Soil Sampling Strategies: Choosing the Right Approach for Every Field
Reading time - 10 minutes
Most fields are sampled the same way regardless of what's happening in the soil. Here are seven smarter, zone-based strategies - and how the entire cycle, from sampling map to VRA, now runs in one product
Michał Pradzynski
OneSoil Chief Agronomist
Every autumn, farmers hear the same advice: test your soil.
What nobody explains is that soil sampling isn't one thing. A farmer with three small uniform fields, a farmer who spotted a weak patch from the combine, and a farmer building variable-rate maps need completely different sampling plans.
So don't start with "how many samples?" Start with "what do I want to know?"
The answer is different for a 2-hectare field in Germany, a 300-hectare field in Ukraine, and a problem spot you only discovered from the combine seat at harvest.
Here are the five most common tasks - and the strategy that fits each one.
What nobody explains is that soil sampling isn't one thing. A farmer with three small uniform fields, a farmer who spotted a weak patch from the combine, and a farmer building variable-rate maps need completely different sampling plans.
So don't start with "how many samples?" Start with "what do I want to know?"
The answer is different for a 2-hectare field in Germany, a 300-hectare field in Ukraine, and a problem spot you only discovered from the combine seat at harvest.
Here are the five most common tasks - and the strategy that fits each one.
📄 First, the vocabulary:
- Core (point / subsample): a single probe of soil taken at one location in the field.
- Sample: 1–20 cores mixed together and sent to the lab as one object. One sample = one lab result.
- Sampling zone: the area of the field whose cores go into one sample.
- Productivity zone: low / medium / high productivity areas built from years of satellite data. Not the same thing as a sampling zone - but they can work together, as you'll see below.
Why zones, not squares (grid)
Traditional grid sampling cuts the field into squares regardless of what's actually happening in the soil.
Zone-based sampling follows the field's real variability instead. When we compared the two approaches on the same field against EM-38 soil scanning data, OneSoil zones showed on average 50% less variability in the soil within each sample.
In plain terms: each lab result actually represents the area it came from, instead of averaging away the differences you're paying the lab to detect.
One map, though, doesn't fit every field. Here are the strategies.
Zone-based sampling follows the field's real variability instead. When we compared the two approaches on the same field against EM-38 soil scanning data, OneSoil zones showed on average 50% less variability in the soil within each sample.
In plain terms: each lab result actually represents the area it came from, instead of averaging away the differences you're paying the lab to detect.
One map, though, doesn't fit every field. Here are the strategies.
Strategy 1: One sample for small regular fields
For: compact, fairly regular fields smaller than ~3 ha.
The simplest case. One sampling zone, one sample - often enough to assess the fertility of the whole field.
Walk in, sample, move to the next field. Fast and cheap, and still structured, documented, and comparable season to season.
However, if you know that a small section of the field is different, yet too small to be managed practically in a different way from the rest of the field (the best example being a corner of the field measuring a few areas that slopes down to a stream or borders a woodland), avoid taking samples from that spot.
The simplest case. One sampling zone, one sample - often enough to assess the fertility of the whole field.
Walk in, sample, move to the next field. Fast and cheap, and still structured, documented, and comparable season to season.
However, if you know that a small section of the field is different, yet too small to be managed practically in a different way from the rest of the field (the best example being a corner of the field measuring a few areas that slopes down to a stream or borders a woodland), avoid taking samples from that spot.
Strategy 2: More samples for thin and irregular fields
For: long rectangles, terraces, irregular shapes - even small ones.
A single sample from a long field averages away too much: soil can vary considerably over a distance of 1–2 km, so a field that stretches across the landscape needs more individual samples, even if its total area is modest.
An incorrectly selected fertilizer rate or the wrong pH level can result in losses amounting to hundreds of Eur per hectare.
Generate a productivity map and then use it to divide the field into two or three sampling areas. Yes – the cost of the analyses will be higher, but incorrect information will cost you much more.
A single sample from a long field averages away too much: soil can vary considerably over a distance of 1–2 km, so a field that stretches across the landscape needs more individual samples, even if its total area is modest.
An incorrectly selected fertilizer rate or the wrong pH level can result in losses amounting to hundreds of Eur per hectare.
Generate a productivity map and then use it to divide the field into two or three sampling areas. Yes – the cost of the analyses will be higher, but incorrect information will cost you much more.
Strategy 3: Composite samples on larger fields - for trends, not fertilizer decisions
For: large fields with stable fertility and consistent yields, where the goal is long-term observation.
If you know your field well, one composite sample from a larger area can serve as a low-cost tracking tool.
An important honest caveat: a composite sample should not be the basis for fertilizer decisions - its accuracy isn't sufficient for that.
What it does well is reveal long-term trends: regular tracking allows early detection of soil degradation or a decline in humus levels.
If you know your field well, one composite sample from a larger area can serve as a low-cost tracking tool.
An important honest caveat: a composite sample should not be the basis for fertilizer decisions - its accuracy isn't sufficient for that.
What it does well is reveal long-term trends: regular tracking allows early detection of soil degradation or a decline in humus levels.
Strategy 4: Change monitoring at fixed points
For: advisors and farmers tracking soil dynamics over time.
Designate a specific area of the field and return to the same spots every season - or as often as your monitoring requires.
This area should cover several dozen square metres, be located away from the headland and outside the tramline, in a spot not prone to waterlogging.
It's not a formal field-trial protocol, but it gives a consistent, comparable series of measurements from the same ground, year after year.
Designate a specific area of the field and return to the same spots every season - or as often as your monitoring requires.
This area should cover several dozen square metres, be located away from the headland and outside the tramline, in a spot not prone to waterlogging.
It's not a formal field-trial protocol, but it gives a consistent, comparable series of measurements from the same ground, year after year.
Strategy 5: Emergency and intervention samples
For: the unpleasant surprises you spot from the combine.
It's often only during harvest that you see a section of the field yielding significantly below expectations.
Instead of guessing, use NDVI history: review previous images of the field, find the spot that stands out in vigour, and plan a sample exactly there.
One targeted sample, one specific answer to "what's wrong here?"
It's often only during harvest that you see a section of the field yielding significantly below expectations.
Instead of guessing, use NDVI history: review previous images of the field, find the spot that stands out in vigour, and plan a sample exactly there.
One targeted sample, one specific answer to "what's wrong here?"
Strategy 6: Match sampling zones with productivity zones - three samples per field
For: understanding the field in more detail while keeping lab costs to a minimum.
The "match with VRA zones" option in the OneSoil App generates a sampling map with exactly three samples per field: one for the high-productivity area, one for medium, one for low - all cores from each productivity level mixed into a single sample.
Because you're paying for only three lab results, you can afford far deeper analysis of each: soil texture, organic matter, cation exchange capacity, total nutrient content, element balance. Fewer samples, more knowledge per sample.
When it works best: when one fairly uniform factor is limiting yield (pH is the classic case).
When it doesn't: if your low-productivity areas have different causes - a dry sandy hill here, a waterlogged spot there - mixing them into one sample hides exactly what you need to see.
The "match with VRA zones" option in the OneSoil App generates a sampling map with exactly three samples per field: one for the high-productivity area, one for medium, one for low - all cores from each productivity level mixed into a single sample.
Because you're paying for only three lab results, you can afford far deeper analysis of each: soil texture, organic matter, cation exchange capacity, total nutrient content, element balance. Fewer samples, more knowledge per sample.
When it works best: when one fairly uniform factor is limiting yield (pH is the classic case).
When it doesn't: if your low-productivity areas have different causes - a dry sandy hill here, a waterlogged spot there - mixing them into one sample hides exactly what you need to see.
For sampling companies: this is also a product strategy: sell high-density sampling in year one, then offer productivity-zone monitoring on the same fields in following years as an ongoing service.
Strategy 7: High density sampling - the precision farming standard
For: accurate fertility maps and VRA. The most precise approach to soil diagnostics.
Sampling in plots built on productivity zones — typically one sample per 2–4 ha, with multiple cores per zone - produces fertility maps that reflect the real spatial variation in soil parameters.
This method combines the core requirements of soil testing: delimiting soil zones with genuinely different properties (polygons aligned with OneSoil productivity zones), detecting spatial variation in fertility, and compliance with formal soil sampling standards.
You control everything about the map: zone size, zone complexity, number of cores per zone, a buffer that keeps points away from field borders - and every point can be moved or added manually.
Sampling in plots built on productivity zones — typically one sample per 2–4 ha, with multiple cores per zone - produces fertility maps that reflect the real spatial variation in soil parameters.
This method combines the core requirements of soil testing: delimiting soil zones with genuinely different properties (polygons aligned with OneSoil productivity zones), detecting spatial variation in fertility, and compliance with formal soil sampling standards.
You control everything about the map: zone size, zone complexity, number of cores per zone, a buffer that keeps points away from field borders - and every point can be moved or added manually.
The whole cycle in one product
Until now, running any of these strategies meant stitching tools together: one program to draw zones, another for navigation, a spreadsheet for lab results, a third tool to build prescriptions.
That's fine if you're a large company with an established software stack. For a farmer, it usually meant a lab printout, an Excel file, and a note about where the samples were taken.
That gap is now closed. The entire soil sampling workflow runs inside OneSoil:
That's fine if you're a large company with an established software stack. For a farmer, it usually meant a lab printout, an Excel file, and a note about where the samples were taken.
That gap is now closed. The entire soil sampling workflow runs inside OneSoil:
- Create the sampling map in OneSoil Pro on the web - zones from your productivity map, your zone size, your number of cores, editable points.
- Sample in the field with the OneSoil Scouting app. Pick your starting point - the app builds the optimal route and guides you from core to core, keeping you within one zone until its sample is complete. One zone, one numbered box, no mixed-up samples. Points can be skipped (water, unharvested crop) and stay visible as skipped in the field record. And if your team prefers to drive its own pattern inside each zone, a one-point-per-zone map supports exactly that.
- Works offline. The route is calculated on your phone, not on a server. In the field you need GPS, not the internet.
- Load lab results back into OneSoil - no external software, no format gymnastics.
- Build your VRA map from those results and send it to the terminal.
OneSoil Platform combines field monitoring, scouting, productivity analysis, variable-rate application maps, soil sampling, field trials.
AI-powered agronomic recommendations in one connected system.
AI-powered agronomic recommendations in one connected system.
More Articles