Skip to content
Field Notes

The research side of permaculture.studio.

Verification record

Every claim in this article was checked against the abstracts of the papers it cites, by a separate pass that was given those abstracts and nothing else. A claim marked supported carries the sentence that supports it, quoted verbatim below.

References
14
Claims checked
6
Fully supported
4
Checked on
Nov 7, 2026

Claim by claim

  1. A meta-analysis of 62 papers and 1,148 paired samples found nitrogen use efficiency rose 20.6% and phosphorus use efficiency 18.9% under improved management compared with conventional practice.

    Partially supportedMeta-analysisMeta-analysis

    “Compared with conventional practice comparisons, NUE increased by 20.6% (95% CI: 17.4–23.9%) and PUE by 18.9% (95% CI: 15.2–22.7%)”

    While the meta-analysis of 62 papers and 1,148 paired samples reported these unadjusted figures, the claim omits that trim-and-fill analysis adjusted the pooled PUE estimate down to 16.4% due to publication bias.

    Checked by google/gemini-3.8-flash on Nov 7, 2026

  2. Heterogeneity was high (I2 = 74.3% for nitrogen, 71.8% for phosphorus), and trim-and-fill adjustment reduced the pooled phosphorus estimate from 18.9% to 16.4% after Egger's test indicated funnel-plot asymmetry (p = 0.041).

    SupportedMeta-analysisMeta-analysis

    “significant heterogeneity (I 2 = 74.3% and 71.8%, respectively). For P, Egger’s regression test indicated funnel-plot asymmetry (p = 0.041), and trim-and-fill analysis adjusted the pooled PUE estimate from 18.9% to 16.4%”

    The abstract directly confirms the I2 values for nitrogen and phosphorus efficiency, the Egger's test result (p = 0.041), and the trim-and-fill adjustment of the pooled phosphorus estimate from 18.9% to 16.4%.

    Checked by google/gemini-3.8-flash on Nov 7, 2026

  3. The pooled figures combine agronomic efficiency, apparent recovery efficiency, partial factor productivity and partial nutrient balance, which the authors state are not numerically interchangeable; separate analyses by metric gave consistent directions.

    SupportedMeta-analysisMeta-analysis

    “The pooled NUE and PUE values combined agronomic efficiency (AE), apparent recovery efficiency (RE), partial factor productivity (PFP), and partial nutrient balance (PNB); because these metrics are not numerically interchangeable, separate analyses were performed and confirmed consistent directional effects across metrics.”

    The claim accurately reflects the source's findings on the combination of metrics, their lack of numerical interchangeability, and the consistency of directional effects in separate analyses.

    Checked by google/gemini-3.8-flash on Nov 7, 2026

  4. Microdosing, deep placement, starter-band phosphorus and integrated organic-inorganic management showed the largest effects, while conservation tillage showed the smallest.

    SupportedMeta-analysisMeta-analysis

    “Practices such as microdosing (28.4%), deep placement (26.1%), integrated organic–inorganic nutrient management (24.3%), and starter-band placement of P (24.8%) were most effective, whereas conservation tillage had the lowest effect (9.7%).”

    The abstract directly identifies these specific practices as the most effective and conservation tillage as having the lowest effect.

    Checked by google/gemini-3.8-flash on Nov 7, 2026

  5. A modelling study of Chinese livestock production estimated that integrated spatial planning could reduce harmful nitrogen exposure for 90% of the population, but required reallocating 14% of production and cutting livestock numbers by one-third against a 2050 baseline.

    Partially supportedUnclassifiedUnclassified. The study design could not be determined automatically.

    “However, an integrated SP approach could reduce harmful N exposures for 90% of China’s population while ensuring food security. This approach involves reallocating 14% of livestock production, reducing livestock numbers by one-third relative to the 2050 baseline, and implementing advanced technologies to mitigate N pollution hotspots.”

    The claim accurately reflects the population benefit and two intervention requirements, but omits a third required component specified in the abstract: implementing advanced technologies to mitigate nitrogen pollution hotspots.

    Checked by google/gemini-3.8-flash on Nov 7, 2026

  6. In a phosphorus-limited plantation study, leaf phosphorus rose from 0.253 in monocultures to 0.337 in two-species and 0.467 in three-species mixtures, but the authors caution that the three-species result rested on six plots.

    SupportedModel or simulationModel or simulation. This comes from a model, not from measurement.

    “Leaf phosphorus increased from 0.253 in monocultures ( n = 33) to 0.337 in two-species ( n = 15) and 0.467 in three-species mixtures ( n = 6; +84.6%; bootstrap 95% CI = 0.161–0.262). The three-species result should be interpreted with caution given the small sample size and limited species combinations.”

    The abstract directly confirms the reported leaf phosphorus values across monocultures and mixtures, noting that three-species mixtures included only six plots and advising caution due to the small sample size.

    Checked by google/gemini-3.8-flash on Nov 7, 2026

What this does and does not prove

It proves that every DOI cited resolves to a real registered work whose returned DOI matches the one requested, that none of them is retracted, and that each claim has a named sentence in a cited abstract behind it.

It does not prove that the paper as a whole supports the claim. An abstract can overstate its own findings, and a sentence can be read out of context. That is why the quote is printed here rather than summarised: if the check is wrong, you can see that it is wrong.

Evidence grading is automatic and conservative. Anything it cannot classify is marked unclassified rather than flattered upward.