Peramalan Curah Hujan Sebagai Upaya Mitigasi Bencana Menggunakan Seasonal Autoregressive Integrated Moving Average

Fayyadh Ghaly
Amelia Susrifalah
Yenni Kurniawati

Jurnal Matematika dan Statistika serta Aplikasinya (Jurnal MSA)
Vol 13 No 1 (2025): VOLUME 13 NO 1 TAHUN 2025

Rainfall prediction is important in disaster mitigation to reduce impacts such as drought, flood, and landslide. Rainfall data that has a seasonal pattern requires an appropriate forecasting method, one of which is SARIMA. This study predicts rainfall at the Deli Serdang Climatology Station, North Sumatra, based on monthly observation data for 2018–2023, showing a seasonal pattern with a 12-month cycle. The best model obtained is SARIMA (0,0,1) (0,0,1)12 with a MAPE of 19.5%, indicating a prediction accuracy of 80.5%. The forecasting results indicate a decrease in rainfall in the first semester of 2024, which is in the medium rainfall category. These findings can support disaster risk mitigation strategies and natural resource management planning related to climate change. The SARIMA model also has the potential to be applied in further climatology studies.