Articles | Volume 9, issue 1
https://doi.org/10.5194/soil-9-21-2023
https://doi.org/10.5194/soil-9-21-2023
Original research article
 | 
11 Jan 2023
Original research article |  | 11 Jan 2023

Shapley values reveal the drivers of soil organic carbon stock prediction

Alexandre M. J.-C. Wadoux, Nicolas P. A. Saby, and Manuel P. Martin

Related authors

Mapping soil organic carbon fractions for Australia, their stocks and uncertainty
Mercedes Román Dobarco, Alexandre M. J-C. Wadoux, Brendan Malone, Budiman Minasny, Alex B. McBratney, and Ross Searle
Biogeosciences Discuss., https://doi.org/10.5194/bg-2022-207,https://doi.org/10.5194/bg-2022-207, 2022
Revised manuscript accepted for BG
Short summary
Multi-source data integration for soil mapping using deep learning
Alexandre M. J.-C. Wadoux, José Padarian, and Budiman Minasny
SOIL, 5, 107–119, https://doi.org/10.5194/soil-5-107-2019,https://doi.org/10.5194/soil-5-107-2019, 2019

Related subject area

Pedometrics
How well does digital soil mapping represent soil geography? An investigation from the USA
David G. Rossiter, Laura Poggio, Dylan Beaudette, and Zamir Libohova
SOIL, 8, 559–586, https://doi.org/10.5194/soil-8-559-2022,https://doi.org/10.5194/soil-8-559-2022, 2022
Short summary

Cited articles

Arrouays, D., Deslais, W., and Badeau, V.: The carbon content of topsoil and its geographical distribution in France, Soil Use Manage., 17, 7–11, 2001. a
Arrouays, D., Jolivet, C., Boulonne, L., Bodineau, G., Saby, N. P. A., and Grolleau, E.: A new projection in France: a multi-institutional soil quality monitoring network, Comptes Rendus de l'Académie d'Agriculture de France (France), 2002. a
Batjes, N. H.: Total carbon and nitrogen in the soils of the world, Europ. J. Soil Sci., 47, 151–163, 1996. a
Beucher, A., Rasmussen, C. B., Moeslund, T. B., and Greve, M. H.: Interpretation of convolutional neural networks for acid sulfate soil classification, Front. Environ. Sci., 9, 679, https://doi.org/10.3389/fenvs.2021.809995, 2022. a, b, c
Breiman, L.: Random forests, Mach. Learn., 45, 5–32, 2001. a
Download
Short summary
We introduce Shapley values for machine learning model interpretation and reveal the local and global controlling factors of soil organic carbon (SOC) stocks. The method enables spatial analysis of the important variables. Vegetation and topography determine much of the SOC stock variation in mainland France. We conclude that SOC stock variation is complex and should be interpreted at multiple levels.