How reliable are icing forecasts? This research presents a statistical evaluation of several algorithms used to predict icing risk, a significant hazard for aviation. By verifying the algorithms against pilot reports from the Winter Icing and Storm Program, the study aims to assess their accuracy and identify areas for improvement. The verification process utilized gridded forecasts of temperature and humidity from the European Centre for Medium-range Weather Forecasts (ECMWF) global model. The results, presented in terms of percentage of detection and total forecast surface area, highlight the inherent trade-off between these two criteria. The study also revealed that incorporating vertical velocity into the algorithms reduces the detection percentage but decreases the area where icing is forecast to an even greater extent. The research underscores the challenges in accurately forecasting icing conditions. The findings provide valuable insights for refining existing icing algorithms and improving aviation safety by mitigating the risks associated with in-flight icing.
A statistical verification of forecast icing algorithms is a relevant contribution to Meteorological Applications, which focuses on practical applications of meteorological research. The journal often features studies aimed at improving weather forecasting and risk assessment, making this paper a fitting addition to its content.