Water activity

Two bags of green coffee can hold exactly the same amount of water and behave completely differently in a warehouse. One arrives clean after four months at sea; the other arrives flat, or mouldy. Water activity is the measurement that explains the difference.

Moisture content answers how much water is in the bean. Water activity — written Aw or aw, and always a number between 0 and 1 — answers how free that water is. Some of the water in a bean is chemically bound up and effectively unavailable; the rest is loose, and loose water is what mould grows in and what chemical reactions run on. As the instrument maker AQUALAB puts it, moisture content is “a quantitative measurement [that] monitors only the amount of water in a substance. Water activity, on the other hand, is a qualitative measurement” — it “assesses the activity (or energy status) of the water.”¹ The Norwegian importer Nordic Approach says the same thing from the buyer’s side: water activity “measures how much of that water is free to support microbial growth and chemical reactions.”²

Formally, Aw is the ratio of the vapour pressure of water in the bean to the vapour pressure of pure water at the same temperature. In practice that means Aw is the same number as the relative humidity the bean would create in a sealed jar, divided by 100. A bean sitting at 0.60 Aw is a bean that would hold the air around it at 60% RH. That equivalence is why the measurement is so useful: it tells you what the coffee is doing to its own packaging, not just what is inside it.

How it is measured

Green coffee Aw is read on a benchtop instrument, using one of two sensor types:¹

• Dew-point sensors, which chill a mirror until water condenses on it and calculate the vapour pressure from that. This is the reference method.
• Electrical-properties sensors, which infer the water’s state from changes in an electrical signal.

Either way the reading takes minutes, needs no oven, and destroys nothing — which is why it has spread through mills, importers’ labs and arrival-sampling rooms far faster than any standard requiring it. This wiki’s Coffee supply chain note records the resulting routine: the mill measures at origin, the importer measures again on landing, and the gap between the two readings is how a buyer works out whether a lot was shipped tired or got tired at sea.

The Cafe Imports study

The largest publicly described body of coffee water-activity data belongs to the American green importer Cafe Imports, and it is the reason the measurement entered specialty vocabulary at all.

The study was run by Ian Fretheim, the company’s Director of Sensory Analysis, with US sensory lab manager Megan Person. It started in 2012, accumulated “more than 25,000 water activity readings”, and was written up as a 42-page report published on 11 February 2019 after roughly twelve months of organising the data.³ Fretheim’s stated motive was blunt: he measured Aw “in the hope that it would provide a more insightful measurement than moisture content, specifically with regard to predicting the shelf life of coffees.”³

That framing — Aw as a shelf-life predictor, competing with moisture content rather than supplementing it — is the whole argument in one sentence, and it is the part of the study that is firmly on the public record.

To verify — the study's actual numbers were not read

The specific conclusions widely attributed to Cafe Imports — that top-scoring coffees clustered around 0.42–0.47 Aw, that good coffee ran up to about 0.57–0.58, and that best practice is to buy green below 11% moisture and store it cool and dry so Aw stays under 0.60 at all stages, limiting non-enzymatic browning and lipid oxidation — could not be confirmed here. The Cafe Imports blog post announcing the study gives no numbers at all; the figures come from a Barista Hustle summary, which returned HTTP 402 (paywalled) on fetch, and from an SCA seminar video.⁴ The 42-page report itself was not read. Treat every one of these digits as reported at second hand.

What counts as a safe number

Here the record gets genuinely messy, and this wiki cannot give a single answer.

Nordic Approach, fetched directly, is deliberately loose: “Specialty coffee is generally considered stable when aw is below ~0.65–0.70, though optimal levels can vary by origin, processing, and storage conditions.”² AQUALAB goes further and refuses the question outright — “there is no one perfect water activity level for great coffee,” because the right level depends on “the type of coffee you want to produce.”¹

Against that, the numbers actually in circulation across the trade are these, and they do not agree with each other:

Figure	Where this wiki found it	Status
Aw ≤ 0.70 — “the SCA specialty limit”	Trade summaries of the Cafe Imports and Sustainable Harvest work54	Not a published SCA rule — see below
Aw ≤ 0.65	Fine Robusta, Cambodian implementation	One roaster’s blog, flagged in that note
Aw ≤ 0.60	Specialty coffee, quoted as specialty green practice	Trade practice, not a standard
~0.65–0.70	Nordic Approach glossary	Confirmed at source, stated as approximate2

The honest summary is that specialty green coffee is generally bought and stored somewhere between 0.40 and 0.65 Aw, that below about 0.60 is the target most quality-focused importers describe, and that no published rule fixes any of it.

To verify — there is no SCA water activity limit, because there is no SCA green standard

The claim that “the SCA sets an upper limit of 0.70 Aw for specialty grade” is repeated across the trade, but the SCA’s own standards page — checked directly for this note — still lists SCA-110 Green Coffee: Grading, Specifications, and Test Methods under “Standards in Development”, not among the published standards.⁶ The published list covers tasting (SCA-102 to SCA-105), evaluator competencies (SCA-710) and equipment certification only. This wiki’s Green coffee grading and Specialty coffee notes record the same gap from the defect-counting side. Do not quote 0.70 Aw as a current SCA requirement. It is trade convention attributed to a standard that has not been published.

To verify — the wiki's own figures disagree

Three notes here now carry three different ceilings: 0.65 for Cambodian Fine Robusta, 0.60 as specialty green practice in Specialty coffee, and ~0.65–0.70 from Nordic Approach above. None of them is wrong at its own source; they are different bodies, different species and different degrees of caution. But nobody should read a single “specialty water activity limit” off this wiki, and anyone writing a contract should name the standard they mean.

To verify — the risk-range table, and the danger of drying too hard

A widely circulated technical report from the importer Sustainable Harvest is the source for a three-band risk model: 0–0.35 Aw “dangerous” because very dry beans expose their lipids to oxygen and go rancid; 0.35–0.65 the safer working band; 0.65–1.0 “dangerous” for mould, moisture uptake and fast quality loss — with the conclusion that 0.70 Aw is too high for specialty coffee in essentially every instance. The report PDF was retrieved for this note but returned as unreadable binary, so none of it has been read at source.⁵ The counter-intuitive half — that over-drying causes its own oxidation problem — is the interesting claim and the one most in need of a second source.

To verify — microbial thresholds and roasted-coffee physics

Two more sets of numbers travel with this topic and were not verifiable within this note’s fetch budget. First, the general food-microbiology thresholds: pathogenic bacteria said to stop below ~0.90 Aw, spoilage moulds below ~0.70, most microorganisms below ~0.60. These are standard food-science figures, not coffee-specific ones, and the FDA technical guide on water activity returned HTTP 403 when fetched. Second, an SCA seminar is reported to give coffee a monolayer value near 0.61 Aw and a glass-transition range of 0.538–0.760 Aw, above which roasted coffee becomes physically unstable; the research report behind this note could not even resolve the name of the researcher credited. Both sets are recorded as claims only.

Why the number moves

Green coffee is hygroscopic — it trades water with the air around it until the two are in balance. That is the practical consequence of the RH/100 equivalence above: a bean at 0.60 Aw parked in a 75% RH warehouse will climb, and one in a 40% RH warehouse will fall. Water activity is therefore not a fixed property of a lot. It is a running conversation between the coffee and its environment, which is exactly why importers re-measure on arrival rather than trusting the mill’s certificate.

That is also the argument for hermetic packaging. As Coffee supply chain records, a jute sack offers no moisture barrier at all, while a GrainPro liner or vacuum pack seals the coffee off from ambient humidity — the difference between a lot whose Aw is fixed at the mill and one whose Aw is decided by the weather in every port it passes through.

To verify — the storage conditions

A commonly quoted storage recommendation of ~20 °C (68 °F) and ≤75% relative humidity for green coffee traces, in the research behind this note, to an SCA educational video rather than a written standard. Plausible and consistent with the hygroscopic logic above, but unread at source and recorded here as a claim.

Where it sits against everything else

Water activity is the third green-coffee number, alongside the two this wiki already documents:

• Defect count — what is physically wrong with the beans, judged by eye.
• Moisture content — how much water is in them, typically targeted around 11% at the dry mill and capped near 12–12.5% before storage (see Washed process and Natural process).
• Water activity — how much of that water can actually do damage.

The first two are old; the third is recent, and it is the only one of the three that predicts what happens after the coffee changes hands. A lot can pass its defect count and sit comfortably inside the moisture range and still carry enough free water to fade in the warehouse — which is precisely the failure mode Fretheim set out to catch.³

Why it matters

For anyone paying auction money, it is insurance. A Best of Panama or Cup of Excellence lot is bought on a cup score recorded at origin and drunk months later on another continent; water activity is the cheapest available proxy for whether the coffee that arrives is the coffee that was scored. See Green coffee auctions and Green coffee buying for the transaction that depends on it.

For the wider trade it is a quiet argument about what quality control should measure. Moisture content is a legacy of an era when the question was “will this rot in the hold?” Water activity assumes the coffee will survive and asks a harder question — how long will it still taste like this? That is a specialty question, and the reason the measurement spread through importers’ labs years before any standards body got around to writing it down.

Related

• Green coffee grading — the defect count that sits beside this measurement
• Specialty coffee · Fine Robusta — the two grade definitions that quote water activity ceilings, and disagree
• Coffee supply chain — arrival sampling, GrainPro liners and the origin-vs-destination reading
• Dry milling · Washed process · Natural process · Honey process — where drying decides the number in the first place
• Coffee packaging · Coffee freshness and degassing — the same water-and-oxygen problem on the roasted side
• Green coffee buying · Green coffee auctions — the transactions this number protects
• Coffee Value Assessment — the SCA framework where physical condition is a formal, separate assessment
• Specialty Coffee Association — the body whose green standard, SCA-110, is still unpublished
• Q Grader · Cupping — the tasting half of quality control
• Barista Hustle — the education publisher whose summary carries most of the circulating figures
• Cafe Imports — the importer that ran the study, and the rest of what it publishes

Official links

• Official site — SCA Coffee Standards (SCA-110 green coffee, still in development)

Footnotes

1. AQUALAB — The Role of Water Activity in Coffee Quality — water activity instrument manufacturer; confirmed via WebFetch; source of the qualitative-vs-quantitative distinction, the dew-point and electrical-sensor methods, and the “no one perfect water activity level” statement ↩ ↩² ↩³

2. Nordic Approach — What is Coffee Water Activity? — Norwegian specialty green importer; confirmed via WebFetch; source of the free-water definition and the ~0.65–0.70 stability range ↩ ↩² ↩³

3. Cafe Imports — Water Activity in Specialty Green Coffee: A Long Term Observational Study by Ian Fretheim — green coffee importer, 11 February 2019; confirmed via WebFetch; source of Fretheim’s title, Megan Person’s role, the 2012 start, the 25,000+ readings, the 42-page report and the shelf-life motive ↩ ↩² ↩³

4. Barista Hustle — Can You Please Summarise Cafe Imports’ Water Activity Paper? — coffee education publisher; returned HTTP 402 (paywalled) on fetch; the numbers attributed to it are unread at source ↩ ↩²

5. Sustainable Harvest — Understanding Water Activity — importer’s technical report; retrieved but returned as unreadable binary, so nothing in it was read directly; cited via the research report only ↩ ↩²

6. SCA — Coffee Standards — Specialty Coffee Association; confirmed via WebFetch; SCA-110 Green Coffee still listed under Standards in Development ↩