Kajian Metode Analisis Spektral pada Peramalan Curah Hujan

Abstract

Spectral analysis can be used to reveal periodic patterns contained in rainfall data. Data analysis using statistical tools and spectral analysis is applied to investigate and interpret the temporal course of the available rainfall data set. The results of spectral analysis using harmonic analysis and periodogram methods on monthly rainfall data for Serdang Bedagai Regency for 10 years show that rainfall in the area has a monsoonal pattern and is influenced by a seasonal cycle with a period of 12 months. The weather in Serdang Bedagai Regency is influenced by monsoon winds which cause the area to have annual oscillations. In the context of rainfall forecasting, the results of spectral analysis in the form of periods can be used as estimates of forecasting model parameters. Analysis of historical rainfall data carried out in spectral analysis can help in verifying whether the model used is accurate enough in capturing annual periodic patterns and anomalies that may arise.