Prediction of a Pilot’s Invisible Foe: The Severe Low-Level Wind Shear

Severe low-level wind shear (S-LLWS) in the vicinity of airport runways (25 knots or more) is a growing concern for safety civil aviation. By comprehending causes S-LLWS events, aviation can be enhanced. rare occurrence, but it hazardous approaching and departing aircraft. This study introduced self-paced ensemble (SPE) framework Shapley additive explanations (SHAP) interpretation system classification, prediction, LLWS severity. Doppler LiDAR- PIREPs-based data from Hong Kong International Airport were obtained, trained, evaluated to predict The SPE was also compared state-of-the-art tree-based models, including light gradient boosting machine, adaptive boosting, classification regression tree models. does not require prior treatment; however, SMOTE-ENN utilized treat highly imbalanced training In terms prediction performance, outperforms all Using SHAP analysis, interpreted. It determined that “runway 25LD”, “mean hourly temperature”, speed” most significant contributors occurrence S-LLWS. optimistic projections events at runway 25LD during periods low-to-moderate temperatures relatively medium-to-high speeds. Similarly, majority took place on runway. Without need augmentation preprocessing, coupled with could effectively an invaluable resource policymakers air traffic analysts.