Predicting Sea Level Rise Using Artificial Intelligence: A Review

• Published in: Archives of computational methods in engineering 2023-09, Vol.30 (7), p.4045-4062
• Main Authors: Bahari, Nur Amira Afiza Bt Saiful, Ahmed, Ali Najah, Chong, Kai Lun, Lai, Vivien, Huang, Yuk Feng, Koo, Chai Hoon, Ng, Jing Lin, El-Shafie, Ahmed
• Format: Article
• Language: English
• Subjects: Algorithms; Artificial intelligence; Coastal structures; Deep learning; Engineering; Error reduction; Feature extraction; Forecasting; Hydrology; Lead time; Machine learning; Mathematical and Computational Engineering; Mathematical models; Ocean currents; Optimization; Review Article; Satellite altimetry; Sea level; Sea surface temperature; Wind direction
• ISSN: 1134-3060; 1886-1784
• DOI: 10.1007/s11831-023-09934-9

Forecasting sea level is critical for coastal structure building and port operations. There are, however, challenges in making these predictions, resulting from the complicated processes at various periods. This study discussed the continual development of the application and forecasting approaches for sea level rise, in standard and advanced modeling versions. To date, the tide gauge and satellite altimetry are the commonly used approaches for sea level measurement. Tide gauges are mostly deficient in typical offshore circumstances; but however, this may be compensated for with satellite altimetry, a complementing technique. With technological improvement, sea level measurement may be forecasted using a variety of computer science approaches known as artificial intelligence, including machine learning and deep learning; capable of extracting information and formulating relationships from the given dataset. Its potential and extensive advantages led to a sharp growth in its recognition among hydrologists. The most successful techniques for enhancing these approaches include hybridization, ensemble modeling, data decomposition, and algorithm optimization. These advanced techniques are a prominent study area and a viable strategy for determining intelligent forecasts of sea level rise with sufficient lead time. For improved performance, the modeling requires incorporating numerous input parameters, such as precipitation, wind direction, ocean current, and sea surface temperature; for better representing the process, thus reducing forecast error and uncertainty. Deep learning is more effective and enhances existing machine learning models for forecasting future sea level rise due to its automatic feature extraction and memory-storing capability.