Alexcandro Hibertus Sianipar, khoirul Fajri

Penerapan Metode SARIMA untuk Peramalan Frekuensi Curah Hujan Bulanan di Sumatera Utara

Introduction

Penerapan metode sarima untuk peramalan frekuensi curah hujan bulanan di sumatera utara. Peramalan frekuensi curah hujan bulanan di Sumatera Utara menggunakan SARIMA untuk mitigasi bencana. Model $SARIMA(0,1,1)(0,1,1)_{12}$ memprediksi peningkatan curah hujan Okt-Des 2026, mendukung peringatan dini banjir.

Abstract

Variabilitas iklim ekstrem di Provinsi Sumatera Utara menuntut strategi mitigasi bencana yang presisi, mengingat tingginya risiko bahaya hidrometeorologi di wilayah tersebut. Penelitian ini bertujuan untuk membangun model peramalan curah hujan yang tangguh (robust) dengan mengintegrasikan data deret waktu (2010–2025) dari Copernicus Data Space Ecosystem menggunakan pendekatan Seasonal Autoregressive Integrated Moving Average (SARIMA). Melalui prosedur iteratif Box-Jenkins dan evaluasi kriteria informasi (AIC), model $SARIMA(0,1,1)(0,1,1)_{12}$ teridentifikasi sebagai estimator terbaik yang mampu menangkap pola musiman dan tren stokastik secara akurat, dengan validasi Root Mean Square Error (RMSE) sebesar 82,11 mm. Proyeksi untuk tahun 2026 mengindikasikan eskalasi curah hujan yang signifikan selama periode Oktober hingga Desember, yang menandai fase kritis bagi potensi banjir. Temuan ini memberikan dasar ilmiah kuantitatif bagi pembuat kebijakan untuk memperkuat sistem peringatan dini dan mengoptimalkan manajemen risiko bencana nasional.

Review

This paper, "Penerapan Metode SARIMA untuk Peramalan Frekuensi Curah Hujan Bulanan di Sumatera Utara," addresses a highly pertinent issue concerning climate variability and hydrometeorological disaster mitigation in North Sumatra. The authors set out to develop a robust rainfall forecasting model, a critical endeavor given the region's vulnerabilities. The study employs the widely recognized Seasonal Autoregressive Integrated Moving Average (SARIMA) methodology, utilizing a comprehensive time-series dataset spanning 2010 to 2025 from the Copernicus Data Space Ecosystem. The application of the Box-Jenkins procedure, coupled with the Akaike Information Criterion (AIC) for model selection, indicates a standard and rigorous approach to time series analysis. Through meticulous iterative analysis, the study successfully identified the $SARIMA(0,1,1)(0,1,1)_{12}$ model as the optimal estimator. This model is reported to accurately capture both seasonal patterns and stochastic trends inherent in the monthly rainfall data, which is crucial for effective hydrological forecasting. The model's robustness was validated with a Root Mean Square Error (RMSE) of 82.11 mm. While an RMSE of this magnitude provides a quantitative measure of error, its practical significance would benefit from further context, such as comparison to historical rainfall averages or other benchmark models, to fully ascertain its 'robustness' in real-world application. The core finding—a projected significant escalation of rainfall in North Sumatra from October to December 2026—highlights a critical period for potential flood risks. The quantitative scientific basis provided by this research is a significant contribution to regional climate adaptation and disaster risk reduction efforts. By pinpointing specific periods of increased flood potential, the findings offer invaluable intelligence for policymakers and disaster management agencies in North Sumatra. This information can directly inform the strengthening of early warning systems and the optimization of national disaster risk management strategies, thereby enhancing community resilience against hydrometeorological hazards. Overall, this study presents a commendable effort to translate complex climatic data into actionable insights, making a clear case for its practical relevance and potential impact on public safety and resource planning.

Full Text

Comments

You need to be logged in to post a comment.