Most bridges require a narrowing of the channel for structural reasons, often including shallower water.  This means that the water passes through a little faster. Ice does not form as easily in faster-moving water, so that could explain the effect.

Another factor could be that the bridges were themselves warmer than the surrounding air because of absorbing light from th sky or heat from traffic. It then would radiate heat down to the water, enough to delay freezing.

If there were any salt on the bridges to prevent ice on the road surface, then salt solution dripping on any ice below could destroy it.  Any weakened or separated bits of ice soon would wash away.Other forms of pollution on the bridge surface could have a similar effect.

Air movement below the bridge could be concentrated to disturb the water surface, also preventing ice collection in such spots.

It might help to decide which effects were most important, if you compared the patterns of ice formation on other spots along the river, say where there were sources of warmth or changes of flow rate.  

Good luck!

Jon

I would be happy to go for the bridge will trap a pocket of air, this air will provide a bucket load of insulating properties and so, it helps the cut not freeze up.

The canal water, and indeed the ground generally, radiates heat. The amount of radiation is proportional to the difference in the absolute temperature (to power 4) of the emitting and receiving bodies. The water and the bridge were probably at about 0 Celsius, but a cloudless sky is much colder. Hence frost or ice on exposed surfaces, but sheltered surfaces are warmer. Similarly a car parked overnight near a wall may have clear windows on the wall side, but frozen on the far side.