The In-flight Icing Diagnostic Algorithm (IIDA) recently became an operational tool for use in anticipating the existence of icing conditions aloft. Objective verification was an important component of the development of IIDA. In addition, an in-depth quality assessment was a critical aspect in the decision process to transition IIDA from an experimental to an operational product. Results of this quality assessment are presented and discussed.

The quality assessment involved two seasons of diagnoses and observations. The evaluations were carried out using the Real-Time Verification System (RTVS) and a post-analysis verification system. As the only widely-available observations of icing conditions, pilot reports (PIREPs) were used as the verification data. The verification methods that were used reflect the many years of work that has been devoted to this problem. The results were stratified by altitude, region, and reported icing severity.

Results indicate that IIDA performs better than other icing forecast algorithms overall, and better than the operational icing forecasts (AIRMETs) when they are evaluated in the same context as IIDA. Overall, the IIDA detection rates (for Yes and No PIREPs) are better than for the other forecasts, and IIDA diagnoses generally cover smaller volumes of airspace. Persistence analyses indicate that the IIDA diagnoses have useful skill out to about six hours. Initial verification results for the forecast version of IIDA (the Integrated Icing Forecast Algorithm; IIFA) are also presented. These results suggest that IIFA also has skill in identifying icing regions.