@evogytis and @Pierre_F I think it’s fair to say that because of lack of sampling in the early days of the outbreak (and from the reservoir), based on the sequence data alone it’s difficult to say whether there was only one ‘spill-over’ event or multiple. Because the reservoir (let’s say bats for the sake of argument, but could be different animal) remains unsampled, a hypothetical scenario that would also be consistent with the data would be e.g. an infected bat in, say, october 2013 starts infecting other bats that then starts infecting people - hence the early diversity we see on the tree could be reflective of (recent) diversity in the bats, and not in humans.
We know from earlier outbreaks, however, that if we sample EBOV from non-human primates (sadly no good data from bats), then the EBOV population is quite diverse (because of older ancestry). So based on those earlier observations, the epidemiological data available from the 2013-2016 EVD epidemic, and the sequence data, I agree with @evogytis that the most parsimonious explanation would be a single introduction. The data is certainly in agreement with this, but unfortunately can’t prove (or disprove) it. One thing we can say with almost certainty, however, is that once this outbreak got going no further spill-over events were observed.