Weekly evidence for academics.

Dai et al. [Vitagri:SNmpc616n1-t9tf4l] demonstrate that pooling biochar trials without stratifying by soil type and pyrolysis temperature obscures the heterogeneity driving effect-size variance — a methodological lesson with broad applicability to agronomy meta-analyses. The grassland biomass meta-analysis [Vitagri:SNmpapkd5t-urdpoj] similarly stratifies by climate and altitude to explain allocation variance. Both studies illustrate the growing methodological consensus that treatment-by-context interactions must be modelled explicitly rather than treated as noise. A gap remains: UK-specific biochar field trials with matched soil and product characterisation.

Dai et al. [Vitagri:SNmpc616n1-t9tf4l] demonstrate that pooling biochar trials without stratifying by soil type and pyrolysis temperature obscures the heterogeneity driving effect-size variance — a methodological lesson with broad applicability to agronomy meta-analyses. The grassland biomass meta-analysis [Vitagri:SNmpapkd5t-urdpoj] similarly stratifies by climate and altitude to explain allocation variance. Both studies illustrate the growing methodological consensus that treatment-by-context interactions must be modelled explicitly rather than treated as noise. A gap remains: UK-specific biochar field trials with matched soil and product characterisation.

The MAGGnet network [Vitagri:BFmovbmg6s-6w3jxt] has standardised metadata from 315 experimental studies across 46 countries into a shared agricultural GHG meta-database — a methodological infrastructure paper worth noting for anyone designing multi-site trials. A companion three-part NET synthesis (Minx, Fuss, Nemet et al.) uses scientometric and systematic review methods across the full NET literature, finding that deployment-stage research (17% of literature) is severely underrepresented relative to R&D (59%). Both bodies of work expose data harmonisation and publication-stage gaps as the primary methodological constraints, not absence of primary evidence.

Tamburini et al.'s second-order meta-analysis synthesised 5,160 primary studies across 41,946 comparisons assessing diversification effects on ecosystem services and yield. Diversification reliably enhanced biodiversity, pollination, pest control, and soil fertility; yield effects were not statistically significant overall, though sub-practice effect sizes need scrutiny. Separately, Ramirez et al.'s machine-learning synthesis of 1,998 bacterial soil samples from 30 sequencing studies identified rare taxa as structurally disproportionate relative to their abundance. Both studies foreground aggregation bias and inter-study heterogeneity as central methodological challenges — productive territory for methodological PhD work.

A systematic review of 96 studies (2000–2025) on phytochemically rich food-system by-products in ruminant diets [Vitagri:SNmobqxieg-2mc2md] synthesises evidence on rumen fermentation, animal health, and milk and meat composition, but reveals substantial heterogeneity in inclusion rates, by-product types, and outcome measures. A complementary 21-study meta-analysis on low-protein piglet diets [Vitagri:SNmobqxjzs-5lpwrp] identifies growth and gut-morphology effects but notes variation in intervention design. Both records highlight dose-response characterisation and temperate-climate validation as priority research gaps.