Threshold Value of Atmospheric Stability Indices During Thunderstorm Events at Minangkabau International Airport

Thunderstorms are a significant challenge for aviation operations, especially in tropical regions such as West Sumatra. This study aims to determine threshold values for six atmospheric stability indices—Convective Available Potential Energy (CAPE), K-Index (KI), Lifted Index (LI), Showalter Index (SI), Severe Weather Threat Index (SWEAT), and Total-totals Index (TTI)—to predict thunderstorm events at Minangkabau International Airport (MIA). Radiosonde and daily synoptic reports from 2018–2022 were analyzed using Rawinsonde Observation Programs (RAOB) and Statistical Package for the Social Sciences (SPSS) with a dummy regression approach. The model was validated using a confusion matrix, measuring accuracy, precision, and recall. Results show that the use of locally calibrated thresholds leads to higher and more consistent accuracy, precision, and recall values compared to global benchmarks, due to better adaptation to local weather parameters such as vertical humidity, mid-layer temperature, and wind structure. KI, SI, and TTI showed high sensitivity (recall >88%), while LI and CAPE performed moderately. Monthly variation in index performance was observed, with KI, SI, and TTI dominant in the wet and transition seasons, and SWEAT effective in the dry season when shear-driven convection increases. Thus, locally calibrated indices are recommended for thunderstorm early warning systems in aviation.