Results of field experiment on nitrogen application 2019_Cover_OneSoil Blog

High Vegetation Index Zone? More Nitrogen. OneSoil's Field Experiment with Variable-Rate Nitrogen Application

We develop all of our apps by working side-by-side with farmers and using the results of our experiments to guide us. During the 2019 season, we tested various methods of nitrogen application on fields in Ukraine. Our precision farming specialist, Usevalad Henin, explains how we did it and what our conclusions were.
The OneSoil web app has a tool for calculating the variable-rate application of nitrogen fertilizers. Users often ask why we don’t set the nitrogen application rates ourselves. We don’t do that because nitrogen is a mobile soil element, and, in order to determine the rates correctly, farmers have to analyze the field themselves. To help analyze the fields properly, we conduct field experiments and share the results with you. This is how the article How to Calculate Nitrogen Fertilizers for Variable-Rate Application appeared in our blog.

During the 2018−2019 season, we also conducted several experiments on the variable-rate application of nitrogen fertilizers. Ivan Humenyuk, Director of Innovation at a farm in central Ukraine, answered our call to action. Thanks to Ivan’s support and enthusiasm, we were able to organize and conduct a scientifically sound experiment on two fields of winter wheat. The results of this experiment were surprising, even for us.

Let’s start at the beginning. We’ll talk about nitrogen application methods, discuss the experiment and data analysis principles, and study the results for the first and the second field. In the end, we’ll decide which nitrogen application method performs the best.

The 3 most popular nitrogen application methods

Heavy nitrogen application in low vegetation index zones
In this case, we're going on the assumption that a lack of nitrogen limits plants' development and, therefore, reduces the vegetation index. By increasing the nitrogen rate, we can increase the NDVI vegetation index and the crop yield as a result.

Here's how nitrogen rates are distributed:
  • A large dose is applied to zones with a low vegetation index.
  • A medium dose is applied to zones with a moderate vegetation index.
  • A small dose is applied to zones with a high vegetation index.
      Heavy nitrogen application in high vegetation index zones
      This method assumes two things. First, if you apply a high rate of fertilizer in zones with a high vegetation index, the yield will be even larger. Secondly, if a high rate of fertilizer is used in zones with a low vegetation index, nothing will happen — crop yield won't increase.

      That's why:
      • A small dose is applied to zones with a low vegetation index.
      • A medium dose is applied to zones with a moderate vegetation index.
      • A large dose is applied to zones with a high vegetation index.
          Heavy nitrogen application in moderate vegetation index zones
          According to this method, increasing the nitrogen application rate in zones with a high NDVI index might lead to crop lodging. The assumption here goes that in zones with a low vegetation index, nitrogen is not the culprit limiting yield; other factors are. But in moderate vegetation index zones, increasing nitrogen fertilizer doses may lead to significantly higher crop yield.

          That's why:
          • A small dose is applied to zones with a low vegetation index.
          • A large dose is applied to zones with a moderate vegetation index.
          • A small dose is applied to zones with a high vegetation index.

            How we conducted the experiment and analyzed the data

            We tested two fields located in central Ukraine's Kirovograd Region. The soils were chernozem, and the crop was winter wheat. The relief was heterogeneous, with a significant difference in elevation everywhere, which causes a difference in crop yield.

            In Field 1, we varied only the first round of nitrogen fertilization. In Field 2, we varied both the first and second applications. We applied fertilizers on February 22 (here's the full timeline of field operations). To do so, we used UAN 23% liquid nitrogen fertilizer, so we'll use liters going forward. The fields were harvested on July 10, using a combine with an on-board computer and yield monitoring system.

            The crop yield data from the combine often has various data outliers and discrepancies that arise during technical operations. To eliminate their impact on the experiment results, we identified homogeneous zones in the fields that were covered by 2 to 4 combine runs, and we calculated the average yield values for these plots. In these areas, there were 8 to 16 zones (replications) with similar rates in each. We considered this data, too, when conducting our analysis.
            Homogeneous regions with similar fertilizer rates on the field 1_OneSoil Blog
            Homogeneous zones (replications) with similar fertilizer rates in Field 1

            Field 1

            The field is 60 hectares in area with an elevation difference of 133 to 152 meters. The field was divided into three zones with high, moderate, and low vegetation indices that were equal in area. We determined the zones using the most recent NDVI image on the day we applied nitrogen fertilizer.
            Vegetation zones on field 1 _OneSoil Blog
            Vegetation index zones in Field 1 using satellite images as of June 14
            The field was divided into five parts to test variable-rate nitrogen fertilizer application methods. We had two control zones and one for each of the three nitrogen application methods. The two upper parts of the field were selected as control zones because their vegetation index was homogeneous. The remaining zones were set up so that there were areas with high, moderate, and low vegetation indices in each of them.
            First control zone in Field 1
            Second control zone in Field 1
            Testing zone for the first method in Field 1
            Testing zone for the second method in Field 1
            Testing zone for the third method in Field 1
            Results for Field 1
            Control zone. The field's control zone used an application rate of 70 liters and had lower NDVI index values than the control area with a 250-liter rate.

            The crop yield analysis showed that the zone with an application rate of 250 liters has an average yield of 8.2 tons per hectare (122 bu/ac), while the zone with a 70-liter rate yielded 7.7 tons per hectare (115 bu/ac). The application of an additional 180 liters of nitrogen fertilizers increased the amount of wheat harvested by 0.5 tons per hectare (7 bu/ac). That means that when starting with similar conditions, crop yield increased 0.2 to 1.5 t/ha (3 to 22 bu/ac), and areas with the highest vegetation index experienced the highest increase in crop yield.
            The low vegetation index zone. In this zone, we used three nitrogen fertilizer application rates: 96, 132, and 270 l/ha. The zone was located on a slope, and we think that the terrain limited plant development. The NDVI index did not rise with an increased nitrogen application rate, which was also confirmed by the crop yield rates:

            • In the zone with an application rate of 270 l/ha, crop yield was 7.81 t/ha (116 bu/ac).
            • In the zone with an application rate of 132 l/ha, crop yield was 7.66 t/ha (114 bu/ac).
            • In the zone with an application rate of 96 l/ha, crop yield was 7.82 t/ha (116 bu/ac).

            That means that crop yield didn't increase in line with increased fertilizer application rates. This speaks to the fact that the soil's nitrogen content isn't what's limiting the vegetation index in the zone with a low index. Something else is. We can only determine what those factors are by conducting a detailed morphometric analysis and soil test.
            The moderate vegetation index zone. For zones with a moderate vegetation index, two nitrogen application rates were used: 168 and 240 l/ha. The zone with an application rate of 168 liters had a crop yield of 7.78 t/ha (116 bu/ac). The zone with an application rate of 240 liters had a crop yield of 7.94 t/ha (118 bu/ac).

            The analysis of separate field zones (replications) showed that crop yield increased slightly (0.1 to 0.18 t/ha or 1.5 to 2.7 bu/ac) when the fertilizer application rate was increased. In other words, there is a steady trend towards increasing crop yield.
            The high vegetation index zone. In the high vegetation index zone, we used three nitrogen fertilizer application rates: 110, 132, and 270 l/ha.

            • In the zone with an application rate of 110 l/ha, crop yield was 8.45 t/ha (126 bu/ac).
            • In the zone with an application rate of 132 l/ha, the crop yield was 8.39 t/ha (125 bu/ac).
            • In the zone with an application rate of 270 l/ha, the crop yield was 8.78 t/ha (131 bu/ac).

            The analysis of separate field zones (replications) showed that there was no increase in crop yield when the nitrogen application rate was changed from 110 to 132 l/ha. But crop yield rose when we increased the rate to 270 l/ha. That means that crop yield increase — between 0.2 to 0.7 t/ha (3 to 10 bu/ac) — can be seen when analyzing both the whole zone and separate areas of it (replications).

            As we see, the most effective method of nitrogen application was the second one — we apply the highest nitrogen rate in the zone with a high vegetation index. In the zone with a low vegetation index, crop yield is limited by terrain-related factors, such as moisture distribution, soil texture, and humus content. That's why increasing the nitrogen fertilizer dose didn't have the desired effect. And in the zone with a moderate vegetation index, the rise in crop yield was rather small so it's doubtful that the increased fertilizer dose is reasonable here.

            Field 2

            Field 2 is four kilometers away from Field 1. The soil and the terrain are similar: it's 87 ha in area and has a difference in elevation between 120 to 151 meters. The first round of nitrogen fertilization was performed in the early spring, and the second one was made at phenological phase 31 BBCH.

            As in the previous experiment, the field was divided into five zones, with two control zones and three zones of equal area to test nitrogen application methods.
            Vegetation index zones in Field 2
            Five zones and corresponding application rates in Field 2
            Five zones, each divided into two parts and corresponding rates in Field 2
            Before the second round of nitrogen fertilization, we divided each of the three experimental zones into two parts. We applied 110 l/ha of nitrogen fertilizers in one part and 190 l/ha in the other. The goal was to compare how the change in fertilizer rates influences zones with equal initial metrics. As such, we tested only the zones with high and low vegetation indices because we weren't able to identify a zone with a moderate vegetation index due to huge variability in NDVI index rates in some areas.
            Results for field 2
            Control zone. In this field, the control zone was not homogeneous: within its borders, there were areas with high, moderate, and low vegetation indices. The nitrogen application rates we used during the first fertilization were 300 and 60 l/ha. For areas with low and moderate vegetation indices, we didn't register any rise in crop yield when increasing the doses during the first fertilization. In the areas with a high vegetation index, crop yield increased by 0.31 t/ha (5 bu/ac).

            For the second round of nitrogen fertilization, we divided our control zone into two parts with application rates of 0 and 150 l/ha. In the areas with a low vegetation index, there was no increase in crop yield. In areas with a moderate vegetation index, we saw a rise between 0.1 and 0.3 t/ha (1 and 4 bu/ac), and 0.3 and 0.6 t/ha (4 and 9 bu/ac) in areas with a high vegetation index.

            The analysis of separate field zones (replications) showed a similar result: for the absolute majority of zones with moderate and low vegetation indices, increasing the fertilizer rate didn't result in higher crop yield. Zones with a high vegetation index saw an increase in crop yield of 0.15 to 0.6 t/ha (2 to 9 bu/ac) when applying an additional 240 liters of nitrogen.

            The low vegetation index zone.
            We applied 3 rates — 120, 165, and 300 l/ha — during the first round of nitrogen fertilization and two other rates — 110 and 190 l/ha — during the second round of nitrogen fertilization.
            1st fertilization of 120 l/ha
            1st fertilization of 165 l/ha
            1st fertilization of 300 l/ha
            6.42 t/ha (95 bu/ac)
            6.42 t/ha
            6.42 t/ha
            6.41 t/ha (95 bu/ac)
            6.54 t/ha (97 bu/ac)
            6.48 t/ha (96 bu/ac)
            Average crop yield after the 2nd fertilization of 110 l/ha
            Average crop yield after the 2nd fertilization of 190 l/ha
            1st feeding 120 l/ha
            1st feeding 165 l/ha
            1st feeding 300 l/ha
            According to this data, there was no significant growth in crop yield when increasing the fertilizer rates. As for Field 1, we think that the limiting factor was the terrain because the low vegetation index zone was located on a slope.
            The moderate vegetation index zone. We used three different rates: 120, 165, and 300 l/ha. During the second round of nitrogen fertilization, we didn't vary the fertilizer dose for areas with a moderate vegetation index because we didn't see any sense in doing so.
            1st fertilization of 165 l/ha
            1st fertilization of 200 l/ha
            1st fertilization of 300 l/ha
            7.91 t/ha (118 bu/ac)
            7.98 t/ha (119 bu/ac)
            8.29 t/ha (123 bu/ac)
            Average crop yield after the 2nd fertilization of 150 l/ha
            1-ая подкормка 120л/га
            1-ая подкормка 165 л/га
            1-ая подкормка 300 л/га
            We can see that crop yield increased by 0.3 t/ha (4 bu/ac) when the application rate was increased up to 300 liters.

            The analysis of separate field zones (replications) also showed a distinct trend towards an increased crop yield when raising the nitrogen application rate. Some areas saw up to 0.5 t/ha (7 bu/ac) more yield.
            The high vegetation index zone. We've reached the most interesting part of our experiment. In these zones, crop yield increased when we upped the nitrogen fertilizer rates. The results are provided in the table.
            1st fertilization of 120 l/ha
            1st fertilization of 200 l/ha
            1st fertilization of 300 l/ha
            8.83 t/ha (131 bu/ac)
            9.0 t/ha (134 bu/ac)
            9.06 t/ha (135 bu/ac)
            8.81 t/ha (131 bu/ac)
            8.98 t/ha (134 bu/ac)
            9.23 t/ha (137 bu/ac)
            Average crop yield after the 2nd fertilization of 110 l/ha
            Average crop yield after the 2nd fertilization of 190 l/ha
            1st feeding 120 l/ha
            1st feeding 165 l/ha
            1st feeding 300 l/ha
            The analysis of separate field zones (replications) showed that in some areas with increased fertilizer application rates, crop yield grew significantly — up to 0.8 t/ha (12 bu/ac).

            Thank you all for reading this far! We know it hasn't been easy. Let's cut to the chase: our conclusions.

            What did the experiment show?

            1
            Field zones with a low NDVI vegetation index did not respond to an increased nitrogen fertilizer rate in either the first or second rounds of fertilization.
            2
            Field zones with a moderate NDVI vegetation index responded to higher nitrogen fertilizer rates, but not by much.
            3
            Field zones with a high NDVI vegetation index responded well to increased fertilizer application rates. But remember that when using high nitrogen doses in these zones, we should also use plant growth regulators!
            We conclude that the most effective method of variable-rate application for nitrogen fertilizers was heavy nitrogen application in high vegetation index zones. Keep in mind that this result is valid for chernozemic soils in a moderate climate zone. On soils with different terrain and located in other climate zones, the experiment could have turned out differently.
            We'd wanted to try variable-rate fertilizer application at our farm for a long time but couldn't make up our minds. The experiment results are within the margin of error. We need to achieve repeatability and establish a clearer relationship between crop yield and the nitrogen application rate. We'll definitely use these experiment results in our future work.
            Ivan Humenyuk
            I hope that variable-rate nitrogen application experiments will continue around the world and that we come to a consensus five or so years from now. For now, we recommend analyzing your field and determining limiting factors before employing variable-rate nitrogen application.

            Timeline of Field Operations

            Results of nitrogen experiment_timeline of field operations_OneSoil Blog
            Usevalad Henin, OneSoil co-founder_OneSoil Blog
            Usevalad Henin
            Usevalad is an expert in GIS and agricultural chemistry. He has been developing precision farming tools since 2013. He is also the co-founder of OneSoil.
            Related Articles
            Comments