Analysis of Upper-Level Aviation Turbulence Observational Data and Performance Evaluation of Prediction Models (KTG and GKTG) over the Korean Peninsula for 5 years (2019∼2023)

This study conducted a statistical analysis of turbulence events using in-situ eddy dissipation rate (EDR) observation data collected above 20,000 feet over the Korean Peninsula from 2019 to 2023. Additionally, the performances of two operational turbulence prediction models from the Korea Aviation Meteorological Office, which are the Korean Aviation Turbulence Guidance (KTG) and Global KTG (GKTG) systems, were evaluated against the same insitu EDR observations. The verification process employed the probability of detection (POD) method, allowing for an objective assessment of prediction accuracy. The statistical analysis results of in-situ EDR data revealed that data was primarily collected along flight routes, with observations increasing over time. The moderate-or-greater (MOG) turbulence was observed frequently during the daytime, and was most prevalent in spring. The comparison of the characteristics of PIREP and in-situ EDR observations showed similar seasonal trends. The evaluation results compared with these EDR data showed both KTG and GKTG demonstrated performance levels applicable for aviation industry professionals (0.809 and 0.805 respectively). Accuracy at the MOG threshold was evaluated by analyzing the correlation between the area under the curve (AUC) and true skill statistic (TSS) values. The KTG model showed relatively higher accuracy. Based on these results, model improvements were implemented, enhancing the TSS of GKTG by 0.22 through threshold adjustment. Considering that models currently used by the Korea Aviation Meteorological Office mainly focus on predicting clear air turbulence and mountain wave turbulence, future research should aim to improve prediction capabilities for convectively-induced and near-cloud turbulence.