Quantifying snowfall rates using underwater sound

Abstract

It is well known that rainfall rate can be estimated by analyzing the spectral character and level of the underwater sound that it generates. There have been a few field reports of increased sound levels associated with the occurrence of snow; these reports suggest an increase in sound levels above 20 kHz. It has previously been demonstrated that snowflakes can generate both a small impact sound and a separate high frequency pulse that is consistent with the sound generated by a resonant bubble. One aspect that these earlier studies have not explored is the dependence of this sound generation on the type or intensity of snowfall. We report on observations of sound generated by snow falling into a tank of water, quantifying snowfall rate using an Optical Scientific precipitation gauge. Recorded signals allowed analysis of frequencies between 1 and 50 kHz. Using the classification scheme of the International Association of Hydrological Sciences for snow, seven distinct snow types, as well as rain and freezing rain, were observed with a range of precipitation rates. Snow types that produced a signal included column, needle, irregular crystal, graupel and ice pellet. Snow types for which no signal was detected were plate, stellar crystal and spatial dendrite. The previously reported rise in high frequency sound could not be distinguished unambiguously in the present data. A small peak at around 12 kHz was seen in spectra of some snow types similar to the characteristic 14 kHz peak seen in some rain-generated sounds. There was a clear correlation between sound level and snowfall rate at frequencies above 10 kHz.