Classic airframe icing depends on the presence of supercooled water drops (SCWD) These only exist in OATs below zero, so there's one limit. By the time the OAT is down to about -20º to -25º most SCWD will have frozen, and only small SCWD giving light rime ice will exist, so there's another limit. Finally, by about -45º all water will have frozen, so no icing at all at these low OATs.
The formation of Cb type cloud requires a steep environmental lapse rate, and to get zero surface temperatures and a rapid drop as you climb you have to be in pretty extreme conditions, not in stable cold highs over land but in Arctic and Antarctic conditions.
In temperate climates you would probably have a surface temp in the low pluses, and that would mean a zero degree height in the cloud in the low thousands, and the end of serious icing at about 15,000ft in the winter and say 8000 to 20,000 in the summer. In the tropics, where the zero degree height is about 14-16,000ft the icing would be from 16 to about 30,000ft. Roughly! But in Cb the OAT is all mixed up as the updrafts and downdrafts occur.
Stratocu does not depend on instability for its formation, it often occurs in a stable flow with loe surface temperatures, so the icing can occur lower down. I, too, have experience of severe icing in Sc at 2000ft in the UK in winter.
So the answer is that icing depends on SCWD and the OAT, and given the requirements for the formation of Cb the figures first given are practical values for temperate climates. Icing in Sc can occur at very low height, and, of course, freezing rain can occur right down to deck level.
Dick W