Variable-Rate Nitrogen Application Experiment: Failed

In this experiment, I tested variable-rate nitrogen application for winter wheat during the first fertilization. Previously, I thought that nitrogen should only be applied by vegetation index zones. That's not quite right.

It's important to look at productivity zones and compare them to the actual vegetation index in different parts of the field before applying nitrogen fertilizers. If the vegetation index map doesn't intersect with the productivity zones map in any way, it's better not to use variable-rate nitrogen application or to try to apply fertilizers based on productivity zones.

Now let's go through things one by one.

Nitrogen is a mobile nutrient. Its quantity in the soil is constantly changing. So the amount of nitrogen fertilizers depends on plants' current condition. The vegetation index is a suitable tool for providing an overall assessment of the crops' conditions. This is the index we rely on in OneSoil when creating a variable-rate application map.

In the OneSoil web app, you can identify zones with low, moderate, and high vegetation indices on the fertilization date, set fertilizer dose, and create a prescription map for fertilizer application equipment. The app is completely free.

How to apply fertilizer by vegetation zones? There's no right answer. Figuring out which zone needs more nitrogen and which one needs less depends entirely on the region's soil types and climate. I enter three rates in each test vegetation zone and analyze the result to determine the best method for a particular field.

In my experience, regions with arid climate require more nitrogen to be applied to zones with a high vegetation index, whereas regions with sufficient moisture supply need to be fertilized more in zones with a moderate vegetation index.

On September 21, 2019, we sowed winter wheat. The first nitrogen fertilization was planned for February 25, 2020. In this part of Ukraine, February saw almost no snow. We can assume that the plants never wintered.

I used field vegetation data as of February 22 to create a prescription map for Mars. You can check your field's vegetation index for any day in the OneSoil web app. Completely free. Sign up, go to the 'Fertilizers' section, open the 'Nitrogen' tab, select the date, and there you have it!

Looking ahead, the Mars vegetation map as of February 22 has a lot in common with the productivity zones map. The zones' locations in the field are the same every year. High-productivity areas continue to have high yield, while low-productivity areas have low yield every year.

The prescription map was created manually. The average fertilizer application rate on the farm is 200 liters of UAN 32 per ha. I decided to test the application rate from 80 to 300 liters per ha. So for each low, moderate, and high vegetation-index zone, I included areas with three nitrogen rates: 80, 200, and 300 l/ha. Some test areas had a control of 0 liters per ha applied.

Yield analysis after the harvest. In mid-July, the field was harvested. This is the map we got thanks to the combine harvesters equipped with yield monitoring. In a very general way, it resembles the vegetation index map I used to create the prescription map for equipment. I've already mentioned that the productivity zones in the field are stable, so this was to be expected.

I delimited homogeneous areas of the field to assess the impact on yield from different nitrogen rates in all vegetation index areas. These are zones over the entire area of which the values of yield and application rate are approximately the same. On average, the combine passed through each zone four times.

This field is located next to Mars, so the fieldwork schedule coincided. In Saturn, winter wheat was also sowed in September, and the first fertilization was planned for February 24.

Once again, I used vegetation index data from the OneSoil web app to create a prescription map for this field. The nitrogen application map was created manually based on this data. There were three nitrogen rates for each zone: 80, 200, and 300 liters of UAN 32 per ha.

Analyzing yield after the harvest. Unlike the first field, the yield map I downloaded from the onboard computer wasn't even close to resembling the vegetation map. Areas with high vegetation indices had low yield, while areas with low vegetation indices became areas with high yield.

Usually, the vegetation index map during early plant development stages has similarities with the actual yield. But this time, something went wrong, and the yield map was the opposite of the vegetation map. To find the reason and analyze test results, I studied the field's productivity zones and determined the limiting factor.

First, I analyzed field vegetation data for the last four years and created a productivity zone map. Saturn's productivity zones turned out to be as stable as those on Mars. At the same time, the productivity zones map quite accurately reflects the actual yield I received in the experiment.

I then studied the field relief and slope map. It turned out that the most productive field areas are in low-lying areas and drainage hollows; low productivity zones are located on hills and areas with a convex surface.

I assume that before the satellite shots were taken (Note from OneSoil: NDVI is calculated using satellite images), plants in high-productivity zones may have suffered from temperature changes. Usually, during the freezing period, the air temperature in low-lying areas is a few degrees lower than on hills. This kind of weather was observed in central Ukraine in late February. This is probably why the vegetation index map made on February 22 is so different from the productivity zones.

Since the field vegetation map doesn't reflect productivity zones and the actual yield, it wasn't quite correct to apply nitrogen fertilizers by vegetation index. I made a mistake. I should have first compared the field's vegetation index and productivity and only then selected the experiment scheme.

So it only makes sense now to analyze Saturn Field's experiment results by comparing actual yield to productivity zones. In each productivity zone, I highlighted areas with homogeneous yield values and collected the average values in the table.