Soil Colour Can Predict Organic Matter in Morocco Farmland

Farmers have long judged their soil by looking at it. Darker ground commonly indicates more fertile land, whereas light, dusty fields may point to problems. But could this straightforward visual signal take the place of costly chemical tests?

New research indicates that it could. Scientists found that soil colour by itself can estimate organic matter content in semi-arid farmland accurately enough for routine monitoring.

Should the findings be confirmed on larger scales, this method could substantially lower testing costs, cut hazardous laboratory waste and offer farmers a quicker means of monitoring soil fertility and carbon storage.

Soil colour meets chemistry

Soil collected from orchards and plantations on semi-arid farms in Morocco showed colour patterns that closely matched measured organic matter concentrations.

Using these samples, Dr. Yassine Bouslihim of the National Institute of Agricultural Research (INRA) showed that straightforward colour measurements repeatedly corresponded with laboratory carbon results.

The central link between colour and organic matter persisted even when the soils were dampened and became darker, although its precision fell slightly.

This consistency indicates that the signal may be robust enough for practical monitoring, while also prompting questions about the extent to which moisture and local soil conditions define the method's limitations.

Why organic matter matters

Soil organic matter consists of carbon-rich material left behind by plants and microscopic soil organisms. Although it may seem insignificant, it is crucial to maintaining productive farmland.

As time passes, organic matter enables soil to retain water and nutrients. It also supports the formation of small, stable aggregates that improve aeration and root development. Healthy soil is friable, taking in water rather than allowing it to run off.

Organic matter is not permanent, however. High temperatures, intensive cultivation and continual cropping can gradually degrade it. Growers may only recognise the loss when soil turns compacted, dry and less productive.

Affordable, regular testing lets farmers spot this deterioration early and alter crop residues, introduce cover crops or adjust irrigation schedules before yields start to fall.

How soil colour becomes data

To investigate whether colour could indicate organic matter concentrations, the researchers used a handheld instrument known as a colourimeter.

This device directs light at the soil and records the colour reflected back, converting fine differences in brown, red and yellow tones into figures that a computer can assess.

The instrument measures a small, level soil area of roughly 8 mm across, meaning that aggregates, stones and uneven surfaces may influence the reading. Careful sample preparation was therefore essential. Levelling the surface and taking repeated measurements made the data more consistent.

Moisture created a further complication. Because wet soil is naturally darker, this effect can obscure distinctions that are more readily identified in dry soil.

To limit this issue, the team concentrated on “hue”, meaning the dominant colour group, such as red, orange or yellow, instead of brightness alone. Hue remained more consistent in damp samples.

Nevertheless, water lowered overall accuracy. Dried soil produced the most dependable comparisons, although the fundamental connection between soil colour and organic matter was still apparent in wet samples.

Identifying fields that need full testing

The researchers applied a computer technique called Random Forest, which combines numerous small decision trees rather than depending on one equation. Averaging its many predictions limits the influence of unusual samples.

For dry soil, the model accounted for about 60 percent of variation in organic matter concentrations. That proportion fell to about 52 percent when the soil was moist.

At this accuracy level, the scan is most useful as a screening method: it can rapidly flag fields likely to require comprehensive chemical tests, rather than fully replacing laboratory work.

Lower costs, fewer chemicals

Conventional soil testing may be costly because it involves chemical reagents, laboratory labour and the disposal of hazardous waste. These expenses can mount rapidly.

The study's financial model found that a scanner-based process would cover its initial cost after about 1,261 samples.

A laboratory handling 5,000 samples annually could reduce costs by 96 percent, recoup its investment in under four months and achieve a five-year return of about 940 percent.

“We found that digital color analysis isn’t just a technical alternative; it is an economic revolution,” said Dr. Bouslihim.

Safety is another benefit. Standard wet-chemistry tests for soil carbon frequently use chromium-based reagents and strong acids, which produce hazardous waste.

Certain chromium compounds include chromium(VI), a toxic form associated with lung cancer that demands careful handling and disposal.

Digital scanning eliminates these chemicals altogether. Its principal environmental impact is electricity consumption and routine calibration. This lowers risks for technicians and makes disposal easier, which could be particularly important for smaller laboratories with limited safety equipment.

Training the soil colour model

Before colour can be used alone, every laboratory must create a local reference dataset. This involves scanning soil samples and matching the readings against reliable chemical results to train the computer model.

This process teaches the system how local minerals, agricultural practices and climate affect soil colour. Standardised preparation, including levelling samples and repeating scans, reduces background noise that may otherwise conceal genuine patterns.

After it has been established, the approach could be used for routine soil monitoring. Periodic chemical tests would still be required, however, to confirm that the system remains accurate over time.

Soil colour method needs more testing

INRA's dataset included only 74 samples, so the current model retains substantial uncertainty, particularly where different soils have similar colours.

Increasing the sample set to include a broader variety of soil types would help distinguish genuine organic signals from iron minerals, salts or recently applied fertilisers that can alter the surface colour.

Including simple moisture measurements or combining colour data with other rapid sensors could improve predictions further without reverting to full chemical analysis. Until larger studies assess the method outside Morocco, laboratories should use it as an indication rather than a definitive judgement.

Even so, the findings are genuinely encouraging. Converting soil colour into useful data shows that rapid testing and safety can go together.

Wider trials will establish how broadly the approach can be applied, but it already gives farmers a less expensive, quicker way to track changes below their feet.