Improving pedotransfer functions for predicting soil mineral associated organic carbon by ensemble machine learning

• Published in: Geoderma Vol. 428; p. 116208
• Main Authors: Xiao, Yi, Xue, Jie, Zhang, Xianglin, Wang, Nan, Hong, Yongsheng, Jiang, Yefeng, Zhou, Yin, Teng, Hongfen, Hu, Bifeng, Lugato, Emanuele, Richer-de-Forges, Anne C., Arrouays, Dominique, Shi, Zhou, Chen, Songchao
• Format: Journal Article
• Language: English
• Published: Elsevier B.V 15.12.2022; Elsevier
• Subjects: algorithms; carbon dioxide; climate; Earth Sciences; Environmental Sciences; Europe; Forward recursive feature selection; Global Changes; LUCAS Soil; Model ensemble; model validation; nitrogen; particulate organic carbon; pedotransfer functions; prediction; sand; Sciences of the Universe; Soil carbon fractions; soil erosion; soil minerals; soil organic carbon; topsoil; Europe; Model ensemble; Forward recursive feature selection; LUCAS Soil; Soil carbon fractions; carbon dioxide; nitrogen; Europe; soil organic carbon; topsoil; climate; soil minerals; pedotransfer function; particulate organic carbon; sand; model validation; prediction; model; soil erosion; algorithm

•We evaluated the potential of pedotransfer function in MAOC prediction.•Forward recursive feature selection performed well in model parsimony and performance.•Cubist had better model performance than Random Forest and Gradient Boosted Machine.•Model ensemble improved model performance and robustnes...