One season of coffee husk biochar lifted wheat yields in Ethiopia

In a single field season in Bita District, southwestern Ethiopia, coffee husk biochar applied with blended NPSB fertilizer raised bread wheat grain yield on acidic soil. The best treatment was 5 tonnes of biochar per hectare with 100 kilograms of NPSB per hectare, on plots that had also been limed. It yielded 3.59 tonnes per hectare and returned the largest net benefit reported in the study [1].

The limits matter as much as the result. One season, one site, three replications per treatment. The abstract reports no yield for the unfertilized control, so from the abstract alone there is no way to say how much of that 3.59 tonnes came from the biochar rather than from the fertilizer, the lime, or the field itself. The statistical claim is limited to a significant interaction between biochar and NPSB on grain yield [1].

What was actually done

The trial ran in the 2023 main cropping season as a factorial design: three levels of coffee husk biochar (2.5, 5 and 7.5 tonnes per hectare) crossed with three rates of blended NPSB fertilizer (50, 100 and 150 kilograms per hectare), under limed and unlimed conditions. Two positive controls, 5 tonnes per hectare of biochar alone and 100 kilograms per hectare of NPSB alone, and one negative control with no fertilizer at all. Treatments were laid out in a randomized complete block design with three replications [1].

Biochar is what remains when biomass is heated in a low-oxygen environment, a process called pyrolysis: porous, carbon-rich, alkaline. In acid soils it can raise pH a little, hold onto nutrients that would otherwise leach, and improve water retention. Coffee husk is the residue left after coffee processing, which makes it locally abundant in coffee-growing regions.

That is the general reasoning. The abstract reports the outcome, not the mechanism. It says nothing about measured changes in soil pH, cation exchange capacity, or water retention. A combination worked better than the parts, statistically speaking; which soil properties moved, and by how much, is not in the abstract.

The numbers, and what the big one means

The headline treatment, 5 tonnes per hectare of biochar with 100 kilograms per hectare of NPSB under limed conditions, returned a net benefit of 276,280 Birr per hectare and a marginal rate of return of 22,076 percent [1]. Under unlimed conditions the same combination gave the highest net benefit in that group, 197,080 Birr per hectare [1]. The study also reports that 7.5 tonnes per hectare of biochar with 100 kilograms per hectare of NPSB under lime gave a net benefit of 221,440 Birr per hectare and a marginal rate of return of 13,292 percent [1].

Those marginal rates of return look absurd until you know what the metric is. It compares a treatment against the next cheapest option and asks how much extra money comes back per unit of extra cost. A figure in the tens of thousands of percent means the additional inputs cost very little relative to the value of the extra grain, at the prices and costs used in this analysis. It is not a profit margin, and it is not transferable: change the price of biochar, the price of wheat, or the cost of labour, and the number moves. The abstract gives no exchange rate, no price year, and no cost breakdown, so the Birr figures cannot be converted into anything meaningful for a reader elsewhere.

The caution from other systems

Biochar is not automatically beneficial. In a greenhouse experiment on pineapple grown in a clayey Ferralsol, charcoal fines applied alone reduced readily available water by up to 16.5 percent. Only when the charcoal was combined with poultry litter did total and readily available water both increase. Modelling suggested the improvement came from changed soil hydraulic properties rather than from increased root growth [3]. That work was done in pots, in a greenhouse, on a different crop and a different soil, and cannot be transferred to wheat in Ethiopia. It does illustrate a pattern: a carbon-rich amendment applied by itself is not reliably an improvement.

There is a second reason to be careful about generalizing to "biochar." Its physical and chemical properties are tied to production variables including pyrolysis temperature and feedstock, and one 2026 paper argues that the initial particle size of the feedstock, which governs heat and mass transfer during pyrolysis, is an under-studied variable [6]. Coffee husk biochar made one way is not the same material as coffee husk biochar made another way. The study reports application rates; the abstract gives no pyrolysis conditions.

Where the evidence runs out

The list of things this study cannot tell you is longer than the list of things it can.

It cannot tell you the size of the effect. Without the control yield in the abstract, there is no baseline to compare 3.59 tonnes per hectare against.

It cannot tell you whether the biochar mattered on its own. The results of the two positive controls, biochar alone and NPSB alone, are not reported in the abstract. The interaction was significant [1], which means the combination behaved differently from what the single inputs would predict, but the direction and size of each single-input effect are not given.

It cannot tell you the mechanism, because no soil measurements appear in the abstract.

It cannot tell you about next season. One cropping season in one district says nothing about whether the effect holds under a different rainfall pattern. Biochar persists in soil, so the second year could plausibly show a different response, in either direction. Nobody has measured that here.

It cannot tell you about other soils or other wheat varieties. The finding is specific to Bita District, to acid soil, and to whatever variety was planted.

And it cannot tell you the lime rate. Lime is central to the best-performing treatment, but the abstract does not state how much was applied or how it was incorporated. Without that, the top treatment is not reproducible from the abstract, even in principle.

What you have is a plausible, locally grounded result from a properly designed but small and short trial, with an economic analysis attached. The direction is consistent with a lot of other work on combining organic amendments with mineral fertilizer in acid soils. The strength of the evidence is one season, one site, three replications, and an abstract that leaves the most useful numbers out.