Inversion of Phytoplankton Pigment Vertical Profiles from Satellite Data Using Machine Learning

Type Article
Date 2021-04
Language English
Author(s) Puissant AgatheORCID1, El Hourany RoyORCID2, Charantonis Anastase Alexandre1, 3, Bowler Chris2, Thiria Sylvie1, 4
Affiliation(s) 1 : Laboratoire d’Océanographie et du Climat Expérimentations et Approches Numériques (LOCEAN), Sorbonne Université, CNRS, IRD, MNHN, 75005 Paris, France
2 : Institut de Biologie de l’École Normale Supérieure (IBENS), École Normale Supérieure, CNRS, INSERM, PSL Université, 75005 Paris, France
3 : École Nationale Supérieure d’Informatique pour l’Industrie et l’Entreprise (ENSIIE), 91000 Évry, France
4 : Observatoire de Versailles Saint-Quentin-en-Yvelins (OVSQ), Versailles Saint-Quentin-en-Yvelines University, 78280 Guyancourt, France
Source Remote Sensing (2072-4292) (MDPI AG), 2021-04 , Vol. 13 , N. 8 , P. 1445 (19p.)
DOI 10.3390/rs13081445
Note This article belongs to the Special Issue Application of Multi-Sensor Fusion Technology in Target Detection and Recognition
Keyword(s) machine learning, inversion, ocean colour, phytoplankton, pigment vertical profile, deep chlorophyll maximum, Tara Oceans, MAREDAT, pigments, ITCOMP-SOM, Self Organizing Maps
Abstract Observing the vertical dynamic of phytoplankton in the water column is essential to understand the evolution of the ocean primary productivity under climate change and the efficiency of the CO2 biological pump. This is usually made through in-situ measurements. In this paper, we propose a machine learning methodology to infer the vertical distribution of phytoplankton pigments from surface satellite observations, allowing their global estimation with a high spatial and temporal resolution. After imputing missing values through iterative completion Self-Organizing Maps, smoothing and reducing the vertical distributions through principal component analysis, we used a Self-Organizing Map to cluster the reduced profiles with satellite observations. These referent vector clusters were then used to invert the vertical profiles of phytoplankton pigments. The methodology was trained and validated on the MAREDAT dataset and tested on the Tara Oceans dataset. The different regression coefficients R2 between observed and estimated vertical profiles of pigment concentration are, on average, greater than 0.7. We could expect to monitor the vertical distribution of phytoplankton types in the global ocean.
Full Text
File Pages Size Access
Publisher's official version 19 3 MB Open access
Top of the page