Multi-variate hybrid modeling for pacific ocean acidification: predicting future pH trends and analyzing key biogeochemical drivers

Abstract

Ocean acidification, driven by rising atmospheric carbon dioxide levels, poses a significant threat to the health of marine ecosystems, particularly in the Pacific Ocean. This study employs a multi-variate hybrid machine learning approach to predict future pH trends within the Pacific and to analyze the influence of key biogeochemical drivers on these trends. Hybrid models, strategically combining the strengths of individual algorithms, were developed for predicting several ocean acidification parameters. A performance analysis demonstrated the superior accuracy of hybrid models compared to their counterparts. The predicted pH trends reveal a concerning shift towards increased acidity within the Pacific Ocean, highlighting the urgency of understanding and mitigating its impacts. In-depth analysis was conducted to identify the relative influence of key biogeochemical factors on the changing pH dynamics. This research aims to provide crucial insights for developing targeted mitigation strategies and protecting the vulnerable ecosystems of the Pacific Ocean from the escalating consequences of ocean acidification.