A method to diagnose surface precipitation types is suggested. Most cases o
f freezing rain and ice pellets occur with a layer warmer than ODC extendin
g above a surface-based layer of air colder than 0 degreesC. The procedure
uses predictors proportional to the product of the mean temperature of a la
yer and its depth. These predictors can be seen as areas on aerological dia
grams. A positive area is associated with a layer warmer than 0 degreesC, c
onversely a negative area is associated with a layer colder than 0 degreesC
. The same predictor is used to discriminate snow from rain. A statistical
analysis was applied using the North American aerological stations network
to determine a set of criteria for discriminating freezing rain, ice pellet
s, snow, and rain. Once the criteria are known, the precipitation type can
be easily diagnosed using temperature profiles from upper-air observations
or from numerical weather prediction models. The method has been in operati
onal use at the Canadian Meteorological Centre since 1995.