Bacteria play a critical role in carbon cycling and nutrient remineralization. To reveal potential mechanisms controlling bacterial abundance in the upper 200 m of the South China Sea (SCS), the generalized linear model (GLM), generalized additive model (GAM) and generalized boosted model (GBM) were constructed to address the relationship between bacterial abundance and environmental factors, including geographical variables, biotic variables and water chemistry. GAM and GBM were found suitable for modeling bacterial abundance in the SCS. The predictive performance of GBM was superior to GLM and GAM for bacterial distribution. In addition, bacterial abundance predicted by GBM from environmental parameters was highly consistent with the observations, indicating that GBM was robust to predict bacterial abundance from environmental parameters. Furthermore, the key environmental factors modulating the horizontal and vertical distribution of bacteria were determined based on models. Horizontally, surface bacterial abundance decreased from onshore to offshore, which was primarily regulated by salinity and chlorophyll-a. Vertically, bacterial abundance decreased with depth. Chlorophyll-a was primarily responsible for vertical variability in bacterial abundance in the upper 100 m, where temperature was higher than the optimum temperature (21 °C) for bacterial growth. In contrast, temperature was a dominant factor regulating bacterial abundance below 100 m, where temperature was below 21 °C and positively correlated with BA. Viruses and nutrients played less important roles in regulating bacterial abundance than chlorophyll-a and temperature in the SCS. Our models elucidated environmental regulations on bacterial abundance, which was helpful for us to understand bacterial carbon cycling in the SCS.
Keywords: Bacterial abundance; Generalized additive model; Generalized boosted model; Generalized linear model; The South China Sea.
Copyright © 2021 Elsevier B.V. All rights reserved.