Where Cannabis Roots Live: Root-Zone Systems, Not Labels

R07 Verified Synthesis — Wave 2, systems comparison

A practitioner-record passage is withheld from the public copy.

The question, and why the labels dissolve

R07 asked what the evidence says about growing cannabis in “soil,” in “coco,” or “hydroponically.” As in R06, the literature does not study those categories. It studies root-zone systems — substrate composition and grade, container geometry, drainage, delivery architecture and recirculation, fertigation frequency and dry-back, nutrient-supply mode, root-zone temperature and oxygen, irrigation-water chemistry and the nutrient program — and outcomes attach to that bundle, not the trade label; the sources themselves use “hydroponic” for a synthetic-fertigated peat pot, for stone wool and coir, for drip architecture and for aeroponic propagation [R07-C001; S-219, S-220, S-205, S-221]. R07 therefore re-files every result into four pre-registered meanings of the word and four classes — water culture, inert-substrate fertigation, organic-matter-based (fibrous) soilless substrate, and soil — and treats any cross-system comparison without a root-zone descriptor as system-orphaned [R07-C001].

A practitioner-record passage is withheld from the public copy.

A practitioner-record passage is withheld from the public copy.

The one comparison, and what a bundle is

The keystone is Yep, Gale and Zheng’s 2020 indoor trial at a licensed Ontario facility, the only located controlled comparison of three root-zone systems in drug-type cannabis. The synthetic-fertigated 11 L peat arm (labelled “hydroponic”) out-yielded the other two root-zone systems (an aquaculture-effluent-fertigated 11 L peat pot and the 3 L peat/DWC aquaponic hybrid) by 42–116% in inflorescence biomass across two cultivars, with significant root-zone effects on most growth and physiological traits [R07-C002; S-219]. It cannot support “hydro yields more than soil”: the “hydroponic” arm is a peat container, and the arms differ simultaneously in nutrient source, root volume and delivery architecture — a system-bundle comparison, not a medium effect [R07-C002]. The magnitude carries high confidence (abstract-verified; the paper is paywalled); any explanation of why carries lower confidence because nutrition, root volume, architecture and substrate changed together, and a second researcher’s descriptor of the paper (60-L DWC tubs, other cultivars, a water temperature, a supplementation arm) is do-not-cite [R07-C002; critique §2b-1].

Positively, the same trial shows aquaponic cannabis, with nitrifying bacteria converting fish-derived ammonium to nitrate, was grown to flowering harvest, and organic-liquid fertilization in coir has a measured dose-response and floral harvest [R07-C028; S-219, S-232, S-101]. “You can’t grow organic in hydro or coco” is false as an absolute. The mechanism constraint is real — mineralization needs biology and residence time, which water culture and inert media supply only through a biological compartment — and certification is separate: as of 2026-09-09 California’s OCal certifies cannabis to comparable-to-organic standards and cannabis cannot carry the USDA designation [R07-C027; R06-C004]. Whether an organic root-zone system is competitive is the bundle question again: in that one trial the synthetic-fertigated 11 L peat arm produced 42–116% more inflorescence biomass than the other two arms [R07-C028, C002].

Three outcomes, plus time

R06’s interpretive rule carries forward and gains a fourth term. Cannabinoid concentration, flower yield (per plant or per m², per cycle or per year, at a stated density) and total cannabinoid yield move independently; R07 adds cycle time — days to a defined size, days to harvest, cycles per year — and never merges any of them; sensory claims name their class (chemistry, trained panel, consumer preference, anecdote) [ledger header].

On the time axis the finding is an absence. No controlled cannabis study located within the documented search (cutoff 2026-09-09) reports days-to-size, days-to-harvest or cycles per year by root-zone system with genotype and environment held fixed; “hydro grows faster” has no cannabis datum on its own metric, and no mechanism for “faster” (oxygen, water status, nutrient availability, root-zone temperature) has been measured in cannabis as a system effect [R07-C018]. The corpus’s only cycles-per-year figure is a 2026 New York greenhouse economics model — five two-month autoflower cycles against four three-month photoperiod cycles — which varies flowering strategy, not root zone [R07-C033; S-011].

A practitioner-record passage is withheld from the public copy.

Coir is not one material

The property behind “coco is the best of both worlds” is measurable, and the measurement dissolves the slogan. Across 13 coir dusts from Asia, America and Africa, every measured property differed significantly between and within sources — salinity 39–597 mS/m, cation-exchange capacity 31.7–95.4 cmol(c)/kg — and 17 further sources carried high exchangeable K, Na, Ca and Mg on the adsorption complex [R07-C003; S-222, S-223, S-107]. That is ornamentals substrate science (Spain, the Netherlands), general rather than cannabis-specific, but it means a “coco” property value has no meaning without source, grade (pith, chip, fibre, blend), compression history and container geometry; coir rows name them or are flagged grade-undisclosed.

A practitioner-record passage is withheld from the public copy.

Substrates head to head: what was measured was physics, not names

Substrate response is genotype-specific. In an indoor pot trial of two genotypes across peat, green-fibre and coir at Hohenheim, ‘KANADA’ lost more than half its floral yield in coir (3.84 g/plant) and green fibre (4.94 g/plant) relative to peat (8.56 g/plant), while ‘0.2x’ held stable high yields across all three (9.19 g/plant green fibre to 7.90 g/plant coir) [R07-C007; S-226]. A ranking from one genotype does not transfer; with two genotypes in one lab the generality is medium. A second interaction comes from stone wool: osmotic-stress crop steering at Utah State cut plant height 15% in both cultivars but cut flower yield about 20% in a cultivar-dependent pattern, with no difference in cannabinoid concentration — no compensating potency gain [R07-C014; S-230]. Genotype × system is tested, not assumed.

The one direct three-class comparison in the corpus is unmatched by design, and that is the finding. In a semi-controlled Hohenheim greenhouse, stone wool, peat-perlite and coir ran under medium-specific fertigation schedules; medium altered vegetative growth and assimilate partitioning (leaf-area index and leaf/stem accumulation impaired in peat-perlite; partitioning “marginal but significant”), the effect on CBD concentration was negligible, and the authors attribute the differences mainly to water-holding capacity and irrigation management — the bundle, not the material [R07-C008; S-220]. Abstract-level; no figure is quoted until the publisher PDF is re-read.

Water availability recurs as the correlate across labs. Across five coir/peat/perlite formulations under a common nutrient program in Spain, measured available water was the main factor affecting growth and production, and the two coir-fibre + perlite substrates grew least [R07-C009; S-227]. Within coir at Guelph, the drier, lower-container-capacity formulation delivered 29% higher dry floral weight in flowering, and the authors suggest drier coir may be preferred in flowering [R07-C010; S-228, abstract-level]. Direction differs by stage — flowering favoured drier, the growth study higher available water — so no universal “drier is better” is carried [R07-C010]. Organic-liquid fertilizer-rate optima also differed between the two coir substrates in flowering (212–261 mg N/L) but not in the vegetative trial (pooled optimum 389 mg N/L); this inherits R06-C008 without patch [R07-C011; S-101]. A Thai trial reporting a peat:perlite:vermiculite mix ahead on height, stem diameter, node number and inflorescence yield, and a coir:perlite:vermiculite:vermicompost mix highest in chlorophyll, is abstract-only with n, pot volume and durations unverified; no figure is drawn [R07-C012; S-229].

Water-status responses are not directionally uniform: controlled drought raised THCA and CBDA concentration 12–13% and THCA yield 43% in one Guelph study; high osmotic stress on stone wool reduced yield without a concentration gain; moderate deficit irrigation at Hohenheim cut water use without biomass or CBD-yield loss while severe stress impaired performance — three studies, three genotype sets, three bundles, corroborating heterogeneity, not a pooled effect [R07-C015, C016; S-231, S-230, S-182]. R11 owns the responses; R07 carries only the system-dependence.

A practitioner-record passage is withheld from the public copy.

Delivery architecture is its own axis

The second keystone extends R06-C015 in full text. In medicinal cannabis in 3 L expanded clay (LECA) at 27.5 plants/m², one clone, fertigated with a standard or P/K/Fe-augmented solution (+83% P, +39% K, +860% Fe), neither the augmented regime nor the fertigation system had a statistically conclusive effect on final inflorescence mass, but the recirculating system consistently outperformed drain-to-waste on total THC yield, especially after the ninth week, and the authors propose week 11 as the harvest time for recirculation [R07-C013; S-205]. Three limits travel with that result: the effect is on total THC yield and harvest timing, not inflorescence mass; it is one clone in one inert-substrate system; and the two architectures ran as sequential cycles, not side by side, so cycle is confounded with architecture; no universal recirculation yield advantage is established [R07-C013]. The circulating “+182%” is recirculating-versus-drain-to-waste under the enriched regime; “50.7%” is augmented-versus-control within drain-to-waste at week 11 only [R07-C013; critique §8].

Outdoors, subsurface drip cut applied irrigation water 18.6% relative to surface drip (188.19 vs 231.23 L/m²) in a Hohenheim foil tunnel with three CBD-rich genotypes, with inflorescence yield up 5% and CBD concentration up 9% — an architecture effect on water use in field soil under a tunnel, not transferable to indoor water culture [R07-C017; S-180]. And recirculation is a documented inoculum route: Fusarium oxysporum was detected in the recirculating nutrient solution of hydroponically grown cannabis sampled over three years in British Columbia [R07-C025; S-235]. Architecture thus sits in the matrix on three axes — cannabinoid yield, water use, biological exposure — none a verdict.

Pathogens and safety: no halo for any medium

Soilless media are not pathogen-free in production. Pathogenic Pythium, Globisporangium and Fusarium isolates were recovered from coir and stone wool under symptomatic cannabis in a commercial Connecticut facility, all tested isolates pathogenic to hemp seedlings (incidence 6.7–100%); independently, several Pythium species caused crown and root rot on cannabis in commercial British Columbia greenhouses [R07-C024; S-116, S-234]. Two labs, one finding: M9’s premise — sterile media, fewer root diseases — is rejected. “Soil has fewer diseases” is equally unsupported: no controlled incidence ranking by medium on a common denominator was located [R07-C024]. The temperature × dissolved-oxygen × oxygen-demand × pathogen interaction in reservoirs is unmeasured in cannabis; reservoir-temperature thresholds and chillers are convention [R07-C025, C035]. Under NaCl stress, drug-type cannabis in a hydroponic — but not an aquaponic — solution was negatively affected at 40 mM, while cannabinoid concentrations declined linearly in both; abstract-level pending a full read [R07-C019; S-232].

Worker safety enters the corpus for the first time. Stone-wool fibre carries occupational exposure limits — U.S. occupational guidance and limits include NIOSH and OSHA values (NIOSH REL TWA 3 fibers/cm³ and 5 mg/m³ total; OSHA PEL as nuisance dust), with an ACGIH TLV of 1 f/cc alongside (ACGIH is a non-governmental body); “not classifiable as to carcinogenicity” — as of 2026-09-10; no cannabis-specific exposure measurement for stone wool, peat dust or reservoir biofilm was located [R07-C026; S-236, S-237].

Water, waste and footprint: directions without magnitudes

“Hydroponics uses 90% less water” is a leafy-greens greenhouse-versus-field figure; both researchers traced it there, neither found a cannabis measurement supporting it, and no located cannabis study meters water use by root-zone system at fixed environment [R07-C029; critique §2a-3]. The measured cannabis savings — moderate deficit irrigation and subsurface drip — are irrigation-management effects at fixed medium, not medium effects [R07-C029, C016, C017]; the indoor water balance remains R04’s (about 70% of applied water recovered as HVAC condensate in one metered study; recoverable is not reused) [R07-C029; R04-C018]. Nutrient discharge by architecture has no cannabis measurement (the runoff concentrations one researcher cited to a Water Board order are not in it), and water metrics are never pooled: applied per event, per plant-day, per m²-cycle, withdrawn per kg dry flower, and the transpired, drained, recirculated, condensate-recovered and discharged fractions are different numbers [R07-C029].

A practitioner-record passage is withheld from the public copy.

Footprint rankings reverse. In a per-m³ LCA of media for Spanish rooftop farming a coir-based medium had the lowest burden in 5 of 8 categories while peat scored best socially; a 2026 European baseline shows a single-score reversal between white peat (7.7 mPt/m³) and coir constituents (9.0–10.5 mPt/m³) across indicators; a Latvian peat-extraction LCA is context-specific [R07-C031; S-238, S-240, S-239]. These are per cubic metre of medium, not per kilogram of flower, and the rockwool leg is absent, so no soil/coir/rockwool hierarchy is asserted; no cannabis LCA varies medium at fixed environment, and the per-area versus per-product reversal of R01 and R04 applies [R07-C031; R04-C029].

The system-property matrix

The brief said the report would end not with a winner but with a matrix, and the ledger carries that matrix as structure with no numeric cells [R07-C021; brief §H]. Root-zone systems are mapped on documented property axes, each cannabis cell graded direct / adjacent / general / unmeasured from this ledger only; every unsourced cell a researcher placed (yields per m² per year, cycles per year, kg media per plant, litres per kg flower, embodied carbon, hours to failure) reads unmeasured [R07-C021].

A practitioner-record passage is withheld from the public copy.

“Coco drip drain-to-waste,” “living-soil bed,” “rockwool flood-and-drain,” “DWC” and “peat-mix hand-water” occupy different regions of that space; they are not contestants [brief §H].

What the archive taught, and what R07 says back

The soil-versus-hydro pros-and-cons table never changed between 2016 and 2024: soil is cheaper, easiest to learn and “an easy environment for bugs to spread”; hydro gives “faster veg and bloom times,” is expensive, and “can be unforgiving” — M1, M4 and M7 in the archive’s own words, eight years without a number [archive-audit finding 2]; the matrix above is what each adjective was reaching for, and each cannabis cell is unmeasured [R07-C018, C020, C032]. M2, “soil tastes better,” the archive never made [R07-C022]. The 2017 “Basics” skeleton and 2018 plan promised a systems comparison and rockwool and coco tutorials; none was written — as in R06, the curriculum promised the comparison and shipped the taxonomy [archive-audit finding 10].

A practitioner-record passage is withheld from the public copy.

A practitioner-record passage is withheld from the public copy.

The public claims, calibrated

Hydro grows faster and yields more than soil (M1): not supportable as a category claim — no cannabis cycle-time or cycles-per-year datum by root-zone system [R07-C018, C002, C001, C033]. Soil tastes better or has more terpenes (M2): unsupported — no trained-panel or blinded-consumer study; chemistry differs by bundle but does not carry preference [C022, C023]. Coco is the best of both worlds (M3): not evidence-based as “best” [C003, C004, C007, C009, C010]. Hydro is unforgiving, soil is easiest to learn (M4): unverified — a plausible buffering hypothesis with no cannabis failure-velocity dataset [C020, C021]. Hydroponics uses 90% less water (M5): unsupported as a cannabis statistic [C029, C016, C017]. Hydro leads to a lot of waste, rockwool to landfill (M6): direction plausible, magnitude unmeasured [C030]. Soil buffers mistakes, hydro shows problems in hours (M7): unverified in cannabis [C020, C021, C004]. You can’t grow organic in hydro or coco (M8): false as an absolute; certification separate [C028, C011, C027]. Soilless media are sterile, so fewer root diseases (M9): premise rejected by two labs; the opposite also unsupported [C024, C025]. Living soil has a lower footprint than coco or hydro (M10): not supportable universally [C031, C030, C029].

What we do not know

Nineteen named gaps were searched and found empty (cutoff 2026-09-09) or not addressed by either pass [ledger §Named data gaps]. The load-bearing ones: no matched root-zone comparison on days-to-size, days-to-harvest or annual cycle count [C018]; no trained-panel or blinded consumer trial isolating root-zone system [C022]; no spent-media mass-flow inventory [C030]; no matched CAPEX, OPEX, labor and output-per-year dataset [C032]; no failure-velocity data by architecture [C020]; no controlled Pythium/Fusarium incidence ranking by medium on a common denominator [C024]; no multi-cycle living-soil bed versus soilless comparison, re-confirmed in a second independent search [C034; R08-C030]; no root-zone temperature or dissolved oxygen measured by system in flowering cannabis, every located comparison air-temp-only [C025]; no container-volume trial at fixed medium and irrigation [C006]; no buffered-versus-unbuffered coir trial or measured Ca/Mg stripping magnitude [C004]; no nutrient-discharge measurement by architecture [C029]. Irrigation-water chemistry is undisclosed in every located trial, so every quantitative row carries water-undisclosed pending full-text re-reads. Media-derived heavy metals in flower and microbial-compliance outcomes by system were addressed by neither researcher — open, not “searched and empty.” Six rows rest on abstracts (C002, C008, C010, C012, C019, C025); their full texts are owed.

Implications for practice, hedged to the grades

Choose a root-zone system by the properties you need to control, not the label, and expect the literature to have measured few of them in this crop [C001, C021]. Treat coir as a material that varies by source and grade: name the grade, expect exchange-site chemistry on RO water, and assume neither a calcium lockout nor its absence [C003, C004]. Test genotype against system before scaling, since a substrate ranking from one cultivar does not transfer, and manage water availability and container capacity as yield variables whose preferred direction differs between growth and flowering [C007, C014, C009, C010]. Do not expect recirculation to raise inflorescence mass; in the one trial it raised total THC yield, shifted the harvest window and moved a pathogen through the solution [C013, C025]. Take water savings from irrigation management, which has cannabis evidence, not from the medium [C016, C017, C029]. Assume no medium is pathogen-free and none is proven cleaner; test the crop, not the category [C024]. And do not price a system on a cycle count borrowed from another environment or flowering strategy [C018, C033].

For the curriculum and Grow Your 6 — editorial

This paragraph is editorial, not ledgered. The honest version of the archive’s story is that Danny taught “coco is hydro” for eight years because portions of the trade taught it, and that the pros-and-cons table, the 7–10× water figure and the 4–5 vs 1–2 slide are claims a working cultivator inherited without a number — then, in 2026, corrected on the record before the literature was consulted. That is the build-in-public material: the definition taken apart, the table set beside the matrix, the slide re-filed as indoor versus outdoor, “blank medium” retired for coco and kept for stone wool. For a six-plant home grower the transferable lessons are small: start in a clean, well-drained container of known grade; step container size up rather than “drowning” a small plant in a large pot, knowing this is convention, not measured effect; expect overwatering to be the commonest failure; treat “living soil” as a management choice whose substrate still has a composition; and treat every “faster,” “cleaner” or “more forgiving” claim as a question about a property nobody has yet measured in cannabis [C006, C036, C020, C034].

Downstream handoffs

• R09–R12: drought heterogeneity and NaCl stress to R11 [C015, C019]; compound-level substrate chemistry (S-219, S-226, S-233) to R12 [C023]; root restriction to R09/R11 [C005, C006]; the root-zone temperature × DO gap to R11 [gap 9].
• R13–R16: no multi-cycle bed comparison [C034]; coir exchange chemistry against mineral-soil CEC [C004]; mineralization needs biology and residence time [C028]; media reuse and fate [C030]; R08-C030 lineage.
• R20 (IPM): pathogens in soilless media and recirculating solution, no incidence ranking [C024, C025]; the archive’s “soil spreads bugs” with the same absent denominator; stone-wool fibre exposure [C026].
• R21–R24: economics without denominators, cycles per year as a strategy variable [C032, C033]; failure velocity and the matrix as design tool [C020, C021]; the container ladder as convention [C006]; Grow Your 6 material [C036]; OCal [C027]; water metrics never pooled [C029].
• R05: the “4–5 vs 1–2” slide is R05’s axis — cross-reference at next touch, no re-open [C018].
• R08: no new patches; R08-C017 extended [C024], R08-C030 re-confirmed [C034], R08-C007 consistent [C011].

A practitioner-record passage is withheld from the public copy.

What would change these conclusions

A matched multi-genotype root-zone trial reporting days-to-harvest, yield per m² per cycle and per year at stated density, and cannabinoid concentration, with nutrition held to plant-available equivalence and root-zone temperature, dissolved oxygen and irrigation-water chemistry logged; a trained-panel or blinded consumer study on matched systems; a controlled pathogen-incidence ranking by medium; a buffered-versus-unbuffered coir trial with measured Ca/Mg stripping; a spent-media mass-flow inventory and a nutrient-discharge measurement by architecture; a matched cost and labor dataset; and the full texts of the six abstract-level rows. Until then, cannabis root zones are well described as systems and unmeasured as categories.

A practitioner-record passage is withheld from the public copy.