969 / 2024-09-19 22:39:28
Quantitative Impact of Nutrients on Phytoplankton Growth in the Western Pacific, Yellow Sea, and South China Sea: Insights from Incubation Experiments and Machine Learning
Nutrient,Phytoplankton,Machine learning
Session 25 - IGAC-SOLAS: Chemistry and physics at surface ocean and lower atmosphere
Abstract Accepted
Nutrients are pivotal factors influencing the proliferation of phytoplankton communities, yet the underlying mechanisms involve intricate biological and chemical processes, which pose challenges for the accurate quantitative prediction of phytoplankton growth in response to nutrient perturbations across various marine regions. This study leverages on-board bioassay incubation data collected from 2014 to 2017 to investigate the quantitative impact of nutrients on the growth of phytoplankton in the Western Pacific, Yellow Sea and South China Sea. We employed Deep Learning Neural Networks (DLNN) and Random Forest (RF) models to predict the changes in the integrated chlorophyll a (Chl a) concentration under various nutrient limitation following nutrient perturbations, including nitrogen limitation, phosphorus limitation, co-limitation by nitrogen and phosphorus, serial limitation by nitrogen and phosphorus, iron limitation, and co-limitation by iron and phosphorus. These models demonstrated robust predictive capabilities, with the DLNN model exhibiting the highest predictive accuracy. These findings are crucial for quantitatively predicting the response of phytoplankton concentration and marine primary productivity to nutrient inputs.