Causality guided machine learning model on wetland CH4 emissions across global wetlands

Publikation: Bidrag til tidsskrift › Tidsskriftartikel › Forskning › fagfællebedømt

Standard

Causality guided machine learning model on wetland CH4 emissions across global wetlands. / Yuan, Kunxiaojia; Zhu, Qing; Li, Fa; Riley, William J.; Torn, Margaret; Chu, Housen; McNicol, Gavin; Chen, Min; Knox, Sara; Delwiche, Kyle; Wu, Huayi; Baldocchi, Dennis; Ma, Hongxu; Desai, Ankur R.; Chen, Jiquan; Sachs, Torsten; Ueyama, Masahito; Sonnentag, Oliver; Helbig, Manuel; Tuittila, Eeva-Stiina; Jurasinski, Gerald; Koebsch, Franziska; Campbell, David; Schmid, Hans Peter; Lohila, Annalea; Goeckede, Mathias; Nilsson, Mats B.; Friborg, Thomas; Jansen, Joachim; Zona, Donatella; Euskirchen, Eugenie; Ward, Eric J.; Bohrer, Gil; Jin, Zhenong; Liu, Licheng; Iwata, Hiroki; Goodrich, Jordan; Jackson, Robert.

I: Agricultural and Forest Meteorology, Bind 324, 109115, 2022.

Publikation: Bidrag til tidsskrift › Tidsskriftartikel › Forskning › fagfællebedømt

Harvard

Yuan, K, Zhu, Q, Li, F, Riley, WJ, Torn, M, Chu, H, McNicol, G, Chen, M, Knox, S, Delwiche, K, Wu, H, Baldocchi, D, Ma, H, Desai, AR, Chen, J, Sachs, T, Ueyama, M, Sonnentag, O, Helbig, M, Tuittila, E-S, Jurasinski, G, Koebsch, F, Campbell, D, Schmid, HP, Lohila, A, Goeckede, M, Nilsson, MB, Friborg, T, Jansen, J, Zona, D, Euskirchen, E, Ward, EJ, Bohrer, G, Jin, Z, Liu, L, Iwata, H, Goodrich, J & Jackson, R 2022, 'Causality guided machine learning model on wetland CH4 emissions across global wetlands', Agricultural and Forest Meteorology, bind 324, 109115. https://doi.org/10.1016/j.agrformet.2022.109115

APA

Yuan, K., Zhu, Q., Li, F., Riley, W. J., Torn, M., Chu, H., McNicol, G., Chen, M., Knox, S., Delwiche, K., Wu, H., Baldocchi, D., Ma, H., Desai, A. R., Chen, J., Sachs, T., Ueyama, M., Sonnentag, O., Helbig, M., ... Jackson, R. (2022). Causality guided machine learning model on wetland CH4 emissions across global wetlands. Agricultural and Forest Meteorology, 324, [109115]. https://doi.org/10.1016/j.agrformet.2022.109115

Vancouver

Yuan K, Zhu Q, Li F, Riley WJ, Torn M, Chu H o.a. Causality guided machine learning model on wetland CH4 emissions across global wetlands. Agricultural and Forest Meteorology. 2022;324. 109115. https://doi.org/10.1016/j.agrformet.2022.109115

Author

Yuan, Kunxiaojia ; Zhu, Qing ; Li, Fa ; Riley, William J. ; Torn, Margaret ; Chu, Housen ; McNicol, Gavin ; Chen, Min ; Knox, Sara ; Delwiche, Kyle ; Wu, Huayi ; Baldocchi, Dennis ; Ma, Hongxu ; Desai, Ankur R. ; Chen, Jiquan ; Sachs, Torsten ; Ueyama, Masahito ; Sonnentag, Oliver ; Helbig, Manuel ; Tuittila, Eeva-Stiina ; Jurasinski, Gerald ; Koebsch, Franziska ; Campbell, David ; Schmid, Hans Peter ; Lohila, Annalea ; Goeckede, Mathias ; Nilsson, Mats B. ; Friborg, Thomas ; Jansen, Joachim ; Zona, Donatella ; Euskirchen, Eugenie ; Ward, Eric J. ; Bohrer, Gil ; Jin, Zhenong ; Liu, Licheng ; Iwata, Hiroki ; Goodrich, Jordan ; Jackson, Robert. / Causality guided machine learning model on wetland CH4 emissions across global wetlands. I: Agricultural and Forest Meteorology. 2022 ; Bind 324.

Bibtex

@article{69cfe77ad55946f487f6a5d10673a60e,
title = "Causality guided machine learning model on wetland CH4 emissions across global wetlands",
abstract = "Wetland CH4 emissions are among the most uncertain components of the global CH4 budget. The complex nature of wetland CH4 processes makes it challenging to identify causal relationships for improving our understanding and predictability of CH4 emissions. In this study, we used the flux measurements of CH4 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 CH4 emissions at sub-seasonal scale. We found that soil temperature is the dominant factor for CH4 emissions in all studied wetland types. Ecosystem respiration (CO2) 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 CH4 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 CH4 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 CH4 emissions within earth system land models.",
keywords = "Causal inference, Eddy covariance CH emission, Machine learning, Wetlands",
author = "Kunxiaojia Yuan and Qing Zhu and Fa Li and Riley, {William J.} and Margaret Torn and Housen Chu and Gavin McNicol and Min Chen and Sara Knox and Kyle Delwiche and Huayi Wu and Dennis Baldocchi and Hongxu Ma and Desai, {Ankur R.} and Jiquan Chen and Torsten Sachs and Masahito Ueyama and Oliver Sonnentag and Manuel Helbig and Eeva-Stiina Tuittila and Gerald Jurasinski and Franziska Koebsch and David Campbell and Schmid, {Hans Peter} and Annalea Lohila and Mathias Goeckede and Nilsson, {Mats B.} and Thomas Friborg and Joachim Jansen and Donatella Zona and Eugenie Euskirchen and Ward, {Eric J.} and Gil Bohrer and Zhenong Jin and Licheng Liu and Hiroki Iwata and Jordan Goodrich and Robert Jackson",
note = "Publisher Copyright: {\textcopyright} 2022",
year = "2022",
doi = "10.1016/j.agrformet.2022.109115",
language = "English",
volume = "324",
journal = "Agricultural and Forest Meteorology",
issn = "0168-1923",
publisher = "Elsevier",

}

RIS

TY - JOUR

T1 - Causality guided machine learning model on wetland CH4 emissions across global wetlands

AU - Yuan, Kunxiaojia

AU - Zhu, Qing

AU - Li, Fa

AU - Riley, William J.

AU - Torn, Margaret

AU - Chu, Housen

AU - McNicol, Gavin

AU - Chen, Min

AU - Knox, Sara

AU - Delwiche, Kyle

AU - Wu, Huayi

AU - Baldocchi, Dennis

AU - Ma, Hongxu

AU - Desai, Ankur R.

AU - Chen, Jiquan

AU - Sachs, Torsten

AU - Ueyama, Masahito

AU - Sonnentag, Oliver

AU - Helbig, Manuel

AU - Tuittila, Eeva-Stiina

AU - Jurasinski, Gerald

AU - Koebsch, Franziska

AU - Campbell, David

AU - Schmid, Hans Peter

AU - Lohila, Annalea

AU - Goeckede, Mathias

AU - Nilsson, Mats B.

AU - Friborg, Thomas

AU - Jansen, Joachim

AU - Zona, Donatella

AU - Euskirchen, Eugenie

AU - Ward, Eric J.

AU - Bohrer, Gil

AU - Jin, Zhenong

AU - Liu, Licheng

AU - Iwata, Hiroki

AU - Goodrich, Jordan

AU - Jackson, Robert

N1 - Publisher Copyright: © 2022

PY - 2022

Y1 - 2022

N2 - Wetland CH4 emissions are among the most uncertain components of the global CH4 budget. The complex nature of wetland CH4 processes makes it challenging to identify causal relationships for improving our understanding and predictability of CH4 emissions. In this study, we used the flux measurements of CH4 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 CH4 emissions at sub-seasonal scale. We found that soil temperature is the dominant factor for CH4 emissions in all studied wetland types. Ecosystem respiration (CO2) 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 CH4 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 CH4 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 CH4 emissions within earth system land models.

AB - Wetland CH4 emissions are among the most uncertain components of the global CH4 budget. The complex nature of wetland CH4 processes makes it challenging to identify causal relationships for improving our understanding and predictability of CH4 emissions. In this study, we used the flux measurements of CH4 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 CH4 emissions at sub-seasonal scale. We found that soil temperature is the dominant factor for CH4 emissions in all studied wetland types. Ecosystem respiration (CO2) 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 CH4 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 CH4 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 CH4 emissions within earth system land models.

KW - Causal inference

KW - Eddy covariance CH emission

KW - Machine learning

KW - Wetlands

U2 - 10.1016/j.agrformet.2022.109115

DO - 10.1016/j.agrformet.2022.109115

M3 - Journal article

AN - SCOPUS:85135914716

VL - 324

JO - Agricultural and Forest Meteorology

JF - Agricultural and Forest Meteorology

SN - 0168-1923

M1 - 109115

ER -

ID: 323990096