Phytoplankton Size Classes in the Global Ocean at Different Bathymetric Depths

A three-component model is often used to invert the phytoplankton size class (PSC) concentration globally, especially in open oceans. Limited by the assumption of this three-component model, new efforts have been made to explore PSCs in different bathymetric regions. Mass global cruise datasets were gathered and classified into coastal, mixed, and open ocean datasets determined by the variation in bathymetric depth. A new three-component power model was established for coastal water samples (< 50 m), where the determination coefficients ( $R^{2}$ ) were 0.95, 0.37, and 0.44 for microphytoplankton (Micro), nanophytoplankton (Nano), and picophytoplankton (Pico), respectively. We also updated the coefficients of the three-component exponential model in the open ocean (>200 m) and found that the PSC verification results performed better in the ocean north of the 40° S latitude line ( $R^{2}= 0.83$ , 0.70, and 0.64, respectively). A smooth function for the samples in mixed ocean waters (50–200 m) was designed to obtain PSCs with different weights between the three-component power and exponential models with relatively low accuracy ( $R^{2}=0.84$ , 0.37, and 0.14, respectively), which is indicative of the complex conditions in these regions. The global PSC proportion distribution is consistent with the existing knowledge in this field. Finally, we analyzed the performance and limitations of different PSC models based on the entire dataset and found that the proposed model outperformed the others, and the assumption of the three-component exponential model may fail in the Southern Ocean.