I am kind of surprised at the number of "compass is hard to use" comments here. I use a compass and a Home Depot angle locater to set the azimuth and elevation of a 220 yagi that I have mounted on an old Atlas studio boom microphone stand in by back yard to capture the passage of the moon through the SPASUR (Lake Kickapoo) radio fence. I usually try to do the antenna pointing during daylight hours. It is a lot easier to read the compass while sighting along the boom in daylight. Whenever I check my alignment by sighting along the antenna boom when the moon is passing through the fence (i.e. moon is at azimuth of 91.4 or 271.4 degrees relative to Lake Kickapoo and the moon is above the horizon at both Lake Kickapoo and my QTH in Pennsylvania) I find I've got the moon boresighted.
It is a given that you keep the compass away from ferrous objects while taking a reading - that includes non-obvious objects like the wrench or pliers in your back pocket that you might have forgot about :-) Alternatively, if you are going to be climbing a tower or some such thing, you can always calibrate the azimuth angle (that was the original question) by determining the bearing of a distant object then pointing the antenna at that same distant object.
The unmentioned question in the original post is just how accurate does the azimuth angle calibration need to be? The above mentioned 220 yagi is (if I remember correctly) 3db down at +/- 15 degrees so az angle is not critical but a hi gain EME array might have a 3db width of +/- 2 degrees requiring more pointing/azimuth accuracy.
JoeC K3FMA