<![CDATA[Accurate sea level forecasting is essential for coastal resilience, as rising sea levels threaten coastal residents, infrastructure, and ecosystems, heightening the hazards of floods and erosion; yet, conventional models struggle with non-linear trends. This study proposes a hybrid ARIMA-DWT (Autoregressive Integrated Moving Average with Discrete Wavelet Transform) model to enhance prediction accuracy by decomposing sea level data into linear (ARIMA) and non-linear (DWT) components. The hybrid model, utilizing monthly sea level data from Penang, Malaysia (1984–2018; N=408), diminished errors by 20% compared to the standalone ARIMA, attaining out-of-sample RMSE=81.76, MAE=68.40, and MAPE=2.45%. The ARIMA-DWT framework effectively captures long-term climate trends and short-term variations, providing a computationally efficient resource for coastal planners. This method demonstrates enhanced efficacy in situations characterized by significant seasonality, offering practical insights for climate adaptation. This research enhances the existing body of work on hybrid forecasting techniques by providing a scalable framework for areas vulnerable to sea level rise.]]>
Copyrights © 2026
