Quantifying and constraining terrestrial biosphere model (TBM) uncertainty and global carbon budget projections via data assimilation

The ocean and land absorb ~50% of anthropogenic CO2 emissions; therefore, they act as a global C sink. Uncertainty in TBM carbon sink projections is high; some models even predict the land may switch from a C sink to a source by the end of the century. There is an urgent need to address this issue – without accurate sink estimates, we cannot quantify the level of C emissions that will keep us within a 2°C rise in temperature.

A large body of my research to date has been focused on reducing this uncertainty in terrestrial biosphere model projections of vegetation and carbon dynamics from site to global scales using statistical model-data integration, or data assimilation (DA), in combination with a wide range of data streams. These include eddy covariance net CO2 and latent heat (LE) flux measurements, chamber-based CH4 fluxes, atmospheric CO2 concentration data, biomass observations and satellite-derived FAPAR/NDVI and fluorescence data.

  • MacBean et al. (2015) Using satellite data to improve the leaf phenology of a global terrestrial biosphere model, Biogeosciences, 12, 7185-7208.
  • Peylin et al. (2016) A new step-wise Carbon Cycle Data Assimilation System using multiple data streams to constrain the simulated land surface carbon cycle, Geoscientific Model Development, 9, 3321-3346.
  • Thum et al. (2017) The potential benefit of using forest biomass data in addition to carbon and water flux measurements to constrain ecosystem model parameters: case studies at two temperate forest sites, Agricultural and Forest Meteorology, 234, 48-65.
  • Li et al. (2017), Reducing the uncertainty of parameters controlling seasonal carbon and water fluxes in Chinese forests, and its implication for simulated climate sensitivities, Global Biogeochemical Cycles, 31.
  • MacBean, N., C. Bacour, N. Raoult, V. Bastrikov, E. N. Koffi, S. Kuppel, F. Maignan, C. Ottlé, M. Peaucelle, D. Santaren, and P. Peylin (2022), Quantifying and Reducing Uncertainty in Global Carbon Cycle Predictions: Lessons and Perspectives From 15 Years of Data Assimilation Studies with the ORCHIDEE Terrestrial Biosphere Model, Global Biogeochemical Cycles, 36, e2021GB007177.

One focus of these efforts have been to assess how multiple different data streams can be combined in a data assimilation system to optimize parameters related to multiple model variables. This work is reflected and highlighted in the following papers:

  • Bacour et al. (2015), Joint assimilation of eddy-covariance flux measurements and satellite observations within a process-oriented biosphere model, Journal of Geophysical Research Biogeosciences, 120, 1839–1857.
  • MacBean et al. (2016) Consistent assimilation of multiple data streams in a carbon cycle data assimilation system, Geoscientific Model Development, 9, 3569-3588.

A recent focus with collaborators Cédric Bacour and Fabienne Maignan at the LSCE has been on assessing how new satellite datasets of solar-induced chlorophyll fluorescence (SIF) to better constrain (reduce biases and improve uncertainties) model gross carbon uptake (or gross primary productivity – GPP). We have tested a simple approach to using SIF data in a model-data integration framework:

  • MacBean et al. (2018), Strong constraint on modeled global carbon uptake using solar-induced chlorophyll fluorescence data, Scientific Reports, 8, 1973. 
  • Bacour, C., F. Maignan, P. Peylin, N. MacBean, V. Bastrikov, J. Joiner, P. Köhler, L. Guanter, and C. Frankenberg (2019), Differences between OCO2 and GOME2 SIF products from a model-data fusion perspective, JGR-Biogeosciences, 124, 3143–3157.
  • Bacour, C., F. Maignan, N. MacBean, A. Porcar-Castell, J. Flexas, C. Frankenberg, P. Peylin, F. Chevallier and N. Vuichard (2019), Improving estimates of Gross Primary Productivity by assimilating solar-induced fluorescence satellite retrievals in a terrestrial biosphere model using a mechanistic SIF model, JGR-Biogeosciences 124, doi: 10.1029/2019JG005040.

Our ongoing work involves the evaluating the consistency of different SIF products, and the continued development of the ORCHIDEE model to include a mechanistic description of the links between photosynthesis- and fluorescence-related processes.

We continue to develop our technical expertise of how to best develop and configure global carbon cycle DA systems, e.g.:

  • Bastrikov, V., N. MacBean, C. Bacour, D. Santaren, S. Kuppel and P. Peylin (2018), Land surface model parameter optimisation using in-situ flux data: comparison of gradient-based versus random search algorithms, Geoscientific Model Development, 11, 4739-4754.

All these studies have ultimately resulted in an improvement in the predicted land carbon sink as well as a reduction and better quantification of its uncertainty into the future (MacBean et al., 2022). In addition, DA methods allow us to identify model processes that are not well represented or missing if there is still a misfit between the model and the data. For example, MacBean et al. (2015)led to my interest in improving the representation of coupled carbon-water-vegetation dynamics in semi-arid ecosystems in models.

I have also collaborated with the CLM-DART group in developing their state (vs parameter) DA system for constraining global scale carbon stocks:

  • Fox, A.M., T.J. Hoar, J.L. Anderson, A.F. Arellano, W.K. Smith, M.E. Litvak, N. MacBean, D.S. Schimel, and D.J.P. Moore (2018), Evaluation of a Data Assimilation System for Land Surface Models using CLM4.5. Journal of Advances in Modeling Earth Systems, 10, 2471–24942
  • Fox, A., X. Huo, T. Hoar, H. Dashti, W. K. Smith, N. MacBean, J. A. Anderson, M. C. Roby, and D. J. P. Moore (2022), Assimilation of Global Satellite Leaf Area Estimates Reduces Modeled Global Carbon Uptake and Energy Loss by Terrestrial Ecosystems, Journal of Geophysical Research: Biogeosciences, 127, e2022JG006830

This work was/is funded by:

  • MULTIPLY: a European Union H2020 project that will provide a platform for retrieval of multiscale land surface information from European Space Agency SENTINEL satellite data.
  • FLuOR (Fluorescence in ORCHIDEE): a French CNES (Centre National d’Etudes Spatiales) Space Agency TOSCA (Terre solide, Océan, Surfaces Continentales, Atmosphère) Programme project to develop, test and optimize a new fluorescence model in the ORCHIDEE terrestrial biosphere model.
  • ESA FLEX Bridge: a preparatory scientific study for Solar Induced Fluorescence retrieval from satellites and applications in assessment of photosynthesis and stress status in terrestrial vegetation. The final report contributed to the scientific basis leading to the selection of the FLEX satellite as the European Space Agency’s (ESA) 8th Earth Explorer mission.
  • CARBONES: European Union 7th Framework project aiming to provide a 30-year long re-analysis of carbon fluxes and pools over Europe and the globe, in form of a user-friendly on-line interface.
  • GEOCARBON: European Union 7th Framework project with the ultimate aim to lay the foundations for an operational Global Carbon Observing and Analysis System in support to both science and policy.
  • ExpeER: a major European Union Infrastructure project in the field of Ecosystem Research. ExpeER integrated existing national infrastructures to improve their research capacity whilst at the same time facilitating access to key experimental and observational platforms as well as analytical and modelling facilities for the benefit of the international research community.
  • UKPopNet: a UK Natural Environment Research Council (NERC) and English Nature funded project to study how changes in biodiversity will affect the sustainability of ecosystems, landscapes and livelihoods. Part of this work was focused on how to model and upscale CO2 and CH4 fluxes at an upland peatland site in Wales.