Light: intensity, daily light integral, spectrum, far-red and UV

Ask a small grower what matters most in a tent and many will say the lamp. Then the questions start. How bright should it be? Is the expensive "full spectrum" panel worth it? Should I add UV in the last weeks, or far-red diodes? Forums answer with confidence, and lamp marketing with even more.

This article sets those questions beside the research. CannaTinker's grow guide, built from UK small-grow practice, gives clear targets. In bloom it aims for 750–900 µmol·m⁻²·s⁻¹ at the canopy top on a 12/12 day, and it dims to 60–70 % in the last days before harvest. The lighting research is unusually good by cannabis standards, but it was done in commercial rooms, chambers and greenhouses, often on high-CBD hemp, and measured yield and cannabinoids, not taste. It informs tent practice without automatically overruling it.

A note before we start: in the UK, growing cannabis, even with a prescription, requires a Home Office licence. This article is educational and describes published research.

Three numbers worth knowing

Much confusion about grow lights comes from mixing up three quantities.

PPFD (photosynthetic photon flux density) is how many usable photons land on a square metre each second, in micromoles (µmol·m⁻²·s⁻¹). "Usable" traditionally means wavelengths from 400 to 700 nm, the band known as PAR. It is what a quantum sensor reads at one spot.

DLI (daily light integral) is the day's total: how many moles of those photons a square metre receives over the whole light period (Faust and Logan, 2018). For a steady lamp the arithmetic is simple: DLI = PPFD × hours of light × 3,600 ÷ 1,000,000, or PPFD × hours × 0.0036.

Efficacy (µmol/J) describes the lamp rather than the plant: how many micromoles of 400–700 nm photons it emits for each joule of electricity (Kusuma, Pattison and Bugbee, 2020). Watts tell you what a lamp costs to run; µmol/J tells you how much light you get for that cost. Lumens, the unit on household bulbs, weight light by the sensitivity of the human eye and are not suitable for plants (Kusuma, Pattison and Bugbee, 2020).

How much light? What the intensity trials found

Up to 1,800 without a ceiling

The most cited study is by Rodriguez-Morrison, Llewellyn and Zheng (2021) at the University of Guelph. They worked inside a commercial indoor facility in Ontario, with plants of the CBD-dominant cultivar 'Stillwater' growing in deep-water culture basins. After two weeks of vegetative growth at about 425 µmol·m⁻²·s⁻¹, plants flowered for 81 days on a 12-hour day under LED fixtures dimmed to give canopy PPFDs from 120 to 1,800. Air temperature was about 25 °C and CO₂ was not enriched: about 437 ppm during the day, close to outdoor air.

The result was striking. Dry flower yield rose in a straight line from 116 to 519 g/m² as average PPFD rose from 120 to 1,800, about 4.5 times as much. It did not level off. The authors concluded that, even at ambient CO₂, the supply of photons was "still limiting whole-canopy photosynthesis" at the top of the range, equivalent to a DLI of about 78. Top buds also became denser, and a larger share of the plant's dry mass ended up as flower.

Potency did not follow. There were no light-intensity effects on the concentration of any cannabinoid measured. Total terpenes rose modestly, from 8.85 to 12.7 mg per gram of dry flower. More light meant more flower, not stronger flower.

There was one warning sign. Fv/Fm, a standard indicator that falls when leaves are under light stress, declined as the light on individual leaves rose. And single-leaf photosynthesis levelled off well below 1,800 even while whole-plant yield kept climbing. A single leaf is a poor guide to a whole canopy.

A high-THC cultivar, without added CO₂

Llewellyn and colleagues (2022) repeated the question on a high-THC cultivar, 'Meridian' (over 20 % THC), in a commercial facility in Ontario. Plants grew in 0.63 m² plots at about 10 plants per square metre, inside light-tight enclosures, with no supplemental CO₂ and a daytime temperature of about 26 °C. They flowered for 45 days at canopy PPFDs of 600, 800 or 1,000.

Raising PPFD from 600 to 1,000, 1.6 times as much light, gave 1.6 times as much dry flower. Cannabinoid and terpene concentrations did not change. By the authors' regression, every extra 100 µmol·m⁻²·s⁻¹ added about 51 g/m² at their planting density.

Two hemp studies

Two newer studies point the same way, with caveats. Hahm and colleagues (2025) grew the hemp cultivar 'Queen Dream' in a closed plant factory in Korea at 200, 400 or 600 µmol·m⁻²·s⁻¹ for 35 days of flowering. Flower dry mass at 600 was 205 % higher than at 200, and total CBD concentration rose by about 37 %. Here potency did rise, but with one plant per treatment in each of three runs, the base is small.

Collado and Hernández (2025) worked in a winter greenhouse in North Carolina. The high-CBD cultivar 'Suver Haze' grew on 1.2 × 1.2 m plots under sunlight plus LED supplements of 150 to 700 µmol·m⁻²·s⁻¹, for a 73-day crop. Raising the LED supplement 4.7-fold increased flower yield 4.1-fold, from 131 to 543 g/m². However, in flowers below the third node on the main stem, the main cannabinoids fell by about 13 % as total light increased. More light grew more lower flowers, and these were somewhat less potent.

Reading the four studies together

On yield, the agreement is strong: across CBD and THC types, in a commercial room, a plant factory and a greenhouse, more light gave proportionally more flower, with no plateau found up to 1,800. On potency, the results are mixed: no change in the two Guelph studies, a rise in one small hemp study, and a fall in the lower flowers in one greenhouse study. Nobody found that more light reliably makes stronger flower.

Why a tent hits a ceiling before the plant does

If yield keeps climbing to 1,800, why does the tutorial stop at 900? Because in a tent it is not the plant that sets the limit. It is the lamp's output and the heat it produces.

Over one square metre, average PPFD is roughly the number of µmol/s that reach the canopy. A lamp's output is its power in watts multiplied by its efficacy. Llewellyn and colleagues (2022) note that most modern horticultural LED fixtures exceed 2.5 µmol/J. At that efficacy, the wall power needed to deliver the tutorial's bloom target is:

• 750 µmol·m⁻²·s⁻¹ over 1 m²: about 300 W
• 900 µmol·m⁻²·s⁻¹ over 1 m²: about 360 W

These are minimums, assuming every photon lands on the canopy; light lost on walls means a real tent needs more. At 2.0 µmol/J the same targets need 375–450 W. A 200 W lamp at 2.5 µmol/J emits at most about 500 µmol/s, so it cannot average 750 across a full square metre, whatever its marketing says. Reaching the 1,800 of the Guelph trial would need over 700 W of very efficient LED in a single square metre.

Then there is heat. Practically all of the electricity a lamp draws ends up as heat inside the tent, either directly from the fixture and driver or when the light is absorbed by leaves, walls and pots. The research rooms were climate-controlled and held 21–27 °C. A home tent depends on an extraction fan and the room around it. Llewellyn and colleagues (2022) list heat management among the extra costs of more light, even at commercial scale. In a small enclosed tent, a lamp powerful enough for 1,500 µmol·m⁻²·s⁻¹ would be hard to keep cool without serious ventilation. That, rather than any biological ceiling, is why 750–900 is a sensible bloom target for a tent.

The daily light integral, with the tutorial's numbers

PPFD is a snapshot. DLI is what the plant actually receives across the day, and Rodriguez-Morrison and colleagues (2021) note that DLI can correlate with yield better than instantaneous PPFD. Converting the tutorial's targets gives the following.

Two things stand out. First, switching from 18 to 12 hours of light cuts the day's total by a third, so the higher PPFD in flower mostly keeps the DLI where veg left it. The Guelph authors made the same point about their own plants: about 425 µmol·m⁻²·s⁻¹ for 18 hours in veg gave a DLI of about 27.5, so the third of their plants flowered below about 640 actually received less daily light after the switch (Rodriguez-Morrison, Llewellyn and Zheng, 2021).

Second, the tutorial's bloom DLI of 32–39 is well below what is possible. Faust and Logan (2018) report that LEDs alone can now deliver 40–50 mol·m⁻²·d⁻¹. The sunniest places on their US maps reach 60–65 in June. The Guelph trial went to about 78. A tent at 750–900 is giving plants a good, ordinary summer day, not an extreme one.

DLI is also handy for comparing schedules: a 20-hour veg day needs less PPFD than an 18-hour day for the same total. Faust and Logan also describe expressing DLI per unit of leaf area rather than floor area. At the same DLI per square metre of floor, small plants with little leaf area receive more light per leaf than a full canopy does, so larger plants need a higher DLI to get the same light per leaf.

Efficacy: where the money goes

Kusuma, Pattison and Bugbee (2020) report that the best LED fixtures in 2014 managed 1.7 µmol/J. By 2020, independently tested white + red fixtures reached 2.5 to 2.8 µmol/J and blue + red ones 3.0, against 1.72 for a 1000 W double-ended high-pressure sodium (HPS) lamp. Typical fixtures were at 2–3 µmol/J. With current LED technology, they calculated practical limits of about 3.4 µmol/J for white + red and 4.1 for blue + red.

A fixture is less efficient than its LEDs. Efficacy falls about 10 % as the diodes heat from 25 to 85 °C. Drivers (power supplies) are 85–95 % efficient. Protective covers cost about another 10 %. With typical parameters, a fixture delivers about two thirds of its LEDs' rated efficacy. The authors advise asking for results from a certified third-party laboratory.

The study that links efficacy to harvests is by Westmoreland, Kusuma and Bugbee (2021) at Utah State University. They grew the medical hemp cultivar 'Trump' (CBD about 8–12 %, THC well under 1 %) in three trials in a walk-in growth room divided into five 1 m² sections with white reflective walls. That is five reflective boxes not unlike a tent, though enriched to 1,000 ppm CO₂. Plants flowered at 750 or 900 µmol·m⁻²·s⁻¹ under an HPS lamp or one of four LED fixtures (1.72 to 2.51 µmol/J). PPFD was kept equal across fixtures by raising the lamps as plants grew. Corners were within 15 % of the average.

Dry flower yields ranged from 500 to 750 g/m², and yields were higher at 900 than at 750. The most efficient white + red LED (2.51 µmol/J) yielded 4.6 % less per square metre than HPS. But it produced 27 % more flower per dollar of electricity. At $0.10 per kWh, a mole of photons cost 1.6 cents from HPS and 1.1 to 1.3 cents from the LEDs. As their title says, efficacy matters more than spectrum for cost.

For a tent, the arithmetic is direct. A 300 W lamp on for 12 hours uses 3.6 kWh a day, roughly 200 kWh over eight weeks of flower. A lamp of 2.0 µmol/J needs 25 % more electricity than one of 2.5 for the same light, and produces 25 % more heat to remove.

Spectrum: how much blue?

Many growers believe more blue light gives shorter plants, tighter buds and higher potency, and some choose "cool" 5000–6500 K lamps or add blue diodes for that reason. These are reports from practice, not measurements.

The Westmoreland trial tested this directly, because its five fixtures differed mainly in blue fraction: 4 % (HPS), 10 % (3000 K white and a white + red), 16 % and 20 % (5000 K white). As blue rose from 4 to 20 %, yield fell in a straight line by about 12 %, consistently across all three trials. CBD and THC concentrations were not affected, and the effect on plant height was not statistically significant (Westmoreland, Kusuma and Bugbee, 2021).

The authors discuss an earlier study by Magagnini and colleagues, which they describe as follows. At about half the intensity, LEDs with 14 % and 24 % blue raised cannabinoid concentration compared with HPS, but total cannabinoid yield did not differ. Westmoreland and colleagues suggest that photoreceptors may be more sensitive to spectrum at lower light levels, which could explain why colour seemed to matter there and not at 900.

Physiology suggests why blue is no bonus. Measuring single lettuce leaves, Liu and van Iersel (2021) found that "absorbed blue photons were used inherently least efficiently" for photosynthesis. Green light, often dismissed, was used as efficiently as red at high light, probably because it spreads more evenly through the leaf. Kusuma, Pattison and Bugbee (2020) add that blue photons are about 20 % less photosynthetically efficient than those from common 660 nm red LEDs, and that 5–30 % blue is typically used in fixtures to restrain stem stretch.

White light with some green also makes pests, disease and nutrient problems easier to spot; Westmoreland and colleagues note that white LEDs typically have a colour rendering index above 70, against 43 for HPS.

UV: the late-flower stress tip

A popular tip says to add a UV-B lamp in late flower because stress "boosts THC and resin". These are claims from practice, not measurements.

The strongest modern test is the same Llewellyn (2022) study. Alongside the intensity treatments, plants at 600 µmol·m⁻²·s⁻¹ received either UV-A from 385 nm LEDs (50 µmol·m⁻²·s⁻¹, 12 hours a day for all 45 days) or UV-A + UV-B from a fluorescent lamp (3 µmol·m⁻²·s⁻¹, 5 hours a day for the last 20 days). Neither UV treatment changed yield, or cannabinoid or terpene concentrations, in the flowers. The only effect was in the small "sugar leaves" around the buds: the UV-A + UV-B treatment gave about 30 % more THC in its neutral form there, but no change in total THC. The authors noticed, without counting, what looked like more trichomes on those leaves. Their conclusion was that they "saw no commercially-relevant benefits" from UV. Meanwhile, raising PPFD from 600 to 1,000 gave 1.6 times the flower.

The positive evidence is older. A review cites a 1987 study by Lydon and colleagues in which UV-B for 3 hours a day raised THC in leaves and buds (Bilodeau et al., 2019). Llewellyn and colleagues point out that the cultivars of that era had much lower cannabinoid content. Modern high-THC genotypes may already be close to their genetic ceiling. They also describe an earlier Guelph study in which short-wave UV-B (287 nm) had only harmful effects. Doses are also reported so inconsistently that studies are hard to compare.

UV is not harmless to people either. In the Llewellyn trial, workers were not allowed in the research area while the UV-B lamps were on.

Far-red and how to compare lamps

Far-red light (about 700–750 nm) sits just beyond the traditional PAR band. Lamps with extra 730 nm diodes are marketed for faster finishing and bigger yields, and some growers run far-red alone briefly at lights-off so plants "sleep" sooner.

The key research is by Zhen and Bugbee (2020). They measured whole-canopy photosynthesis of 14 crop species, including lettuce, tomato, cucumber, soybean, wheat and maize, but not cannabis, with plants placed in a 100-litre measurement chamber. When far-red was added to white light, up to about 40 % of the background, it raised canopy photosynthesis as much as the same number of ordinary 400–700 nm photons. On its own, far-red did very little. Above about 40 % the gains faded, and 746 nm diodes were less effective than 711 or 723 nm ones.

The explanation is the Emerson effect, which goes back to work by Robert Emerson in the 1940s and 1950s. Plants run two photosystems in series. Shorter wavelengths tend to over-excite one, and far-red mainly drives the other, so a modest amount of far-red rebalances them. Zhen and Bugbee therefore proposed that PAR should be redefined as 400–750 nm. Their measurements lasted hours, not a crop cycle, and far-red also shapes plants. Bilodeau and colleagues (2019) review how a low red to far-red ratio triggers shade-avoidance responses such as stem stretch. They also note that one cannabis cultivar, 'G-170', showed no flowering response to changes in that ratio.

This matters when comparing lamps. A standard PAR meter reads 400–700 nm and ignores far-red. If one manufacturer quotes output or efficacy over 400–700 nm and another over 400–750 nm (sometimes labelled "extended PAR" or ePAR), the second number will look better without the lamp being comparable. Kusuma, Pattison and Bugbee (2020) restrict efficacy to 400–700 nm except for far-red LEDs. When comparing, check that both figures cover the same range. For reference, the fixtures in the Westmoreland trial, including HPS, emitted 2–6 % far-red.

What this means for a 1×1 tent

These are not a recipe, but reasonable readings of the evidence for a single square metre.

• Treat 750–900 as a sensible bloom target, not a ceiling. The research shows more light means more flower up to 1,800, but a tent is limited by lamp power and heat. Raise intensity gradually after the switch to 12/12 rather than all at once.
• Size the lamp honestly. For 750–900 over a full square metre, expect to need at least 300–360 W from the wall at 2.5 µmol/J, more at lower efficacy. Ask for third-party efficacy figures at full power, measured over 400–700 nm.
• Measure, do not guess. A PPFD map from the manufacturer only means something if it was made at your hanging height over 1 m². Check the centre and the corners. Reflective walls help, as they did in the Utah chambers.
• Watch leaves and temperature. Pale, bleached or upward-curling top leaves, and a hot canopy, are signs to raise or dim the lamp. The tutorial's 35–45 cm between lamp and canopy in stretch is a practical buffer.
• Think in DLI when changing schedules. The switch to 12 hours cuts daily light by a third. Late-flower dimming to 60–70 % brings DLI down to about 19–27. No study has tested whether this protects terpenes, and in the Guelph trial terpenes were slightly higher, not lower, under more light. Dimming remains grower practice, worth treating as such.
• Do not pay extra for blue or UV. An ordinary white LED already contains about 10–20 % blue, and the evidence favours efficacy over colour. UV gave no benefit in the best modern trial and is hazardous to eyes and skin; if you try it, treat it as an experiment.
• Treat far-red as a modest extra. A few per cent of far-red is sensible, and those photons do contribute. It is not a switch for bigger yields, and in a height-limited tent extra stretch is not welcome.

What we know, and what we don't

We know, with good confidence, that cannabis flower yield rises roughly in proportion to light up to very high intensities. This held for CBD and THC types, in commercial rooms, a plant factory and a greenhouse, and in the most-cited case without CO₂ enrichment. We know that lamp efficacy drives running cost more than spectrum does, and that the arithmetic of PPFD, DLI and µmol/J is plain physics that applies to any tent.

We know less about quality. Potency responses differ between cultivars and canopy positions, and taste was never measured. The blue-light results come mainly from one hemp cultivar under CO₂ enrichment. The UV results rest on one modern cultivar and two UV schedules, against older studies on lower-potency plants. Far-red's contribution to photosynthesis is well documented in other crops, but no source here tests it on cannabis through to harvest.

And we know almost nothing directly about tents. No study used a home tent, a small lamp, a warm, poorly ventilated space or late-flower dimming. The research supplies the right questions and some firm numbers; applying them in one square metre still takes careful observation.

Sources

Bilodeau, S.E., Wu, B.-S., Rufyikiri, A.-S., MacPherson, S. and Lefsrud, M.G. (2019) 'An update on plant photobiology and implications for cannabis production', Frontiers in Plant Science, 10, 296. doi:10.3389/fpls.2019.00296.

Collado, C.E. and Hernández, R. (2025) 'Vegetative and reproductive stage lighting interactions on flower yield, water use efficiency, terpenes, and cannabinoids of Cannabis sativa', Scientific Reports, 15, 43641. doi:10.1038/s41598-025-27437-4.

Faust, J.E. and Logan, J. (2018) 'Daily light integral: a research review and high-resolution maps of the United States', HortScience, 53(9), pp. 1250–1257. doi:10.21273/HORTSCI13144-18.

Hahm, S., Bok, G., Kim, S.-J., Kim, B., Lee, Y.-J., Kim, S. and Park, J.S. (2025) 'High light intensity enhances cannabinoid biosynthesis through concerted gene expression in hemp (Cannabis sativa) flowers', Frontiers in Plant Science, 16, 1687794. doi:10.3389/fpls.2025.1687794.

Kusuma, P., Pattison, P.M. and Bugbee, B. (2020) 'From physics to fixtures to food: current and potential LED efficacy', Horticulture Research, 7, 56. doi:10.1038/s41438-020-0283-7.

Liu, J. and van Iersel, M.W. (2021) 'Photosynthetic physiology of blue, green, and red light: light intensity effects and underlying mechanisms', Frontiers in Plant Science, 12, 619987. doi:10.3389/fpls.2021.619987.

Llewellyn, D., Golem, S., Foley, E., Dinka, S., Jones, A.M.P. and Zheng, Y. (2022) 'Indoor grown cannabis yield increased proportionally with light intensity, but ultraviolet radiation did not affect yield or cannabinoid content', Frontiers in Plant Science, 13, 974018. doi:10.3389/fpls.2022.974018.

Rodriguez-Morrison, V., Llewellyn, D. and Zheng, Y. (2021) 'Cannabis yield, potency, and leaf photosynthesis respond differently to increasing light levels in an indoor environment', Frontiers in Plant Science, 12, 646020. doi:10.3389/fpls.2021.646020.

Westmoreland, F.M., Kusuma, P. and Bugbee, B. (2021) 'Cannabis lighting: decreasing blue photon fraction increases yield but efficacy is more important for cost effective production of cannabinoids', PLoS ONE, 16(3), e0248988. doi:10.1371/journal.pone.0248988.

Zhen, S. and Bugbee, B. (2020) 'Far-red photons have equivalent efficiency to traditional photosynthetic photons: implications for redefining photosynthetically active radiation', Plant, Cell & Environment, 43(5), pp. 1259–1272. doi:10.1111/pce.13730.