It is recommended that copper tubing not be exposed to the elements when installed in a very cold environment. Tubing should be installed under ceiling lagging to insulate and protect it from draughts. (The wind chill factor is the main cause of failure under these conditions.)
Pipes, fittings and components should, when necessary, be protected against freezing using thermal insulation. The insulation provided should be appropriate to the minimum temperatures of the geographical area. Ref: SANS 10252-1: 2012.
Special care must be taken when plumbing is installed in a Cold Storage area. Tube should not come in contact with the wall and floor areas.
Freezing itself does not necessarily mean that the pipes will burst. It is not the distension of the tubes by the ice which will cause a burst, but rather its distension by the water under pressure which causes the problem. Bursting, for example, does not occur at the first part to freeze, on the contrary, bursting invariably occurs where the water is still present as a liquid and usually, therefore, one of the last parts of the system to actually freeze. The pressure to produce bursting is generated by the change from water to ice and if the coldest point is in the middle of a straight length so that a plug of ice is first formed at this point with say, a fitting or stop tap at one side, then the chances are that the plug of ice will be moved along the tube by the pressure rather than a pressure build-up occurring.
The effect of pressure is to depress the freezing point slightly, so that under a mild frost nothing would happen in the case of a smooth bore copper tube. On the other hand, of course, if the first point to freeze was where there was a fitting, or stop cock, and there was a length of pipe which was subjected to freezing temperatures with a solid closure at the far end, then eventually the pressure build-up would be such as to burst the tube. Under these conditions the relative bursting point of a galvanized iron pipe will be higher than that of a corresponding copper pipe, so that marginally less failures might occur with galvanised iron. On the other hand, the chances of a plug of ice moving in a galvanized iron pipe is remote, so that from this point of view more failures would be likely to occur with galvanized iron than with copper.
Summarizing the position, you have to emphasize that layout and exposure conditions are extremely important and that the cause of bursting is pressure generated by the conversion of water to ice and that this pressure can be transmitted hydraulically in the usual way. The freezing itself is not a danger, except for the generation of this pressure, it is not the distension of the tube by the ice but rather the distension of the tube by water under pressure which causes the problem. Depending on the exposure conditions, bursts are less likely to occur under some circumstances with copper because of the smoothness of the bore.