Reading a soil test: four questions for the farm

Suppose two farms receive a soil-health score of 72. The number is invented. On one farm, the priority is getting water into a compacted part of a field. On the other, the question concerns declining organic matter across a rotation. The matching number tells neither farmer what to do next. It may make the farms look comparable while hiding the reasons they were tested.

That is the attraction and the weakness of a score. It compresses an untidy collection of observations into something that fits a dashboard, certificate or sales presentation. The compression can be useful if the reader understands what went into it. It becomes misleading when the number begins to stand for every function of soil, across every place and use.

The US Natural Resources Conservation Service describes soil health through indicators, including physical, chemical and biological properties. It explicitly rejects the idea that one outcome, such as yield, determines the whole condition. A soil can support a good crop while still presenting another resource concern. Several measurements may be needed because several functions matter.

This is a useful starting point for the expanding market in soil information. The better report is not necessarily the one with the greatest number of tests or the most elaborate colored dial. It is the one connecting appropriate observations to a question someone can act on, while retaining enough context to interpret change.

1. What job is the soil being asked to do?

In a field, soil helps support roots, retain and move water, cycle nutrients and provide habitat. These functions interact, but their current limitations need not be identical. A management plan should identify the problem before choosing the most fashionable measurement.

Consider infiltration. Water remaining on a surface after rain can prompt a question about how it enters the soil. The answer needs attention to the place and conditions, not merely a laboratory carbon value taken elsewhere in the field. A measurement of one property can inform the investigation without replacing observations of the function causing concern.

Or consider rooting. A crop may encounter a physical limitation at a particular depth while a surface sample looks reassuring. An average taken across several locations may soften the very contrast the farmer needs to locate. Spatial variation is not always statistical clutter; sometimes it is the beginning of the diagnosis.

NRCS’s soil-health testing program links testing with conservation planning and evaluation over time. Its approach places the measurements within a process for identifying resource issues and assessing management. That is a different purpose from collecting a score simply to label a field good or bad.

The distinction changes the choice of evidence. A question about compaction, a question about nutrient availability and a question about carbon storage may require different measurements and timescales. They can share a sampling visit, but they should not lose their identities when the results arrive in the same envelope.

For a buyer promoting a soil-friendly supply chain, this can be inconvenient. A universal number makes comparison easy. Yet a more specific account is often more informative: which problem was identified, which practice was changed and what happened to the relevant function? A less tidy answer can be a more credible one.

2. Compared with which soil?

Soils have histories that precede the current management regime. Their properties reflect parent material, climate, organisms, landscape position and time. A farm does not begin with a blank substrate whose entire condition can be credited to the latest practice.

NRCS’s account of soil formation describes these influences and the ways soil surveys help interpret land. The point for a performance comparison is simple: two measurements need a relevant reference. A difference between soils can arise from what they are as well as from how they have been managed.

Texture is especially important to the comparison. It should not be confused with structure. Texture concerns the mineral particle-size mixture; structure concerns how particles are arranged. A farmer can affect structure through management without turning one soil’s fundamental texture into another’s. Ranking the first against an unrelated reference can imply an unrealistic target.

The May 2025 NRCS technical note on testing describes contextual interpretation through soil peer groups, including texture and regional climate. Its discussion of SHAPE illustrates why a result should be judged in relation to relevant conditions. A percentile within such a group is not a universal percentage of all possible soil health.

That difference matters when numbers travel from a technical report to a public claim. “Better than many comparable samples” and “ninety percent healthy” do not say the same thing. The second can imply an absolute scale that the method never established. A reader should be told what the score means before being invited to admire it.

A useful reference can also be the same field over time, provided the observations remain comparable. It gives the farmer a way to investigate change under their own conditions. But the comparison still needs a record of management, weather, location and method; a before-and-after pair is not automatically an explanation of why the difference occurred.

The carbon percentage needs another line in the ledger

Soil carbon illustrates how a familiar number can conceal a missing quantity. A carbon concentration describes the share of carbon in the sampled soil material. A carbon stock concerns the amount held in a defined area and depth. To move between them, the amount of soil represented also matters.

FAO’s soil-organic-carbon stock protocol includes concentration, bulk density, depth and the coarse-fragment fraction in the calculation. It is a reminder that the percentage printed on a report does not by itself give the stock. Comparing stocks requires a consistent and suitable sampling and calculation approach.

Give two imaginary soils the same 2% organic-carbon concentration and examine one hectare to a depth of 20 centimetres, ignoring stones. At a dry bulk density of 1.3 tonnes per cubic metre, the volume represents 2,600 tonnes of soil and 52 tonnes of carbon. At 1.0 tonne per cubic metre, it represents 2,000 tonnes of soil and 40 tonnes of carbon. The assumed concentrations match; the calculated stocks differ because the represented soil masses differ. These are illustrative inventories, not measurements of sequestration or an argument for compacting soil to claim carbon gains.

It follows that a photograph of darker soil or a rise in one surface percentage is not a complete carbon-removal account. A credible assessment has to specify the material and depth being compared, the method used and the uncertainty. Management that changes soil density creates another reason to examine comparability rather than treat a fixed volume as an unchanging mass.

That does not make carbon measurement useless. It makes it a measurement task. For a farmer, a sample may help understand organic matter and management. For a program making quantified climate claims, the demands are broader. Using the same word carbon in both conversations should not conceal the different work required.

3. When was the measurement made?

Soil is active, and some indicators respond to recent conditions. A snapshot taken at one moment can be informative without being a permanent description. The calendar belongs in the report alongside the test method.

The NRCS soil-quality measurement guide emphasizes comparable reference conditions and recording management history and sampling locations. Its discussion of timing also explains why a comparison should avoid confusing a recent disturbance or seasonal difference with a lasting management result. The document is older guidance, not a new 2026 study.

For example, a biological measurement can reflect the conditions under which it was taken and analyzed. If a later result differs, the investigation should consider those conditions before attributing the entire change to a new practice. A test is an observation made through a procedure, not a direct transcript of everything living in the field.

Carbon stocks can require a different patience. NRCS’s testing program notes that monitoring management-related changes commonly takes several years, often more than five. A program demanding a large, certain annual improvement may therefore set an expectation poorly matched to the measurement and the natural variability being observed.

This is where management records become valuable. What crop was grown? What disturbance or amendment occurred? Which part of the field was sampled, and when? Those details turn a collection of numbers into a sequence that can be investigated. Without them, the report can become an expensive guessing exercise.

Repeated observations also need a plan. Sampling only after a result looks disappointing, or changing methods when another seems more favorable, weakens comparison. The purpose is to learn whether a management decision is achieving its intended result, including when it is not. A monitoring process that cannot retain an unwelcome finding is a promotion process instead.

4. What decision changes because of the result?

This is the question that gives the report its value. If no possible finding would alter management, further investigation or the claim being made, it is worth asking why the test was commissioned. Measurement should inform a choice, not decorate one already made.

A farmer may use the results to prioritize a part of a field, reconsider a practice or investigate an unresolved resource issue with an adviser. The appropriate response depends on the diagnosis and the operating conditions. A broad principle does not supply an identical prescription for every farm, rotation or climate.

Programs supporting better management should recognize that complexity. A change that helps one objective can bring costs or practical constraints elsewhere. Equipment, labor, timing and access to suitable inputs enter the decision. Recommending a practice without those considerations can leave the farmer responsible for implementing a plan that exists only on paper.

Different land uses need their own interpretations too. NRCS publishes separate soil-health resources for rangeland, including a 2025 in-field assessment guide. That is a useful example of resisting a single universal template. A soil’s function and the management choices available depend partly on the system it belongs to.

At policy scale, the same discipline helps prevent a monitoring program from being mistaken for a completed restoration. The European Commission’s Soil Monitoring Law overview records entry into force in December 2025, national transposition due in December 2028 and first reporting due in December 2031. Those milestones describe the construction of a framework; they do not mean Europe’s soils became healthy when the legislation took effect.

The framework can improve knowledge and support action, but observations and outcomes remain distinct. A larger database is a resource for better decisions. Whether those decisions improve soil functions has to be investigated over appropriate periods and places. The dashboard should leave room for that work rather than announce its result in advance.

Keep the useful disagreement

Return to the two farms with the invented score of 72. The most valuable part of their reports may be the disagreement among indicators: a reassuring result beside a persistent limitation, or a change that appears in one measure before another. Compressing everything into a score can hide the next question.

A good account would retain that detail while explaining it plainly. It would identify the function under investigation, the reference used, the time and method of observation, and the management decision the evidence supports. A summary could still sit at the top, but the farmer should be able to see what the summary leaves out.

The next soil test should arrive with the field history, sampling method and a decision someone is prepared to reconsider. For the two fictional farms, the same score could then lead to quite different work. One needs to understand compaction and water movement; the other needs evidence across its rotation. Keep those questions beside the number. The result becomes an aid to managing a particular soil, instead of a verdict that sounds precise while leaving its owner without a next step.