Skip to content
Lead: Fertilizer advice for cannabis circulates in every forum. Solid advice is rare. Around a dozen reliable studies from the past six years now provide concrete optima for N, P and K, but they also show how strongly genotype, photoperiod, growing medium and water regime feed into the result. If you understand that, you save money and get more out of your plants.
What this is about
For a long time, cannabis fertilization was received wisdom: a bit of the Coffman and Gentner paper from 1977, plus experience from forums and substrate manufacturer data sheets whose recommendations rested on hearsay. That has changed since roughly 2019. Several research groups, above all the Volcani Institute in Israel (Bernstein lab) and the University of Guelph in Canada (Zheng lab), have run controlled experiments on medical drug-type cannabis. Agronomically, that is a far more reliable evidence base than anything publicly available before.
For this article we evaluated twelve studies published between 2019 and 2025, six of them as full-text deep dives. Three core findings emerge. And just as important: three methodological limitations you should know about before you copy any recommendation straight into your setup.
Cannabis fertilization by the numbers: N, P and K
The most robust evidence exists for nitrogen in the vegetative phase. Saloner and Bernstein (2020) tested five N levels (30, 80, 160, 240, 320 ppm) on the THC/CBD cultivar “Annapurna” over 32 days. At 30 and 80 ppm the plants showed clear deficiency symptoms: chlorosis, stunted growth, less root mass. From 160 ppm upwards the optimum was reached. More did not add yield. It produced the first signs of inefficient uptake. For veg, 160 ppm N is therefore the point beyond which more stops paying off, a recommendation that appears robust across different cultivars and setups.
For the flowering phase, Bevan, Jones and Zheng (2021) is the methodologically most interesting study. They used response surface methodology (a statistical approach that models interactions between variables) to determine the optimal combination of N, P and K in a hydro setup (deep water culture). The trial ran 100 plants of the THC cultivar “Gelato” over eight weeks of flower, in a Health Canada certified production room. The optimum came out at 194 mg/L N and 59 mg/L P. The interesting part: potassium was not the limiting factor within the range tested (60 to 340 mg/L), and the model could not identify a clear optimum for K. Translated into everyday practice: N and P are the dials you fine-tune in flower; K needs to stay in range, but it does not need obsessive optimization.
With potassium, things become cultivar-dependent again. Saloner, Sacks and Bernstein (2019) tested two cultivars in parallel. “Royal Medic” (fairly compact) reaches its optimum at 175 ppm K, and pushing further degrades performance. “Desert Queen” (the larger growth type of the two) still benefits from an increase to 240 ppm. The study is one of the few that explicitly tests the genotype-by-fertilization interaction, and it gives the direct reason why a universal K recommendation for all cultivars cannot work.
More fertilizer does not mean more cannabinoids
One assumption from practice says that generous feeding gives you more THC. Bernstein et al. (2019) looked into it, with a surprising result. They tested the effect of increased N, P, K and humic acid supply on the cannabinoid profile of the cultivar “NB100” across the full vegetative and flowering period. Increased K supply lowered THC and CBG in the inflorescence, raising P produced inconsistent effects, and humic acid mainly changed the mineral profile rather than the cannabinoids. The authors are careful with their conclusion: the relationship between cannabinoid content and nutrient supplementation is complex and non-linear. Anyone who keeps feeding beyond the optimum shifts the profile, usually not for the better.
Caplan, Dixon and Zheng (2019) add a complementary perspective: controlled drought stress in late flower increased cannabinoid content, with no meaningful reduction in floral biomass reported. Water and nutrient regimes are therefore directly coupled; if you optimize fertilization, think the irrigation schedule through with it.
Interactions that forums often leave out
Four interactions belong on the radar of CSCs and growers.
First, photoperiod: Saloner and Bernstein (2020) tested explicitly under 18/6, that is in veg. Bevan et al. (2021) tested at 12/12, that is in flower. The N optima of the two studies do overlap (160 to 194 mg/L), but the underlying logic differs: in veg, N feeds mass growth; in flower, it competes with cannabinoid metabolism. A feeding schedule that ignores the photoperiod switch leaves either yield or quality on the table.
Second, genotype: the K study shows it most clearly, but it applies to N and P just as much. When you bring in a new cultivar, treat recommendations from studies on other cultivars as a starting value, not a target value, and watch the plants.
Third, growing medium: Bevan et al. (2021) pointed to Caplan et al. (2017), who found a higher N optimum (212 to 261 mg/L) for organically based substrates. The explanation is mundane but important: organic N sources are not immediately plant-available. The nominal dose has to be higher to bring effective availability up to the level of a mineral solution. That is also the core of the review by Ahmadi et al. (2024) on chemical versus organic fertilization systems. That review mainly covers industrial hemp, but the mechanisms transfer to drug-type cannabis. Malík and Tlustoš (2025) and Nemati et al. (2021) add current reviews on growing media.
Fourth, the microbiome. Ahmed and Hijri (2021) review the state of knowledge on mycorrhizae, PGPR and endophytes in cannabis. Their verdict is sobering: plausibly effective, barely established causally. And commercial “myco” products often do not contain the species that actually work in cannabis. If you experiment with biostimulants, treat it as a trial and not as a standard recommendation.
Methods: what the studies deliver and what they do not
Three limitations to keep in mind. First: small sample sizes. The Bernstein and Saloner studies work with n = 5 plants per treatment. That is tight agronomically, often sufficient statistically, but sensitive to outliers. Bevan et al. (2021), with n = 100 plants in total, is the methodologically cleanest study in the selection.
Second: a narrow cultivar base. Taken together, the twelve studies cover maybe ten to twelve different cultivars. Compared to the range of cultivars on the market, that is very little. If your cultivar is not in that selection, the rule is: take the recommendation as a starting value, then observe and adjust.
Third: the lab-to-practice gap. All the studies cited here run in controlled growth rooms or certified production facilities. CSC cultivation is usually more variable: temperature, light quality and water quality fluctuate more. A cultivar optimum from a study is a corridor, not a point.
Two methodological developments are useful for practice. Schober, Präger and Graeff-Hönninger (2024, University of Hohenheim) and Sanaeifar et al. (2024) tested hyperspectral imaging for non-destructive assessment of nutrient status in cannabis. The sensor hardware is likely to drop into the low four-figure euro range over the next few years. For CSCs running more than one grow room, that could become an interesting lever for spotting deficiencies before they become visible.
What this means for your club
If you want to review your feeding schedule, these are the most reliable default values from current research. They hold for mineral nutrient solutions, for drug-type cannabis, in substrate or hydro setups: 160 ppm N in veg, 190 to 200 mg/L N and roughly 60 mg/L P in flower, and no overdosing on K (in the 150 to 200 ppm range you are on the safe side agronomically in most cases, and depending on the cultivar higher works too). With organic fertilization, plan for nominal doses 30 to 50 percent higher, because availability is delayed. Whenever you introduce a new cultivar, document the first two cycles closely, then fine-tune the schedule. And stay skeptical about biostimulants: the effect on cannabinoids has not been reliably demonstrated yet.
Above all: more is not better. The most common misconfiguration in practice is overfeeding. It does not just cost money, it can shift the cannabinoid profile for the worse.
Tell us how your schedule looks and whether the numbers match your experience. And if you run a cultivar we have not found a study on yet, let us know. We will take it into the next research cycle.
Take-aways
- 160 ppm N in veg, roughly 190 mg/L N and 60 mg/L P in flower are the most reliable ballpark figures for mineral nutrient solutions.
- K is rarely the limiting factor in flower, but it is strongly cultivar-dependent: cultivars with a larger growth type need more.
- More feeding does not give you more THC. Increased K supply can even lower the cannabinoid profile.
- Photoperiod, growing medium, water regime and genotype interact: feeding schedules from forums that ignore these variables are incomplete.
- Microbial biostimulants are a promising lever, but one that is barely established for cannabis: treat them as a trial, not as a standard.
Sources
- Bernstein, N., Gorelick, J., Zerahia, R. & Koch, S. (2019). Impact of N, P, K, and Humic Acid Supplementation on the Chemical Profile of Medical Cannabis. Frontiers in Plant Science. doi:10.3389/fpls.2019.00736
- Saloner, A. & Bernstein, N. (2020). Response of Medical Cannabis to Nitrogen Supply Under Long Photoperiod. Frontiers in Plant Science. doi:10.3389/fpls.2020.572293
- Saloner, A., Sacks, M. M. & Bernstein, N. (2019). Response of Medical Cannabis Genotypes to K Supply Under Long Photoperiod. Frontiers in Plant Science. doi:10.3389/fpls.2019.01369
- Bevan, L., Jones, M. & Zheng, Y. (2021). Optimisation of Nitrogen, Phosphorus, and Potassium for Soilless Production of Cannabis sativa in the Flowering Stage Using Response Surface Analysis. Frontiers in Plant Science. doi:10.3389/fpls.2021.764103
- Caplan, D., Dixon, M. & Zheng, Y. (2019). Increasing Inflorescence Dry Weight and Cannabinoid Content in Medical Cannabis Using Controlled Drought Stress. HortScience. doi:10.21273/hortsci13510-18
- Ahmed, B. & Hijri, M. (2021). Potential impacts of soil microbiota manipulation on secondary metabolites production in cannabis. Journal of Cannabis Research. doi:10.1186/s42238-021-00082-0
- Jin, D., Jin, S. & Chen, J. (2019). Cannabis Indoor Growing Conditions, Management Practices, and Post-Harvest Treatment: A Review. American Journal of Plant Sciences. doi:10.4236/ajps.2019.106067
- Schober, T., Präger, A. & Graeff-Hönninger, S. (2024). A non-destructive method to quantify the nutritional status of Cannabis sativa using in situ hyperspectral imaging. Computers and Electronics in Agriculture. doi:10.1016/j.compag.2024.108656
- Sanaeifar, A., Yang, C., An, M. et al. (2024). Noninvasive Early Detection of Nutrient Deficiencies in Greenhouse-Grown Industrial Hemp Using Hyperspectral Imaging. Remote Sensing. doi:10.3390/rs16010187
- Ahmadi, F., Kallinger, D., Starzinger, A. et al. (2024). Hemp Cultivation: Chemical Fertilizers or Organic Technologies, a Comprehensive Review. Nitrogen. doi:10.3390/nitrogen5030042
- Malík, M. & Tlustoš, P. (2025). Soilless Growing Media for Cannabis Cultivation. Agriculture. doi:10.3390/agriculture15181955
- Nemati, R., Fortin, J.-P., Craig, J. et al. (2021). Growing Mediums for Medical Cannabis Production in North America. Agronomy. doi:10.3390/agronomy11071366
