@article{fdi:010080581,
  title = {{N}eural network approaches to reconstruct phytoplankton time-series in the global ocean  [+ {C}orrection 2022, art. 5669, 7 p.]},
  author = {{M}artinez, {E}lodie and {B}rini, {A}. and {G}orgues, {T}homas and {D}rumetz, {L}. and {R}oussillon, {J}. and {T}andeo, {P}. and {M}aeze, {G}. and {F}ablet, {R}.},
  editor = {},
  language = {{ENG}},
  abstract = {{P}hytoplankton plays a key role in the carbon cycle and supports the oceanic food web. {W}hile 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. {W}ith the aim of reconstructing this longer-term phytoplankton variability, a support vector regression ({SVR}) approach was recently considered to derive surface {C}hlorophyll-a concentration ({C}hl, a proxy of phytoplankton biomass) from physical oceanic model outputs and atmospheric reanalysis. {H}owever, those early efforts relied on one particular algorithm, putting aside the question of whether different algorithms may have specific behaviors. {H}ere, 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}). {T}he {MLP} outperforms the {SVR} to capture satellite {C}hl (correlation of 0.6 vs. 0.17 on a global scale, respectively) along with its seasonal and interannual variability, despite an underestimated amplitude. {A}mong deep learning algorithms, neural network such as {MLP} models appear to be promising tools to investigate phytoplankton long-term time-series.},
  keywords = {phytoplankton time-series reconstruction ; ocean color ; neural networks ; support vector regression ; multi-layer perceptron ; physical predictors ; {MONDE}},
  booktitle = {},
  journal = {{R}emote {S}ensing},
  volume = {12},
  numero = {24},
  pages = {art. 4156 [14 ]  [+ {C}orrection 2022, art. 5669, 7 p.]},
  year = {2020},
  DOI = {10.3390/rs12244156},
  URL = {https://www.documentation.ird.fr/hor/fdi:010080581},
}