The city are in the process of slowly transitioning the ferries to electric (battery powered and hydrogen fuel cell, based on distance). I believe that there's already a fuel cell ship in limited service.
I learned about it by attending the port of SF public meetings where they've recently been discussing the Sea Change Ferry, the first hydrogen fuel cell powered public ferry in the world that they will be starting service with soon. It's so cool, they've designed a floating fuel barge that takes power from the Port side power connections they have that get their power directly from hydroelectric sources. The fuel barge takes that clean electricity and disassociates bay water to generate the hydrogen to power the ferry. The ferry then stops by every so often and picks up more hydrogen.
anyway, in those meetings they've casually dropped "everyone knows that the existing ferries ate worse than driving" like it's common knowledge among the SF port folks. I'm sure they did a study on it that we can find but my 2 cents is that it doesn't really matter if they look better on paper when full or not, etc. What matters is the empirical average, we have the data. Co2 emissions per passenger mile is the metric that matters and what I'm assuming they used to make the comparison. That way the specifics of the ferry or the car, and their loading don't get factored in.
It would be interesting to see if that calculation was with average ferry load, and if there's is any tip over with some reasonable increase in ridership. During rush hour the ferries are packed already, but I find that people in the US leave the car at home only as a last resort.
I'm also thinking that when assessing regional impact of transportation the analysis should be end to end: if a ferry is slightly worse, but then you arrive and are walking around or taking an electric bus/tram, then on the whole it might be lower emissions for that one trip than driving, with the additional external benefit of one less car in an urban environment (fewer harmful particulates, fewer cars around which afford us to reshape public space to give more space to pedestrians, cyclists and public transport, less noise, lower chances of collisions, etc). If people within a city are using PT, and 5% come into the city by ferry, does that provide a benefit in terms of CO2 emissions compared to the status quo? I would be surprised if it didn't, but of course don't have numbers in either direction.