Using a "Partial Delta Yield" Approach to Fine Tuning N Rates
Keith Reid, Soil Fertility Specialist

Nitrogen response research has focused on identifying an exact optimum N fertilizer rate. The drawback to this approach is that it cannot account for all the variations between different soil types, crop rotations, manure use patterns, etc. on different farms. The partial delta yield approach tries to answer a much simpler question: "Am I applying more nitrogen to the crop than I need to?" Over time, this approach should narrow in on an optimum rate for a particular farm, but it will take several years to define an exact number. The key to this approach is to gather the information needed with a minimum of inconvenience, complexity and expense from lost yield.


Any trial involves making some assumptions. I will try to list the assumptions underlying this approach. If these assumptions are not valid on your farm, the results won't necessarily give you the best answers.

1. Yield response to N follows a quadratic or quadratic-plateau pattern (ie. The law of diminishing returns).
2. The optimum N rate is fairly consistent over the whole field, even though the yield may vary.
3. Optimum N rates are fairly consistent from year to year.
4. A farmer's crop rotation and pattern of manure use is fairly consistent from year to year.
5. Most farmers are already applying optimum or above-optimum rates for their crop (although the same approach can be modified for investigating increasing N rates).
6. Farmers are able to vary the rate of nitrogen applied to part of the field, and can measure the yields at the end of the season.

Laying out the strips:

To provide good information, the partial N strips need to be repeated, and need to be located so that they are representative of the field. Each field that is being checked should include a minimum of three partial N strips. If resources are limited, do the comparisons in one field, rather than doing a smaller number of comparisons spread over several fields. The strips should go across any major variations on topography or soil type. If this is not possible, widely contrasting areas in the field should be treated as separate comparisons. Uneven manure applications across the field will make interpretation of the results difficult or impossible.

The partial N rate should be two-thirds of the normal rate applied to the rest of the field. This should provide a large enough difference that the difference in N response is larger than the random variation between adjacent strips in the field, but not so large that yield losses are extreme.

Each partial N strip needs to be wider than the width of the combine header, but as narrow as possible beyond that. They must be laid out so that the strips can be harvested as a full header width with the minimum disruption of the harvest. Location of the strips must be carefully marked with flags or some other identification system.

Physically applying the reduced rate of nitrogen in strips will be easiest with side-dress applicators. Spray application of UAN or airflow application of granular fertilizer can be varied relatively easily, but the strips will be quite wide unless you can convince your applicator to apply a half boom width at the partial rate, and then double back and apply the other half boom width at the full rate. The most difficult system to apply partial rates in will be pull type spinner spreaders - these will require some very careful planning and implementation on the part of the farmer.

Soil nitrate levels can vary from year to year, depending on the amount of nitrate loss over winter, and the amount of N mineralized from organic sources in the spring. We are undertaking a project to determine if a benchmark site, which is sampled each year, can track the relative availability of soil nitrate from year to year, without having to sample the entire field. If this approach is valid, it will be useful to have a history of soil nitrate levels from a benchmark within your field. Choose a site typical of the field, away from headlands or other sources of variation, and collect 6-8 cores, 12 inches (30 cm) deep, for analysis for nitrate. This will provide a baseline for that site.

Measurements at harvest:

a) Weigh wagons or scales
At harvest time, combine each strip with the reduced N rate (or a full header width within that strip), and weigh. Determine the moisture content from a sample. Repeat for an adjacent strip with the full N rate. Measure the length and width of each strip, so that yields per acre can be calculated. Average the yields from the reduced N strips, and compare this to the average of the yields from the full N strips. This method can be used to assess field scale N rates, but cannot assess smaller areas within the field.

b) Combine Yield Monitors
As with any comparison using a combine yield monitor, make sure that the monitor is properly calibrated before attempting to harvest the trial. When harvesting, save each reduced N strip, and the adjacent full rate strip, as a separate load. This will allow much more accurate comparisons of the average values for each strip.

Interpreting the Results:

Compare the average of the yields from the reduced N strips with the average yields from the adjacent full rate strips. DO NOT compare the average of the reduced N strips with the average yield of the rest of field - there will be too much random variation introduced for the comparison to mean anything.

If the yields from the reduced rate strips are the same as the yields from the full rate strips (within two or three percent), or higher than, you can have a fair degree of confidence that your normal rate of nitrogen fertilizer was too high for that field and that year. If you cannot identify any reason why the response to nitrogen would be lower than normal, you can consider reducing N rates next year.

If there is a moderately higher yield in the strips receiving the full rate of nitrogen fertilizer, the N rate used was probably close to the optimum for that field. If the yield was much higher in the full N strips, it may mean that the rate used was below the optimum for that field. Further work is needed to identify how large a yield discrepancy is needed to indicate that fertilizer rates should be increased.


Back to Top