Assessment of Moving Average (MA) Method for Rainfall Prediction in Yogyakarta, Indonesia

This study investigated a time series model using a moving average (MA) for predicting rainfall trend in seven areas (Panggang, Gedangan, Kedung Keris, Ngawen, Wanagama, Tepus, and Playen) located in the Gunung Kidul Province, Yogyakarta, Indonesia. A database with daily rainfall data covering the period of 2010–2019 obtained from the Central Statistical Body of Gunungkidul District (BPS), Indonesia, were analysed. In this study, the MA was developed using Microsoft Excel. Six performance indicators, including Root Mean Square Error (RMSE) and Index of Agreement (IA), were used to evaluate the goodness-of-fit of the time series model. The results specify that the MA is a reliable model, with the RMSE and IA values in the ranges of 10.7–18.3 and 0.80–0.85, respectively. Small error and high agreement rates between the observed and predicted values indicates that prediction using MA method has good potential to be used as one of the prediction tools for rainfall modelling.