Atmospheric Stability Index Threshold for Very Heavy Rainfall Events in Indonesia

Abstract

Indonesia's maritime tropical climate exhibits high atmospheric variability often triggering extreme rainfall that disrupts flight and increases hydrometeorological disaster risks. This study investigates threshold values of atmospheric stability indices —Lifted Index (LI), K-Index (KI), and Precipitable Water (PW)— as predictors of very heavy rainfall events (>100 mm/day). Daily rainfall data from six meteorological stations (Medan, Padang, Tangerang, Surabaya, Merauke, Kupang) across Indonesia for 1993-2024 was analysed for very high rainfall events. We used the receiver operating characteristic area under the curve (ROC AUC) method to quantify the indices’ ability to characterize and distinguish very heavy and non-very heavy rainfall, objectively identifying the optimal threshold. The results show spatially varying thresholds reflecting dominant local atmospheric factors, with KI dominant in Kupang, while LI and PW in Merauke. This suggests that the optimal threshold value is a region specific, emphasizing a local calibration. further, the findings are crucial for enhancing prediction accuracy and extreme weather warnings at airports.