The Automated Surface Observing System (ASOS) detects significant amounts of surface icing, but currently reports only those icing events associated with freezing rain. Freezing rain represents about 25 percent of all the ice detected by the system; the majority of ice detected by the ASOS is associated with freezing drizzle, rime, or hoarfrost, and is not reported in the standard METAR/SPECI code. It has been found that the use of multiple sensors available on the ASOS permits the reliable association of surface icing with its source. This paper describes a multi-sensor algorithm which could allow the ASOS to report the source of surface icing, with particular emphasis on the identification of freezing drizzle. Two case studies (one during the January 1998 northeastern ice storm) illustrate the performance of the algorithm relative to manual observations