<![CDATA[One of the climate components that is always changing is rainfall, which affects agriculture, water resource management, and hydrometeorological disaster mitigation. The rainfall characteristics of Langkat Regency are variable and influenced by seasonal patterns, thus requiring a prediction technique that can systematically balance the elements of trend and season. The objective of this research is to identify seasonal patterns and create a model to predict rainfall in Langkat Regency using the Seasonal Autoregressive Integrated Moving Average (SARIMA) method. The data used comes from historical monthly rainfall data from the Meteorology, Climatology, and Geophysics Agency (BMKG). Descriptive analysis, data stationarity examination, transformation and differencing, model parameter identification through autocorrelation function and partial autocorrelation function, parameter estimation, and residual diagnostic testing are all parts of the research. The Akaike Information Criterion and Bayesian Information Criterion values are used to select the best model. This is also based on the prediction error rates calculated using Mean Absolute Error, Root Mean Square Error, and Mean Absolute Percentage Error. Next, the best SARIMA model is used to predict the upcoming rainfall. It is expected that this research will produce an accurate prediction model to describe the seasonal rainfall patterns of Langkat Regency. The research results can be used to assist in agricultural planning, water resource management, and anticipating floods and droughts in Langkat Regency.]]>
Copyrights © 2026
