Causal Drivers of Land-Atmosphere Carbon Fluxes from Machine Learning Models and Data
- Mozhgan Askarzadehfarahani
, - Mozhgan A Farahani,
- Allison E Goodwell
Mozhgan A Farahani
Department of Civil Engineering, University of Colorado Denver
Allison E Goodwell
Prairie Research Institute, Prairie Research Institute, University of Illinois at Urbana-Champaign, Department of Civil Engineering, University of Colorado Denver
Abstract
• Information theory measures describe individual and joint causal relationships in observed versus modeled vertical carbon dioxide fluxes. • Three machine learning models overestimate unique information from sources at the expense of synergistic, or joint information. • Regionally trained models have improved functional but not predictive performances, indicating a trade-off.10 May 2024Submitted to ESS Open Archive 13 May 2024Published in ESS Open Archive
