Sea level variability and modeling in the Gulf of Guinea using supervised machine learning

Area of Science:

• Oceanography
• Climate Science
• Machine Learning

Background:

• Rising sea levels pose a significant threat to low-lying coastal regions like the Gulf of Guinea (GoG).
• Understanding historical sea level variability and influencing factors is crucial for effective coastal planning and mitigation.

Purpose of the Study:

• To analyze mean sea level anomaly (MSLA) trends in the GoG from 1993-2020.
• To investigate the links between sea level variability and oceanic/atmospheric forcings.
• To evaluate machine learning models for optimizing sea level projections.

Main Methods:

• Analysis of MSLA trends across three distinct periods (1993-2002, 2003-2012, 2013-2020).
• Investigation of interannual sea level variability and its correlation with large-scale oceanic and atmospheric phenomena.
• Performance evaluation of supervised machine learning techniques (Random Forest Regression, Gradient Boosting Machines) for sea level modeling.

Main Results:

• A consistent rise in MSLA linear trends was observed across the GoG, with a total trend of 88 mm (1993-2020).
• The highest decadal trend (38.7 mm) occurred during 2013-2020, with a 100% increment in 2003-2012.
• Random Forest Regression and Gradient Boosting Machines achieved 97% and 96% accuracy, respectively, in reproducing interannual sea level patterns.

Conclusions:

• Sea level rise in the GoG is accelerating, influenced by regional physical forcings and large-scale climate phenomena.
• Advanced machine learning models offer high accuracy for predicting regional sea level changes.
• These findings are vital for developing robust coastal management and climate adaptation strategies in the GoG.