The short answer. On every farm we monitor we take physical soil samples and analyse them for organic carbon by dry combustion in an ISO-accredited laboratory. Our AI then ties those samples to satellite imagery of the same fields at 100 readings a hectare, a value every 10 metres, so the result is a map and a field total rather than a list of sample points. The error of that step is measured against samples the AI never saw and deducted in full. Nothing is estimated from a list of practices you say you follow. Your result is measured from your soil.
Reviewed by David Wright, founder and CEO of Ecometric. Last reviewed September 2026.
Step one: physical samples on your fields
Measurement starts at set-up: your SBI number and a short call to agree field boundaries and sampling access. The sampling is then designed field by field, so that every field you are paying to monitor has its own result rather than a share of a farm average.
Each sample goes to an ISO-accredited laboratory and is analysed by dry combustion, which directly measures the amount of CO2 released by the sample as it is combusted at over 1,000 degrees. That laboratory value is the anchor for everything that follows.
Step two: satellite imagery fills the variation between samples
One sample tells you about one spot. You already know your fields are varied: the headland, the light patch, the corner that goes first in a dry year. A standard soil test averages all of that into one number, and the variation disappears.
Our AI ties your laboratory results to 10-metre resolution satellite imagery of the same fields. The output is 100 readings a hectare, a value every 10 metres, giving a map and a more accurate field total rather than a list of sample results. The satellite never replaces the samples. It fills in the variation between them.
Step three: the error is measured, then deducted
Some samples are held back and the AI never sees them. Its estimate at those points is compared with the laboratory result, and the difference is the error of the method on your farm, in that year, on your soil. That error is deducted in full before anything is reported. It is why every result we send carries its accuracy, and why the change on the report is the change we can stand behind in front of an auditor. There is a separate guide on how the accuracy is stated.
Step four: change against your own baseline
The first round sets your baseline, field by field. Every round after it is measured the same way, at the same point in the year, because soil carbon has to be sampled at the same point in the year for the change to be real. The report shows the change against your own baseline, with the accuracy stated. That is how two farms on the same clay loam, in the same wet year, could be shown going in opposite directions: one lost 28 tonnes a hectare and the other gained 7. It was only visible because both were measured.
Timing matters as much as method, which is why sampling happens once a year in the autumn–winter window.
Measured, not modelled
Activity based soil carbon models allocate sequestration rates according to planned management practices. Without measured soil carbon change there is no way of telling whether a farming system is improving or in decline. Nothing on our reports is estimated from a list of practices you say you follow.
The method went through four rounds of independent scientific review to earn Regen Registry accreditation and its own credit class. It is audited by Carbon Direct and Nova-Cert, verified to ISO 14065 and rated by BeZero in the global top 30%. Every round is measured to that full Registry standard whichever tier you are on, so stepping up to emissions reporting or credits later is a decision rather than starting again.
What you get
A report, field by field, showing change against your own baseline with the accuracy stated, and a map of each field at 10-metre resolution. The results are yours: if you stop working with us you take your baseline and every measurement with you. And because we track annual change in soil carbon, management practice, yield and emissions across every project we monitor, we can tell you not just what happened on your farm but what worked on farms like it.
Questions we are asked about the method
Do you use a probe or a sensor instead of a laboratory?
No. Every result is anchored to laboratory dry-combustion analysis of physical samples from your fields. We do not replace that with a probe reading or a spectral estimate.
How many samples do you take?
The sampling design is set field by field from the variability the imagery shows, so the number depends on the holding and the number of fields.
Is it the same method every year?
Yes. The same fields, the same method, the same point in the year. That is what makes the change between rounds real rather than an artefact of when or how the soil was sampled.
Can I measure soil carbon with a home test kit?
A kit or a loss-on-ignition result gives a rough organic matter reading for the sample it tests. Measuring a change in soil carbon stock needs laboratory analysis by dry combustion, the bulk density of the soil and enough samples to cover the variation across each field, which is why every result we report is anchored to laboratory samples. Soil carbon testing methods compared.
How do I get my fields measured?
Send us your SBI number and your hectares and we will send a written quote within two working days. Minimum project area is 150 hectares. Get a quote.
Part of our guide to soil carbon measurement for farms. Related: soil carbon testing for your farm · how the accuracy is stated · why once a year · the science behind the method · how the sampling is designed.
