Usability of NWP Model Liquid Water Output for In-Flight Icing Forecasts

In this study we explore the usability of NWP model liquid water output to support the identification of in-flight icing as well as the icing severity calculation of the German icing forecast and diagnosis system ADWICE (Advanced Diagnosis and Warning System for Aircraft Icing Environments). The latter is in operational use by the German Weather Service (DWD). Further developments are in process and made in cooperation with the University of Hannover. ADWICE diagnoses but also prognoses three- dimensional air spaces where in-flight icing might occur. Icing severity which depends on ambient air temperature, droplet size distribution and super-cooled liquid water content (LWC) is estimated by a simple parcel uplift approach in combination with the NWP model LWC and other microphysical assumptions. To improve that calculation and the identification of regions with an icing risk the question has to be answered whether current liquid water content output of the COSMO-EU NWP model is useful for these purposes. As a first step that quantity was determined at PIREP locations over the Eastern US where during twelve test cases great numbers of pilot reports were made available. The icing information given in the PIREPs is used as “truth” data and has been compared to the generated fields of LWC. The results of this comparison show too small amounts of LWC in the neighborhood of aircraft icing observations. Furthermore the distribution of LWC is not consistent with the icing observations mentioned in the PIREPs. To explore the matter further, an idealized simulation test case was done in order to analyze the COSMO-EU cloud microphysical scheme in a more detailed way. It was found that the freezing of cloud water was parameterized too fast and too strong, resulting in too low values of LWC. Improvements in the COSMO-EU cloud microphysical scheme are, therefore, needed. Generally, not only surface precipitation, as it is the standard, but also cloud liquid water should be used to evaluate the microphysical scheme. This requirement seems to hold not only for the COSMO-EU model but for many NWPs.