Data fusion enhances the accuracy of soil organic carbon estimation by using high accuracy surface modeling
文献类型：期刊论文
| 作者 | Zhou, Wei1,2,3; Wang, Ting2; Peng, Yao2; Yu, Wenping2; Sun, Xiaofang1; Tian, Yongzhong2; Li, Saibo3; Du, Zhengping3; Yue, Tianxiang2,3 |
| 刊名 | SOIL & TILLAGE RESEARCH
 |
| 出版日期 | 2026-03-01 |
| 卷号 | 257页码:106945 |
| 关键词 | High accuracy surface modeling Soil organic carbon Multisource remote sensing Machine learning Tibetan Plateau |
| ISSN号 | 0167-1987 |
| DOI | 10.1016/j.still.2025.106945 |
| 产权排序 | 2 |
| 文献子类 | Article |
| 英文摘要 | Having the ability to accurately and effectively obtain soil organic carbon (SOC) spatial information is critical for assessing soil carbon sequestration capacity and mitigating climate change. However, there remains a significant research gap in the collaborative application of multi-source data and their impact on model estimation accuracy. This gap limits the ability to assess soil carbon pools accurately. Therefore, we propose a data-model fusion framework that uses three types of multi-source data-environmental variables, optical remote sensing, and synthetic aperture radar (SAR)- along with three machine learning algorithms to predict SOC. We conducted data fusion of SOC field observation data and model simulations using high-accuracy surface modeling (HASM). The results showed that: (1) The data VII combination, which incorporates all three data types, paired with support vector machine (SVM), random forest (RF), and Extreme Gradient Boosting (XGBoost) models, obtained higher prediction accuracy (R2 increased by 4 % - 53 %, and RMSE decreased by 18 % - 25 %) compared to other data combinations. (2) After data fusion using HASM, simulation accuracy improved significantly (R2 increased by 18 % - 22 %, and RMSE decreased by 2 % - 12 %). Additionally, the spatial distribution pattern was more reasonable, with corrections made to previously underestimated and overestimated SOC content. This study demonstrates that multi-source data fusion combined with machine learning techniques can achieve optimal results for SOC prediction. This approach provides an accurate and novel method for estimating SOC at national and global scales and offers scientific guidance for the spatial planning of terrestrial carbon sink strategies. |
| URL标识 | 查看原文 |
| WOS关键词 | LAND-USE ; PREDICTION ; RESOLUTION |
| WOS研究方向 | Agriculture |
| 语种 | 英语 |
| WOS记录号 | WOS:001612811500001 |
| 出版者 | ELSEVIER |
| 源URL | [http://ir.igsnrr.ac.cn/handle/311030/217741]  |
| 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 | Zhou, Wei; Yue, Tianxiang |
| 作者单位 | 1.Qufu Normal Univ, Sch Geog & Tourism, Rizhao 276826, Peoples R China 2.Southwest Univ, Sch Geog Sci, Chongqing 400715, Peoples R China; 3.Chinese Acad Sci, State Key Lab Resources & Environm Informat Syst, Inst Geog Sci & Nat Resources Res, Beijing 100101, Peoples R China; |
| 推荐引用方式 GB/T 7714 | Zhou, Wei,Wang, Ting,Peng, Yao,et al. Data fusion enhances the accuracy of soil organic carbon estimation by using high accuracy surface modeling[J]. SOIL & TILLAGE RESEARCH,2026,257:106945. |
| APA | Zhou, Wei.,Wang, Ting.,Peng, Yao.,Yu, Wenping.,Sun, Xiaofang.,...&Yue, Tianxiang.(2026).Data fusion enhances the accuracy of soil organic carbon estimation by using high accuracy surface modeling.SOIL & TILLAGE RESEARCH,257,106945. |
| MLA | Zhou, Wei,et al."Data fusion enhances the accuracy of soil organic carbon estimation by using high accuracy surface modeling".SOIL & TILLAGE RESEARCH 257(2026):106945. |
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。