How Bad Is Weed for the Planet? Depends What You Measured.
R04 Synthesis for Everyone — the draft public article
From the Cannabis Cultivation Knowledge Engine. Every factual statement traces to a verified claim in the R04 ledger (IDs like C016); sources in the plain-English ledger.
You have heard the numbers. Cannabis uses one percent of America’s electricity. A single joint is like driving a car for miles. Weed farms drink twice the water of vineyards and are draining the salmon streams. Or, from the other side: outdoor cannabis is basically a vegetable, packaging is the only real problem, and legal farms are clean by definition.
Every one of those claims has a source. Almost none of them survives being asked a simple question: measured on what?
Why there is no single answer
An environmental footprint is not a property of a plant. It is a property of a measurement. Change the boundary (just the electricity bill, or the whole farm, or everything through the dispensary), change the unit (a kilo of flower, a gram of THC, a square metre of canopy), change the impact category, or change the kind of grow, and the number changes, often by more than tenfold. Cannabis numbers can only be compared when all four match [C001].
That would be fine if there were many studies to compare. There aren’t. The entire life-cycle literature for cannabis cultivation is one modeled indoor study, one Canadian model, one real-data study of outdoor pots in Quebec, a drying study that borrows the Quebec data, a master’s thesis, and some hemp work. Nobody has done the accounting for a greenhouse, a light-dep hoop house, or plants grown in the ground [C002]. And the studies that exist share authors, data and assumptions, so when they agree they are often agreeing with themselves [C003].
The famous numbers, traced
“One percent of US electricity” comes from a 2012 model built on 2011 assumptions. It was a reasonable estimate of its moment. It has been repeated as a permanent fact ever since, and nobody has produced a current figure [C004].
“A joint equals X miles of driving” can be calculated, and the calculation is the lesson. On stated inputs, a gram grown in a warehouse works out to roughly six to thirteen miles; a gram grown outdoors in pots works out to a few hundred yards. The equivalency belongs to the grow and the power grid, not to the joint [C005].
“Six gallons a plant a day” was adopted from a 2010 county draft ordinance so that researchers could model stream impacts. The “twice the water of grapes” comparison is that same assumption run through a calculation [C009]. Real measurement exists, but only in experiments: subsurface drip used about a fifth less water than surface drip in one tunnel trial [C010].
Water: the question that finally got a scaled answer
For a decade the argument was binary. Cannabis is dewatering the salmon streams, or cannabis is a rounding error. A 2026 study across fourteen Northern California watersheds did the obvious thing and compared cannabis with the other water users in the same catchments. Cannabis was the biggest user in 1.6 percent of catchment-months. Homes and other farms were each the biggest user about a third of the time. Cannabis pressure on streams showed up only at monthly scale, mostly in dry years [C016].
That does not erase the older finding. In four specific watersheds studied in 2015, modeled cannabis demand exceeded summer low flow in three of them, in the worst cases by a lot [C017]. Both things are true. Cannabis water use is concentrated in particular places at particular times, where it adds to bigger users. Timing and location, not gross consumption, are the story, which is why California now requires growers to store winter water rather than divert streams in summer [C014].
Indoors, a different fact: in the one metered study, about seventy percent of the water given to plants came back as dehumidifier condensate. Recoverable water is not reused water, and no one has published a complete water balance for a commercial indoor facility [C018].
Wildlife: real poison, wrong generalization
The strongest evidence of environmental harm in the whole literature is wildlife poisoning around trespass grows on public land. Most tested fishers and spotted owls carried rat poison; some died of it; a banned insecticide turned up at the sites and in the streams below them [C023, C026]. Those grows are a specific population, criminal operations on public land, and no one knows what share of cannabis they represent.
The generalization fails in both directions. A study of owls near houses found most of them exposed too, with no clear link to cannabis and a strong link to the suburban fringe. Rat poison has more than one source [C024]. And licensed farms are not proven clean; they are unmonitored. Nobody has measured pesticide fate on a legal farm [C026]. California’s rule does not ban synthetic pesticides; it permits products whose ingredients are exempt from residue limits and whose labels don’t forbid the use, and even those can fail a lab test [C027].
The surprise in the dirt
The one outdoor study with real farm data produced the most counterintuitive result in the report. The plots given more fertilizer had the lower carbon footprint per kilo of flower. They yielded more, so they needed about half as many plants, and each plant sits in a pot of peat-based mix that is, by itself, sixty-five to seventy-five percent of the whole footprint [C029, C030]. Reusing that mix, in a model, cut a substantial slice of the emissions. Indoors, potting mix barely registers; the electricity dwarfs it.
So “organic” or “sun-grown” carries no footprint verdict. What matters is the media, the nitrogen, the timing of water, and where the farm sits [C029, C030, C017, C019].
What nobody has measured
Nobody has weighed the industry’s plastics, lamps, spent media, plant waste or packaging, so “packaging is the biggest problem” is unsupported, and so is “packaging doesn’t matter” [C033]. The story that California forces growers to landfill plant waste is false; the rule permits composting and tilling on site [C034]. No regulation and no certification has ever been evaluated for whether it changed an environmental outcome [C036, C037].
The takeaway
Ask “how bad is cannabis for the planet” and the honest answer is a map. Indoors, it is electricity, and the size of the number depends on the grid. Outdoors in pots, it is the potting mix. In Northern California, it is water timing in dry years in particular catchments. On public land, it is poison, from a population nobody can size. Everywhere, it is a set of blanks that a knowledge base should keep visible: greenhouse and in-ground footprints, an indoor water balance, nutrient leaching in the field, pesticide fate on legal farms, and a real count of how much land the crop actually occupies.
Anyone who hands you a single number has already chosen the boundary for you. Ask which one.