The paper you cited doesn't supply evidence- according to the abstract it's their best guess based on experience with other coronaviruses.

The evidence is actually scant but there's a decent discussion of it in the Guardian article I linked below and the paper linked from it.