Yield per square metre indoors: what the studies actually measured

One controlled trial has pushed cannabis further under artificial light than any other published study, and still found no ceiling: dry flower yield rose from 116 g/m² at an average canopy photosynthetic photon flux density (PPFD) of 120 µmol·m⁻²·s⁻¹ to 519 g/m² at 1,800 µmol·m⁻²·s⁻¹, climbing in an unbroken straight line across that range ^[1]. That is one real number, from one real trial, and it answers a narrower question than most people ask of it. Almost everything else circulating as "yield per square metre" is either an unsourced marketing claim or a very different kind of evidence: what investigators recorded when they measured real, if illicit, dismantled indoor grows. This page reports what each source actually measured, what "per square metre" has to mean before either number is useful, and how to turn your own canopy's light reading into an expected range instead of a hopeful guess.

Lowest tested light, dry yield

116g/m²

120 µmol·m⁻²·s⁻¹ APPFD, controlled trial [1]

Highest tested light, dry yield

519g/m²

1,800 µmol·m⁻²·s⁻¹ APPFD, controlled trial, no plateau reached [1]

Netherlands forensic-survey average

505g/m²

86 raided indoor grows, regression point estimate [3]

Belgian trial range, by light and cultivar

125–483g/m²

400–600 W/m², four cultivars, 11-week cycle [4]

What the denominator has to mean before the number means anything

"Grams per square metre" only means something once you know which square metre. Three different areas get called "per m²" in cultivation writing, and they don't produce comparable numbers:

• Canopy area: the footprint actually occupied by leaf and flower under light. This is what a controlled trial like Rodriguez-Morrison's reports, built from a fixed planting density (0.09 m² per plant, roughly 11.1 plants/m²) ^[1].
• Floor area harvested: canopy plus the walkways and gaps woven through it.
• Total licensed or room area: floor area harvested plus everything else in the building, mother plants, seedling space, drying rooms, corridors.

A literature and grower-interview review prepared for a US state cannabis regulator found that ancillary space, everything that isn't flowering canopy, "can easily be half as large as the canopy area that is harvested," meaning as little as two-thirds of a real facility's floor plan is ever growing flower at any one time ^[2]. Quote a yield figure against total floor area instead of canopy area and you can understate the true canopy number by a third or more without anyone lying about anything; quote a canopy figure as if it were achievable across a whole room and you overstate what the room will actually produce by a similar margin. Every figure quoted below is per square metre of canopy, stated explicitly wherever it might otherwise be ambiguous.

The one trial built to find a ceiling, and it didn't find one

Rodriguez-Morrison, Llewellyn and Zheng grew a single cannabis cultivar, 'Stillwater', in deep water culture under eight target canopy light intensities from 200 to 1,600 µmol·m⁻²·s⁻¹, with actual measured average PPFD (APPFD) across individual plants ranging from 120 to 1,800 µmol·m⁻²·s⁻¹ ^[1]. The flowering photoperiod was 12/12, CO₂ was left at ambient (437 ± 39 ppm during the light period, no enrichment), and air temperature was held close to 25.3 °C (77.5 °F) throughout, with RH close to 60.5% during the light period and 53.1% during the dark period. Dried inflorescence weight, at roughly 11% final moisture, rose from 116 g/m² at 120 µmol to 519 g/m², 4.5 times more, at 1,800 µmol, and the paper describes the relationship as linear across the whole tested range, with no sign of levelling off even at the top ^[1].

At the trial's own planting density, those totals work out to roughly 10.4 g/plant at the low end and 46.7 g/plant at the high end, computed from the reported density rather than published directly. That density, about eleven small plants per square metre grown to a single main cola each, is far higher and more uniform than almost any home tent or commercial room: closer to a research-grade screen-of-green than to a handful of larger, trained plants. Treat the g/m² figures as what a very dense, very even canopy can do under that specific light ramp, not as a number any grower gets simply by matching the PPFD.

One more figure from the same trial is worth carrying forward: energy-use efficacy at the 900 µmol treatment worked out to 0.54 g of dry flower per kWh of lighting electricity ^[1]. That number reappears in grams per watt: a useful metric or not, where it's worth more than it is here.

Why no real room runs at 1,800 µmol·m⁻²·s⁻¹

The paper doesn't explain why growers can't simply run every tent at the top of this range, because it wasn't built to answer that question: the trial held temperature and RH steady by design, so heat removal was never the limiting factor in the experiment itself. In a real room, it usually is. Roughly full midday sun delivers somewhere in the region of 2,000 µmol·m⁻²·s⁻¹; putting that much light on a canopy electrically means putting a proportional amount of waste heat into the same airspace, since only part of a fixture's electrical input leaves as usable photons and the rest, plus most of what the leaf absorbs and doesn't use, ends up as heat the room has to remove. Push canopy PPFD that high without matching cooling and dehumidification capacity and leaf temperature climbs, vapour pressure deficit (VPD) drifts out of range, and stomata close to protect the plant long before the light itself becomes the constraint; see what VPD actually measures for why that happens.

Ambient CO₂ is the other candidate limit, and it's worth being precise about what this particular trial does and doesn't show on that point: yield kept rising in a straight line to 1,800 µmol at ordinary ambient CO₂ (437 ppm), with no enrichment and no sign that CO₂ availability was capping the response within this cultivar and setup ^[1]. That argues against CO₂ being the hard ceiling here, at least at this scale and photoperiod. The practical ceiling most real rooms hit first is a plant-room engineering one, dehumidification and heat-removal capacity, not a fixed biological one. A room that can hold VPD and temperature in range at 1,000 µmol may simply not be able to at 1,800, whatever the photosynthetic machinery could theoretically do.

What police raids and grower-run trials add that a lab can't

A single-cultivar research trial answers "how does yield respond to light, all else held constant." It says nothing about what typical growers, working with typical mistakes, typical cultivars and typical equipment, actually produce. Three sources fill that gap, and they are not all the same kind of evidence.

Toonen, Ribot and Thissen (2006) is the genuine article: a regression model built from 86 samples recorded by law enforcement from dismantled illicit indoor cannabis plantations across the Netherlands, not a staged experiment ^[3]. Its point estimate for dried female flower was 505 g/m² of grow-room floor area as police recorded it (a median raided room ran 15 plants/m²), not a strict canopy measurement; in these small, densely packed illicit setups the two usually run close together, but it's worth naming given how much the denominator matters (see above) ^[3]. The regression model's detail is as reported in a later literature review ^[2]. This is what real, mostly unoptimised, non-research grows actually produced when investigators measured what was left behind, which is exactly why it's useful: it isn't selected for best practice.

Vanhove, Van Damme and Meert (2011) sits between the two: a controlled, full-factorial trial (four cultivars, two planting densities at 16 and 20 plants/m², two lighting power densities at 400 and 600 W/m², an 11-week flowering cycle), but one deliberately built and benchmarked against Belgian police data on typical illicit setups ^[4] (figures as reported by ^[2]). Converted to metric, yields ranged from roughly 125 g/m² to 339 g/m² under 400 W/m² and roughly 231 g/m² to 483 g/m² under 600 W/m², climbing to about 438 g/m² for the two highest-yielding cultivars specifically at 600 W/m² ^[4]. Belgian police reviewing the study's own figures noted that they sat below the average yield found in common illicit indoor plantations, which is a useful sanity check in the other direction: even a fairly generous controlled trial can undershoot what real grows achieve ^[4]. Planting density, 16 against 20 plants/m², made no measurable difference to yield per square metre over that narrow range; lighting power did, substantially.

Potter and Duncombe (2012) tested lighting power density directly: three levels (270, 400 and 600 W/m² under high-pressure sodium, HPS, lamps), three weeks of vegetative growth followed by eight weeks of flower at 12/12 ^[5]. They reported yield per watt of installed lighting rather than yield per square metre directly: 0.9 to 1.6 g of dry flower per watt, with the highest efficiency at the lowest power density and efficiency falling as power density rose ^[5]. Applying those reported ratios to their own tested power densities gives an approximate total yield of roughly 430 g/m² at 270 W/m² and roughly 540 g/m² at 600 W/m², a calculation from the reported g/W figures, not a total the paper states directly. The same trial found that individual leaf and flower potency weren't changed much by irradiance; what changed was the ratio of flower to leaf tissue, so overall plant potency rose with light mostly because there was proportionally more flower, not because any given gram of flower got stronger ^[5].

The same regulatory review's authors pooled their sixteen grower interviews with a wider literature set that included Toonen's and Vanhove's own figures among many others: trimmed to remove outliers, the blended indoor estimate averaged around 40 g per square foot per harvest, about 431 g/m², sitting in the upper half of the roughly 116–519 g/m² range this page cites overall rather than the middle of it. That it lands high is unsurprising given how it was built, and it's a reminder that self-reported and anecdotal figures tend to skew high ^[2].

The table below keeps each source's own numbers separate rather than folding them into one blended figure.

Turning your own light reading into an expected range

This won't give you an exact number; nothing published will. It gives you a defensible starting range from your own canopy PPFD, using the one relationship that's actually been measured.

1. Measure or estimate canopy-level PPFD with a quantum sensor at several points across the canopy, at the height the top of the flowers will reach; see PAR, PPF and PPFD explained if you're not sure what you're reading. Say your reading averages 700 µmol·m⁻²·s⁻¹ at a 12/12 flowering photoperiod.
2. Convert to daily light integral (DLI): DLI = PPFD × photoperiod hours × 0.0036. At 700 µmol and 12 hours, that's 700 × 12 × 0.0036 = 30.2 mol·m⁻²·d⁻¹. The DLI calculator does this automatically and compares it against typical stage targets.
3. Estimate expected yield using a straight line drawn between the trial's own two reported points (116 g/m² at 120 µmol, 519 g/m² at 1,800 µmol): yield ≈ 116 + 0.24 × (PPFD − 120) g/m². At 700 µmol, that gives roughly 116 + 0.24 × 580 ≈ 255 g/m². This is a rough interpolation of the reported range, not the paper's own regression equation, which wasn't published in a form that could be reproduced here; treat it as an order-of-magnitude estimate, not a forecast.
4. Adjust down for density and uniformity if your canopy isn't as dense or as evenly lit as the trial's roughly 11 plants/m² single-cola screen. A handful of larger, trained plants filling the same footprint less completely will usually land below that estimate, not above it, everything else being equal.

What moves yield beyond the light meter

Light explains a large share of the spread above, but not all of it, and none of the studies here held everything else constant against each other. Cultivar mattered enough in Vanhove's trial that the gap between the best and worst of four varieties under identical light and density was larger than the gap between their two lighting levels ^[4]; see genetics and cultivar selection for how much of that is knowable in advance. Vegetative time and plant size at the switch to 12/12 change how much canopy is already established when flowering starts, covered in flower week 1: the flip. Planting density didn't move Vanhove's numbers over a narrow 16-to-20-plants/m² range, but density interacts with training method well outside that range; see yield and plant density. Nutrient and irrigation management, and canopy training, each have their own literature and their own pages rather than a repeat of it here: start with nutrition and training and canopy management.

Where home and commercial rooms actually land

How much canopy you're allowed to run in the first place is a separate question, set by your own jurisdiction's plant-count, licensing and canopy rules, not by anything on this page; see the law section, or, for Horus's launch market, Malta's CHRA and ARUC rules.

When this number needs revisiting

This range rests on one controlled trial from 2021 and three forensic or grower-style studies from 2006 to 2012, none of them testing LED fixtures released since, CO₂ enrichment at home scale, or genetics bred specifically for light response. Revisit this page if a new controlled PPFD-response trial is published on a different cultivar or under CO₂ enrichment, if home-scale CO₂ supplementation becomes common enough to change what "typical home" means, or if LED efficacy moves far enough that heat load, not light level, stops being the practical ceiling described above. As of September 2026, none of those has happened yet.