New progress has been made in global phytoplankton pigment concentration inversion

The reporter learned from the Institute of Oceanography of the Chinese Academy of Sciences on May 29 that the research group of Li Xiaofeng, a researcher at the institute, has made important progress in inverting the global phytoplankton pigment concentration based on deep learning algorithms, and the research results were published in the top journal in the field of remote sensing “Environmental Remote Sensing” (IF=13.85). 

Phytoplankton is not only the primary producer of the ocean, but also a vital carrier in the process of marine biogeochemical cycling, and its community structure is related to the change of marine ecological environment, which is an important indicator factor for understanding the ecological evolution driven by marine dynamic processes. The pigment concentration of phytoplankton is an important basis for their classification and analysis of their community structure. The optical absorption information related to phytoplankton pigment concentration can be obtained by using marine optical remote sensing, but due to the variable optical characteristics of seawater and the “packing effect” in the optical absorption process of phytoplankton, it is difficult to invert the concentration of phytoplankton at the same time on a global scale.

Global distribution of the ratio of phytoplankton pigment concentration to total chlorophyll concentration Photo courtesy of the research group

It is understood that based on the long-term collection of on-site HPLC data and MODIS satellite remote sensing data, the study constructs a global phytoplankton pigment concentration matching dataset, which for the first time realizes the inversion of the concentration of 17 phytoplankton pigments in the global ocean, and obtains the distribution of different phytoplankton taxa in the global ocean. 

In the process of constructing the deep learning model, the research team fully considered the influence of other substances in seawater on the inversion of phytoplankton pigment concentration, and used the residual network and multi-scale pyramid structure to realize the acquisition of complex nonlinear relationships and multi-scale feature learning when multiple pigment concentrations were inverted at the same time. The global phytoplankton pigment concentration inversion model can be used to study the change process of long-term marine phytoplankton taxa, and reveal the influence of large-scale marine dynamic processes on the structure of marine phytoplankton communities. 

The results show that the deep learning algorithm can effectively invert the phytoplankton pigment concentration at large spatiotemporal scales, so as to analyze the phytoplankton community dynamics in the global ocean. During the 2015/2016 El Niño event, prokaryotic predominant seas extended from 180°E to 150°W to the east. From 2003 to 2021, prokaryotic abundance was positively correlated with El Niño intensity, but negatively correlated with the abundance of phytoplankton as a whole. 

The first author of the paper is Li Xiaolong, a senior engineer of the institute, and the co-authors include master’s student Yang Yi and Nagoya University professor Joji Ishizaka, and the corresponding author is Li Xiaofeng. (Source: China Science News, Liao Yang, Wang Min)

Related paper information:

Source link

Related Articles

Leave a Reply

Your email address will not be published. Required fields are marked *

Back to top button