Improving soil organic carbon mapping in intensively cultivated Mollisols: A geographically adjusted Gaussian mixture model incorporating tillage practices and spectral indicators
文献类型：期刊论文
| 作者 | Tan, Qiuyuan1,5,6; Geng, Jing4,6; Yu, Yong6; Pei, Jie4,6; Fang, Huajun2,3,5 |
| 刊名 | SOIL & TILLAGE RESEARCH
 |
| 出版日期 | 2026-02-01 |
| 卷号 | 256页码:106911 |
| 关键词 | Soil organic carbon Digital soil mapping Tillage management Geographically Adjusted Gaussian Mixture Model Mollisols |
| ISSN号 | 0167-1987 |
| DOI | 10.1016/j.still.2025.106911 |
| 产权排序 | 2 |
| 文献子类 | Article |
| 英文摘要 | Understanding how tillage and land management practices affect the spatial distribution of soil organic carbon (SOC) is essential for maintaining soil fertility and promoting sustainable agricultural production. While digital soil mapping (DSM) using machine learning has become widespread, its performance in intensively cultivated regions is often constrained by the limited inclusion of management-related variables and by assumptions of spatial stationarity-despite the fact that soil-environment relationships vary significantly across space, especially in landscapes shaped by human activity. This study presents a novel hybrid modeling framework-the Geographically Adjusted Gaussian Mixture Model (GAGMM)-that integrates probabilistic environmental clustering with spatial proximity weighting to improve SOC prediction across Mollisols croplands in Northeast China. To reflect management-induced SOC variability, a comprehensive suite of predictors was assembled, including direct tillage indicators (crop type, nitrogen fertilizer input, residue return) and indirect proxies such as vegetation phenology, net primary productivity (NPP), and bare-soil spectral indices. GAGMM was benchmarked against conventional global, crop-type partitioned, and unadjusted clustering models using Random Forest (RF) and Extreme Gradient Boosting (XGBoost) algorithms as base learners. The GAGMM-XGBoost model yielded the highest accuracy (R2 = 0.783, RMSE = 0.772 g/kg), outperforming the global XGBoost model by 12.50 %. Variable importance analysis indicated that tillage-related indicators accounted for the largest share of SOC variation (57.95 %), followed by climatic and spectral features. These findings highlight the significance of incorporating management practices in SOC modeling and demonstrate the value of GAGMM as a flexible approach for supporting data-driven soil management in intensively farmed landscapes. |
| URL标识 | 查看原文 |
| WOS关键词 | SPECTROSCOPY ; SHRINKAGE ; SELECTION ; CROPLANDS ; SCALE ; LAND |
| WOS研究方向 | Agriculture |
| 语种 | 英语 |
| WOS记录号 | WOS:001598668700001 |
| 出版者 | ELSEVIER |
| 源URL | [http://ir.igsnrr.ac.cn/handle/311030/219780]  |
| 专题 | 资源与环境信息系统国家重点实验室_外文论文 |
| 通讯作者 | Geng, Jing |
| 作者单位 | 1.Chinese Acad Sci, Inst Geog Sci & Nat Resources Res, State Key Lab Resources & Environm Informat Syst, Beijing 100101, Peoples R China; 2.Zhongke Jian Inst Ecoenvironm Sci, Jian 343000, Peoples R China 3.Chinese Acad Sci, Inst Geog Sci & Nat Resources Res, Key Lab Ecosyst Network Observat & Modeling, Beijing 100101, Peoples R China; 4.Sun Yat Sen Univ, Key Lab Comprehens Observat Polar Environm, Minist Educ, Zhuhai 519082, Peoples R China; 5.Univ Chinese Acad Sci, Beijing 100049, Peoples R China; 6.Sun Yat Sen Univ, Sch Geospatial Engn & Sci, Zhuhai 519082, Peoples R China; |
| 推荐引用方式 GB/T 7714 | Tan, Qiuyuan,Geng, Jing,Yu, Yong,et al. Improving soil organic carbon mapping in intensively cultivated Mollisols: A geographically adjusted Gaussian mixture model incorporating tillage practices and spectral indicators[J]. SOIL & TILLAGE RESEARCH,2026,256:106911. |
| APA | Tan, Qiuyuan,Geng, Jing,Yu, Yong,Pei, Jie,&Fang, Huajun.(2026).Improving soil organic carbon mapping in intensively cultivated Mollisols: A geographically adjusted Gaussian mixture model incorporating tillage practices and spectral indicators.SOIL & TILLAGE RESEARCH,256,106911. |
| MLA | Tan, Qiuyuan,et al."Improving soil organic carbon mapping in intensively cultivated Mollisols: A geographically adjusted Gaussian mixture model incorporating tillage practices and spectral indicators".SOIL & TILLAGE RESEARCH 256(2026):106911. |
入库方式： OAI收割
来源：地理科学与资源研究所
浏览0
下载0
收藏0
其他版本
除非特别说明，本系统中所有内容都受版权保护，并保留所有权利。