How to create a prescription map for variable-rate seeding

And how OneSoil can help with it.
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.
Last year here at OneSoil, we conducted a number of experiments on variable-rate seeding to find out how it can increase productivity in each hectare of a field. In this post, we'll share our insights into which data can be helpful when doing variable-rate seeding. More importantly, we'll look at how to put together prescription maps for equipment, fast and free.
Get all new articles straight to your inbox!
Last year here at OneSoil, we conducted a number of experiments on variable-rate seeding to find out how it can increase productivity in each hectare of a field. In this post, we'll share our insights into which data can be helpful when doing variable-rate seeding. More importantly, we'll look at how to put together prescription maps for equipment, fast and free.

Why you need variable-rate seeding

There're two ways you can increase a field's productivity: increasing yield or reducing the cost of seeds. Variable-rate seeding means we sow a different number of seeds in different parts of a field. We increase the rate in some parts of the field and decrease it in others. If we do everything right, we get a higher yield and maybe even reduce seed costs.

Variable-rate seeding is financially advantageous.

What information you need to create a seeding map

It's important to correctly prescribe seeding rates for different parts of a field. This prescription depends on the soil's fertility and, as a result, on the field's productivity. These are the main parameters we use when building seeding maps.

Here's the data we might use for this purpose:
  • soil nutrient test,
  • soil brightness and relief,
  • yield data for multiple years,
  • vegetation data.

Soil nutrient test

A soil test gives us valuable information about the soil's properties. It tells us the soil's phosphorus, potassium, and other macronutrient and micronutrient content. It also gives insight into the soil's acidity level and organic matter content. We can then use this data to analyze the soil's fertility. But using soil test results has its downsides.
It's not precise. There's one clear methodology of how to conduct the soil test itself. However, scientific circles still dispute whether grid sampling or zone sampling is the better way to take samples. That means that relying on a soil test's data when assessing fertility can be risky.

It's expensive. It costs about USD $20–50 to test just one soil sample. On average, we need to take one soil sample from each 0.3 ha of a field to create a true map of its chemical properties. If we have a 100 ha field, we need to take 300 soil samples. The whole test will cost us $6,000–15,000 (and that's without counting the cost of taking the samples themselves). In our experience, these investments in soil testing don't provide a return, even if we reach a higher yield.
It's not precise. There's one clear methodology of how to conduct the soil test itself. However, scientific circles still dispute whether grid sampling or zone sampling is the better way to take samples. That means that relying on a soil test's data when assessing fertility can be risky.

It's expensive. It costs about USD $20–50 to test just one soil sample. On average, we need to take one soil sample from each 0.3 ha of a field to create a true map of its chemical properties. If we have a 100 ha field, we need to take 300 soil samples. The whole test will cost us $6,000–15,000 (and that's without counting the cost of taking the samples themselves). In our experience, these investments in soil testing don't provide a return, even if we reach a higher yield.

Soil brightness and relief

These are the two factors that most often tell us about the soil's organic nutrient content and moisture level. Organic nutrient content and moisture, in turn, influence the soil's fertility.
It's not accurate. Over the past two years, we've manually analyzed several hundred fields. In each case, the relief and organic nutrients' impact on the soil's fertility and productivity was different. Our experience has certainly shown that the relief had the biggest impact on the distribution of organic nutrients in the soil. That, in turn, affected the soil's fertility. However, in some fields, the organic nutrient content didn't depend on the relief at all. That could be because of a low acidity level, for example.

In other words, the relief and soil brightness don't always correctly reflect soil fertility and field zone productivity.
It's not accurate. Over the past two years, we've manually analyzed several hundred fields. In each case, the relief and organic nutrients' impact on the soil's fertility and productivity was different. Our experience has certainly shown that the relief had the biggest impact on the distribution of organic nutrients in the soil. That, in turn, affected the soil's fertility. However, in some fields, the organic nutrient content didn't depend on the relief at all. That could be because of a low acidity level, for example.

In other words, the relief and soil brightness don't always correctly reflect soil fertility and field zone productivity.

Yield data for multiple years

In this case, we need the yield data for at least the past 3 years. The perfect scenario is if the weather conditions and crops growing in the field varied during these 3 years. We need this information to make sure the productivity zones are stable and to avoid making mistakes when deciding sowing rates.

The actual crop yield data, in this case, is the best source to look at when assessing the field's productivity.
Limited access to historical yield data. Productivity zones are not always stable. If they change every year because of the weather or crop's sensitivity to growing conditions, we need to look at yield data for the past 5 or even 7 years. Oftentimes, we simply don't have this data.

Calibration issues with combines. If you have one combine working in the field, we can collect all its data and calibrate it in the office. It gets more complicated when there are several combines. With several combines, we have to calibrate all the equipment perfectly to get the real yield data. If we don't track the actual data that each combine collected in the field, we won't be able to calibrate it well in the office.
Limited access to historical yield data. Productivity zones are not always stable. If they change every year because of the weather or crop's sensitivity to growing conditions, we need to look at yield data for the past 5 or even 7 years. Oftentimes, we simply don't have this data.

Calibration issues with combines. If you have one combine working in the field, we can collect all its data and calibrate it in the office. It gets more complicated when there are several combines. With several combines, we have to calibrate all the equipment perfectly to get the real yield data. If we don't track the actual data that each combine collected in the field, we won't be able to calibrate it well in the office.

Vegetation data

During the key growth stages, there's a strong correlation between plants' vegetation and the actual yield we get. That's why we can also use vegetation data for multiple years to assess productivity.
You can monitor plants' vegetation and track the key growth stages in our web and mobile apps.
In 2019, we conducted experiments on variable-rate seeding in 40 fields with three spring crops: soybeans, sunflowers and corn. In each experiment, we used 4 to 5 years of vegetation relative data values to build productivity maps. What does that mean? We didn't know the actual grain weight that we got at each point of the field, but we could see that Part A of a particular field was 17% more productive than Part B of that same field.
Yield map and vegetation map of a cornfield in the milk stage
This is the fastest and easiest method to use. In just a few clicks, the OneSoil platform lets you get the last 5 years of aggregate vegetation data for each field. We'll automatically choose images for days when the correlation between the modeled crop yield and vegetation was strongest. Then we'll take that data to build a prescription map.

How to create a seeding map in the OneSoil web application

1
First, go to the 'Sowing rate' tab and select a field.
2
Specify the crops that grew in this field in previous years. Do this to improve productivity mapping accuracy.
3
Enter sowing rates. You can use either kilograms or seeds per hectare, or pounds and seeds per acre. Enter three rates, one for each of the three productivity zones: high, moderate, and low. We recommend determining the rates with a difference of at least 10,000 seeds per hectare.
4
Choose the type of onboard computer.
5
Press 'Download file', and that's it! Now you've got a prescription file.
All you need to do now is upload this file on the onboard computer and start sowing.
This is how you can create a prescription map for variable-rate seeding in a few steps

Keep this in mind when using variable-rate seeding

Our team's experiments in previous years showed us that it's not only important to correctly assess a field's productivity zones, but also to choose the right hybrid.
In 23 cornfields in central Ukraine, high productivity zones saw a yield increase for the majority of hybrids planted in them when the seeding rate was increased. The yield didn't change in low productivity zones when the rate was decreased.

In soybean fields, we noticed a slight yield increase when applying the same approach to seeding rates. However, in the 14 sunflower fields with different soil types and located in various climatic conditions, the yield didn't depend on the sowing rate at all.
In 23 cornfields in central Ukraine, high productivity zones saw a yield increase for the majority of hybrids planted in them when the seeding rate was increased. The yield didn't change in low productivity zones when the rate was decreased.

In soybean fields, we noticed a slight yield increase when applying the same approach to seeding rates. However, in the 14 sunflower fields with different soil types and located in various climatic conditions, the yield didn't depend on the sowing rate at all.
In 2020, we'll continue studying how different spring crop hybrids respond to variable-rate seeding. We'll also try to cover a wider range of hybrids and seeding rates.

Variable-rate seeding explained by Usevalad Henin
Illustrations created by Vanya Uvarov and Dasha Sazanovich
Text edited by Tanya Kavalchuk
Layout by Anton Sidorov
Create a seeding map with OneSoil!
Tell us how you create seeding maps:
Usevalad Henin
Related Articles
People, fields and technology
Get all new articles straight to your inbox! You can read them even in the field.