Marine stratocumulus clouds are highly abundant, and they exert a significant cooling effect on the Earth's atmosphere. Improved understanding of their evolution and abundance has been targeted as critically important for constraining future climates. Droplets in these clouds, if sufficiently large and abundant, can collide to be converted into drizzle drops, acting as a sink for cloud moisture thereby limiting stratocumulus cloud lifetimes. Currently, the primary tool for ground-based detection of the presence of drizzle in stratocumulus is Doppler radar, although the longer wavelengths of the electromagnetic pulses generally limit drizzle detection to larger drops. The current study presents an innovative approach that extends radar detection to fine drizzle by using a machine-learning algorithm trained with aircraft in-situ measurements, identifying skewness in the Doppler radar signal as important for discriminating fine drizzle presence. Based on this method, the authors conclude that drizzle is far more common in marine stratocumulus clouds than previously thought, a result that represents a potentially major advance in our understanding of marine stratocumulus properties. Moreover, the study demonstrates the great potential of machine-learning for extending the capabilities of well-established atmospheric measurement techniques.