The articles refer to the cracked member as a "beam" but, in reality, I don't think it is a beam. The structure of that type of bridge relies primarily on the main arches to carry the weight -- of the bridge structure itself, of the roadway, and of the traffic. Think of it like at attic truss in a house. The lower, horizontal member that carries the ceiling isn't a "beam," it's the bottom chord of a truss. The nature of the members of a truss is that they are supposed to act mostly in tension or compression, and major loads are supposed to be applied only at intersections of multiple members.
You can see that in these photos. Look behind the riveted gusset plate and you can see several other parts converging in where the cracked "beam," the vertical member, and the other "beam" come together. That said, truss members do act to a lesser degree as beams for localized loads. Again using the house attic truss, the loads of the roof go along the top chord members until the loads get transferred into the truss verticals and diagonals. Likewise the weight of the ceiling -- each short section of bottom chord acts like a short, local beam to carry the weight of that section of ceiling to the points of intersection; at the points on intersection ("panel points") all loads are converted into either tension or compression in the truss members.
The cracked beam should have been subject almost entirely to tension or compression. That's not a tension crack -- that's a shear crack. For long beams, the type of stress away from the ends is bending, and the type of stress at the ends is shear. Joints in steel structures are often reinforced near joints to handle shear stresses. The fact that this member failed -- apparently in shear -- so close to the end of that gusset plate suggests to me that the entire design of the bridge may be flawed. I don't know what the two highway departments are doing but if I were the engineer hired to look at this thing, I would be re-running ALL the numbers for the design of the entire bridge. I would not be surprised if we read in a few weeks that those gusset plates are too small (too short) and that they all have to be replaced.