Sea Surface Temperature Forecasting in the South China Sea: A Comparative Study of ARIMA and Expert Models Using MODIS Satellite Data

Abstract

Accurate surveillance of sea surface temperature (SST) is necessary for interpreting oceanographic trends, the dynamics of marine ecosystems, and climate variability in this region. This study applies an integrated approach combining satellite imagery and statistical forecasting to analyze historical SST trends and provide robust forecasts for 2025-2030. Monthly SSTs for 2010-2024 were derived from Moderate Resolution Imaging Spectroradiometer (MODIS) imagery, then extracted and aggregated into a time-series dataset. ARIMA with different parameter configurations ((1,1,1), (2,1,1), and (2,1,2)) and Expert Modeler were used to forecast SST for the selected region in the South China Sea. The results indicate steady growth in SST, with an increment range from 0.260°C to 0.390°C. The Expert Modeler provides a better forecast, with a Stationary R-Squared value of 0.969 across all percentiles. Such results provide valuable information on regional climate and are useful for marine resource management, fishing, and environmental practices.