Research Brief — R08: Living Soil & Indoor No-Till Feasibility

Report ID: R08 (Wave 2, pulled forward per RESEARCH_QUEUE) · Brief version: v1.0 — FROZEN August 31, 2026 on Danny’s explicit confirmation. Both research runs execute this version and must cite it; the critique stage records it. Date drafted: August 31, 2026 (v0.9 initial draft; v0.95: five refinements from ChatGPT’s first editorial review; v1.0: four fixes from ChatGPT’s read of the full draft — comparator integrity, tier taxonomy unification, compost-denominator guard, HpLVd rewording) · Drafted by: Claude (editor) · Protocol: v1.3 Assigned to: ChatGPT deep research + Gemini deep research (independently, same frozen brief; both runs must cite the brief version they ran)

Assignment

Produce an evidence review of living-soil and no-till cultivation systems in indoor (and, secondarily, controlled greenhouse) cannabis production. The governing question is not ideological (“is living soil better?”) but feasibility-and-evidence: which biological processes claimed for these systems are supported, which production outcomes have direct evidence, which remain extrapolations from field agriculture, and under what operating conditions does indoor no-till become more or less technically and commercially credible?

Keep three claims-classes strictly separate throughout — this is the report’s central discipline:

1. Biologically plausible — a mechanism exists in soil science (AMF symbiosis, aggregation, microbially mediated nutrient cycling).
2. Agronomically demonstrated — the mechanism has produced a measured production outcome in a relevant cropping system.
3. Commercially feasible indoors — the system survives contact with indoor cannabis economics. Feasibility standard: a commercial-feasibility discussion is incomplete unless it addresses, at minimum, yield per area per time, labor burden, input cost, crop-cycle repeatability, failure/recovery cost, sanitation/reset burden, and market-value/differentiation potential. “A grower successfully ran beds indoors” is existence evidence, not competitiveness evidence — never equate the two.

Grower discourse collapses these into one another; this report must not.

Methodology requirements

Practices, not labels (the methodological spine — R01-C033). Benefits observed in “organic/regenerative” systems attribute to specific practices (carbon inputs, continuous living cover, diversity, reduced disturbance), not to labels. Evaluate practices and their mechanisms, never brand-words. Corollary already in the ledger (R01-C035): in meta-analysis, continuous living cover drives infiltration gains (+35% cover crops, +59% perennials) while no-till alone shows no overall infiltration effect — “no-till” the label is not the active ingredient even in field soil. Carry that skepticism indoors.

Define the system before evaluating it. “Living soil,” “no-till,” “organic,” “regenerative,” “soil food web,” “KNF/JADAM,” and “reusing a container of amended media” are not interchangeable. Open with an operational taxonomy of what indoor practitioners actually run — at minimum: (a) raised/fabric beds with undisturbed soil across cycles; (b) large containers with top-dressing and minimal disturbance; (c) recycled/re-amended media that is mixed between cycles (this is media reuse, not no-till — say so); (d) “living soil” bottle-fed hybrids. State, for each evidence source, which system it actually tested. A claim about (c) does not support (a).

Comparator integrity. For comparative studies, characterize what actually differs between treatments. “Living soil vs. conventional,” “organic vs. synthetic,” or “no-till vs. tilled” is not an attributable treatment contrast unless the underlying practices and inputs are specified — such comparisons frequently change media composition, fertilizer chemistry, microbial inoculation, irrigation, container volume, nutrient concentration, and amendment schedule simultaneously. When multiple variables change together, report the observed system-level difference but do not assign causality to soil biology, no-till, fertilizer source, or any other single component without evidence isolating it. This applies to academic treatment labels exactly as it applies to practitioner labels — R01-C033’s discipline does not stop at the peer-review boundary.

The unit of no-till persistence. “No-till” for one cycle, for several cycles, and as an effectively permanent bed with only top-dressing/transplant disturbance are meaningfully different systems. Wherever possible, record the number of consecutive crop cycles and/or duration of bed reuse behind every claim — this variable may explain much of the contradictory evidence, and a claim demonstrated at cycle 2 does not transfer to cycle 10.

Evidence tiers by transfer distance. Rank and label all evidence: T1 direct indoor/greenhouse cannabis studies; T2 cannabis studies in other environments; T3 controlled-environment horticulture and container/greenhouse soil-biology studies in comparable crops; T4 field-agriculture no-till and soil-food-web literature. Practitioner and commercial self-report is not a fifth tier: tag it ANECDOTAL as a separate source-status label, in addition to whatever system and question it speaks to. T4 can establish mechanism plausibility only. AMF colonization, microbial diversity, aggregate stability, and nutrient cycling in field soils do not automatically transfer to a 30–200 gallon indoor bed on a 60–75 day flowering cycle under irrigation control — the burden of proof for transfer sits with the claim, and each claim must carry its tier.

The indoor constraint set (explicit, falsifiability guard). For every feasibility claim, address which of these constraints the evidence actually covers: root-zone volume (beds vs. pots); drainage and irrigation strategy (hand, drip, blumat); environmental control interactions; pest/pathogen persistence and sanitation risks across cycles (e.g. fungus gnats, root aphids, Fusarium/Pythium carryover in undisturbed media; separately, persistent propagation/sanitation risks such as HpLVd, which is a transmission/sanitation problem rather than a soil-persistence one); nutrient loading, salinity, and element accumulation over repeated cycles; compaction and structure decline; crop turnover speed and sanitation protocols; labor per cycle; and whether the soil is actually left substantially undisturbed between cannabis cycles. Without this, “no-till works indoors” is unfalsifiable. Absence of evidence on a constraint is a reportable finding.

Separate yield/quality from economics. A system can be biologically functional and lose commercially (labor, floor-space efficiency, consistency, cycle flexibility, remediation risk) — or be somewhat throughput-inefficient and still win commercially. Economic backdrop from the ledger: wholesale price collapse across mature markets (R02-C025) and the disproportionate exposure of high-OpEx indoor production to price compression (R05-C029) make differentiation arguments (input cost, quality premium, brand, resilience) genuinely interesting — but they must be evidenced, not inferred from enthusiasm. Note also R05-C012: in one outdoor cannabis LCA, potting media accounted for up to ~75% of global warming potential. That makes media reuse a real sustainability inquiry — but word it narrowly: does repeated media/bed reuse materially reduce environmental burden once amendment inputs, transportation, energy, replacement frequency, and yield effects are included? “Reusing soil = lower footprint” without that accounting is a label-based inference — exactly what R01-C033 warns against.

Provenance for quantitative estimates. Every major quantitative estimate carries the 10-field provenance record: geography, period, population, metric, functional unit, system, product boundary, method, directional bias, confidence.

Pre-registered provenance trap — denominator-free grower numbers. Expect this literature to be full of grower claims and commercial case studies reporting yield or input savings without denominator information: “30% less fertilizer” is nearly meaningless without the baseline regime, crop cycle, area, substrate volume, yield outcome, and what counted as fertilizer. Such figures are admissible as practice evidence or hypotheses, never as canonical quantitative claims, unless their provenance is reconstructable. Flag them as such rather than dropping or laundering them.

Key questions

A. Definitions & practitioner landscape

1. Build the operational taxonomy above. What do commercial and home-scale indoor living-soil/no-till operations actually do, cycle to cycle? Which named methodologies (soil-food-web, KNF, Korean/JADAM, “Clackamas Coot” style mixes, ROLS) map to which practices?
2. What is known about adoption: how many licensed producers run living-soil/no-till indoor systems? Named commercial examples with verifiable public documentation are valuable; label operator self-reports as anecdotal.

B. Soil biology indoors (mechanism tier)

3. What does the evidence say about establishing and maintaining functional soil biology (AMF, bacterial/fungal communities, micro/mesofauna) in containers and indoor beds — in cannabis where studied, in container/greenhouse horticulture otherwise? Does cannabis form functional AMF symbioses, and does high-P fertility suppress them (extend R01-C034)?
4. What happens to soil biological and physical properties in reused, undisturbed indoor media across multiple cycles — succession, stability, or decline? What does the evidence say about compaction, aggregation, and structure in containers without tillage?
5. Nutrient cycling capacity: can mineralization from amendments + biology actually meet cannabis’s documented nutrient demand (especially flowering N/P/K rates) without soluble inputs, at what media volume per plant, and on what evidence tier?
6. Media composition: what does container-media science say about compost fraction, aeration porosity, water-holding, and salinity in mixes intended for multi-cycle reuse? (Practitioner doctrine commonly specifies high compost fractions — e.g. “30%+ compost in soil mass” in the teaching archive — evaluate against evidence, preserving each source’s denominator basis per the M10 guard: dry mass, wet mass, or volume, never silently converted.)

C. Production outcomes (agronomy tier)

7. Yield and consistency: any direct comparisons of living-soil/no-till vs. soluble-fed media indoors (cannabis or comparable CEA crops)? Report ranges, functional units, and confounds (genetics, environment) — not single anecdote-derived numbers.
8. Quality — treated as five distinguishable things, never one bucket: cannabinoid concentration, terpene concentration and profile, sensory quality (aroma, flavor, appearance), contamination/safety, and consumer preference/market premium. Is there measured evidence for differences on any of these attributable to living-soil practice (as opposed to genetics, environment, harvest timing)? A paper showing altered terpene composition does not become “living soil produces better cannabis” — report which quality dimension moved, in which direction, on what evidence tier.
9. Contamination and compliance: microbial testing outcomes (total yeast/mold, Aspergillus), heavy-metal accumulation in reused media, and pesticide-residue considerations for living systems. Testing-lab datasets welcome.

D. Failure modes & operating conditions (required output category)

10. Document the characteristic failure modes of indoor living-soil/no-till, distinguishing acute failure (a crop-loss event: pathogen outbreak, irrigation failure, infestation) from cumulative drift (salinity, nutrient imbalance, disease pressure, structural change, or pest reservoirs accumulating across cycles while the system still looks fine). Actively search for longitudinal evidence — multi-cycle observations, however imperfect — rather than treating one successful harvest as proof of system stability. Under what conditions do practitioners abandon or reset these systems (parallel to R05-C030’s failure-mode profiling)?
11. Conversely: operating conditions associated with reported success — bed volume, cover/companion planting indoors, irrigation type, cycle cadence, amendment regime, IPM approach. Label the evidence tier of each.

E. Economics (feasibility tier)

12. Input-cost structure vs. bottle-fed programs (amendments vs. soluble nutrients; media amortization vs. replacement); labor deltas; floor-space and turnover penalties or gains. Jurisdiction- and scale-sensitivity flagged as in R05.
13. The differentiation case: given commodity price compression (R02-C025, R05-C029), what verifiable evidence exists that living-soil positioning commands a price premium, brand value, or resilience advantage? Market data preferred; marketing claims labeled as such.

F. Synthesis targets

14. Produce the three-tier map: mechanisms supported / production outcomes directly evidenced / extrapolations — and the conditions under which indoor no-till becomes more or less credible.
15. The 5–10 most repeated public claims on this topic and what the evidence says about each (see pre-registered myths).

Pre-registered myth checks

Evaluate at minimum the following, worded so that absence of evidence can remain absence of evidence (“unsupported” ≠ “the opposite is true”; unknown is an acceptable verdict — v1.3 requirement (e)):

• M1. “You never need bottled nutrients in a living soil.”
• M2. “The soil gets better every cycle, indefinitely.”
• M3. “No-till eliminates nutrient deficiencies.”
• M4. “Living soil automatically improves terpene/cannabinoid expression.”
• M5. “Synthetic nutrients kill the soil / sterilize the microbiome.”
• M6. “Living soil is inherently incompatible with commercial indoor production” (the inverse myth — treat with the same rigor).
• M7. “Compost teas reliably prevent disease / boost yield.”
• M8. “No-till saves money because you never buy soil again.”
• M9. “Living soil buffers pH for you — pH management becomes unnecessary.” (from the teaching archive — see archive-audit.md)
• M10. “The soil should be 30%+ compost.” (the archive’s lone quantitative living-soil spec, stated there as “30%+ Compost in soil mass”; test against container-media evidence on compost fraction, aeration, and salinity. Denominator guard: container-media recipes are frequently expressed by volume, and mass vs. volume produce radically different formulations given the bulk-density differences between compost, peat/coir, and aeration amendments — preserve whether each source reports compost fraction by dry mass, wet mass, or volume, and do not compare percentages across bases without conversion data)

Known controversies to handle carefully

• The soil-food-web commercial ecosystem (courses, consultants, microscopy services) is a marketing engine; label practitioner-guru sources as anecdotal regardless of production values.
• The inverse tribe (hydro/salt-fed) is equally motivated; “living soil can’t be consistent” needs evidence too.
• Field no-till literature is politically loaded (carbon markets, regenerative branding) — import mechanisms, not slogans (see R01-C023: the tillage/organic trade-off is real, the comparative ranking is disputed).
• Cannabis-direct studies may be sparse across whole question areas; a well-documented evidence gap is a first-class finding, not a failure of the run.

Standing source & citation requirements (PROTOCOL v1.3 — verbatim)

• Every citation supporting a substantive claim must carry a resolvable DOI, stable URL, or independently retrievable identifier — an unresolvable citation makes the claim unsourced.
• Citation-identity integrity: for every substantive source, the authors, title, publication year, venue, volume/pages where applicable, and DOI or stable identifier must resolve to the same bibliographic object — a DOI that resolves to a different real paper fails identity verification. If you cannot fully resolve the tuple, flag it explicitly as a partial citation rather than completing it from memory.
• Exact claim support: a source supporting the general direction of a finding does not automatically support its stated magnitude, population, geography, time period, or causal wording — preserve these distinctions in candidate claims.
• Corrections inherit the complete verification burden: a replacement citation or corrected statistic remains unverified until its existence, identity, and claim support are checked.
• When fact-checking public claims, “the myth is unsupported” and “the opposite is true” are different findings — do not overcorrect; unknown is an acceptable verdict.
• For each major claim, tag the evidence tier (T1 indoor/greenhouse cannabis · T2 other cannabis · T3 adjacent CEA/container horticulture · T4 general/field agriculture), the ANECDOTAL source-status flag where it applies, and the system definition the claim applies to.
• Report ranges and disagreements between sources, not averaged-away conclusions; list data gaps explicitly.

Candidate claims for editorial review

Conclude with 15–30 concise, testable candidate claims with supporting citation(s), each carrying its evidence tier and the system definition it applies to. Do not assign final evidence status or confidence; the editor will.

Search-stopping rationale

End with an explicit statement of why searching concluded: databases/queries covered, saturation observed, and named question areas where searching stopped because the literature is genuinely absent (distinguish “not searched” from “searched and empty”).

What “done” looks like

A structured report answering A–F with cited, tier-labeled evidence; the failure-modes/operating-conditions section (D) as a distinct output; the myth verdicts (calibrated); a top-10–15 strongest-sources list with one-line reliability notes; an explicit data-gap list; candidate claims; and the search-stopping rationale. Uncertainty findings are as valuable as positive findings.

Editor’s note (repo workflow): outputs land in raw-inputs/chatgpt-YYYY-MM-DD.md and raw-inputs/gemini-YYYY-MM-DD.md. Claude then runs verification & critique (critique.md), claim extraction (ledger.md, IDs R08-C001+), synthesis, and content extraction per PROTOCOL v1.3. Corpus anchors: R01-C033 (practices-not-labels), R01-C034 (AMF), R01-C035 (infiltration/no-till null), R01-C023 (tillage dilemma), R05-C012 (potting-media footprint), R05-C029 + R02-C025 (economic backdrop), R05-C030 (failure-mode profiling precedent).