Forecasting sea surface temperature anomalies using the SARIMA ARCH/GARCH model

Abstract

The areas that are close to the Indian Ocean can give a significant impact on climate change that caused by SSTA. Based on the records from NOAA website, there are data about points in the Indian Ocean. Moreover, climate change on earth is influenced by several parameters, one of them is called SSTA. Therefore, it is necessary to forecast SSTA in the future to minimize the impact of climate change. With a purpose to predict data that contains seasonal patterns, the method that can be used is SARIMA method with ARCH or GARCH. This study aims to determine the results of SSTA forecasting for the period from August to December 2018. This study uses the 4°N90°E point, which is a point located close to Aceh Province. The data used are daily data obtained from the NOAA website, from July 2010 to July 2018. Based on the research results, the best model for SSTA forecasting is the SARIMA model (2, 1, 1)(0, 1, 1)-GARCH(1, 1) with forecasting accuracy values of 3.67% MAPE, 0.032 MAE, and 0.016 RMSE. The MAPE values that are smaller than 10% indicate that the SARIMA (2, 1, 1)(0, 1, 1)-GARCH(1, 1) model is very good for forecasting SSTA in the future.