Precision agriculture
Soil sampling: when to sample, how deep to go, and grid vs management zones
Soil is usually sampled after harvest and before fertiliser is applied, typically from the 0–20 cm cultivated layer. A composite sample combines a dozen or more individual cores. The quality of the result depends first and foremost on where that soil came from — even the best laboratory cannot compensate for poor sampling. Below we cover timing, technique, grid versus management-zone sampling, and how to turn the results into a fertiliser plan.
Why test your soil and how often
A soil test answers two questions: how much phosphorus, potassium and magnesium the crop can find in the soil on its own, and whether soil pH is locking them up. Without that, fertiliser rates are set out of habit or from a table for an average field that does not exist in practice.
P, K and Mg levels change slowly, so a full test is usually repeated every 4 years (guidance gives a range of 3–5 years). pH is worth checking more often, especially on light soils and under heavy nitrogen use, which acidifies the soil. Mineral nitrogen (Nmin) is a separate story: it is tested every year in early spring, because it changes from week to week.
Remember too that results are only comparable over time if you sample at a similar time of year, to the same depth and in the same places. That is why it pays to record zone boundaries and sampling points, not just the results.
When to take soil samples
The best time is between field operations: after the previous crop has been harvested and before autumn fertilising or liming, or in early spring before the first fertiliser application. The soil should be moist but not waterlogged. Very dry soil can prevent the probe from collecting enough material from the full depth, while wet soil sticks together and is difficult to mix evenly.
| Situation | When to sample | Notes |
|---|---|---|
| P, K, Mg levels and pH | From harvest to autumn, or in early spring | Always before fertilising and liming |
| After mineral P and K fertiliser | Usually no sooner than 2–3 months later | Fertiliser granules inflate the result |
| After liming | Usually after several months, ideally after a season | Lime reacts with the soil slowly, so an earlier pH reading is misleading |
| After manure or slurry | Once ploughed in and a few months have passed | Avoid old heap sites |
| Mineral nitrogen (Nmin) | Early spring, before the first N application | Separate samples by depth, fast chilled transport |
| Frozen, flooded or very dry soil | Do not sample | Results and depth will be unreliable |
If you test a field regularly, stick to the same time of year. Potassium levels and pH can differ slightly between autumn and spring, and what you are comparing is the trend, not a single number.
Equipment and sampling depth
For manual sampling, you need a depth-marked soil probe or auger, a clean plastic bucket for mixing, sample bags or boxes, and a marker. Nmin sampling requires a longer probe and chilled storage for transport. On larger areas, automated samplers mounted on a quad or pick-up can take cores at a consistent depth.
- Arable land: usually 0–20 cm, i.e. the plough layer; some guidance and labs use 0–30 cm. What matters is that it is always the same.
- Grassland: shallower, most often 0–10 cm, because that is where most of the roots are.
- Nmin: in layers, typically 0–30 cm and 30–60 cm, and for some crops 60–90 cm as well, each layer as a separate sample.
- Min-till and no-till: nutrients can accumulate near the surface, so shallow cores may overstate nutrient levels. Consistent full-depth sampling is especially important.
Do not use galvanised buckets or rusty spades if you are ordering micronutrients. Zinc and iron from tools can get into the sample.
How to take a representative soil sample, step by step
The composite sample that goes to the lab is usually about half a kilogram of soil. It has to describe several hectares, which means millions of kilograms of soil. That is why the number of cores and how you spread them across the area matter.
- Divide the field into areas with similar soil, topography and fertiliser history. Typically one composite sample describes 1 to 4 ha of uniform field.
- Plan your route: zigzag or diagonally across the whole zone, so the cores cover it evenly rather than just one corner.
- Avoid headlands and a strip of a dozen or so metres from field boundaries, ditches, tracks, old manure and lime heap sites, hollows where water stands, and wheelings.
- Push the probe in vertically to full depth. Take usually 15–20 cores per composite sample, and even more on more variable fields.
- Tip all cores into a clean bucket, break up the clods and mix thoroughly.
- From the mixed soil, put about 0.5 kg into a bag and label it with the field number, zone and depth. Scatter the rest back on the field.
- Samples for P, K, Mg and pH can be air-dried. Keep Nmin samples cool and get them to the lab as quickly as possible, following its instructions.
Grid or management zones: how to divide the field
Traditionally a field is divided into blocks of a few hectares with one sample from each. Precision agriculture uses two approaches: a dense grid (e.g. one sample per 1–3 ha, regardless of what is happening in the field) or management zones, i.e. parts of the field that behave similarly season after season.
| Criterion | Regular grid | Sampling zones |
|---|---|---|
| Number of samples | High, rises directly with area | Lower, depends on field variability, not hectares |
| Sampling and analysis cost | Higher | Usually lower for similar usefulness |
| Accuracy | Good with a dense grid, poor with a sparse one | Good, if the zones match real soil differences |
| What it shows | Point values, the map is built by interpolation | An average for each zone with similar conditions |
| When it makes sense | First test, fields with no history, research, suspected patchy levels from old heaps | Fields with visible yield variation, repeat tests, preparing VRA maps |
A grid assumes nothing in advance and, when dense enough, will catch even local anomalies. Its weakness is cost: with a sparse grid, a sample may land in an atypical spot and distort a sizeable part of the map. Zones take advantage of the fact that soil differences show up in the crop. If part of a field has yielded worse for years, it usually has different soil, water or nutrient levels and is worth testing separately.
In practice the two approaches complement each other. You can do a denser grid once to get to know the field, then run later tests by zone and check whether the boundaries still make sense.
How to define management zones from historical satellite imagery
A single satellite image captures one moment and can be misleading — for example after frost or uneven emergence. Management zones are therefore built from several seasons of data. If the same area repeatedly appears weaker or stronger on NDVI and related index maps, the pattern is likely persistent and may be linked to soil, drainage or topography.
Nirby uses the same approach: it analyses patterns visible in satellite monitoring, divides the field into management zones and lets you set the target zone size. It then places a sampling point in each zone and calculates an efficient route through the field. The mobile app guides the sampler from point to point, with sample numbers following the route order.
To be clear about what Nirby does not do: it does not take samples or run analyses. You take the samples yourself or hire a field sampling company, and the analysis is done by a lab, e.g. one of the OSChR (Polish regional agrochemical stations) or a commercial laboratory. Nirby comes back in once the results are ready.
What to order from the lab and how to read the results
A basic analysis covers pH and available phosphorus, potassium and magnesium. Add the rest as needed: micronutrients for sensitive crops (e.g. boron for oilseed rape and beet, zinc for maize), organic matter if you want to track it, and Nmin in spring before nitrogen goes on.
| Parameter | What for | What to watch |
|---|---|---|
| pH (usually in KCl) | Liming decisions, availability of other nutrients | Compare results using the same method |
| Phosphorus (P) | Phosphate fertiliser rate | Reported as P or P₂O₅, check the unit |
| Potassium (K) | Potash fertiliser rate | Reported as K or K₂O, the class depends on soil category |
| Magnesium (Mg) | Magnesium fertilising, choice of lime | With low Mg, consider magnesian lime |
| Micronutrients (B, Cu, Mn, Zn, Fe) | Sensitive crops, deficiency symptoms | Optional, not every year |
| Nmin | Spring nitrogen rate | Separate samples by depth, sampled every year |
The method matters. In Poland the Egner-Riehm method has long been the standard for phosphorus and potassium, and labs increasingly also offer Mehlich-3, which measures several nutrients from one extract. Results from the two methods are not directly comparable, so pick one and stick with it for later tests. On the order form, also give the agronomic soil category (from very light to heavy), because the boundaries of the nutrient classes and liming needs depend on it.
Results usually come as figures in mg per 100 g of soil plus a class: very low, low, medium, high, very high. The class matters more than the figure itself, because it tells you whether to fertilise to crop need, above it (building up levels) or below it (drawing down reserves). Nirby accepts results by the Egner or Mehlich-3 method, assigns them to sampling zones and creates nutrient maps for phosphorus, potassium, magnesium and pH.
From results to a fertiliser plan and a VRA map
A nutrient map on its own fertilises nothing. The value comes when you turn each zone's result into a rate: crop requirement for the target yield minus what the soil already supplies, taking into account the previous crop and organic manures. A zone in the low class gets more, a zone in the high class gets less or none.
In Nirby, the same zones you sampled become the basis of the fertiliser plan. The system calculates recommendations for each zone and takes into account the Nitrates Directive requirements and ekoschematy (Polish eco-schemes under the CAP), and you convert the finished plan into a variable rate application map, exported as ISOXML or Shapefile to a compatible ISOBUS terminal.
Common mistakes and a checklist
- Too few cores. Five cores across several hectares is a lottery, not a sample.
- Sampling straight after fertilising or liming. The result will show fertiliser that has not yet reacted with the soil.
- Inconsistent depth. Shallow cores inflate nutrient levels, especially under min-till.
- Cores near headlands and tracks. Soil there is compacted, contaminated and fertilised differently.
- Changing lab method between tests. Egner and Mehlich-3 give different figures for the same soil.
- No record of sampling locations. Without zone boundaries and points, you cannot compare successive tests.
- Results filed away and forgotten. A soil test only creates value when the results inform a decision, such as fertiliser or lime rates.
Before heading out to the field, run through a short checklist:
- Timing: after harvest or in early spring, a few months after fertilising and liming, soil moist and not frozen.
- Field division: zones or grid recorded on a map, number of samples known.
- Equipment: probe marked for depth, clean bucket, bags, marker, cool bag for Nmin.
- Technique: usually 15–20 cores per sample, zigzag across the zone, away from headlands and heaps.
- Lab order: pH, P, K, Mg, micronutrients if needed, same method as last time, soil category stated.
- After results: assign to zones, nutrient maps, fertiliser plan and VRA map.
If you want zones, sampling routes and results in one place alongside your fertiliser plan, take a look at soil testing in Nirby.
Frequently asked questions
How do you take soil samples for testing?
Using a soil probe, take usually 15–20 cores from the 0–20 cm layer, walking a zigzag across a uniform part of the field and avoiding headlands, tracks and old heap sites. Mix the cores in a clean bucket, put about 0.5 kg into a labelled bag and send it to the lab. One composite sample typically describes 1 to 4 ha.
When is the best time to take soil samples?
Ideally after the previous crop is harvested, before you spread fertiliser and lime, or in early spring before the first fertiliser application. After mineral fertiliser, usually wait 2–3 months, and longer after liming. The soil should be moist, but not frozen, flooded or very dry. For repeat tests, stick to a similar time of year.
How deep should soil samples be taken?
On arable land usually from the 0–20 cm layer, although some labs and guidance use 0–30 cm. On grassland, sample shallower, most often 0–10 cm. Mineral nitrogen (Nmin) samples are taken in layers, typically 0–30 and 30–60 cm. The most important thing is that depth is constant across the field and between tests.
How often should you test your soil?
Testing for phosphorus, potassium, magnesium and pH is usually done every 4 years, and guidance gives a range of 3–5 years. pH is worth checking more often on light soils and under heavy nitrogen use. Mineral nitrogen is tested every spring, because its level changes quickly.
Is it better to sample on a grid or by management zones?
A grid needs no prior knowledge of the field and, when dense, will catch local differences, but it means many samples and a higher cost. Zones defined from multi-year satellite imagery or yield maps group parts of the field with similar conditions, so they describe variability well with fewer samples. On fields with visible yield differences, zones are usually more practical.
