skip to main content
DOE PAGES title logo U.S. Department of Energy
Office of Scientific and Technical Information

Title: Predicting carbon dioxide and energy fluxes across global FLUXNET sites with regression algorithms

Abstract

Spatio-temporal fields of land–atmosphere fluxes derived from data-driven models can complement simulations by process-based land surface models. While a number of strategies for empirical models with eddy-covariance flux data have been applied, a systematic intercomparison of these methods has been missing so far. In this study, we performed a cross-validation experiment for predicting carbon dioxide, latent heat, sensible heat and net radiation fluxes across different ecosystem types with 11 machine learning (ML) methods from four different classes (kernel methods, neural networks, tree methods, and regression splines). We applied two complementary setups: (1) 8-day average fluxes based on remotely sensed data and (2) daily mean fluxes based on meteorological data and a mean seasonal cycle of remotely sensed variables. The patterns of predictions from different ML and experimental setups were highly consistent. There were systematic differences in performance among the fluxes, with the following ascending order: net ecosystem exchange (R2 < 0.5), ecosystem respiration (R2 > 0.6), gross primary production (R2> 0.7), latent heat (R2 > 0.7), sensible heat (R2 > 0.7), and net radiation (R2 > 0.8). The ML methods predicted the across-site variability and the mean seasonal cycle of the observed fluxes very well (R2 > 0.7), while the 8-day deviationsmore » from the mean seasonal cycle were not well predicted (R2 < 0.5). Fluxes were better predicted at forested and temperate climate sites than at sites in extreme climates or less represented by training data (e.g., the tropics). Finally, the evaluated large ensemble of ML-based models will be the basis of new global flux products.« less

Authors:
 [1];  [2];  [3];  [4];  [5];  [6];  [2];  [7];  [8];  [9];  [10];  [11];  [12];  [13];  [1]
  1. Univ. of Tuscia, Viterbo (Italy)
  2. Max Planck Institute for Biogeochemistry, Jena (Germany)
  3. Woods Hole Research Center, Falmouth, MA (United States)
  4. Japan Agency for Marine-Earth Science and Technology, Yokohama (Japan); National Institute for Environmental Studies, Tsukuba (Japan)
  5. Univ. de Valencia, Paterna (Spain)
  6. Univ. of Tuscia, Viterbo (Italy); Sapientia Hungarian Univ. of Transylvania, Miercurea Ciuc (Romania)
  7. McMaster Univ., Hamilton, ON (Canada)
  8. European Commission, Ispra (Italy)
  9. Univ. College, Cork (Ireland)
  10. ETH Zurich, Zurich (Switzerland); International Livestock Research Institute (ILRI), Nairobi (Kenya)
  11. Univ. of Granada, Granada (Spain)
  12. Friedrich Schiller Univ. Jena, Jena (Germany)
  13. ETH Zurich, Zurich (Switzerland)
Publication Date:
Research Org.:
Oregon State Univ., Corvallis, OR (United States)
Sponsoring Org.:
USDOE
OSTI Identifier:
1361528
Grant/Contract Number:  
FG02-04ER63917; FG02-04ER63911
Resource Type:
Accepted Manuscript
Journal Name:
Biogeosciences (Online)
Additional Journal Information:
Journal Name: Biogeosciences (Online); Journal Volume: 13; Journal Issue: 14; Journal ID: ISSN 1726-4189
Publisher:
European Geosciences Union
Country of Publication:
United States
Language:
English
Subject:
59 BASIC BIOLOGICAL SCIENCES; 54 ENVIRONMENTAL SCIENCES

Citation Formats

Tramontana, Gianluca, Jung, Martin, Schwalm, Christopher R., Ichii, Kazuhito, Camps-Valls, Gustau, Ráduly, Botond, Reichstein, Markus, Arain, M. Altaf, Cescatti, Alessandro, Kiely, Gerard, Merbold, Lutz, Serrano-Ortiz, Penelope, Sickert, Sven, Wolf, Sebastian, and Papale, Dario. Predicting carbon dioxide and energy fluxes across global FLUXNET sites with regression algorithms. United States: N. p., 2016. Web. doi:10.5194/bg-13-4291-2016.
Tramontana, Gianluca, Jung, Martin, Schwalm, Christopher R., Ichii, Kazuhito, Camps-Valls, Gustau, Ráduly, Botond, Reichstein, Markus, Arain, M. Altaf, Cescatti, Alessandro, Kiely, Gerard, Merbold, Lutz, Serrano-Ortiz, Penelope, Sickert, Sven, Wolf, Sebastian, & Papale, Dario. Predicting carbon dioxide and energy fluxes across global FLUXNET sites with regression algorithms. United States. doi:10.5194/bg-13-4291-2016.
Tramontana, Gianluca, Jung, Martin, Schwalm, Christopher R., Ichii, Kazuhito, Camps-Valls, Gustau, Ráduly, Botond, Reichstein, Markus, Arain, M. Altaf, Cescatti, Alessandro, Kiely, Gerard, Merbold, Lutz, Serrano-Ortiz, Penelope, Sickert, Sven, Wolf, Sebastian, and Papale, Dario. Fri . "Predicting carbon dioxide and energy fluxes across global FLUXNET sites with regression algorithms". United States. doi:10.5194/bg-13-4291-2016. https://www.osti.gov/servlets/purl/1361528.
@article{osti_1361528,
title = {Predicting carbon dioxide and energy fluxes across global FLUXNET sites with regression algorithms},
author = {Tramontana, Gianluca and Jung, Martin and Schwalm, Christopher R. and Ichii, Kazuhito and Camps-Valls, Gustau and Ráduly, Botond and Reichstein, Markus and Arain, M. Altaf and Cescatti, Alessandro and Kiely, Gerard and Merbold, Lutz and Serrano-Ortiz, Penelope and Sickert, Sven and Wolf, Sebastian and Papale, Dario},
abstractNote = {Spatio-temporal fields of land–atmosphere fluxes derived from data-driven models can complement simulations by process-based land surface models. While a number of strategies for empirical models with eddy-covariance flux data have been applied, a systematic intercomparison of these methods has been missing so far. In this study, we performed a cross-validation experiment for predicting carbon dioxide, latent heat, sensible heat and net radiation fluxes across different ecosystem types with 11 machine learning (ML) methods from four different classes (kernel methods, neural networks, tree methods, and regression splines). We applied two complementary setups: (1) 8-day average fluxes based on remotely sensed data and (2) daily mean fluxes based on meteorological data and a mean seasonal cycle of remotely sensed variables. The patterns of predictions from different ML and experimental setups were highly consistent. There were systematic differences in performance among the fluxes, with the following ascending order: net ecosystem exchange (R2 < 0.5), ecosystem respiration (R2 > 0.6), gross primary production (R2> 0.7), latent heat (R2 > 0.7), sensible heat (R2 > 0.7), and net radiation (R2 > 0.8). The ML methods predicted the across-site variability and the mean seasonal cycle of the observed fluxes very well (R2 > 0.7), while the 8-day deviations from the mean seasonal cycle were not well predicted (R2 < 0.5). Fluxes were better predicted at forested and temperate climate sites than at sites in extreme climates or less represented by training data (e.g., the tropics). Finally, the evaluated large ensemble of ML-based models will be the basis of new global flux products.},
doi = {10.5194/bg-13-4291-2016},
journal = {Biogeosciences (Online)},
number = 14,
volume = 13,
place = {United States},
year = {2016},
month = {7}
}

Journal Article:
Free Publicly Available Full Text
Publisher's Version of Record

Citation Metrics:
Cited by: 35 works
Citation information provided by
Web of Science

Save / Share:

Works referenced in this record:

Ecosystem carbon dioxide fluxes after disturbance in forests of North America
journal, January 2010

  • Amiro, B. D.; Barr, A. G.; Barr, J. G.
  • Journal of Geophysical Research, Vol. 115
  • DOI: 10.1029/2010JG001390

Forecasting the cyanotoxins presence in fresh waters: A new model based on genetic algorithms combined with the MARS technique
journal, April 2013


Evaluating the Land and Ocean Components of the Global Carbon Cycle in the CMIP5 Earth System Models
journal, September 2013


'Breathing' of the terrestrial biosphere: lessons learned from a global network of carbon dioxide flux measurement systems
journal, January 2008

  • Baldocchi, Dennis
  • Australian Journal of Botany, Vol. 56, Issue 1
  • DOI: 10.1071/BT07151

Terrestrial Gross Carbon Dioxide Uptake: Global Distribution and Covariation with Climate
journal, July 2010


Carbon dioxide exchange over multiple temporal scales in an arid shrub ecosystem near La Paz, Baja California Sur, Mexico
journal, May 2012


Improving canopy processes in the Community Land Model version 4 (CLM4) using global flux fields empirically inferred from FLUXNET data
journal, January 2011

  • Bonan, Gordon B.; Lawrence, Peter J.; Oleson, Keith W.
  • Journal of Geophysical Research, Vol. 116, Issue G2
  • DOI: 10.1029/2010JG001593

Random Forests
journal, January 2001


The ERA-Interim reanalysis: configuration and performance of the data assimilation system
journal, April 2011

  • Dee, D. P.; Uppala, S. M.; Simmons, A. J.
  • Quarterly Journal of the Royal Meteorological Society, Vol. 137, Issue 656
  • DOI: 10.1002/qj.828

Effects of climate extremes on the terrestrial carbon cycle: concepts, processes and potential future impacts
journal, May 2015

  • Frank, Dorothea; Reichstein, Markus; Bahn, Michael
  • Global Change Biology, Vol. 21, Issue 8
  • DOI: 10.1111/gcb.12916

Multivariate Adaptive Regression Splines
journal, March 1991


Large-scale gaussian process multi-class classification for semantic segmentation and facade recognition
journal, January 2013

  • Fröhlich, Björn; Rodner, Erik; Kemmler, Michael
  • Machine Vision and Applications, Vol. 24, Issue 5
  • DOI: 10.1007/s00138-012-0480-y

NDWI—A normalized difference water index for remote sensing of vegetation liquid water from space
journal, December 1996


The effect of driving climate data on the simulated terrestrial carbon pools and fluxes over North America: Effect of climate on terrestrial carbon pools
journal, June 2013

  • Garnaud, C.; Sushama, L.; Arora, V. K.
  • International Journal of Climatology, Vol. 34, Issue 4
  • DOI: 10.1002/joc.3748

Decomposition of the mean squared error and NSE performance criteria: Implications for improving hydrological modelling
journal, October 2009


The random subspace method for constructing decision forests
journal, January 1998

  • Tin Kam Ho,
  • IEEE Transactions on Pattern Analysis and Machine Intelligence, Vol. 20, Issue 8
  • DOI: 10.1109/34.709601

Uncertainty in eddy covariance measurements and its application to physiological models
journal, July 2005


Overview of the radiometric and biophysical performance of the MODIS vegetation indices
journal, November 2002


Refinement of rooting depths using satellite-based evapotranspiration seasonality for ecosystem modeling in California
journal, November 2009


A Guided Hybrid Genetic Algorithm for Feature Selection with Expensive Cost Functions
journal, January 2013


Diagnostic assessment of European gross primary production
journal, October 2008


Recent decline in the global land evapotranspiration trend due to limited moisture supply
journal, October 2010

  • Jung, Martin; Reichstein, Markus; Ciais, Philippe
  • Nature, Vol. 467, Issue 7318
  • DOI: 10.1038/nature09396

Global patterns of land-atmosphere fluxes of carbon dioxide, latent heat, and sensible heat derived from eddy covariance, satellite, and meteorological observations
journal, January 2011

  • Jung, Martin; Reichstein, Markus; Margolis, Hank A.
  • Journal of Geophysical Research, Vol. 116
  • DOI: 10.1029/2010JG001566

An overview of MODIS Land data processing and product status
journal, November 2002


Comparison of the data-driven top-down and bottom-up global terrestrial CO 2 exchanges: GOSAT CO 2 inversion and empirical eddy flux upscaling : DATA-DRIVEN TERRESTRIAL CO
journal, July 2015

  • Kondo, Masayuki; Ichii, Kazuhito; Takagi, Hiroshi
  • Journal of Geophysical Research: Biogeosciences, Vol. 120, Issue 7
  • DOI: 10.1002/2014JG002866

Comprehensive comparison of gap-filling techniques for eddy covariance net carbon fluxes
journal, December 2007


Global products of vegetation leaf area and fraction absorbed PAR from year one of MODIS data
journal, November 2002


River flow forecasting through conceptual models part I — A discussion of principles
journal, April 1970


A new assessment of European forests carbon exchanges by eddy fluxes and artificial neural network spatialization
journal, April 2003


Effect of spatial sampling from European flux towers for estimating carbon and water fluxes with artificial neural networks: SAMPLING EFFECT ON FLUXES UPSCALING
journal, October 2015

  • Papale, Dario; Black, T. Andrew; Carvalhais, Nuno
  • Journal of Geophysical Research: Biogeosciences, Vol. 120, Issue 10
  • DOI: 10.1002/2015JG002997

Evaluation of terrestrial carbon cycle models for their response to climate variability and to CO 2 trends
journal, April 2013

  • Piao, Shilong; Sitch, Stephen; Ciais, Philippe
  • Global Change Biology, Vol. 19, Issue 7
  • DOI: 10.1111/gcb.12187

On the separation of net ecosystem exchange into assimilation and ecosystem respiration: review and improved algorithm
journal, September 2005


The Carbon Dioxide Exchange
journal, August 2010


First operational BRDF, albedo nadir reflectance products from MODIS
journal, November 2002


Assimilation exceeds respiration sensitivity to drought: A FLUXNET synthesis
journal, February 2010


Reduction in carbon uptake during turn of the century drought in western North America
journal, July 2012

  • Schwalm, Christopher R.; Williams, Christopher A.; Schaefer, Kevin
  • Nature Geoscience, Vol. 5, Issue 8
  • DOI: 10.1038/ngeo1529

Optimization of mixed refrigerant systems in low temperature applications by means of group method of data handling (GMDH)
journal, September 2015

  • Shirmohammadi, Reza; Ghorbani, Bahram; Hamedi, Mohsen
  • Journal of Natural Gas Science and Engineering, Vol. 26
  • DOI: 10.1016/j.jngse.2015.06.028

Modeling and measuring the effects of disturbance history and climate on carbon and water budgets in evergreen needleleaf forests
journal, December 2002


Uncertainty analysis of gross primary production upscaling using Random Forests, remote sensing and eddy covariance data
journal, October 2015

  • Tramontana, Gianluca; Ichii, Kazuito; Camps-Valls, Gustau
  • Remote Sensing of Environment, Vol. 168
  • DOI: 10.1016/j.rse.2015.07.015

A full greenhouse gases budget of Africa: synthesis, uncertainties, and vulnerabilities
journal, January 2014


Optical remote sensing and the retrieval of terrestrial vegetation bio-geophysical properties – A review
journal, October 2015

  • Verrelst, Jochem; Camps-Valls, Gustau; Muñoz-Marí, Jordi
  • ISPRS Journal of Photogrammetry and Remote Sensing, Vol. 108
  • DOI: 10.1016/j.isprsjprs.2015.05.005

Validation of the land-surface temperature products retrieved from Terra Moderate Resolution Imaging Spectroradiometer data
journal, November 2002


Estimation of net ecosystem carbon exchange for the conterminous United States by combining MODIS and AmeriFlux data
journal, October 2008


A continuous measure of gross primary production for the conterminous United States derived from MODIS and AmeriFlux data
journal, March 2010

  • Xiao, Jingfeng; Zhuang, Qianlai; Law, Beverly E.
  • Remote Sensing of Environment, Vol. 114, Issue 3
  • DOI: 10.1016/j.rse.2009.10.013

Characterization of forest types in Northeastern China, using multi-temporal SPOT-4 VEGETATION sensor data
journal, October 2002


Prediction of Continental-Scale Evapotranspiration by Combining MODIS and AmeriFlux Data Through Support Vector Machine
journal, November 2006

  • Yang, F.; White, M. A.; Michaelis, A. R.
  • IEEE Transactions on Geoscience and Remote Sensing, Vol. 44, Issue 11
  • DOI: 10.1109/TGRS.2006.876297

Developing a continental-scale measure of gross primary production by combining MODIS and AmeriFlux data through Support Vector Machine approach
journal, September 2007

  • Yang, Feihua; Ichii, Kazuhito; White, Michael A.
  • Remote Sensing of Environment, Vol. 110, Issue 1
  • DOI: 10.1016/j.rse.2007.02.016

Global vegetation gross primary production estimation using satellite-derived light-use efficiency and canopy conductance
journal, June 2015

  • Yebra, Marta; Van Dijk, Albert I. J. M.; Leuning, Ray
  • Remote Sensing of Environment, Vol. 163
  • DOI: 10.1016/j.rse.2015.03.016

Global estimates of evapotranspiration and gross primary production based on MODIS and global meteorology data
journal, July 2010


How errors on meteorological variables impact simulated ecosystem fluxes: a case study for six French sites
journal, January 2012


    Works referencing / citing this record:

    Large‐Scale Droughts Responsible for Dramatic Reductions of Terrestrial Net Carbon Uptake Over North America in 2011 and 2012
    journal, July 2018

    • He, Wei; Ju, Weimin; Schwalm, Christopher R.
    • Journal of Geophysical Research: Biogeosciences, Vol. 123, Issue 7
    • DOI: 10.1029/2018jg004520

    A data-driven approach to identify controls on global fire activity from satellite and climate observations (SOFIA V1)
    journal, January 2017

    • Forkel, Matthias; Dorigo, Wouter; Lasslop, Gitta
    • Geoscientific Model Development, Vol. 10, Issue 12
    • DOI: 10.5194/gmd-10-4443-2017

    Carbon–water flux coupling under progressive drought
    journal, January 2019


    Understanding the Land Carbon Cycle with Space Data: Current Status and Prospects
    journal, February 2019

    • Exbrayat, Jean-François; Bloom, A. Anthony; Carvalhais, Nuno
    • Surveys in Geophysics, Vol. 40, Issue 4
    • DOI: 10.1007/s10712-019-09506-2

    A global moderate resolution dataset of gross primary production of vegetation for 2000–2016
    journal, October 2017


    Coupling the Canadian Terrestrial Ecosystem Model (CTEM v. 2.0) to Environment and Climate Change Canada's greenhouse gas forecast model (v.107-glb)
    journal, January 2018

    • Badawy, Bakr; Polavarapu, Saroja; Jones, Dylan B. A.
    • Geoscientific Model Development, Vol. 11, Issue 2
    • DOI: 10.5194/gmd-11-631-2018

    Understanding the Land Carbon Cycle with Space Data: Current Status and Prospects
    journal, February 2019

    • Exbrayat, Jean-François; Bloom, A. Anthony; Carvalhais, Nuno
    • Surveys in Geophysics, Vol. 40, Issue 4
    • DOI: 10.1007/s10712-019-09506-2

    Large‐Scale Droughts Responsible for Dramatic Reductions of Terrestrial Net Carbon Uptake Over North America in 2011 and 2012
    journal, July 2018

    • He, Wei; Ju, Weimin; Schwalm, Christopher R.
    • Journal of Geophysical Research: Biogeosciences, Vol. 123, Issue 7
    • DOI: 10.1029/2018jg004520

    Deep learning and process understanding for data-driven Earth system science
    journal, February 2019


    A global moderate resolution dataset of gross primary production of vegetation for 2000–2016
    journal, October 2017


    Carbon–water flux coupling under progressive drought
    journal, January 2019


    A data-driven approach to identify controls on global fire activity from satellite and climate observations (SOFIA V1)
    journal, January 2017

    • Forkel, Matthias; Dorigo, Wouter; Lasslop, Gitta
    • Geoscientific Model Development, Vol. 10, Issue 12
    • DOI: 10.5194/gmd-10-4443-2017

    Coupling the Canadian Terrestrial Ecosystem Model (CTEM v. 2.0) to Environment and Climate Change Canada's greenhouse gas forecast model (v.107-glb)
    journal, January 2018

    • Badawy, Bakr; Polavarapu, Saroja; Jones, Dylan B. A.
    • Geoscientific Model Development, Vol. 11, Issue 2
    • DOI: 10.5194/gmd-11-631-2018

    Summarizing the state of the terrestrial biosphere in few dimensions
    journal, January 2020

    • Kraemer, Guido; Camps-Valls, Gustau; Reichstein, Markus
    • Biogeosciences, Vol. 17, Issue 9
    • DOI: 10.5194/bg-17-2397-2020