Simulation and Analysis of Sea-Level Change from Tide Gauge Station by Using Artificial Neural Network Models

Abstract

Sea level change is one of the most certain results of global warming. Sea level change would increase erosion in coastal areas, result in intrusion into water supplies, inundate coastal marshes and other important habitats, and make the coastal property more vulnerable to erosion and flooding. This situation coincides with the massive socio-economic development of the coastal city areas. The coastal areas of the East Coast of Peninsular Malaysia are vulnerable to sea-level change, flooding, and extreme erosion events. The monthly Mean Sea Level (MSL) change was simulated by using two Artificial Neural Network (ANN) models, Feed Forward- Neural Network (FF-NN) and Nonlinear Autoregressive Exogenous- Neural Network (NARX-NN) models. Both models did well in recreating sea levels and their fluctuating patterns, according to the data. The NARX-NN model with architecture (5-6-1) and four lag options, on the other hand, got the greatest results. The findings of the model's mean sea level rise simulation show that Kuala Terengganu would have a growing and upward trend of roughly 25.34 mm/year. This paper shows that the eastern coast of Malaysia is highly vulnerable to sea-level rise and therefore, requires sustainable adaptation policies and plans to manage the potential impacts. It recommends that various policies, which enable areas to be occupied for longer before the eventual retreat, could be adapted to accommodate vulnerable settlements on the eastern coast of Malaysia.