Thursday, 16 January 2025
Hall C (New Orleans Ernest N. Morial Convention Center)
Deep convection can cause various weather hazards such as convectively-induced turbulence, lightning, icing threats, and downbursts, which pose significant concerns to aviation operation as they can result in economic losses and structural damage as well as inflight injuries. To provide a safe guidance for flying aircraft to avoid deep convective areas (DCAs), this study developed a multi-model and -diagnostic ensemble (MMDE)-based probabilistic forecast system using global numerical weather prediction (NWP) models such as the Unified Model and Korean Integrated Model operated by the Korea Meteorological Administration. To predict the target predictands for DCAs that are defined as the areas where the 15 dBZ Echo Top Height (ETH) exceeds flight level (FL) 250 (about z = 7-8 km) and FL350 (about z = 10-11 km) from the ground-based radar data, this study employed a total of 22 predictors, which consists of physical variables computed directly from cumulus parameterization schemes (CPS) and dynamical mesoscale forcings responsible for DCAs, derived from two NWP models. The individual DCA diagnostics used in this study showed various forecasting skills against the ground-based radar ETH observation during June to September in 2022. The Area Under the Curve (AUC) values of the DCA diagnostics vary depending on the spatial resolution and CPS used in the underlying NWP models. Using these diagnostic ensemble members, the MMDE-based probabilistic forecast product are developed. For an objective evaluation, we examined various statistical metrics such as AUC, true skill statistics, Brier score and reliability diagram in Korean Airspace. Additionally, we also investigated some case studies with abnormal flight trajectories to avoid DCAs. The evaluation results showed that newly proposed MMDE-based probabilistic forecast outperforms both individual diagnostics and those based on the single NWP model. This could be eventually very useful as a decision support tool for DCA avoidance guidance in the air traffic management in Korea and East Asia.
Key Words: Deep convective area, Multi-model and -diagnostic ensemble, Probabilistic forecast
Acknowledgment: This research is supported by the Korea Meteorological Administration Research and Development Program under Grants KMI2022-00410.
- Indicates paper has been withdrawn from meeting
- Indicates an Award Winner