I published recently a study, where I used atmospheric CO2 observations to evaluate two different soil carbon models, coupled to a land surface model (paper here, FMI science news in English, in Finnish).
Heterotrophic respiration is important part of the global carbon cycle, but it’s direct observation is somewhat impossible at global scale. Even at site scale it is a laborious measurements, since the soil respiration consists of both autotrophic (by plant roots) and heterotrophic (by soil organisms) components. Belowground things are difficult, as soil is very heterogeneous. There are challenges and uncertainties are large in modelling the soil at global scale.
A pretty common summary of a land surface model study goes along the lines: “Process X is important for the global carbon cycle. We added/modified its presentation in a land surface model. A related observation was better now captured by the model, improvement being yy%. The global GPP decreased/increased xx% because of this. X is therefore important and should be included/improved in land surface models. This study increases our understanding of the global carbon cycle and is therefore relevant to humanity.”
So, I made myself a bit that kind of a study when implementing a new soil carbon model to a land surface model, published in 2011. Nine years later I publish a study saying: “Actually I had a closer look, and the CO2 observations show that it… might need further consideration (or at least new parameterization).”
When I was doing the original study, I was also checking the results against some northern CO2 measurement stations and the new model showed better results. The paper about the work anyhow didn’t include these. My collaborators told me several things that prevent the analysis: not having anthropogenic and ocean fluxes in the simulations and having only point measurements as evaluation data.
In 2012 I was writing a grant proposal taking these points into consideration. A new thing was also the GOSAT satellite observing column XCO2, having in theory a global coverage. I was very inspired by the publications of Gretchen Keppel-Aleks. I got the grant, but because life happens, the work didn’t start before 2017. I’m grateful Julia Nabel and Aki Tsuruta to have done the modelling part and I was just concentrating on the analysis.
It was a whole bunch of data to analyse, but my group leader Sönke Zaehle had long experience in working with CO2 data and was able to give advice. I remember being puzzled, how we could separate the role of the different litter layers vs. environmental drivers between the models. Sönke suggested using a simple box model calculation, that solved it. With some insight you can do things easier (but getting that insight can be painful, I imagine).
As we submitted our paper, it was great that the group of Gretchen Keppel-Aleks published a similar paper in the same journal (Basile et al., 2020), that had more emphasis on the interannual variation. This helped us in answering the reviewers, that the idea is not total nonsense. It’s always a bit challenging, when the research is not a direct continuation of something old. Prof. Jouni Jaakkola gave a very inspiring seminar talk last week at our institute. He talked about a study that was difficult to get published (Jaakkola et al., 1991), because the journals wouldn’t believe the results (that also small amounts of pollutants can increase the risk of children getting sick). I can so much relate to that (despite not thinking any of my studies will become a road sign for future studies, like his did).
Prof. Jaakkola had only a long history in interdisciplinary collaboration. He told that it is the individuals doing the work that matter, on how successful it ends. He also reminded that it is important to remember that different disciplines have their own definitions. That also rings a bell for me. This our study was very strongly based on carbon cycle, but it combined the terrestrial view with the atmosphere-based data. It was a great possibility for me to learn more from the atmosphere.
This was the first time I was using XCO2 data. It’s only through the LCC-project folks and a recent group talk by Aki Tsuruta, that the penny is finally starting to drop for me about the true potential of the Sentinel fleet for the carbon cycle studies. It’s enormous. Cool time to do science now.
References:
Basile, S. J., Lin, X., Wieder, W. R., Hartman, M. D., and Keppel-Aleks, G.: Leveraging the signature of heterotrophic respiration on atmospheric CO2 for model benchmarking, Biogeosciences, 17, 1293–1308, https://doi.org/10.5194/bg-17-1293-2020, 2020.
Jaakkola, J. J., M Paunio, M Virtanen, and O P Heinonen, 1991: Low-level air pollution and upper respiratory infections in children. American Journal of Public Health 81, 1060_1063, https://doi.org/10.2105/AJPH.81.8.1060.