Qing Zhu Fa Li William J. Riley Margaret Torn Housen Chu Gavin McNicol Mingshu Chen Sara Knox Kyle B. Delwiche Huayi Wu Dennis Baldocchi Hongxu Ma Ankur R. Desai Jiquan Chen Torsten Sachs Masahito Ueyama Oliver Sonnentag Manuel Helbig Eeva-Stiina Tuittila Gerald Jurasinski Franziska Koebsch David I. Campbell Hans Peter Schmid Annalea Lohila Mathias Goeckede Mats B. Nilsson Thomas Friborg Joachim Jansen Donatella Zona Eugenie S. Euskirchen Eric Ward Gil Bohrer Zhenong Jin Licheng Liu Hiroki Iwata Jordan P. Goodrich Robert B. Jackson Kunxiaojia Yuan 2022 <p><span>Wetland CH</span>₄<span>&nbsp;emissions are among the most uncertain components of the global CH</span>₄<span>&nbsp;budget. The complex nature of wetland CH</span>₄<span>&nbsp;processes makes it challenging to identify causal relationships for improving our understanding and predictability of CH</span>₄<span>&nbsp;emissions. In this study, we used the flux measurements of CH</span>₄<span>&nbsp;from eddy covariance towers (30 sites from 4 wetlands types: bog, fen, marsh, and wet tundra) to construct a causality-constrained machine learning (ML) framework to explain the regulative factors and to capture CH</span>₄<span>&nbsp;emissions at sub-seasonal scale. We found that soil temperature is the dominant factor for CH</span>₄<span>&nbsp;emissions in all studied wetland types. Ecosystem respiration (CO</span>₂<span>) and gross primary productivity exert controls at bog, fen, and marsh sites with lagged responses of days to weeks. Integrating these asynchronous environmental and biological causal relationships in predictive models significantly improved model performance. More importantly, modeled CH</span>₄<span>&nbsp;emissions differed by up to a factor of 4 under a +1°C warming scenario when causality constraints were considered. These results highlight the significant role of causality in modeling wetland CH</span>₄<span>&nbsp;emissions especially under future warming conditions, while traditional data-driven ML models may reproduce observations for the wrong reasons. Our proposed causality-guided model could benefit predictive modeling, large-scale upscaling, data gap-filling, and surrogate modeling of wetland CH</span>₄<span>&nbsp;emissions within earth system land models.</span></p> application/pdf 10.1016/j.agrformet.2022.109115 en Elsevier Causality guided machine learning model on wetland CH4 emissions across global wetlands article