. You don't seem to have much with ultrasonic sensors.
I operated and worked on, in the past, SONAR, equipment.
I did make an ultrasonic 'imaging' thingy that ran well for a bit more then a year before I switched over to using LIDAR.
Me, if I was to make a snow level detector, I'd use the method used by SNOTEL sites.
A more trustworthy measurement is called Snow Liquid Equivalent (SWE), which shows the amount of liquid that would be present if you melted a column of snow. The way that SWE is measured is truly ingenious. A large rubber bladder is filled with antifreeze liquid and placed on the ground. As snow accumulates on the rubber bladder, the weight of the snow presses down and forces some anti-freeze liquid out of the bladder and through a measurement tube. Scientists correlate the amount of anti-freeze forced out of the bladder with the weight of the snow above. This measurement is usually very accurate, but it does not provide the exact data skiers are looking for. Skiers can estimate new snowfall by multiplying the change in SWE by about 15, which is an average snow-to-liquid ratio for western states. For example, if SWE increased by 0.5 inches during a storm, this could mean about 7.5 inches of new snow (15 x 0.5 = 7.5). For heavier, wetter snow, multiply by about 10. For drier, fluffier snow, multiply by about 20.
I do not think that measuring snow is the best use of an ultrasonic device due to, in the least, environmental conditions.