Neural Network Approaches to Reconstruct Phytoplankton Time-Series in the Global Ocean

Type Article
Date 2020-12
Language English
Author(s) Martinez ElodieORCID3, Brini Anouar3, Gorgues Thomas3, Drumetz LucasORCID2, Roussillon Joana3, Tandeo Pierre2, Maze GuillaumeORCID1, Fablet Ronan2
Affiliation(s) 1 : Laboratoire d’Océanographie Physique et Spatiale (LOPS), IUEM, University Brest-CNRS-IRD-Ifremer, 29200 Brest, France
2 : IMT Atlantique, Lab-STICC, UMR CNRS 6285, 29200 Brest, France
3 : Laboratoire d’Océanographie Physique et Spatiale (LOPS), IUEM, University Brest-CNRS-IRD-Ifremer, 29200 Brest, France
Source Remote Sensing (2072-4292) (MDPI AG), 2020-12 , Vol. 12 , N. 24 , P. 4156 (14p.)
DOI 10.3390/rs12244156
WOS© Times Cited 1
Note This article belongs to the Special Issue The Ocean Colour Essential Climate Variable: Advances, Applications and Aspirations
Keyword(s) phytoplankton time-series reconstruction, ocean color, neural networks, support vector regression, multi-layer perceptron, physical predictors
Abstract

Phytoplankton plays a key role in the carbon cycle and supports the oceanic food web. While its seasonal and interannual cycles are rather well characterized owing to the modern satellite ocean color era, its longer time variability remains largely unknown due to the short time-period covered by observations on a global scale. With the aim of reconstructing this longer-term phytoplankton variability, a support vector regression (SVR) approach was recently considered to derive surface Chlorophyll-a concentration (Chl, a proxy of phytoplankton biomass) from physical oceanic model outputs and atmospheric reanalysis. However, those early efforts relied on one particular algorithm, putting aside the question of whether different algorithms may have specific behaviors. Here, we show that this approach can also be applied on satellite observations and can even be further improved by testing performances of different machine learning algorithms, the SVR and a neural network with dense layers (a multi-layer perceptron, MLP). The MLP outperforms the SVR to capture satellite Chl (correlation of 0.6 vs. 0.17 on a global scale, respectively) along with its seasonal and interannual variability, despite an underestimated amplitude. Among deep learning algorithms, neural network such as MLP models appear to be promising tools to investigate phytoplankton long-term time-series.

Full Text
File Pages Size Access
Publisher's official version 14 3 MB Open access
Supplementary File 1 3 321 KB Open access
Top of the page

How to cite 

Martinez Elodie, Brini Anouar, Gorgues Thomas, Drumetz Lucas, Roussillon Joana, Tandeo Pierre, Maze Guillaume, Fablet Ronan (2020). Neural Network Approaches to Reconstruct Phytoplankton Time-Series in the Global Ocean. Remote Sensing, 12(24), 4156 (14p.). Publisher's official version : https://doi.org/10.3390/rs12244156 , Open Access version : https://archimer.ifremer.fr/doc/00667/77871/