A grower in a temperate maize rotation applies certified premium biochar at 10 t/ha after reading promising trials from Kenya. Two seasons later the maize is flat; the legume in the rotation is up modestly. The grower asks the obvious question: was the trial wrong, or was the soil different? Both, and in a way that the literature now lets us predict.
The global biochar yield-response distribution is bimodal-by-context, and the literature is now precise about which side of the distribution any given field sits on. Four meta-analyses are load-bearing here: Jeffery et al. (2017), 1,125 paired observations, grand mean +10 percent1; Ye et al. (2020), 264 comparisons with fertiliser and biochar separately controlled 3; Schmidt et al. (2021), a systematic review of 26 global meta-analyses converging on ~13 percent2; and Aurangzeib et al. (2025), focused on acidic-soil physical-property and yield gains 5. The grand mean is real, but it is not the design parameter.
How much does biochar move yields?
Three independent meta-analytic streams converge on the same headline. Jeffery et al. (2017) report a grand mean grain-yield response of +10 percent across 1,125 paired comparisons drawn from 109 published studies, with a 95 percent confidence interval that excludes zero 1. Liu et al. (2013) on a smaller early dataset reported +11 percent 4. Schmidt et al. (2021), in a systematic review of 26 separate meta-analyses comprising roughly 14,000 paired observations, place the grand mean at +13 percent and show that 24 of the 26 underlying meta-analyses report a positive central tendency 2.
Ye et al. (2020) tighten the design: they include only studies that ran a proper separate-controls design (biochar alone, fertiliser alone, biochar + fertiliser, untreated control). On that stricter slice they find +15.8 percent when biochar is applied with NPK fertiliser, and an effectively zero +0.7 percent (not significant) when biochar is applied alone 3. The implication is decisive: the literature's headline number is conditional on a fertilisation regime that most trials run by default. Read naïvely, the grand mean overstates what biochar does as a standalone amendment.
Publication bias is the standing caveat. The biochar literature, like much applied agronomy, tilts toward published positives; Jeffery et al. (2015) and Schmidt et al. (2021) both estimate the bias at 2–4 percentage points on the grand mean72. The conditional structure within the meta-analyses (which contexts respond, which do not) is the load-bearing finding, not the grand mean itself.
Where it works, where it doesn't
Five splits explain most of the variance in the meta-analyses. None is independent of the others, but they are the cleanest cuts.
Soil pH. Jeffery et al. (2017) report +30 percent on acidic soils (pH < 5.5), +9 percent on neutral soils, and no significant effect on alkaline soils (pH > 7.5) 1. Aurangzeib et al. (2025), restricting to acidic soils across climates, raise the figure to +48 percent for crop yield with a –12 percent reduction in bulk density and a +18 percent gain in CEC5. The mechanism is liming and CEC supply, and it is most actionable where it is most missing.
Soil texture. Sandy soils respond strongly: +19 to +28 percent in yield, alongside +25 percent in plant-available water capacity (Razzaghi et al. 2020) 14. Fine-textured fertile silts and clays sit near zero; they already retain water and cations adequately, so biochar adds little marginal supply.
Climate. The title of Jeffery et al. (2017) is unambiguous: “Biochar boosts tropical but not temperate crop yields.” The split is +25 percent tropical, statistically zero temperate 1. The operative variables are the underlying soils (tropical weathered Ultisols and Oxisols are typically acidic, low-CEC, low-organic-matter) and the climate's interaction with biochar's water-retention properties.
Crop type. Legumes respond +21 percent on average; maize and oil-seed crops cluster +12 to +28 percent; rice is modest at +6 to +10 percent and often not statistically distinct from zero 3 4. Wheat and barley sit between, with negative responses documented under high application rates on fertile soils.
Fertilisation regime. This is the cleanest interaction. Across Ye et al. (2020), biochar + NPK delivers +15.8 percent versus untreated control; biochar alone delivers +0.7 percent; biochar + NPK versus NPK alone is +6 to +15 percent depending on soil 3. The defensible reading is that biochar is a soil-conditioning multiplier on nutrient supply, not a nutrient source. The single most common failure mode in field deployment is applying biochar without addressing the fertility constraint.
Why context decides: four mechanisms
The mechanism that dominates in a given context is what predicts the response. Four operate at meaningful scale, and most fields are governed by one or two.
CEC and pH lift dominates in acidic tropical soils. Biochar's negative surface charges (modest on fresh char, much higher on aged char) sorb cations (K⁺, Ca²⁺, Mg²⁺, NH₄⁺); its alkalinity raises pH and reduces aluminium toxicity. On a weathered Ultisol with pH 4.8 and a CEC of 6 cmolc/kg, a wood biochar at 10 t/ha typically lifts pH by 0.3–0.8 units and CEC by +20 to +40 percent (Atkinson et al. 2010, Joseph et al. 2021) 68. The yield response follows directly from the nutrient stoichiometry.
Water retention dominates in sandy and semi-arid systems. Razzaghi et al. (2020) meta-analyse 50 studies and find a +25 percent improvement in plant-available water capacity on coarse-textured soils, against +1 percent on clays 14. Omondi et al. (2016) report a global bulk-density decrease of –9 percent with biochar amendment 15. The yield-translation pathway runs through drought resilience: in years where rainfall is adequate, biochar contributes little; in dry years, it can carry 10–20 percent of the yield.
Nitrogen immobilisation risk is the dominant negative mechanism on fertile temperate soils. Fresh, high-VOC biochar, particularly low-temperature char from straw and manure, can immobilise mineral N for one to three growing seasons through microbial assimilation of labile carbon (Cornelissen et al. 2018, Borchard et al. 2019) 9 18. Where N is already limiting, the result is a mild yield depression. The mitigation is co-application with N, or aging the biochar before deployment.
Microbiome and rhizosphere effects are slower and less directly yield-linked. Biederman and Harpole (2013) report microbial biomass +28 percent under biochar; Lehmann et al. (2011) document AMF colonisation +52 percent16 11. The yield translation is variable and often delayed; the soil-health translation is more reliable.
One temporal caveat sits across all four. Biochar ages in soil. Surface oxidation over one to three years increases CEC roughly five-fold and shifts the dominant mechanism, particularly for the CEC and N-immobilisation pathways (Cheng et al. 2008) 19. Field trials shorter than two seasons systematically underestimate steady-state response.
A response-surface decision tool
Move the dials below. The factor effects are calibrated by inspection from the interaction tables in Jeffery (2017), Schmidt (2021), Ye (2020), and Aurangzeib (2025) 1 2 35, combined additively on the log scale and exponentiated back to a percentage. The output is a magnitude-of-effect estimate with a deliberately wide confidence band; it is not a fitted predictive model, and not a substitute for a site-specific trial.
The structural lesson is the same as the prose: the same biochar at the same rate spans roughly –10 percent to +50 percent depending on context. Two readings are particularly useful. First, set the “fertilisation” dial to none: the predicted response collapses across all other settings, reproducing the central Ye et al. (2020) finding that standalone biochar is statistically zero. Second, hold soil constant at acidic + sandy + tropical and vary the crop: legumes and vegetables hold the high end; rice and wheat sit lower. The crop-by-crop variance is the second-largest source of difference, after the fertilisation regime.
What biochar to use: feedstock, temperature, rate
Three properties decide the response within any given context.
Feedstock separates the population. Manure-derived biochar averages the highest yield effect (~+42 percent in Schmidt 2021's feedstock analysis) due to its embedded nutrients 2. Wood-based biochar sits in the middle (+12 to +22 percent) and is the workhorse for permanent-removal claims thanks to its low ash, high carbon, and consistent H/Corg. Crop-residue biochar (straw, rice husk) is intermediate. Biosolids and sewage-sludge biochar runs negative on average; Schmidt (2021) reports a –28 percent mean yield response driven by heavy-metal load, salt, and PAH residues 2. The first-order decision is feedstock; everything else is second-order.
Pyrolysis temperature reshapes the chemistry. Low temperatures (300–450 °C) yield biochar with higher CEC and more reactive surface groups, but also more volatile organic compounds and lower long-term stability. High temperatures (600–700 °C) produce highly aromatic, stable carbon with low CEC and low VOC content but reduced nutrient retention. The sweet spot for combined agronomic and removal claims is 500–650 °C on wood feedstocks, where H/Corg drops below the EBC C-Sink threshold of 0.4 while preserving moderate CEC (Crombie et al. 2013, Mašek et al. 2018) 10 13. The permanence framing of this temperature window is detailed in the companion accounting article.
Application rate is the dose-response axis. Most published trials run between 5 and 20 t/ha. Jeffery et al. (2017) find that yield response rises with rate up to roughly 20–30 t/ha in tropical contexts, then plateaus1. On temperate fertile soils, the same dose-response inverts: rates above 20–30 t/ha consistently produce yield depressions of 3 to 15 percent, driven by N immobilisation and, at extreme rates, pH overshoot (Schmidt et al. 2021) 2. The practical operating window for most contexts is 5 to 15 t/ha, with the upper bound on tropical acidic soils and the lower bound on fertile temperate.
Pyrolysis × feedstock properties
The grid below interpolates published property tables for the four main feedstock classes across the 300–700 °C pyrolysis range. Bilinear interpolation between reference points from Crombie et al. (2013) and Mašek et al. (2018)10 13; values are reference centres, not site guarantees. Note the biosolids row in particular: across the full temperature range, the verdict warns explicitly against agronomic claims.
A productive way to read the grid: pick wood, slide the temperature from 300 to 700 °C. H/Corg drops from ~0.75 down through the EBC C-Sink threshold of 0.4 at around 500 °C; CEC declines along the way, from ~38 to ~15 cmolc/kg. The reverse trade-off between biological reactivity and permanence sits at the centre of feedstock and temperature selection.
Soil health beyond yield
Many soil-health gains accrue even where yield does not move. The temperate-zone grower whose maize is flat is not necessarily seeing no work done; the work is often not visible at harvest. The meta-analytic record on non-yield endpoints is substantially more consistent than on yield itself.
Physical structure. Bulk density drops –9 percent on average across the global meta-analytic record (Omondi et al. 2016) 15. Aggregate stability rises +30 to +46 percent on amended soils, particularly when biochar is co-applied with compost. Plant-available water capacity gains +25 percent on coarse soils, with a long tail of effects under drought (Razzaghi et al. 2020)14.
Chemistry. CEC gains +20 to +40 percent at typical rates on low-CEC soils, with five-fold amplification of CEC over one to three years as the biochar surface oxidises (Cheng et al. 2008) 19. pH lift is site-dependent: on acidic soils, a 0.3 to 0.8 unit increase is typical; on alkaline soils, the same intervention pushes pH further out of the productive range and is contraindicated. Phosphorus availability rises on acidic and neutral soils; on alkaline soils the meta-analytic effect is null (Joseph et al. 2021)8.
Biology. Microbial biomass rises +28 percent (Biederman and Harpole 2013) 16; arbuscular mycorrhizal colonisation +52 percent (Lehmann et al. 2011) 11. Enzyme activities (alkaline phosphomonoesterase, dehydrogenase) rise +40 to +46 percent under combined biochar and AMF treatments. The biological response is the slowest to manifest and the most variable across trials, but it is the most consistent component of the soil health response across climates.
Greenhouse gas side effects. Soil N₂O fluxes drop a meta-analytic 12 to 21 percent under biochar (Borchard et al. 2019) 18. The effect is most reliable on fertilised croplands and weakens or disappears in low-N systems. Methane responses are smaller and less consistent; nitrate-leaching reductions are well-documented and routinely 20–30 percent.
Practical implications for project design
The conditional structure of the evidence base translates directly into three project-design decisions.
Pre-screen the context. Acidic tropical or subtropical + sandy or weathered + maize, legume, vegetable or oil-seed + NPK-fertilised systems sit on the high-response side of the distribution. Fertile temperate silt + cereal + an already-good N status sits on the low- or no-response side. Do not budget yield gains for the second category; design the value capture around soil-health and removal-credit endpoints instead.
Spec the biochar to the response. Wood-based, mid-high temperature (500–650 °C), 10–20 t/ha is the default for tropical acidic + NPK contexts. On temperate fertile soils where the agronomic case is weak, the case for biochar pivots to the removal credit and the soil-conditioning side benefits; pick the feedstock and temperature that optimise H/Corg and durability rather than CEC. Biosolids-derived biochar carries a documented negative yield expectation; even where the removal-credit accounting permits its use, agronomic claims should not be attached. The EBC v10.5 and IBI v2.1 standards specify the operating envelope12 17.
Design trials to age. First-season trials systematically underestimate the steady-state effect. Two to three growing seasons is the minimum horizon for an agronomic claim; five years is what the long-term datasets begin to stabilise around. Trials that include a biochar-alone arm alongside the biochar + NPK arm (the Ye et al. 2020 design) are the only ones that cleanly separate the conditioning effect from the fertilisation effect 3. Report full biochar characterisation alongside the yield numbers (H/Corg, ash, CEC, PAH and heavy-metal screen, feedstock) or the result is uninterpretable across contexts.
The conditional yield case, summarised
- Biochar's global mean yield effect is real (~10–13 percent) but conditional on soil × climate × crop × fertilisation regime. Acidic tropical or sandy systems on NPK carry the average; fertile temperate silts do not.
- Standalone biochar without fertiliser is statistically zero across the cleanest meta-analytic slice (Ye 2020). The literature's headline number assumes NPK co-application.
- Mechanism diagnosis (CEC + pH lift, water retention, N-immobilisation risk, microbial) predicts response better than any single biochar property. Most fields are governed by one or two mechanisms.
- Feedstock is the first-order decision. Biosolids char runs negative on average; wood and manure carry the agronomic upside. Mid-high pyrolysis temperature (500–650 °C) is the sweet spot for combined yield and durability claims.
- Soil-health gains (water capacity, CEC, microbiome, reduced N₂O) accrue even where yield does not move. On the temperate-zero soils, the value capture pivots to the removal credit, not the harvest.
References
- 1.Jeffery, S., Abalos, D., Prodana, M., Bastos, A.C., van Groenigen, J.W., Hungate, B.A., Verheijen, F. (2017). Biochar boosts tropical but not temperate crop yields. Environmental Research Letters, 12(5), 053001. doi:10.1088/1748-9326/aa67bd
- 2.Schmidt, H.P., Kammann, C., Hagemann, N., Leifeld, J., Bucheli, T.D., Sánchez Monedero, M.A., Cayuela, M.L. (2021). Biochar in agriculture — A systematic review of 26 global meta-analyses. GCB Bioenergy, 13(11), 1708–1730. doi:10.1111/gcbb.12889
- 3.Ye, L., Camps-Arbestain, M., Shen, Q., Lehmann, J., Singh, B., Sabir, M. (2020). Biochar effects on crop yields with and without fertilizer: a meta-analysis of field studies using separate controls. Soil Use and Management, 36(1), 2–18. doi:10.1111/sum.12546
- 4.Liu, X., Zhang, A., Ji, C., Joseph, S., Bian, R., Li, L., Pan, G., Paz-Ferreiro, J. (2013). Biochar's effect on crop productivity and the dependence on experimental conditions — a meta-analysis of literature data. Plant and Soil, 373(1), 583–594. doi:10.1007/s11104-013-1806-x
- 5.Aurangzeib, M., Khan, S. et al. (2025). Does biochar improve acidic soil physical properties and crop yield under varying climatic and soil conditions? A global comprehensive meta-analysis. Soil Use and Management, 41(1). doi:10.1111/sum.13183
- 6.Atkinson, C.J., Fitzgerald, J.D., Hipps, N.A. (2010). Potential mechanisms for achieving agricultural benefits from biochar application to temperate soils: a review. Plant and Soil, 337(1), 1–18. doi:10.1007/s11104-010-0464-5
- 7.Jeffery, S., Bezemer, T.M., Cornelissen, G., Kuyper, T.W., Lehmann, J., Mommer, L., Sohi, S.P., van de Voorde, T.F.J., Wardle, D.A., van Groenigen, J.W. (2015). The way forward in biochar research: targeting trade-offs between the potential wins. GCB Bioenergy, 7(1), 1–13. doi:10.1111/gcbb.12132
- 8.Joseph, S., Cowie, A.L., Van Zwieten, L., Bolan, N., Budai, A., Buss, W. et al. (2021). How biochar works, and when it doesn't: a review of mechanisms controlling soil and plant responses to biochar. GCB Bioenergy, 13(11), 1731–1764. doi:10.1111/gcbb.12885
- 9.Cornelissen, G., Jubaedah, Nurida, N.L., Hale, S.E., Martinsen, V., Silvani, L., Mulder, J. (2018). Fading positive effect of biochar on crop yield and soil acidity during five growth seasons in an Indonesian Ultisol. Science of the Total Environment, 634, 561–568. doi:10.1016/j.scitotenv.2018.03.380
- 10.Crombie, K., Mašek, O., Sohi, S.P., Brownsort, P., Cross, A. (2013). The effect of pyrolysis conditions on biochar stability as determined by three methods. GCB Bioenergy, 5(2), 122–131. doi:10.1111/gcbb.12030
- 11.Lehmann, J., Rillig, M.C., Thies, J., Masiello, C.A., Hockaday, W.C., Crowley, D. (2011). Biochar effects on soil biota — a review. Soil Biology and Biochemistry, 43(9), 1812–1836. doi:10.1016/j.soilbio.2011.04.022
- 12.European Biochar Certificate (EBC). (2025). Guidelines for a Sustainable Production of Biochar, version 10.5E. Carbon Standards International.
- 13.Mašek, O., Buss, W., Roy-Poirier, A., Lowe, W., Peters, C., Brownsort, P., Mignard, D., Pritchard, C., Sohi, S. (2018). Consistency of biochar properties over time and production scales: a characterisation of standard materials. Journal of Analytical and Applied Pyrolysis, 132, 200–210. doi:10.1016/j.jaap.2018.02.020
- 14.Razzaghi, F., Obour, P.B., Arthur, E. (2020). Does biochar improve soil water retention? A systematic review and meta-analysis. Geoderma, 361, 114055. doi:10.1016/j.geoderma.2019.114055
- 15.Omondi, M.O., Xia, X., Nahayo, A., Liu, X., Korai, P.K., Pan, G. (2016). Quantification of biochar effects on soil hydrological properties using meta-analysis of literature data. Geoderma, 274, 28–34. doi:10.1016/j.geoderma.2016.03.029
- 16.Biederman, L.A., Harpole, W.S. (2013). Biochar and its effects on plant productivity and nutrient cycling: a meta-analysis. GCB Bioenergy, 5(2), 202–214. doi:10.1111/gcbb.12037
- 17.International Biochar Initiative. (2015). Standardized Product Definition and Product Testing Guidelines for Biochar That Is Used in Soil (IBI Biochar Standards), v2.1.
- 18.Borchard, N., Schirrmann, M., Cayuela, M.L., Kammann, C., Wrage-Mönnig, N., Estavillo, J.M., Fuertes-Mendizábal, T., Sigua, G., Spokas, K., Ippolito, J.A., Novak, J. (2019). Biochar, soil and land-use interactions that reduce nitrate leaching and N₂O emissions: a meta-analysis. Science of the Total Environment, 651, 2354–2364. doi:10.1016/j.scitotenv.2018.10.060
- 19.Cheng, C.H., Lehmann, J., Engelhard, M.H. (2008). Natural oxidation of black carbon in soils: changes in molecular form and surface charge along a climosequence. Geochimica et Cosmochimica Acta, 72(6), 1598–1610. doi:10.1016/j.gca.2008.01.010