Preprints
https://doi.org/10.5194/bg-2022-211
https://doi.org/10.5194/bg-2022-211
 
02 Nov 2022
02 Nov 2022
Status: this preprint is currently under review for the journal BG.

A differentiable ecosystem modeling framework for large-scale inverse problems: demonstration with photosynthesis simulations

Doaa Aboelyazeed1, Chonggang Xu2, Forrest M. Hoffman3,4, Alex W. Jones5, Chris Rackauckas6, Kathryn E. Lawson1, and Chaopeng Shen1 Doaa Aboelyazeed et al.
  • 1Civil and Environmental Engineering, The Pennsylvania State University, University Park, PA 16802, USA
  • 2Earth and Environmental Sciences Division, Los Alamos National Laboratory, Los Alamos, NM, 87544, USA
  • 3Computational Sciences & Engineering Division and the Climate Change Science Institute, Oak Ridge National Laboratory, Oak Ridge, Tennessee, USA
  • 4Department of Civil and Environmental Engineering, University of Tennessee, Knoxville, Tennessee, USA
  • 5SciML Open Source Software Organization, https://sciml.ai
  • 6Computer Science and Artificial Intelligence Laboratory (CSAIL), Massachusetts Institute of Technology, Cambridge, MA, 02139, USA

Abstract. Photosynthesis plays an important role in carbon, nitrogen, and water cycles. Ecosystem models for photosynthesis are characterized by many parameters that are obtained from limited in-situ measurements and applied to the same plant types. Previous site-by-site calibration approaches could not leverage big data and faced issues like overfitting or parameter non-uniqueness. Here we developed a programmatically differentiable (meaning gradients of outputs to variables used in the model can be obtained efficiently and accurately) version of the photosynthesis process representation within the Functionally Assembled Terrestrial Ecosystem Simulator (FATES) model. This model is coupled to neural networks that learn parameterization from observations of photosynthesis rates. We first demonstrated that the framework was able to recover multiple assumed parameter values concurrently using synthetic training data. Then, using a real-world dataset consisting of many different plant functional types, we learned parameters that performed substantially better and dramatically reduced biases compared to literature values. Further, the framework allowed us to gain insights at a large scale. Our results showed that the carboxylation rate at 25 °C (Vc,max25), was more impactful than a factor representing water limitation, although tuning both was helpful in addressing biases with the default values. This framework could potentially enable a substantial improvement in our capability to learn parameters and reduce biases for ecosystem modeling at large scales.

Doaa Aboelyazeed et al.

Status: open (until 22 Dec 2022)

Comment types: AC – author | RC – referee | CC – community | EC – editor | CEC – chief editor | : Report abuse

Doaa Aboelyazeed et al.

Doaa Aboelyazeed et al.

Viewed

Total article views: 256 (including HTML, PDF, and XML)
HTML PDF XML Total BibTeX EndNote
192 60 4 256 1 2
  • HTML: 192
  • PDF: 60
  • XML: 4
  • Total: 256
  • BibTeX: 1
  • EndNote: 2
Views and downloads (calculated since 02 Nov 2022)
Cumulative views and downloads (calculated since 02 Nov 2022)

Viewed (geographical distribution)

Total article views: 266 (including HTML, PDF, and XML) Thereof 266 with geography defined and 0 with unknown origin.
Country # Views %
  • 1
1
 
 
 
 
Latest update: 09 Dec 2022
Download
Short summary
Photosynthesis is critical for life and is affected by a changing climate. Many parameters come into play when modeling, but traditional calibration approaches have faced many issues. Our framework trains coupled neural networks to provide parameters to a photosynthesis model. Using big data, we independently found parameter values that were correlated with those in the literature while giving higher correlation and reduced biases in photosynthesis rates.
Altmetrics