Verified Synthesis — R05: Indoor vs. Greenhouse/Light-Dep vs. Sun-Grown Cannabis
Version: v1.0 — SIGNED OFF by Danny, August 28, 2026
Date: August 25, 2026 · Editor: Claude · Researchers: ChatGPT Deep Research, Gemini Deep Research
Ledger: ledger.md — inline citations like [R05-C001] refer to graded, sourced ledger claims. Full citations in SOURCE_LIBRARY.md.
Executive summary
Ask most people in cannabis which cultivation environment is “best” and you’ll get a confident answer. Ask the scientific literature and you get a more interesting one: the three broad environments — indoor, greenhouse, and outdoor sun-grown — solve different problems and externalize different costs, and the public debate is far more confident than the underlying evidence warrants.
Where the evidence is genuinely strong, it is strong in one direction: energy and carbon. Fully indoor cultivation carries a dramatically larger climate burden per kilogram of dried flower than sun-grown production — modeled at roughly 2,300 to 5,200 kg CO2e/kg for US indoor facilities depending almost entirely on where the facility sits [R05-C001], versus reductions on the order of 90% for outdoor systems in comparative models [R05-C003]. Modern LED lighting genuinely helps but does not close this gap, because lighting is only part of a coupled system in which climate control and dehumidification remain enormous loads [R05-C004, R05-C005]. This is the best-established finding in the entire report [R05-C002].
Where the evidence is weakest is exactly where the marketing is loudest: product quality. The belief that indoor flower is inherently more potent is not established by any direct evidence independent of genetics, management, and lab variation [R05-C020]. The counter-belief that sun-grown flower universally has richer terpene profiles is also not established — it rests on essentially one small, confounded (but genuinely fascinating) matched-clone study [R05-C019, R05-C021]. The indoor price premium is real, but it measures market perception and grading convention, not chemistry [R05-C022].
And the environmental story has a twist that “sun-grown = sustainable” marketing does not survive: in the one outdoor cannabis life-cycle assessment with field data, potting media accounted for up to 75% of the outdoor system’s global warming potential [R05-C012]. Outdoor cultivation has the smallest energy footprint, but what you grow in, what you feed the plants, and — above all — when you take water from which watershed [R05-C013] determine whether it is actually sustainable [R05-C011].
Finally, the honest structural finding: “greenhouse” is not one thing. A passive hoop house and a sealed, supplementally-lit glasshouse are different systems with radically different footprints [R05-C008], and nobody has yet published the subtype-by-subtype carbon numbers that would let us rank them precisely [R05-C009]. Anyone who gives you those numbers today is extrapolating.
What the evidence establishes
Energy and carbon: the settled direction
The modern anchor is Summers, Sproul & Quinn’s 2021 national life-cycle assessment (S-001), which modeled US indoor production at 2,283–5,184 kg CO2e per kg of dried flower — the low end in hydro-powered, mild-climate Washington, the high end where fossil-heavy grids meet punishing HVAC loads [R05-C001]. The single largest lever is not grower skill or fixture choice; it is location — grid carbon intensity and climate.
Every subsequent independent analysis points the same way. A Canadian comparative model found outdoor production cutting carbon on the order of 90% [R05-C003]; a 2025 industry-wide analysis modeled US cannabis at roughly 44 Mt CO2e per year — comparable to millions of homes — with a shift toward outdoor cultivation reducing emissions by as much as 76% in its scenario [R05-C010]. The direction of this comparison is unusually consistent for such a young literature [R05-C002].
Two things make this more than a static fact. First, LEDs help but do not fix it. Measured fixture efficacy improved from about 1.7 μmol/J (HPS) to 2.5 μmol/J (modern LED) [R05-C004] — a real gain — but direct field metering of working indoor facilities found HVAC alone consuming ≈45% of facility energy, and dehumidifiers performing at only 40–65% of their rated efficiency under real cultivation conditions [R05-C004]. Counterintuitively, swapping HPS for LED can even increase winter heating demand in cold climates, because the old lights were also space heaters [R05-C005]. Which load dominates — cooling, dehumidification, or heating — is itself a function of climate [R05-C006].
Second, the famous numbers have vintages. The most-quoted indoor carbon figure in popular coverage traces to Evan Mills’ 2012 “Up in Smoke” study — built on magnetic-ballast HPS, uninsulated rooms, and a 2010 grid [R05-C033]. Quoting it as current is an error. The elegant footnote: the author of the definitive 2025 industry analysis is the same Evan Mills. The person who defined the HPS-era baseline published its modern replacement — and the conclusion, thirteen years of technology later, still stands in direction if not in detail.
Greenhouse is a ladder, not a rung
California’s own licensing code concedes the point, splitting “mixed-light” cultivation into wattage tiers [R05-C008]. The useful ladder runs: passive hoop house → light-deprivation greenhouse → supplemental-light greenhouse → sealed high-tech greenhouse. Each added layer of control drags the system’s energy profile toward indoor’s. A passive hoop is nearly outdoor; a sealed glasshouse running LEDs, chillers, and CO2 enrichment is nearly a warehouse with free daylight. Treating “greenhouse” as one category — as most published LCAs still do — produces misleading averages [R05-C009].
Light deprivation itself is settled mechanism: photoperiod-sensitive cannabis flowers in response to sufficiently long uninterrupted darkness, so blackout curtains that artificially extend the night induce flowering months before autumn would — enabling multiple harvests per season [R05-C025]. How many is a function of cultivar, climate, propagation scheduling, and facility, with 2–4 typical in commercial practice.
The externality map
Outdoor cultivation’s environmental risks are real but specific: the damage concentrates in timing and siting, not gross consumption. Peak irrigation demand lands exactly on the seasonal low-flow period of Mediterranean-climate watersheds: in the four studied Northern California basins, modeled cannabis demand ranged from 2–23% of annual seven-day low flow in the least-impacted watershed to 34–165% and 36–173% in two others, and exceeded 100% of a near-zero minimum in the fourth — a genuine threat to salmonid habitat, and the reason off-season water storage is the standard mitigation [R05-C013]. (Corrected 2026-09-06: the original sentence gave “up to 23%,” the least-impacted watershed’s ceiling, as the study’s ceiling; independently second-checked by ChatGPT — see verification/R05-C013-correction-2026-09-06.md.) The dramatic pesticide-contamination findings in the literature come overwhelmingly from illicit and trespass grows and must not be generalized to licensed farms [R05-C017]. And the potting-media finding [R05-C012] reframes the whole question: an outdoor container grow that trucks in peat-based media every cycle may forfeit much of its sustainability advantage — a result that points directly at soil reuse, no-till, and locally-sourced amendments, the subjects of this corpus’s later reports.
What is probable but not settled
- The matched-clone chemistry study. Zandkarimi et al. (2023) grew genetically identical clones indoor (artificial media, electric light) and outdoor (living soil, sun) and found: broadly comparable primary cannabinoids; greater terpene abundance and diversity outdoors, dominated by sesquiterpenes (β-caryophyllene, α-humulene, α-bergamotene, α-guaiene, germacrene B); unusual C4/C6 cannabinoid homologues only in outdoor samples; and more oxidized, degraded cannabinoids indoors [R05-C019]. This is a genuinely important result — and it is one study, two cultivars, with media confounded with environment. It establishes that environment can materially alter cannabis chemistry [R05-C018]. It does not establish that outdoor flower is better.
- Water-use efficiency vs. water use. In a controlled greenhouse trial, adding supplemental light raised total evapotranspiration while cutting water per gram of flower by ~37% [R05-C015] — a clean illustration that absolute use and efficiency can move in opposite directions, and why single-metric water claims mislead [R05-C014].
- Condensate recovery. Field metering suggests ~70% of water applied in sealed facilities returns as recoverable HVAC condensate [R05-C007] — controlled environments spend energy to gain water recapture that open fields give up to the sky.
- Economics. Industry surveys across years agree on the ordering — outdoor cheapest, greenhouse middle, indoor most expensive per pound — while disagreeing on the numbers, which are hostage to jurisdiction, energy prices, taxes, and time [R05-C027]. The sharpest recent evidence: a 2026 New York greenhouse analysis in which the same facility flipped from −$1.48/ft²/yr (autoflower strategy) to +$7.18/ft²/yr (photoperiod) on yield and planting-material assumptions alone [R05-C028]. More cycles ≠ more profit. What is structural: high-OpEx indoor production is the first casualty of wholesale price compression [R05-C029].
Traditional practice & practitioner knowledge
Some of the most reliable knowledge in this domain has never been formally studied, and we label it honestly rather than dressing it in borrowed citations. The characteristic failure modes of each system are practitioner consensus: outdoor growers fear autumn rain, frost, and — increasingly — wildfire smoke taint (a mechanism currently borrowed from wine-grape research); light-dep growers fear a torn curtain or stalled motor, because a single light leak during the dark period can trigger hermaphroditism; indoor growers fear the power outage or chiller failure that cooks a sealed room in hours [R05-C030]. The indoor price premium itself is best understood as traditional market practice — a grading convention built on bag appeal, density, and THC-number shopping — rather than a chemical judgment [R05-C022].
Disputes & open questions
Our two researchers disagreed in instructive ways (full record: critique.md). Where one asserted that light-dep is “the ultimate environmental and quality sweet spot,” the evidence supports only “often an attractive middle path, unquantified at the subtype level” [R05-C009]. Where one asserted outdoor flower “passes compliance testing at high rates,” no dataset controlling for state-by-state testing differences actually exists [R05-C023]. Both the myth (“outdoor can’t pass testing”) and the counter-myth (“outdoor passes fine”) are currently unsourced. Contamination is real in every system — it just arrives by different doors: ambient spores and drift outdoors; humidity pockets, recirculating pathogens, and Hop Latent Viroid indoors; pesticides documented in illicit grows on both sides of the roof [R05-C023]. Heavy-metal risk follows the media and inputs, not the environment [R05-C024].
What we don’t know
Stated plainly, because these gaps are findings:
- What share of North American cannabis is actually grown in each environment. License counts and licensed area exist; production shares do not [R05-C031].
- Subtype-specific greenhouse footprints. The ladder is real; its rungs are unquantified [R05-C009].
- Whether the terpene findings generalize beyond two cultivars and one confounded comparison [R05-C021].
- Comparative contamination rates under matched testing rules [R05-C023].
- Reliable per-plant or per-kg water figures — the famous 22.7 L/plant/day is a 2010 county-ordinance assumption, not a measurement [R05-C014].
- Physical/commercial quality (density, bag appeal, aroma retention) under blinded, genetics-matched comparison — essentially unstudied [R05-C022].
Then vs. now — the archive delta
This report re-tested what Danny taught in the 2016–2020 Crop Circle and [course] classes (archive-audit.md):
- “Full control of environment = medical quality” (2016) → retired in its strong form. Controlled environments demonstrably buy consistency and physical protection; they do not demonstrably buy superior chemistry [R05-C020, R05-C019]. The 2016 claim was honest market observation — indoor did command the premium — but the premium measures perception [R05-C022].
- The four-garden harvest taxonomy (2016) → survives almost untouched. Full-term outdoor (1 harvest) / light-dep (multiple) / single-room indoor (4–5) / perpetual — a decade later, the professional literature’s cadence tables match it nearly line for line [R05-C025, R05-C026].
- “Light dep up to 5×” → revised to “multiple, typically 2–4, context-dependent” [R05-C025].
- The HPS-first lighting list (2016) → a time capsule. LED was one bullet among six; it is now the default, and the honest 2016 con — “can be costly with electricity” — has been quantified into tonnes of CO2 per kilogram [R05-C001, R05-C004].
A practitioner-record passage is withheld from the public copy.
Implications for cultivation practice
For the home grower, the honest translations are: choose your environment for the problems you want to solve (control and scheduling vs. energy cost vs. seasonal exposure), not for a chemical superiority no one has demonstrated [R05-C020, R05-C021]. If sustainability motivates you, the sun is the single biggest lever [R05-C002] — but so is what fills your containers and how many times you refill them [R05-C012], which is precisely where living soil and media reuse enter the story. If you run light-dep, engineer the blackout like the single point of failure it is [R05-C030]. If you grow outdoors in a dry-summer climate, store water in the wet season [R05-C013]. And treat every confident universal claim about potency, terpenes, or cleanliness — in either tribal direction — as unproven until someone runs the studies [R05-C021, R05-C023].
Sources
Anchor sources: S-001 (Summers 2021), S-002 (Zandkarimi 2023), S-003/S-004 (Desaulniers Brousseau 2024 ×2), S-005 (Mills 2025), S-006 (Mills 2012), S-007 (PG&E/ETCC field study), S-008 (Bauer 2015), S-010 (Collado 2024), S-011 (Vergara & Ruterbories 2026), S-012 (Backer 2019), S-013–S-014 (CA DCC, Health Canada), S-015–S-016 (industry cost surveys), S-017–S-023 (contamination literature), the author’s archive (Danny’s teaching archive), S-025 (Westmoreland/Kusuma/Bugbee 2021). Full citations, verification status, and the refuted-citation log: SOURCE_LIBRARY.md.
Changelog
- v1.0 (2026-08-28) — Danny’s sign-off. Canonical synthesis; content extraction proceeds from this version.
- v0.1 (2026-08-25) — Initial draft from ledger v0.1 (33 claims).