In recent posts, I've discussed how density isn't all that important in transit demand, how the idea of "supporting transit" is problematic, and how different people have different goals for transit, and density affects these goals differently. Because my own goals (see the top of this page) are wrapped up in a feedback loop based on mode share, my most intermediate goal is getting people out of their cars.
Transit mode share, in fact, is where density is least relevant. This may seem surprising, but only if you believe that density is the only way to control the relative value that people get from various modes. The transit boosters who worry about density actually believe that it's possible in the short term to increase the value of transit by throwing more money at it, but that that's unsustainable in the long run. Their big blind spot is that we actually have quite a bit of control over the value of driving, if we can find the political will.
This brings us back to the Magic Formula for Transit Ridership:
1. Give transit its own right-of-way and good terminals
2. Make it hard to use cars
3. Make it expensive to use cars
Many transit advocates have enough exposure to the concept of (3) in the form of congestion pricing and gasoline prices, but they seem very resistant to considering step (2), probably because they don't want to be accused of wanting to take anyone's car away. The Very Serious People are all afraid to talk about decreasing the size of the road network.
But what if, while Spain was building all those high-speed rail lines, they didn't also pump billions into a truly gigantic highway network? If drivers faced constant congestion on old highways, wouldn't we expect higher ridership on the trains? Wouldn't we also expect that if the highway network was old and small enough, but the train network was the size it is today, eventually there would be enough demand for the trains that they would be completely profitable - operations and capital?
You can do the same thought experiment with any place. No matter how sparsely populated it is, just subtract some roads while keeping the rail and/or bus network constant, and eventually the place "supports transit." Take Wyoming. Now imagine it without interstate highways. Would that be enough to support restored passenger service on the train lines? How about if we turn all the roads to gravel?
Back in 2010, I had a similar discussion on Human Transit about the supposed convenience of cars. A lot of people had problems with the idea that convenience was dependent on the quality of the infrastructure, but I think I showed that if you throw enough money at any transportation system you can make it feel convenient to its users.
Similarly, if you make the car infrastructure shitty enough and expensive enough, you can make transit feel like a bargain. Density may make it politically easier to support transit expansion or harder to support road expansion, but that's not a matter of "the density to support transit," it's "the density to make it likely enough that transit will receive more political support than roads," which is not the same thing at all.
If you're really not convinced, I challenge you to come up with a place, or a route, where you can't increase transit ridership by taking away roads or increasing prices. If you want data, I have density and mode share figures for all of the municipalities and census-designated places in the New York Combined Statistical Area. Go for it.