R04 Ledger in Plain English — The Environmental Footprint of Cannabis Cultivation

Companion to Claim Ledger v1.0 (signed off 2026-09-07). Each entry restates the claim for a general reader, says how solid it is in plain words, and points at the evidence. Full citations, grading, boundary lines, and evidence-type tags live in ledger.md and SOURCE_LIBRARY (S-166–S-197).

How to read “how solid”: Solid = well-documented, verified records or full-text confirmation. Promising = real evidence, but one study, one site, or model-dependent. Honestly unknown = nobody has defensibly measured it, and saying so is the finding.

The map, not the number (C001–C003)

• C001 — THE HEADLINE: there is no single “environmental footprint of cannabis.” Solid. Every number out there belongs to a particular boundary (just electricity? the whole farm? through to the store?), a particular unit (per kilo, per gram of THC, per square metre, per plant), a particular impact category, and a particular kind of grow. You can only compare numbers that share all four. The honest product is a map of what was measured on what, with the blanks marked.
• C002 — The whole cannabis life-cycle literature fits on one hand. Solid. One modeled US indoor study, one Canadian industry model, one real-data outdoor study of pots in Quebec, a 2026 drying study that reuses the Quebec numbers, a Washington master’s thesis, and some hemp work. Nobody has published a life-cycle assessment of a supplemental-light greenhouse, a light-dep hoop house, or cannabis grown in the ground.
• C003 — Those studies lean on each other more than their count suggests. Solid. Same authors, same inventories, the same water-reporting lineage, overlapping wildlife samples, the same Colorado field campaign, the same background emission factors. When “several studies agree,” check whether they share a dataset.

Energy and carbon (C004–C008)

• C004 — “Cannabis uses 1% of US electricity” is a 2012 model, not a current fact. Solid on provenance. It came from a model built on 2011 assumptions (about 20 TWh a year, about $6 billion). A 2018 industry report re-used it with voluntary facility data and market guesses. Nobody has established a current national share.
• C005 — “A joint equals X miles of driving” has no universal answer. Promising as arithmetic, useless as a fact. On stated inputs, an indoor gram works out to roughly 5.6–12.8 miles and an outdoor-pot gram to roughly 0.15–0.27 miles at the EPA’s car factor. The spread is the point: the answer depends on the grow and the grid, not the joint.
• C006 — New 2025–26 studies add cases, not a new range. Solid. Nothing published since January 2025 replaces the inherited indoor-vs-outdoor carbon figures from R05.
• C007 — Drying is its own carbon story. Promising. In the one drying study, pretreatment cut drying-stage emissions by half to nearly three-quarters, but the whole-production reduction was far smaller. Don’t let the big number stand in for the small one.
• C008 — Cannabis vs tomatoes vs beer: mostly not a fair fight. Solid. Per-area energy comparisons among indoor crops can work; per-kilo comparisons of dry flower against fresh produce are distorted by water content; no boundary-matched per-dose comparison with other consumer goods was verified.

Water (C009–C018)

• C009 — “Six gallons a plant a day” is an old ordinance assumption, and “twice the water of grapes” grew out of it. Solid on provenance. The 22.7-litre figure was adopted from a 2010 county draft ordinance for modeling; the grape comparison is that assumption run through an area-and-season calculation, not a paired experiment.
• C010 — Subsurface drip used 18.6% less water in one tunnel trial, with better yield and CBD. Promising. One site, one season, three CBD-type genotypes, Germany.
• C011 — The only field evapotranspiration measurement is low-density hemp. Promising, and it can’t rescue any per-plant number, because field ET includes the soil and the weeds.
• C012 — Moderate drought kept indoor yield in one experiment. Promising; leaf-level, not facility-level.
• C013 — Among 901 enrolled Northern California growers, wells were the most common reported water source (58%), then streams (22%) and springs (16%). Solid, with a catch: that’s a count of sites, not a share of water, and compliant sites reported wells far more often than non-compliant ones.
• C014 — Storage shifts when water is taken, but most farms didn’t have enough of it to skip summer stream diversion entirely; California has required dry-season forbearance since 2019. Solid.
• C015 — Unpermitted farms pulled more water in the dry season than permitted ones, in models. Promising. Scenario modeling shows permitting plus storage could halve dry-season extraction in many watersheds.
• C016 — THE WATER HEADLINE: cannabis was the biggest water user in only 1.6% of catchment-months. Promising, high confidence as graded. Across fourteen upper watersheds, homes and other farms were the top user far more often; cannabis pressure showed up only at monthly scale, mostly in dry years. Cannabis water use is concentrated in place and time, and it adds to bigger users. It is neither “devastating the watersheds” nor “negligible.”
• C017 — In four specific watersheds, modeled cannabis demand exceeded summer low flow in three of them. Solid. The least-hit watershed ran 2–23% of low flow; two others 34–165% and 36–173%; the fourth blew past a minimum that was near zero. (The corpus used to quote just “23%.” That was wrong, and it’s fixed.)
• C018 — Indoors, about 70% of the water you give plants comes back out as dehumidifier condensate, in the one metered study. Promising. Recoverable is not reused, and no one has published a full water balance for a commercial indoor facility.

Land, habitat, wildlife (C019–C025)

• C019 — Mapping found 4,428 grows in Humboldt County, clustered, often remote, often on steep slopes, some near threatened-fish streams. Solid, as siting proxies. Nobody measured the erosion or the fish.
• C020 — Per acre cleared, cannabis fragmented forest more than logging did, though logging cleared more land overall. Promising; the exact multipliers are behind a paywall.
• C021 — Nobody can tell you what share of farmland cannabis occupies. Honestly unknown. License sizes are maximums, big-outdoor licenses have no cap, and no state publishes planted acreage. Any percentage you see was made up.
• C022 — Licensed farms sit near fewer unlicensed ones after legalization, but that doesn’t prove licensing protects habitat. Promising as an association only.
• C023 — The strongest direct evidence of wildlife poisoning comes from trespass grows on public land. Solid, and bounded. Most tested fishers and spotted owls carried rodenticide; some died of it; carbofuran was found. But those are tested carcasses, not whole populations, and nobody knows what share of cannabis trespass grows represent.
• C024 — Owls near houses were poisoned too, with no clear link to cannabis. Solid. Sixty-two percent of tested barred owls were exposed; the pattern tracked the wildland–urban interface. Rat poison has more than one source.
• C025 — Wildlife on small private farms changes its behavior; whether that matters is unresolved. Promising, low confidence; one shared dataset, plus a dissertation nobody has fully read.

Chemicals and air (C026–C032)

• C026 — Pesticides have been found in streams below trespass grows; nobody has monitored licensed farms. Solid for trespass sites (diazinon, carbofuran, in headwaters below eradicated grows), honestly unknown for legal farms. Product testing is not environmental monitoring.
• C027 — No pesticide is registered for cannabis; California allows products whose ingredients are tolerance-exempt and label-compatible, and even those can fail a residue test. Solid. “Zero synthetic pesticides” and “legal means clean” both misdescribe the rule.
• C028 — Tripling phosphorus tripled nothing useful and multiplied leachate twelve-fold. Promising; one chamber trial on high-CBD hemp. No field study of nutrient leaching from cannabis exists.
• C029 — The higher-fertilizer plots had the lower footprint per kilo. Promising, and counterintuitive: they needed about half as many plants, so about half the potting soil. Measured per gram of THC the gap narrowed but didn’t flip. “Less fertilizer, greener” isn’t established.
• C030 — Potting mix was 65–75% of that outdoor grow’s carbon footprint; reusing it (in a model) cut a substantial slice. Promising; one system, one scenario. Indoors, the mix barely registers.
• C031 — Cannabis terpenes are real and measured; the modeled effect on regional ozone was under one part per billion in the Colorado scenarios. Solid on the measurements, promising on the model. Odor complaints are a separate, real issue.
• C032 — CO₂ enrichment: no one has studied its net climate effect. Honestly unknown.

Waste and rules (C033–C037)

• C033 — Nobody has weighed the industry’s media, plastics, lamps, plant waste, or packaging. Honestly unknown. “Packaging is the biggest impact” is unsupported, and so is “packaging doesn’t matter.”
• C034 — California lets growers compost plant waste on site; the “50/50” rule is about making it unrecognizable, not about landfills. Solid. The “forced into landfills” story is false for California.
• C035 — Fixes rank only within a category, on the same baseline. Solid as a rule. Moving production outdoors cuts carbon in every model, by amounts you can’t pool; drip and storage help water; less phosphorus helps leaching; for habitat and waste there’s nothing measured to rank. There is no honest overall ranking.
• C036 — Rules set requirements, not results. Solid. California just repealed its cultivation electricity-reporting and carbon-offset rules (July 2026); no regulation anywhere has been evaluated for its environmental effect.
• C037 — Certifications certify practices, not footprints. Solid. None verified shows a measured improvement per unit of product.

What nobody knows yet

Greenhouse, light-dep and in-ground LCAs · a full indoor water balance · field nutrient leaching · pesticide fate on licensed farms · any waste or packaging mass flow · a real planted-acreage denominator · whether any rule or certification changes anything · the footprint of a six-plant home grow.