I don't make recommendations that you don't pay me for. And when I do, I use analysis to answer questions about things like these. I analyze trip making behavior and travel demand to assess the viability of a proposed system, line, or connection. I use math and data to make such "predictions", based on observation and measurement of past travel behavior and demand changes in the face of consumers actual reactions and use of new alternatives.

Population and/or demographic trends (as in the population of say Cloverdale, as was mentioned)are only one factor to consider, in creating a "gravity" or "disaggregate" travel demand model, to predict "future" station demand, and the attraction of those trips to other stations on the line. But you also need to look at the costs: travel time cost, fare, wait time, and access from train to final destination. You nee to look at factors such as proposed service frequency, average speed between stops, and total travel times, to determine these things. You also have to look at auto costs like depreciation, fuel, maintenance, insurance (fixed and variable mileage costs). You have to look at congested and uncongested travel times during different time periods, and compare those to transit or rail.
\
Is Cloverdale, "too small" to have a rail station, or be the norther terminus of SMART? I don't know. I haven't seen the data, nor have I seen the math, so I don't know that answer. This area seems to be growing, and congestion is creeping a little further north on 101 every day, despite (maybe even because of...?) the recent improvements, all of which will tend to improve future rail attractiveness.

Another "cost benefit" trade off to think about is the cost of the alternative, or of "doing nothing". If the average vehicle trip on the 101 is 15 miles, and I remove 200 trips going 30 miles, I have now "created" additional freeway capacity for 400 new trips on that same roadway, by removing those trips competing for taht road space. What would it cost in new capacity, O & M, etc. to create that additional auto travel capacity, versus taking it off of the road, to put it on the train?

In order to assess potential travel demand from Cloverdale (and its surrounding "catchment" area) to other locations accessible from SMART, in the 101 corridor. I would need to look at the distribution of commute (home to work) data from the census, and traffic counts on 101 from the ramps, and any Origin and Destination surveys that have been conducted, to determine travel demand to places within the corridor from Cloverdale and stations proposed north of Santa Rosa Airport. That is how this kind of data that is/should be collected to do these things in a professional manner.

If the population or "size" of the market, or the "rural" nature of a location, were the criteria for deciding where to stop the choo-choo to pick folks up, then Metrolink probably wouldn't have built a station at Acton. Acton was built as an emergency station by Seabees after the 93 earthquake, in a "rural" location, with a small population. And, after that, this census tract around Acton (very large rural tract) has one of the highest percentage of home to work transit commutes in all of Southern California.... So size alone may not matter much in predicting ridership.

I have expressed my views, based on my best professional opinion, and observation of intermodal transfers all over the world, that the 10 minute walk from SMART to the ferry terminal will have an extremely negative impact on achieving optimal intermodal transfers between the ferry and SMART. It should be a 3 minute ship to train transfer, and the terminal should be "on-dock" or as close to "on-dock" as possible. Wait time and transfer time penalties are among the biggest negative factors reducing transit utility and thus ridership.