the N rate that gives the MRTN. Based on this, we would set the N rate guideline for corn following soybean in Illinois at 122 to 162 lb N per acre.
This N rate range changes depending on the relative prices of corn and nitrogen. When the N price goes up and corn price doesn't, the range shifts to lower N rates, at the rate of about 10 lb of N for each 6-cent increase in the price of a pound of N ($100 more per ton of anhydrous ammonia). The range also gets narrow as N price increases, reflecting the fact that the financial penalty for being over or under the "best" N rate increases more quickly.
Why a range of suggested N rates, and how do we decide where to be within that range? N rate ranges are used as recommendations in several other states (including Iowa), and they've generally become accepted over time. But the main reason we chose to use them was that the data show that returns to N are similar over a range of rates, so there's no indication that we gain much by trying to peg N rates more exactly. A range also allows people to apply personal approaches to risk and to find rates with which they are comfortable, whether that's toward the middle of the range (a reasonable "default" number), on the high end due to experience or a landlord who requires that the crop "never be deficient," or toward the low end to better prevent movement of leftover nitrate into water or as a fiscally conservative approach.
We have data to take a similar approach for corn following corn, though our database is not as large. Because corn following corn usually yields less without N fertilizer than corn following soybean, the gross and net returns to N application in corn, after subtracting the baseline of yield without N, are usually greater for corn following corn. Using our database of 82 trials for corn following corn and the same prices as we used above, the MRTN for corn following corn is $106.49 per acre, and this occurs at an N rate of 156 lb N per acre. Yield at this N rate is 148 bushels per acre. The RTN is within $1 per acre of the MRTN over the range of 137 to 174 lb N per acre.
These numbers for corn following corn are only about 15 lb N per acre more than those for corn following soybean. What happened to the 40 lb "soybean N credit"? When we do these studies in the same trial in the same field, the difference in the N rate needed to maximize RTN is about 40 lb on average. The main reason we see a smaller difference when using these databases is because many of the trials we are including were done in different fields, so the comparison is less "controlled." It is also possible that corn following corn tends to be in fields of higher productivity, where the N supply from soil organic matter might be larger, so that the need for fertilizer N might be less. Such a difference might be picked up in field trials but not in side-by-side trials.
Besides the need for higher N rates at lower yield levels, corn following corn also differs from corn following soybean in an important way: when corn follows corn, there is a relationship between yield and the N rate it takes to reach that yield. This relationship is not especially strongwe get the same variation due to soil, crop, and weather patterns as we do when corn follows soybean. But it is clearly not due just to chance. If we look at the results from all 82 corn-following-corn trials in our database, we find that the optimum N rate increases by about 0.4 lb for each bushel of increase in yield, starting at about 100 bushels per acre.
One reasonable way to use this information is to subtract the 148 bushels per acre (the yield at MRTN mentioned above) from the yield we expect to get, multiply this difference by 0.4 lb N per bushel, and add (or subtract if we expect less than 148 bu per acre) this amount of N to the N rate at the MRTN (156 lb per acre in this case) as a way to adjust N rate within the suggested range.
Expected yield is best estimated by averaging yields over the past five seasons in a field, but in this case that needs to be yield of corn following corn. If corn has not followed corn in the field or has done so only once or twice, then use the yield average for corn following soybean in the field and subtract 10%. As an example, if the yield average for corn following soybean in a field is 190, use 171 as the expected yield when corn follows corn in that field. Subtracting 148 gives a difference of 23 bushels, and this would give an adjustment of 23 x 0.4 = 9 lb of additional N. That may not seem like much, and it is still possible to use more N and stay within the range, but using expected yield to nudge the N rate up or down is justified by the fact that higher yields of corn following corn do need a little more N.
One of the useful features of this approach is that it allows us to use the database as it exists, but we can easily add data as we generate them from more N response trials. We need to generate enough data so that fields in Illinois, or in each region of Illinois, are well represented in the database we use to calculate new guidelines. This will take hundreds of trials, but in the end we will have much more confidence in the guidelines.--Emerson Nafziger and Bob Hoeft