Soil Carbon on Smallholder Farms

Updated 2026-11-02 · 4 min read · By the akar.earth specialist team

Can smallholder farms earn soil carbon credits? How soil carbon is measured, which methodologies apply, cost and permanence issues, and Indonesia specifics.

Smallholder farms can earn soil carbon credits, but the economics are difficult: soil carbon changes slowly, varies widely within a field and costs a lot to measure. In Indonesia, soil carbon works best as an add-on to tree planting or agroforestry, inside a large aggregated project, rather than as a stand-alone credit.

Soil organic carbon is one of the five carbon pools. On mineral soils, practices such as adding organic matter, reducing tillage, keeping soil covered and planting trees can raise it over years. The challenge is proving a change that is small relative to the existing stock and to natural variation.

Methodology options

How soil carbon is measured

1. Stratify the area by soil type, slope, land use and management history. See plot sampling design.
2. Sample to a fixed depth, commonly at least the top 30 cm, with the depth set by the methodology.
3. Measure carbon concentration in an accredited lab.
4. Measure bulk density, because stock depends on soil mass, not just concentration.
5. Correct for soil mass. Many methods use an equivalent soil mass approach so changes in density do not create false gains.
6. Repeat at the same locations after several years and calculate change with uncertainty.
7. Apply models where allowed, calibrated and validated with local measurements.

The uncertainty deduction can be large for soil carbon because variability is high. That deduction directly reduces credited tonnes.

Why smallholder soil carbon is hard

• Small plots, many farmers. Sampling cost per hectare is high when farms are 0.5 to 2 hectares.
• Slow change. Measurable gains may take five years or more, while farmers need income sooner.
• Permanence. Gains can be lost if practices stop, for example if a farmer returns to burning residues or intensive tillage.
• Leakage of inputs. Moving manure or compost from one field to another does not create new carbon.
• Lab capacity. Reliable soil labs and sample logistics across Indonesia’s islands add cost and delay.

Indonesia specifics

Indonesian smallholder landscapes range from volcanic soils in Java, Bali and parts of Sumatra to weathered tropical soils in Kalimantan, and peat in coastal lowlands of Sumatra, Kalimantan and Papua. Peat needs a separate approach (see peatland REDD+): agricultural soil carbon methods are not designed for it. On mineral soils, coffee and cocoa agroforestry, home gardens and composting programmes are the most realistic settings.

When soil carbon makes sense

Include soil carbon when:

• the project is large enough to share sampling costs across thousands of hectares,
• the practices are likely to continue because they also raise yields,
• the methodology allows modelling with modest sampling,
• buyers value the co-benefits for soil health and resilience.

Exclude it, conservatively, when costs exceed the likely credit value or when practices are unlikely to persist. Most methodologies let you leave out a pool when it is not expected to decrease as a result of the project; check the exact condition in yours.

Cost-saving options

If you do include soil carbon, several design choices reduce cost without weakening integrity:

• Composite samples: mix several cores from the same stratum into one lab sample to lower lab fees, if your methodology allows it.
• Fewer, better strata: stratify on the factors that really drive soil carbon, rather than on every administrative boundary.
• Spectroscopy: some labs use infrared spectroscopy calibrated against conventional analysis to process many samples cheaply; check whether your methodology accepts it.
• Modelling with calibration: use a process model where allowed, with enough measurements to validate it locally.
• Longer intervals: remeasure soil less often than trees, in line with methodology rules.

Discuss these choices with your validator before sampling starts.

Common mistakes

• Measuring carbon concentration without bulk density.
• Sampling different points at baseline and remeasurement.
• Promising farmers soil carbon income in the first years.
• Applying an agricultural soil method to peat.
• Ignoring that the uncertainty deduction may remove most of the claimed gain.

For tree-based options that usually carry the economics, see coffee agroforestry carbon and the agroforestry hub. Our carbon modelling service can test whether soil carbon is worth including.