Even without measures such as circulating water or heating it, there were other physical factors at work slowing a freeze-up of a water tank. The large volume of water in the tank and the high latent heat of fusion of water worked against a complete freeze. Also materials such as wood, air, ice, and snow are thermal insulators which also slow the transfer of thermal energy. A metal water tank, on the other hand, without an insulating gap of air between an outer shell and inner vessel is probably not a good idea.
I've noticed that water tanks along the Canadian Pacific often had the bottom of the tank competely enclosed in a wood structure for good reasons.
It's all physics!