Forecasting of Rainfall Using Seasonal Autoregreressive Integrated Moving Average (SARIMA) Aceh, Indonesia

Abstract

Climate change which has become increasingly erratic in recent decades has become a problem of global warming. So that it has an impact and influence in changing rainfall patterns. A very volatile climate overall can threaten the success of food production. Information about rainfall patterns is very important to agriculture that relies on rainfall as the main source of irrigation. The purpose of this study is to predict rainfall from all time series based on rainfall data for 15 years, 10 years and 5 years. Prediction results were evaluated using the Nash-Sutcliffe Efficiency (NSE) statistical method, RMSE-Observation Standard Deviation Ratio (RSR) and PBIAS. This research was conducted in Aceh Besar District. Indonesia which coincided with Indrapuri District. Analysis of the data used in this study uses the Seasonal Autoregressive Integrated Moving Average (SARIMA) models. The best prediction results are generated from the use of rainfall time series data onto 5 years for 2013-2017 with the evaluation value of the model obtained is in the “Very Good " category. Prediction models for the best rainfall predictions are (0.0.0) and (0.1.2)12 with the respective values of NSE of 0.84, RSR 0.41 and PBIAS - 2.8. So as a whole the closest prediction results in the actual values are obtained from time series rainfall data onto the past five years.