Gap-filling eddy covariance methane fluxes: Comparison of machine learning model predictions and uncertainties at FLUXNET-CH4 wetlands - Archive ouverte HAL Access content directly
Journal Articles Agricultural and Forest Meteorology Year : 2021

Gap-filling eddy covariance methane fluxes: Comparison of machine learning model predictions and uncertainties at FLUXNET-CH4 wetlands

, , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , (1, 2) , , , , , (1, 2) , , , , , , , , , , , , , , , , , , , , ,
1
2
Jeremy Irvin
  • Function : Author
Sharon Zhou
  • Function : Author
Gavin Mcnicol
  • Function : Author
Fred Lu
  • Function : Author
Vincent Liu
  • Function : Author
Etienne Fluet-Chouinard
  • Function : Author
Zutao Ouyang
  • Function : Author
Sara Helen Knox
  • Function : Author
Antje Lucas-Moffat
  • Function : Author
Carlo Trotta
  • Function : Author
Dario Papale
Domenico Vitale
  • Function : Author
Ivan Mammarella
Pavel Alekseychik
  • Function : Author
Mika Aurela
Anand Avati
  • Function : Author
Dennis Baldocchi
Sheel Bansal
  • Function : Author
Gil Bohrer
  • Function : Author
David Campbell
  • Function : Author
Jiquan Chen
Housen Chu
Higo Dalmagro
  • Function : Author
Kyle Delwiche
  • Function : Author
Ankur Desai
  • Function : Author
Eugenie Euskirchen
Sarah Feron
  • Function : Author
Mathias Goeckede
  • Function : Author
Martin Heimann
  • Function : Author
Manuel Helbig
  • Function : Author
Carole Helfter
  • Function : Author
Kyle Hemes
  • Function : Author
Takashi Hirano
  • Function : Author
Hiroki Iwata
  • Function : Author
Gerald Jurasinski
  • Function : Author
Aram Kalhori
  • Function : Author
Andrew Kondrich
  • Function : Author
Derrick Yf Lai
  • Function : Author
Annalea Lohila
Avni Malhotra
  • Function : Author
Lutz Merbold
Bhaskar Mitra
  • Function : Author
Andrew Ng
  • Function : Author
Mats Nilsson
Asko Noormets
  • Function : Author
Matthias Peichl
A. Camilo Rey-Sanchez
  • Function : Author
Andrew Richardson
  • Function : Author
Benjamin Rk Runkle
  • Function : Author
Karina Vr Schäfer
  • Function : Author
Oliver Sonnentag
  • Function : Author
Ellen Stuart-Haëntjens
  • Function : Author
Cove Sturtevant
  • Function : Author
Masahito Ueyama
  • Function : Author
Alex Valach
  • Function : Author
Rodrigo Vargas
  • Function : Author
George Vourlitis
  • Function : Author
Eric Ward
  • Function : Author
Guan Xhuan Wong
  • Function : Author
Donatella Zona
  • Function : Author
Ma. Carmelita R Alberto
  • Function : Author
David Billesbach
  • Function : Author
Gerardo Celis
  • Function : Author
Han Dolman
  • Function : Author
Thomas Friborg
  • Function : Author
Kathrin Fuchs
  • Function : Author
Mangaliso Gondwe
  • Function : Author
Jordan Goodrich
  • Function : Author
Pia Gottschalk
  • Function : Author
Lukas Hörtnagl
  • Function : Author
Franziska Koebsch
  • Function : Author
Kuno Kasak
  • Function : Author
Regine Maier
  • Function : Author
Timothy Morin
  • Function : Author
Eiko Nemitz
  • Function : Author
Walter Oechel
  • Function : Author
Patricia Oikawa
  • Function : Author
Keisuke Ono
  • Function : Author
Torsten Sachs
Ayaka Sakabe
  • Function : Author
Edward Schuur
  • Function : Author
Robert Shortt
  • Function : Author
Ryan Sullivan
Daphne Szutu
  • Function : Author
Eeva-Stiina Tuittila
  • Function : Author
Andrej Varlagin
Joeseph Verfaillie
  • Function : Author
Christian Wille
  • Function : Author
Lisamarie Windham-Myers
  • Function : Author
Benjamin Poulter
Robert Jackson

Abstract

Time series of wetland methane fluxes measured by eddy covariance require gap-filling to estimate daily, seasonal, and annual emissions. Gap-filling methane fluxes is challenging because of high variability and complex responses to multiple drivers. To date, there is no widely established gap-filling standard for wetland methane fluxes, with regards both to the best model algorithms and predictors. This study synthesizes results of different gap-filling methods systematically applied at 17 wetland sites spanning boreal to tropical regions and including all major wetland classes and two rice paddies. Procedures are proposed for: 1) creating realistic artificial gap scenarios, 2) training and evaluating gap-filling models without overstating performance, and 3) predicting half-hourly methane fluxes and annual emissions with realistic uncertainty estimates. Performance is compared between a conventional method (marginal distribution sampling) and four machine learning algorithms. The conventional method achieved similar median performance as the machine learning models but was worse than the best machine learning models and relatively insensitive to predictor choices. Of the machine learning models, decision tree algorithms performed the best in cross-validation experiments, even with a baseline predictor set, and artificial neural networks showed comparable performance when using all predictors. Soil temperature was frequently the most important predictor whilst water table depth was important at sites with substantial water table fluctuations, highlighting the value of data on wetland soil conditions. Raw gap-filling uncertainties from the machine learning models were underestimated and we propose a method to calibrate uncertainties to observations. The python code for model development, evaluation, and uncertainty estimation is publicly available. This study outlines a modular and robust machine learning workflow and makes recommendations for, and evaluates an improved baseline of, methane gap-filling models that can be implemented in multi-site syntheses or standardized products from regional and global flux networks (e.g., FLUXNET).

Dates and versions

insu-03286676 , version 1 (15-07-2021)

Identifiers

Cite

Jeremy Irvin, Sharon Zhou, Gavin Mcnicol, Fred Lu, Vincent Liu, et al.. Gap-filling eddy covariance methane fluxes: Comparison of machine learning model predictions and uncertainties at FLUXNET-CH4 wetlands. Agricultural and Forest Meteorology, 2021, 308-309, pp.108528. ⟨10.1016/j.agrformet.2021.108528⟩. ⟨insu-03286676⟩
77 View
0 Download

Altmetric

Share

Gmail Facebook Twitter LinkedIn More