Analysis of Lightning Characteristics and Hazard Prediction in Karimun Regency Using Maximum Entropy (MaxEnt) Modeling

This study analyzes the characteristics and predicts the lightning hazard zones in Karimun Regency, Indonesia, using the Maximum Entropy (MaxEnt) model. Lightning strike data from 2022 to 2024, consisting of 82,795 events, were analyzed along with environmental variables, including elevation, land cover, and rainfall. The MaxEnt model demonstrated good predictive performance with an Area Under the Curve (AUC) value of 0.654. Internal model analysis revealed that rainfall was the most influential variable in model formation, contributing 79.1%, followed by land cover (12.1%) and elevation (8.8%). Spatially, the distribution analysis showed that approximately 60% of lightning strikes occurred in lowland areas (0-22 meters above sea level), with shrub/vegetation areas experiencing the highest frequency. Interestingly, the statistical correlation between total monthly rainfall and lightning occurrences was consistently weak at all observation stations (R² ≈ 0). Diurnally, lightning activity peaked in the afternoon (13:00-18:59). The final vulnerability map successfully identified high-risk areas concentrated in the subdistricts of Kundur, Kundur Utara, Kundur Barat, Buru, and Karimun. These findings provide valuable insights for lightning risk mitigation strategies, emphasizing that the complex interaction of various environmental factors drives lightning occurrence patterns in the tropical island region.