Whatever you may think about Bus Rapid Transit (and I'm not the biggest fan), I would hope we can get transit supporters to agree on this: it should not take infrastructure from trains.
I've established in previous posts that buses running in mixed traffic do not qualify as "rapid," and that it makes a lot more sense to talk about BRT corridors than BRT lines, so only physically separated corridors really count as BRT. If you want to build true BRT, the next question is where do you get the corridors?
You could build brand-new corridors, and some plans do call for that, such as the Rockland-Westchester corridor Option 3B where the DOT would build a viaduct above or a tunnel below Interstate 287 for the entire corridor from Tarrytown to Port Chester. But corridor acquisition is one of the major costs of transit construction, bus or rail, and the current estimate (PDF, page 3) is that BRT with a brand new viaduct in Westchester would cost $2.55 billion versus $897 million without. Light rail would only be $1.6 billion more, and most of that is probably for two proposed tunnels that are not strictly necessary.
You could acquire your corridor by converting lanes currently dedicated for mixed traffic travel or parking, and that's probably the best option for someone who really cares about effective rapid transit. That's what was done for the famous BRT prototypes in Curitiba and Bogota. More locally, the Fordham Road Select Bus Service (PDF) simply converted travel lanes on Pelham Parkway and parking lanes on Fordham Road to bus lanes. Yes, they're not physically separated, and they're not 24 hours, but they're still widely acknowledged to speed travel time.
You could also take over an existing or abandoned rail corridor, as the State DOT is proposing to do with the Piermont Line between Suffern and Airmont (PDF). It's been done in numerous cases around the US, including Pittsburgh, Los Angeles and Boston, and it's proposed for New Britain, Connecticut. This might possibly be a good idea for a very short segment - say a block or two - but otherwise it's a really bad idea.
Why is it such a bad idea? Because even in dedicated rights-of-way, buses tend not to attract the same numbers as trains: the Pittsburgh busways have never attained anywhere near the ridership of the rail lines that were maintained. If busways do manage to attract ridership, they often don't have the capacity to handle it. In Curitiba, the current busways are maxed out at 18,000 passengers per hour per direction, and in 2007 the government approved construction of an elevated rail system.
If you convert a rail ROW to a busway, attract a lot of riders, and then want to upgrade the ROW to rail again to handle the ridership, you have to shut down the ROW for long enough to lay down all the rail, rebuild the stations and whatever else might be necessary. Congratulations, you've inconvenienced thousands!
I don't care how old and stale the rail right-of-way is, if it's good enough for buses it's good enough for trains. If it's more than two blocks long, don't waste time and money with busways. Just find a way to make the trains work. That was true in Pittsburgh, and it's true in New Britain and Suffern.