Using Crowd-Sourced Precipitation Observations to Build a Winter Surface Hydrometeor Classification Algorithm

The mPING project collected over 200,000 observation of winter surface precipitation type between 19 Dec 2012 and 30 April 2013. These data are used to drive a classification scheme that uses soundings extracted from the High Resolution Rapid Refresh model analysis to generate predicted observed precipitation type based on sounding parameters. The data set limitations, the various methods used to generate the classifier and the classification results on test data are all discussed and presented.