The albedo cliff
Fresh snow reflects perhaps eighty to ninety percent of incoming shortwave radiation. Bare ground or open water reflects something like ten to twenty. That is not a gradual difference; it is close to a step change, and it sits right at the melt threshold.
The consequence is a feedback. Warmth melts snow, exposing darker ground, which absorbs more energy, which produces more warmth. The same loop runs in reverse when snow arrives. Systems with a feedback this strong tend to have sharp transitions rather than smooth ones, and to show hysteresis: the conditions that end a snow season are not quite the conditions that would have started it.
Ageing matters too, and by more than people expect. Snow does not go from ninety to twenty in one step. As grains round and grow, and as dust and soot accumulate at the surface, albedo declines through the seventies and sixties well before the pack is gone. Deposited particulates can advance melt by days to weeks in affected regions, which is a small forcing acting on a large lever.
The practical version: a late dusting of fresh snow on an old pack resets the surface for a while, and it is worth noting whether what you are standing on is new or merely white.