Initial Steps Toward a Next Generation Current Icing Product Algorithm

The Current Icing Product (CIP) provides hourly, three-dimensional diagnoses of icing probability, categorical icing severity, and the potential for supercooled large droplets (SLD). These diagnoses are created by synthesizing observations from geostationary satellite, weather radar, surface weather reports (METAR), voice pilot reports (PIREPs), and lightning networks, and then combining the observations with a numerical weather prediction (NWP) model forecast valid at the diagnosis time. The currently operational CIP provides these outputs at a horizontal grid spacing of 13-km, the same as the WRF-RAP NWP model forecast data used to create the diagnoses. Recently NWP model forecast data with a horizontal grid spacing of 3-km have become available operationally. In addition to new NWP model datasets, a new geostationary satellite dataset has also recently become available from the GOES-R series. The CIP algorithm will require several changes to accommodate the use of these new datasets.

The most challenging aspect of integrating these new datasets into the CIP algorithm is the changes to the underlying NWP model forecast horizontal grid spacing. Currently the CIP maps gridded observational datasets with a horizontal grid spacing less than 13-km to each 13-km grid cell from the NWP model forecast. Since the HRRR NWP model forecasts have a horizontal grid spacing of 3-km, the methods used to assign a single value to each CIP grid cell from the observational datasets need to be modified or changed completely. The changes to these methods will be a major focus of this work. Once these modifications have been made, additional minor adjustments will be required to use the new NWP model data and then a baseline version of the CIP algorithm using the HRRR NWP model data will exist.
Establishing performance of this baseline CIP algorithm is vital for future development. The adapted algorithm with a horizontal grid spacing of 3-km, as well as the currently operational algorithm with a horizontal grid spacing of 13-km will be run for a one month time period in February of 2019. Icing probability diagnoses will be evaluated using icing PIREPs. These results will be compared to results from a similar evaluation of the icing probability forecasts from the Forecast Icing Product (FIP) algorithm, as well as explicit NWP model forecasts of supercooled liquid water. These results will serve as the benchmark for comparing any additional modifications and adaptations made to the next generation CIP algorithm.

This research is in response to requirements and funding by the Federal Aviation Administration (FAA). The views expressed are those of the authors and do not necessarily represent the official policy or position of the FAA.