I don't know if it needed to be done, but considering the work to improve the conveyance of the river and the eventual Gardiner realignment the design seems pretty future proof with the maximum clearance below less supporting obstacles.
The criteria of no work in the water (and no permanent structure in the water) likely came from environmental or hydrology.
They probably wanted to improve the hydraulics of the channel by having no piers to obstruct the water (assuming the existing bridge will be replaced during this one's life and it too can be improved at that time). Not sure if debris (logs, ice) jamming up on the bridge piers is an issue here.
The other thing that's seen is in-water restriction windows to allow various fish, mollusks, etc. to spawn.
Bayfield River has a similar bridge as a single span, and it replaced a 2-span bridge with pier in the water. A 3-span was desired, with piers near the shore, but the in-water work window was small and related to temperature. I don't recall exactly, but it was no work until water was above 21C (or thereabouts) for some type of mussels. Depending on the year, this could have meant late June or even August for start of work. All of a sudden, the in-water work could not be guaranteed to be done in a year and with that uncertainty, a single span bridge was chosen.
Another thing is traffic. Building a pier in line with the DVP median would mean reduction in traffic by a lane in each direction. With very good soil (I doublet it), it could be as short as a several week closure, but realistically more like several months or several seasons. A pier in line with the pier just on the west side of DVP would require in-water work. (for those wondering, the in-water work for the temporary works is for much less loads than the permanent structure, so they could go in much faster, and they must only be designed for a 10- or 25-year flood and not 100 year).
The bridge had to be raised by about 1-foot extra clearance - similar to the above assuming that somewhere in this bridge's life the adjacent one will be removed and then it too can be improved. I am sure the current bridge has been hit a few times in its life. This doesn't really affect the need for, or absence of bridge piers.
Once a decision is made to go single span, this is probably the bridge type that has the shallowest superstructure (distance from top of tracks to underside of bridge structure) for what can span that distance. Basically, the beams in the floor system just take the load to the nearest hanger/cable location and its the arch that actually carries the load across the river. If beams were to span the whole way, the beams themselves would probably be about span/18 deep (5m). Look a the beam spanning the river
now - it's probably 40m span and 2.5m deep.
The one difference in this bridge is that the cables are not vertical. A typical arch, like the
Humber Pedestrian Bridge, has vertical hangers/cables. This type of bridge is more flexible as there is bending in the arches between the hanger points. Not noticeable on a pedestrian bridge, but not suitable for a rail bridge. So this one is a Network Arch, which has cables criss-cross to stiffen it more with some truss-like behaviour.
For the criteria of a single span bridge about 100m long, with clearance concerns below and rail above - this was probably the simplest bridge that could have been built here. (A
cable stay bridge would have been a more expensive option). Possibly a
truss would have been comparable, but those would be much less aesthetically pleasing.
Half-through girders are often used for this purpose (minimizing distance between tracks and bridge underside), but for this span it would be a massive girder that is too large to ship.