Green coffee storage: 21,916 readings put water activity at 0.554
Specialty green coffee averages 0.554 on the water activity scale across 21,916 readings, and the moisture meter explains only 78 percent of that spread.
Two meters sit on the same bench in a green coffee lab. One reads moisture content, a percentage of the bean's weight. The other reads water activity, a figure between 0 and 1 describing how much of that water is loose enough to take part in a reaction. Buyers have leaned on the first for decades because it is quick and cheap. The second is what green coffee storage actually turns on. On the same sack of beans, the two do not always agree.
What the second meter is reading
Not all the water in a bean is the same water. Some is locked into the structure and goes nowhere. Some is free to move, to evaporate, to feed a reaction. The authors of a 2023 storage study put it in one line: a higher reading means more water is available for chemical reactions. It is not a quantity. It is a readiness.
The top of that scale is where biology starts. The same paper cites earlier work putting the germination and growth of fungus at 0.85. Nothing in specialty green coffee comes close, and the interesting range for green coffee beans sits far below it.
Where specialty green coffee actually sits
Cafe Imports has measured it on every sample crossing its lab for six years: around 25,000 readings, 10,000 of them from the same coffees over time. Ian Fretheim's write-up gives the population estimate: 21,916 observations, a mean of 0.554, a standard deviation of 0.057, a median of 0.556, a mode of 0.558. The samples come from 44 countries, with monsooned, decaffeinated and robusta lots left out.
The extremes are wide and nearly empty. The lowest reading in the set is 0.188, the highest 0.761. Microbial growth does not much begin until 0.600, and anything past 0.668, two standard deviations up, is called very uncommon.
That distribution is one merchant's sample stream, not a world average, and the study says so: around 19 percent Colombian, between 65 and 70 percent washed.
Moisture content is a decent predictor, not a good one
Across 20,993 of those samples the correlation between the two readings came out at 88.4 percent, with an R squared of 78.1 percent. Read it generously and 78 percent of the variation in one number is explained by the other. Read it the other way and about a fifth is not.
The gap shows at the edges. Of the 703 coffees sitting at 11 to 11.1 percent moisture, 51 came in above 0.610. Same figure on the meter, different behaviour in the sack. Every coffee harvest travels, and the mismatch starts to matter somewhere along the long reduction from cherry to green coffee.
Drying trials in Costa Rica and Colombia found the two move differently in time. Moisture falls steadily from day to day. The other reading swings as it comes down.
Twelve months, two bags, three temperatures
A 2023 paper in Scientific Reports ran the experiment. Arabica, washed or left natural after picking, packed in either a jute bag or a GrainPro polymer bag, held in chambers at minus 10 °C, 10 °C and 20 °C, and assessed at 3, 6, 9 and 12 months.
The polymer bags held the level much better. In jute the reading fell as the chamber got warmer, and two jute samples at 20 °C, one natural and one washed, had dropped below 0.45 by month six and were taken out of the measurement. Across the rest, the measured ranges were 0.40 to 0.59 for the natural lots and 0.38 to 0.61 for the washed ones, under the 0.70 ceiling the paper cites from SCA guidelines for specialty grade.
The 0.45 floor is not the study's own finding. The paper takes it from earlier literature as the level green beans should stay above in storage.
What the cupping score did after nine months
Scores followed the chambers. A total above 80 points is what makes a coffee specialty coffee, and the lots held at minus 10 °C and 10 °C cleared it at the start, at 3 months and at 6 months. By month 9 several were only close to 80. The grain at 20 °C aged fastest. Roasted beans have a short clock of their own, the few days a fresh roast spends resting before it brews well.
Among the cold chambers, washed or natural mattered more to the result than which bag the lot sat in. The paper tracked the reading and the score side by side for twelve months. It did not show that one caused the other.
So ask a roaster when the coffee landed and how it has been held since, not only when it was roasted. The roast date is printed on the bag. The nine months before it are not.
Taste our coffee from Mexico
Freshly roasted, delivered to your door.
