Providing crops with adequate plant-available nitrogen is very important, especially to dairy farmers, because it has a profound effect on crop yields and quality. While all plant available nutrients can change forms and become more or less available, nitrogen is particularly difficult to manage. Plant-available nitrogen in a given undisturbed field can increase as organic matter decomposes (mineralization), and can decrease via leaching, volatilization, denitrification, or immobilization. To make matters worse, each of these factors are affected by management: tillage, manure incorporation, levels of soil organic matter, timing of manure application, temperature, etc.
All those variables make it hard for farmers to know if the nitrogen they’re applying is actually making a difference. There’s so much lag time between the time you put the fertilizer on the field and the time you cover the bunker silo for the last time during the harvest season, that the cause-effect relationship may be less apparent. You’re left wondering, “Was this terrific yield really the result of the fertilizer I put on, or was it the amount of manure that I put on? Maybe we just had good rain? Or was it the Sna-Koil fertilizer additive I bought?”
The good news is that there are two tools out there to help you figure out how much nitrogen to apply without leaving money on the table. Between them no one should find themselves in a situation where they have no idea how much sidedress N to apply to their 2014 corn crop. Depending on your scale, proper use of these tools (together with common sense) could easily improve your bottom line by tens of thousands of dollars per year, both by increasing yield and by avoiding sidedress-nitrogen applications where they are not needed.
PSNT (Pre-sidedress Nitrate Test)
In the early-1990s the pre-sidedress nitrate test (PSNT) was developed as a direct nitrate measurement that was used to predict the need for sidedress nitrogen on corn (i.e., how much more N you need to apply to attain your realistic yield goal). It is a snapshot of the soil nitrate (NO3-) concentration at the time the samples are collected. The sidedress-N recommendation is based on research that correlates soil nitrate concentrations when corn is 8-12” tall with the total amount of plant-available nitrogen that will be released by soil organic matter over the course of the growing season.
Basically, PSNT is like saying, ‘I measured four cords of wood in the shed in October, and that should get me through an average winter.’ But what if the winter is not average? It might be REALLY cold; winter might drag on and on; someone might help themselves to my wood; the shed could burn down; or the wood might be greener than average. Similarly, the PSNT gives good recommendations in a normal year but may not accurately predict the optimal sidedress nitrogen rate when conditions are abnormal.
While the PSNT typically offsets the cost of sample processing and labor (by far), you may not have used it as often as you could have because it:
- needs to be done during an otherwise busy time of year (starting when the corn is about 6” tall).
- takes a significant amount of time to do properly because fields need to be subdivided and sampled according to soil type, features of the land, and management history.
- requires that samples be taken to a depth of 12”, which can be challenging in stony soils.
- does not always demonstrate a payoff within a week. When fertilizer costs are reduced, the payoff is immediate. Yield benefits are not experienced until harvest and sometime are attributed to other factors.
That last weakness of PSNT is important. Which of the variables was it that gave us a better harvest? Can interacting factors that affect soil fertility dynamics and crop yield even be modeled? The answer is ‘Yes’ on both counts.
Adapt-N
Cornell soil scientists have developed a program (Adapt-N) that models nitrogen behavior in agronomic soils. This model includes as many of the relevant variables as they have data to support, as well as historical and real-time weather data from each site studied. That means that it can model what’s happening to the soil in your field to give you a better idea of how nitrogen needs are changing and what you’ll need to get the best results possible. The table below, from the Adapt-N website, shows you what the program was designed to do. While Cornell maintains control over the evolution of the tool, Adapt-N has been licensed to Agronomic Technology Group. Depending on your scale, Adapt-N will cost about $2-3/acre this year. A recorded webinar explaining the new user interface and fee structure can be found here .
Comparing PSNT and Adapt-N
The PSNT works well in a ‘normal year’ but consistently over-recommended sidedress N by about 30-50 pounds per acre on the studied fields in 2013. This is not surprising because heavy rain just before the tests were collected leached much of the existing nitrate from the top 12” of soil. The logic of the PSNT says, ‘low soil nitrate concentrations now means that soil nitrate concentrations will continue to be low and therefore lots of sidedress-N is necessary to meet yield goals.’ In 2013 sidedress-N recommendations from the PSNT were often over 100 lb of actual N per acre.
Adapt-N had a good sense of whether more or less N was needed, but also seemed to generally under-recommend N by 20-30 lb/ac and sometimes more. With our approach, it was impossible to determine if this was due to weaknesses in the model itself or imperfections in the information that we were feeding into the program (manure analysis, soil organic matter levels, etc).
Recommendations given by any model cannot be better than the data that is fed into the model: garbage in, garbage out. That being the case, neither the PSNT nor Adapt-N should be used without common sense. If either tool generates a recommendation that is significantly outside of what you consider to be reasonable or normal for the conditions in a given field, other measurements should be taken and/or the data you entered into the program should be reconsidered. Obviously, technology has not yet eliminated the need for common sense!
Overall, I believe that Adapt-N will be much better than the PSNT for predicting the need for sidedress N for several practical reasons:
- Adapt N accounts for most major variables known to affect soil nitrogen behavior, whereas the PSNT only considers the nitrate concentration.
- Adapt-N is not blind to the past and therefore has the ability to model the behavior of soil N under unusual environmental conditions whereas the PSNT only gives a snapshot of a particular moment in time.
- Adapt-N uses historical and real-time data to model how much plant-available nitrogen is in the ‘pipeline’ and anticipate when it will become available to the crop; the PSNT measures how much nitrate is in the ‘leaky bucket’ right now.
- If managed well, Adapt-N can/will continue to improve its accuracy over time, while the PSNT will never change.
- Assuming that Adapt-N recommendations are as good as or better than the PSNT, Adapt-N is more adoptable. Data can be entered into the Adapt-N program at any time: early in the spring, at night, on rainy days, etc. Farmers that have nutrient management plans already have much of the data they need to make the program work! The PSNT samples can only be collected during an otherwise very busy time of year.
- Adapt-N allows the user to subdivide fields up into more appropriate management zones with very little extra effort or cost. In contrast, if I am using the PSNT and decide to split a 15 acre field into 3 management zones (which is often appropriate), I have just tripled the amount of time and cost required to test that field. Up until this year, Adapt-N was available at no cost.
Finally, it is important to mention highlighting the limitations of the PSNT is not a criticism of the tool or the scientists who developed it. Those who create tools are usually more aware of its limitations than anyone else! The PSNT was developed using a valid and rigorous process, and continues to be a good tool in a normal year. Many farmers have made/saved money by using it, and many more should have. The PSNT can still be used in the traditional manner or be used to corroborate the data that Adapt-N generates for those occasions when the farmer is looking for assurance that the tool is working.
It would have been nice to show PSNT v Adapt N recommendations for a large range of conditions, rather than only show the data for a highly unusual location-year in which the PSNT would have needed a deeper sample to be accurate. Maybe the story would be the opposite in other years? What we really need to know is how well did Adapt N perform over a wide area and time frame. I’m sure it works best near New York State where the data base is greatest. Folks in the corn belt say they are less than happy with the results.