fauxcelt wrote:It seems to me (and I am not an expert) that a long-term transportation policy for a metropolitan area which is based only on building more roads or expanding existing roads is lopsided and not a good idea unless you are willing to use or even consider other modes of transportation in addition to building more roads just for private vehicles such as cars.It's unsustainable. First, you have "Braess' Paradox" which is a mathematical description of a road network in which adding connections can actually INCREASE total travel time for all road users. (Look it up and do the math yourself--simple algebra--this effect is powerful and real.) Secondly, you have "induced demand" which occurs over time after any transit line or road is build. (Induced demand is a term for the phenomenon of transportation-centered residential and commercial development leading to changes in O/D points for all network users over time.) The difference is that the transit line succeeds based on this induced demand whereas the "new breezy expressway" turns into the "congested, stop&go nightmare" as a result of induced demand. Third, traffic modeling has shown that traffic on roads is more like capillaries than arteries (where blood cells keep getting stuck against the capillary walls). I furthermore suspect that increased road usage does not cause linear increases in travel time (per Braess' model) but rather the travel time blows up geometrically*. Driving on a roadway is a social activity (much as we pretend otherwise) and also limited by human physical and cognitive abilities. Hence the "snarls". Rail deals with increased demand with these things known as "trains". You also have signalling and dispatching, and sidings for separating mixed traffic. (Sidings-er, "local lanes" have been tried on highways, but they don't perform well.) Fourth: maintenance costs, maintenance costs, maintenance costs. I rest my case.
*-if you think about it from a mathematical point of view, I just can't see where you have a linear model--independent variables. the mess looks very recursive to me. The first models all used linear functions, but that's because it's easier to plunk down a linear estimate that sorta works than use a non-linear function for a good fit--and then attempt to manipulate it.