Assessing the impacts of variations in icing relevant observational datasets on diagnoses of in-flight icing conditions

The accurate diagnosis of in-flight icing conditions is directly affected by the quality and age of observational datasets used to infer the microphysical state of the atmosphere. Typically, observational datasets are available at regular discreet intervals. However there are rare occasions when discontinuities, outages, quality issues, and other problems inherent to near real-time datasets occur. Often times these issues are out of the control of the end user. Therefore, it is desirable to have knowledge about how operational products perform under these constraints. One such operational product is the Current Icing Product (CIP) algorithm for diagnosing in-flight icing conditions. The CIP provides an hourly diagnosis of icing probability and severity. Meteorological datasets including satellite, radar, surface station observations (METAR), lightning data, and voice pilot reports (PIREPs) are used in combination with numerical weather prediction (NWP) model data to generate these diagnostics. At a minimum CIP requires satellite, METAR, and NWP model data while lightning, radar, and PIREP data provide supplemental information when available. Satellite, METAR, and PIREP data will be the focus of this study.

Initial results show that the ability of CIP to correctly diagnose positive icing conditions (PODy) degrades as the age of METAR and satellite data increases. The volume of icing probability diagnosed by CIP is also higher in the cases using older data than those using the most up-to-date information, indicating a disconnect between that dataset and the other up-to-date datasets being blended together in CIP. These effects have been noted in the full CIP domain, but results suggest that they may be more pronounced near the edges of large cloud features. Identifying these regions to further investigate the impacts of older data may yield improved results. Older satellite and METAR data may incorrectly represent the existence of clouds over a particular gridpoint within the CIP domain. In reality that gridpoint may actually be clear due to clouds (or other favorable icing conditions) exiting the area. Identifying the frequency with which situations like this occur is a primary goal of this study. In addition to examining the impact of data age, variations in the icing intensity reports from voice PIREPs will also be evaluated. Voice PIREPs of in-flight icing currently offer the only in-situ observations of icing available to CIP. However, additional in-situ observations of parameters indicative of in-flight icing conditions do exist. These observations could provide additional benefit to the CIP, both for the identification of in-flight icing conditions and for the confirmation of ice-free airspace, if made available. The distribution of icing intensity values and concentration of in-flight icing observations will be varied to observe the impact on the resulting icing probability and severity diagnosis by the CIP.