Cannabinoid degradation during storage: what the data shows

Leave cannabis resin in daylight and its acidic THC (THCA) breaks down roughly 40% faster than the same material kept in the dark, at least in the one controlled study that measured it directly ^[1]. That is one of only two datasets in the published literature that put a number on cannabinoid storage decay, and the two do not agree with each other nearly as well as a tidy shelf-life chart would suggest. This page reports what each one actually measured, under what conditions, over what timescale, and what a testing programme should do with figures that disagree by an order of magnitude.

THCA half-life, resin, daylight

≈330days

Lindholst (2010), room temperature

THCA half-life, resin, darkness

≈462days

Lindholst (2010), room temperature

THCA half-life, dissolved extract, daylight

35days

Lindholst (2010), room temperature

THCA half-life, dissolved extract, darkness

91days

Lindholst (2010), room temperature

What actually degrades THC in storage

Nothing evaporates. The mass in the jar stays roughly the same; the molecules in it change shape. Fresh flower carries most of its THC as the acidic precursor, THCA, which converts to the neutral form, THC, mainly through heat during drying and curing (this conversion is called decarboxylation). What happens next, over weeks to years of storage rather than hours of drying, is a slower reaction: neutral THC oxidises, and the dominant known breakdown product is cannabinol (CBN). It's a genuinely different molecule, not a diminished version of THC, and lab panels report it separately for exactly that reason.

CBN is not a stable endpoint either. In the eight-year dataset covered below, CBN content rose for the first two years as THC broke down, then fell back by the eight-year mark: the reaction chain evidently keeps moving past CBN, into products that neither study measured ^[2]. Read a single CBN percentage as a rough clock on how long a sample has been degrading, not as the final resting state of a THC molecule.

Two variables move that clock. Light, specifically the daylight comparison Lindholst ran directly, raises the rate of THCA loss in resin by about 40% relative to the same material kept dark ^[1]. Heat almost certainly does the same by ordinary reaction kinetics, but neither study cited here reports a usable rate constant for it: Lindholst's abstract confirms that 4 °C and −20 °C storage were tested alongside room temperature, without stating how much slower they were, so treat "colder is better" as chemistry, not as one of this page's cited numbers. Oxygen exposure is the third variable growers usually hear about, and it is the weakest-evidenced of the three here: neither study isolated it as a variable, both stored material in relatively enclosed containers, and the case for sealed, low-oxygen storage slowing degradation rests on general oxidation chemistry rather than a dedicated cannabis trial.

Two studies, fifteen years apart, built almost nothing alike

Between them, these are the only two datasets found that put numbers on how cannabinoid content changes with storage time. They do not describe the same material, the same design, or even the same decade of analytical chemistry, which matters for how much weight either number deserves on its own.

Lindholst (2010): four years, two storage forms, daylight against dark

Christian Lindholst, of Aarhus University's Department of Forensic Medicine, tracked cannabinoid levels, THC, CBN, CBD and CBG, in both their neutral and acidic forms, in cannabis resin slabs and in cannabis extracted into organic solvent, stored at room temperature, 4 °C and −20 °C for up to four years ^[1]. The headline result is that acidic THC degrades exponentially through decarboxylation, and that the rate depends heavily on light and on whether the cannabinoid is still inside the plant matrix or already in solution:

Once cannabinoids are pulled into solvent, degradation speeds up roughly ninefold at the same light exposure: an extract's THC half-life in daylight (35 days) is shorter than resin's in darkness measured in months, not weeks. Neutral THC, the abstract notes, degrades "somewhat slower" than the acidic form, without a separate reported rate. The abstract does not state a sample size, a cultivar count, or a cultivar description, which is a real limit on how far a single-author, single-lab result should be generalised; it is the finding this field has, not a finding replicated across labs.

Fettoukh et al. (2025): eight years of seized resin, a much rougher picture

A far larger, more recent dataset comes from a forensic study of 150 cannabis resin (hashish) samples seized by Morocco's Gendarmerie Royale, held in secure storage, opaque polyethylene bags or cardboard boxes, in dark, dry rooms at 20–25 °C ^[2]. This was a cross-sectional design, not a single tracked batch: 30 samples were drawn from storage at each of five points, fresh and after two, four, six and eight years, and each group's cannabinoid content was measured by gas chromatography with flame-ionisation detection, which reports total THC (THCA plus decarboxylated THC together) rather than the two forms separately.

The scale of that early drop is the number worth sitting with: THC fell by about 92% within the first two years of dark, ambient storage, a far steeper loss than Lindholst's 462-day dark half-life would predict for the same interval (that half-life implies roughly 33% of starting THCA still present at two years, not under 8%). Neither figure is wrong; they are answers to different questions. Fettoukh's samples were compressed resin blocks, not the loose resin slabs or dried flower most cultivation and QA work deals with, and because the design is cross-sectional, the "fresh" group and the "eight-year" group are different seizures entirely, with unknown differences in starting cultivar, processing and handling before they ever reached secure storage. A single very degraded early batch, or one unusually potent fresh batch, could shift a 30-sample group average a long way.

Why these two don't agree, and why you shouldn't average them

What survives the disagreement is direction, not magnitude: every dataset found here, and every mechanism understood for how oxidation and decarboxylation work, points the same way. Time degrades THC; light and heat speed it up; dark, cool, sealed storage slows it down. Treat any specific percentage on this page as an illustration of that direction from one named study, not as a universal decay constant for your own flower, cultivar or packaging.

What this means for batch-age QA and potency labels

For a licensed producer or an association, this stops being a curiosity the moment a batch sits in inventory. A cannabinoid-content figure from a release assay is a snapshot, not a fixed property of the product, and pharmacopoeial efforts to standardise cannabis quality testing treat it that way: USP's 2020 Cannabis Expert Panel review frames identity and cannabinoid content as specifications to define and verify against acceptance criteria, not facts established once at intake ^[3].

Two real regulatory approaches show what that means in practice, from two different systems. Malta's ARUC requires lab testing before release and, separately, retesting of any undistributed stock every six months, or every four months if microbial counts run high ^[4]: the regulator has effectively assumed potency drift and built a fixed re-check into the workflow rather than trusting one assay indefinitely. Australia's TGO 93, by contrast, sets a fixed tolerance band at the point of testing: assayed potency for an herbal final dosage form must sit within 80–120% of the amount stated on the label ^[5]. A tolerance band like that has no built-in retest cadence of its own, so whatever schedule sits around it has to account for drift directly.

Put a number on that gap. Using Lindholst's dark-storage THCA half-life for resin (462 days) as an illustrative, conservative model, not a guarantee for flower specifically or for total THC, a batch that assays at 24% THC and then sits in sealed, dark, cool holding for six months would be modelled at roughly 18% (about 76% of the original figure remains, per the chart above). Total THC on a release certificate is mostly the neutral form, which Lindholst's abstract reports degrading somewhat more slowly than the acidic form modelled here, so treat 18% as a worst case, not a precise prediction. Against an 80–120% tolerance band anchored to that original 24% label, an 18% result would already sit below the floor, months before anyone opened the container to look at it. Against Malta's six-month retest cycle, that drift gets caught by design, because potency is re-verified at that point rather than trusted from the original assay date indefinitely.

Three practical rules follow directly. Date every batch record at the point of testing, not at harvest, and keep both dates on file. Treat a release-time potency figure as a number with an implied use-by, not a fixed fact, scaling the interval to how dark, cool and sealed the actual holding conditions are rather than to a generic shelf-life claim. And when inventory will sit for longer than a couple of months in anything short of sealed, dark, cool storage, budget for a measurable drop before the next scheduled test, not after a customer or auditor finds it. What that holding environment should actually look like, container by container, is covered in long-term storage: light, oxygen, temperature and, for containers specifically, in curing containers compared; what a commercial hold room does differently from a jar on a shelf is in commercial curing and holding. None of this changes what a jurisdiction actually requires for testing frequency or label tolerances where you operate, which is set by your own regulator; see the law section for that, or, for Horus's own launch market, Malta's CHRA and ARUC rules.

What's still missing

No published study has tracked one batch of dried, cured flower, the form most cultivation and QA work actually handles, continuously through controlled light, temperature and oxygen conditions over multiple years. The two datasets here cover resin and solvent extract (Lindholst) and compressed forensic hashish (Fettoukh); both are related material, neither is quite the product sitting in a commercial hold room or a home grower's jar. The effect of oxygen exposure specifically, as distinct from light and heat, has not been isolated in a cannabis storage trial that this page could find; it is inferred from general chemistry, not measured here. This page should be revisited if a controlled, flower-specific storage study is published, or if a regulator changes a retest cadence or tolerance band cited above.