Whole cherry goes into a closed vessel fitted with a valve. The microbes produce carbon dioxide, the gas purges the oxygen, and fermentation continues in what is left. Afterwards the fruit stays on for drying, which is the natural half of the name.
The two words describe two different stages. Anaerobic is the tank; natural is the drying. An anaerobic washed is the same tank followed by depulping and a rinse.
What sealing actually changes
Measured through a sealed ferment, pH falls from about 5.8 to 4.4. Enterobacteria dominate the first day at around eighty-five per cent of the bacterial population. Lactic acid bacteria climb through the run but stay low compared with a traditional open tank, and yeast growth is limited, which the authors attribute to the lack of oxygen. Esters and alcohols rise between forty-eight and seventy-two hours, which is the character buyers are paying for.
The word itself is weaker than it sounds. Yeasts are facultative anaerobes: they ferment sugars perfectly well without oxygen and are already doing so in an ordinary tank. Sealing the vessel changes the balance of organisms rather than switching fermentation on.
The gradient problem
One measurement deserves more attention than it gets. In a single sealed tank at seventy-two hours, lactic acid measured 5.06 grams per litre at the bottom against 2.18 at the top, and acetic acid 1.00 against 0.40. More than a twofold difference within one vessel.
Cenicafé's own work on Castillo, comparing whole against depulped and open against sealed, found no significant sensory differences between the conditions. Taken together the message is consistent: the coffee at the bottom of the tank has had a different ferment from the coffee at the top, and the label on the bag cannot tell you which one is in your bag.
The house buys these on the sample, never on the method, and says so rather than charging for the word.