The European Pharmacopoeia monograph for cannabis flower caps loss on drying at 12.0 per cent. Health Canada sets no moisture limit on dried cannabis at all. A lot that clears every Canadian release test can still sit outside the specification a German importer releases against, and nobody finds out until the flower is on a bench in Frankfurt.
Cannabis water activity is the measure of unbound water available in dried flower, reported on a scale from 0 to 1. ASTM D8197 puts the accepted range for dry cannabis flower at 0.55 to 0.65. Above 0.65 there is enough free water for mould to grow. Below 0.55 the flower is over-dried, terpenes go with the water, and a procurement manager can tell by opening the jar.
Water activity and moisture content are not the same number, and export specifications use both. Which one your buyer writes into the contract decides what your dry room has to deliver.
What cannabis water activity actually measures
Water activity measures the vapour pressure of water in the flower against the vapour pressure of pure water at the same temperature. It reports how much of the water in a bud is free rather than bound into plant tissue. Micro-organisms can only use free water. That is why the number predicts mould risk better than weight ever has.
Moisture content is a different measurement entirely. It reports total water as a percentage of weight, bound and unbound together, and it is what pharmacopoeial testing captures under loss on drying. The European Pharmacopoeia cannabis flower monograph, 3028, sets loss on drying at a maximum of 12.0 per cent using general chapter 2.2.32, determined on a 25 g to 50 g sample. The monograph was published in supplement 11.5 of the European Pharmacopoeia and a revised text went out for comment in Pharmeuropa 37.4, with the consultation closing on 31 December 2025.
Two numbers, two failure modes
A batch can pass one and fail the other. Dry a room too fast and the outside of each bud gives up water while the core holds it, so a sample weighed for loss on drying reads inside the limit while water activity climbs over the following fortnight as the bud equilibrates in the bag. Dry too hard and you get the reverse: loss on drying looks excellent, water activity drops under 0.50, and the terpene profile that justified your price has gone up the exhaust, which is the same quality signal buyers read in hand-trimmed versus machine-trimmed flower.
ASTM published both halves of this in 2018. D8196 is the practice for determining water activity in cannabis flower, D8197 the specification for holding it between 0.55 and 0.65. The United States Pharmacopeia covers the same ground from the microbial side in general chapter 1112, which maps water activity to the organisms a product can support. None of these are regulations. All three are what a buyer's quality team reads before it writes your spec sheet.
If you only measure one number, measure water activity. Weight tells you what left the plant. Water activity tells you what is going to happen next.
Where the export limits come from in each market
No regulator in your four main export markets writes a water activity limit into law. Several of them write a moisture limit, and all of them write microbial limits that water activity quietly controls. The gap between those two facts is where Canadian producers get caught.
| Market | Moisture requirement | Water activity requirement | Where it bites |
|---|---|---|---|
| Canada | None set | None set | Microbial limits under section 93(3) of the Cannabis Regulations |
| European Union | Loss on drying, maximum 12.0 per cent (Ph. Eur. 3028) | None set, but routinely specified by importers | Batch release by the importer's Qualified Person |
| Australia | Set by the product specification under TGO 93 | None set | Sponsor's release testing and TGA microbiological expectations |
| New Zealand | Loss on drying, inside the minimum quality standard | None set | Product assessment, and stability across the whole shelf life |
What Canada actually requires
Section 91 of the Cannabis Regulations requires testing for microbial and chemical contaminants on each lot or batch of dried cannabis. Section 93(3) then says those contaminants must sit inside generally accepted tolerance limits established in Schedule B to the Food and Drugs Act and appropriate for the intended use. Schedule B points to pharmacopoeias, which means a Canadian producer picks its own reference standard. Two Licensed Producers can both be fully compliant while releasing against different numbers.
Moisture is absent from that framework because Health Canada regulates contamination outcomes rather than the process variable that drives them. Your European buyer regulates both. Our comparison of export testing requirements across four markets sets out the full panel differences, and the New Zealand minimum quality standard shows how far a receiving regulator can go past the Canadian list.
Read the table one more way. Every cell in the water activity column says none set, and every importer in those markets asks for the number anyway. That is not a contradiction. It means the limit lives in your specification and your quality agreement rather than in a statute, and a specification you did not write is one somebody else wrote for you.
Why the number decides your microbial result
Microbial limits are the reason any of this matters commercially. European Pharmacopoeia chapter 5.1.8 covers the microbiological quality of herbal medicinal products, and for herbal drugs to which boiling water is not added before use it sets a total aerobic microbial count acceptance criterion of 10 to the power of 5 CFU per gram, a total yeast and mould count of 10 to the power of 4 CFU per gram, absence of Escherichia coli in 1 g and absence of Salmonella in 25 g.
One detail decides whether that clause is testable. The bile-tolerant gram-negative limit is not identical between chapter 5.1.4 and chapter 5.1.8, so a specification that says compliant with the European Pharmacopoeia and stops there has not said anything testable. Name the chapter and the edition. Buyers who have been burned already do.
Yeast and mould is a drying problem, not a testing problem
Total yeast and mould is the count that fails, and it fails because water was available while the flower sat in a tote, a bag or a pallet. Bring a lot in at 0.70 water activity and the count you get at week one is not the count the importer gets at week six. Testing at release captures a moment. Water activity captures a trajectory.
That trajectory is what drives the decontamination conversation. A producer who irradiates to bring a number down is paying twice: once for the treatment and again in the terpene loss a pharmacy customer will notice. Fix the dry room instead. We cover the treatment side in our guide to irradiation standards for export flower, and the reviewing side in how to read a cannabis certificate of analysis.
One more consequence follows from the same physics. The European monograph requires the measured content of total THC and total CBD to sit within 10 per cent of the value stated on the label when the flower is prescribed to patients. Potency is reported on a dried basis, so a moisture number that moves between your laboratory and theirs moves the assay result with it. A potency dispute is often a moisture dispute wearing a different coat.
Holding the specification from dry room to importer's bench
Five controls hold the number from harvest to release. Run them in this order, because each one depends on the one above it. Call it the 5-point moisture control path.
- Measure water activity, not only weight. A meter on the dry room floor, run to the ASTM D8196 method, with the reading recorded per lot rather than per room. Weight loss tells you the room is drying. It does not tell you any single bud is done.
- Write both numbers into the specification. A water activity range of 0.55 to 0.65 and a loss on drying limit, each with the method and the pharmacopoeial chapter named. Send that sheet to the importer before they send you theirs.
- Equilibrate before you package. Flower packaged while it is still moving toward the target keeps moving inside the bag, and the bag is where the buyer's retention sample comes from.
- Control the package and the transit. Barrier material, headspace and temperature all move water activity after release. Test at packaging, hold retention samples under the same conditions you promised, and check them on the same schedule as your stability and shelf-life programme.
- Agree the arrival test in writing. Who retests, against which chapter, and what happens on an out-of-trend result. Settle it in the quality agreement, not during the argument that follows a batch rejection at import.
Where producers actually lose the batch
Rarely on the meter. Three failures account for most of it. Sampling from the top of a tote, where airflow has already pulled the surface dry, so the recorded value flatters the lot and the arriving flower misses the export-grade standard the buyer signed for. Reporting a dry room average instead of a lot value, which averages away the one rack that was too close to the dehumidifier. And leaving water activity off the certificate of analysis entirely, which means the first number anyone has seen for that lot is generated by the importer's laboratory, three weeks and one ocean later. At that point you are not presenting data. You are disputing theirs.
AlphaLeaf grows indoor, hand-trimmed cannabis flower in Montreal as a Health Canada Licensed Producer, with refined genetics and controlled drying and curing built around terpene retention rather than throughput. We hold export authorisation under the Cannabis Act, test every lot through an ISO/IEC 17025 accredited laboratory, and keep the lot-level moisture and batch records that EU importers and their Qualified Persons expect to read before release.
If you are qualifying a Canadian supplier, ask for the water activity value on the lot you are buying, not the range the producer aims for. Talk to our team and we will send you ours. One number, one lot. That is the whole test.
Frequently Asked Questions
What is the ideal water activity for dried cannabis flower?
ASTM D8197 specifies a range of 0.55 to 0.65 for dry cannabis flower. Above 0.65 there is enough free water to support mould growth. Below 0.55 the flower is over-dried and loses terpene content and structure, which buyers in medical channels notice immediately.
Does Health Canada set a moisture or water activity limit for dried cannabis?
No. Section 91 of the Cannabis Regulations requires testing for microbial and chemical contaminants on each lot or batch, and section 93(3) requires those contaminants to sit within tolerance limits established in Schedule B to the Food and Drugs Act. Moisture content and water activity are left to the producer's own specification.
What is the loss on drying limit for cannabis flower in the European Pharmacopoeia?
The Cannabis flos monograph, number 3028, sets loss on drying at a maximum of 12.0 per cent, determined using general chapter 2.2.32 on a 25 g to 50 g sample. The monograph was published in supplement 11.5 and a revised version was issued for comment in Pharmeuropa 37.4, with consultation closing on 31 December 2025.
Is water activity the same as moisture content?
No. Moisture content reports total water as a percentage of weight, both bound and unbound, and is measured as loss on drying. Water activity reports only the unbound water available to micro-organisms, on a scale from 0 to 1. A batch can meet a loss on drying limit and still carry a water activity value high enough to grow mould in storage.
Why does water activity affect microbial test results?
Micro-organisms can only use free water, so water activity sets whether yeast and mould can multiply after release. European Pharmacopoeia chapter 5.1.8 sets a total yeast and mould acceptance criterion of 10 to the power of 4 CFU per gram for herbal drugs to which boiling water is not added. A lot released at high water activity can pass at release and fail weeks later at the importer.
Should water activity appear on a cannabis certificate of analysis?
It is not required in Canada, but importers in Germany, Australia and New Zealand ask for it during supplier qualification. Include the lot value, the method used, and the date it was measured. A certificate without it leaves the importer's own laboratory to produce the first water activity number anyone has seen for that lot.
