D_001 — Neobiome food production: regenerative / permaculture / biodynamic only
Decision
Neobiome models and recommends food production based on regenerative agriculture, permaculture, biodynamic, organic, and biointensive systems only. Industrial / conventional (high-input, high-tillage, monoculture) production is not adopted as a model benchmark or a design recommendation. Conventional data is retained only as a contrast baseline and as evidence of the harms that justify the regenerative selection.
This governs the Neobiome Intelligence F1_Food slice (the veg_yield input must be sourced from regenerative/biointensive evidence) and the thesis food-system argument (stated explicitly, with the supporting data below).
Rationale
-
Long-term sustainability over short-term gains. Intensive conventional vegetable production carries documented environmental harms in NZ: continuous rotary-hoeing / deep-ripping lowers soil carbon and biodiversity, and excessive fertiliser drives nitrate leaching — NZ synthetic-nitrogen fertiliser sales rose >600% between 1991 and 2019, conflicting with freshwater-health policy. LIT_056 Regenerative agriculture (cover crops, minimal tillage, green manures, soil-carbon restoration) is the documented sustainable counter, gaining scientific momentum in NZ. LIT_056 LIT_013
-
Quality over quantity. Regenerative/biodynamic systems prioritise nutrient density, soil health, biodiversity, and dietary diversity — and NZ’s real vegetable self-sufficiency gap is diversity, not mass: domestic production of legumes (0.57 servings/day) and dark-green-leafy vegetables (0.03 servings/day) falls far short of dietary needs. LIT_056
-
Yield: a contested question — and not the basis for this decision. The yield evidence is held here in balance, deliberately, so the specification is not justified on a yield-superiority claim.
- Evidence that regenerative can match or exceed conventional: at the small-plot / community-garden scale, biointensive market gardens match or beat conventional per-area yields on intensive-husbandry crops (salad, chard, kale, leeks) through close spacing, deep compost / no-till, succession and season extension — e.g. salad 3.1 vs 0.6, chard 3.0 vs 0.8, leeks 3.3 vs 2.1 kg/m². CR_021
- Evidence that it does not: at the whole-system level organic horticulture averages 10–32% lower than conventional (636-ratio meta-analysis), and the advantage reverses on bulk staple crops (potato 2.1 vs 4.5, carrots 2.2 vs 6.1 kg/m²) — the very crops a self-sufficient community must grow. The headline biointensive cases report gross revenue, not measured yield, and exclude storage crops. In NZ specifically, regenerative productivity benefits were characterised as anecdotal and unevidenced as of 2021. CR_021 LIT_013
- Basis for the specification: Neobiome adopts regenerative production as a researcher’s value decision grounded in long-term environmental sustainability, soil health, biodiversity and dietary quality — not on a claim of yield or production efficiency. The yield evidence is kept on the record (above) to inform the model’s
veg_yieldinput and to keep the thesis critically honest; it is deliberately not the justification for the choice.
Implications
- Model (
F1_Food):veg_yieldmust be sourced from regenerative/permaculture/biointensive evidence. Conventional sources (OT_069, LIT_056) are recorded as the contrast baseline only; their yields are not adopted. - Yield basis for the model (two figures held; conservative adopted): the regenerative
veg_yieldis supported by two values from the same evidence base, kept side-by-side per the both-sides approach above:- ~21 t/ha (2.08 kg/m²) — adopted for the model (conservative). The unweighted mean of the path-adjusted median per-crop yields in the primary AMOS table (LIT_058, Table 5.1) — the report’s own recommended (median) basis.
- 25 t/ha (2.5 kg/m²) — upper estimate. The judgment-weighted synthesis central (CR_021), leaning toward higher-value greens. Plausible but unsubstantiated by area weights. Both rest on the same single non-peer-reviewed UK survey, so this is an aggregation-method choice, not a source-authority one; the median basis is more defensible, hence the model uses ~21 t/ha (LIT_058). ⚠ UK data (NZ localisation = RT_242), and a whole-diet “staples haircut” still applies on top (the community must grow bulk staples — potatoes/carrots/onions — where the regenerative advantage reverses).
- Research targets: RT_061 is re-scoped to NZ small-plot regenerative yield; a companion RT covers the international biointensive/permaculture yield literature (e.g. Jeavons GROW BIOINTENSIVE, Fortier, Coleman, no-till market-garden studies) + NZ regenerative-horticulture data.
- Future food-slice refinement: score diet diversity (the five MDD-W groups; legumes + dark-green-leafy as the binding gap), not just total vegetable mass.
- Thesis: state the specification explicitly, supported by the sustainability-vs-short-term-gains and quality-vs-quantity evidence above.
Supporting sources
- LIT_056 — Curran-Cournane & Rush (2021): the environmental-footprint critique of intensive NZ veg production + the diet-diversity gap.
- OT_069 — Reid & Morton (2019): the conventional per-crop yield baseline (contrast only).
- LIT_013 — NZ regenerative-agriculture evidence base (productivity/soil/nutrient benefits anecdotal as of 2021).
- CR_021 — deep-research synthesis: the regenerative
veg_yieldfigure (≈25 t/ha, medium confidence) and the contested yield evidence (both strengthening and contradicting parity), held in balance for critical honesty.
Connections
Links to
Referenced by