Journal of Integrative Agriculture ›› 2025, Vol. 24 ›› Issue (11): 4395-4414.DOI: 10.1016/j.jia.2025.02.049

• • 上一篇    下一篇

破碎农业景观中土壤有机碳制图：多类别遥感变量的有效性和可解释性

  

  • 收稿日期:2041-11-13 修回日期:2025-02-25 接受日期:2025-01-20 出版日期:2025-11-20 发布日期:2025-10-17

Mapping soil organic carbon in fragmented agricultural landscapes: The efficacy and interpretability of multi-category remote sensing variables

Yujiao Wei1, Yiyun Chen1, 2#, Jiaxue Wang1, Peiheng Yu3, Lu Xu4, Chi Zhang1, Huanfeng Shen1, Yaolin Liu1, 2, 5, Ganlin Zhang6, 7, 8   

  1. 1 School of Resource and Environmental Sciences, Wuhan University, Wuhan 430079, China

    2 Key Laboratory of Digital Cartography and Land Information Application Engineering, Ministry of Natural Resources, Wuhan 430079, China

    3 Key Laboratory of Land Surface Pattern and Simulation, Institute of Geographic Sciences and Natural Resources Research, Chinese Academy of Sciences, Beijing 100101, China

    4 School of Geology and Geomatics, Tianjin Chengjian University, Tianjin 300384, China

    5 Duke Kunshan University, Kunshan 215316, China

    6 State Key Laboratory of Soil and Sustainable Agriculture, Institute of Soil Science, Chinese Academy of Sciences, Nanjing 210008, China

    7 University of Chinese Academy of Sciences, Beijing 100049, China

    8 Key Laboratory of Watershed Geographic Science, Nanjing Institute of Geography and Limnology, Chinese Academy of Sciences, Nanjing 210008, China

  • Received:2041-11-13 Revised:2025-02-25 Accepted:2025-01-20 Online:2025-11-20 Published:2025-10-17
  • About author:Yujiao Wei, E-mail: weiyujiao@whu.edu.cn; #Correspondence Yiyun Chen, E-mail: chenyy@whu.edu.cn
  • Supported by:
    This study was supported by the National Key Research and Development Program of China (2022YFB3903302).

PDF

摘要：

土壤有机碳 (SOC) 的准确空间分布对于指导农业管理和改善土壤碳封存至关重要，尤其是在破碎的农业景观中。虽然遥感提供了关于异质农业景观的空间连续环境信息，但其与 SOC 的关系仍不清楚。在本研究中，我们假设多类别遥感衍生变量可以增强我们对复杂景观条件下 SOC 变化的理解。以中国云南杞麓湖流域为案例区，基于从灌区采集的 216 个表层土壤样本，我们应用极端梯度提升 (XGBoost) 模型来研究植被指数 (VI)、亮度指数 (BI)、水分指数 (MI) 和光谱变换（ST，主成分分析和缨帽变换）对 SOC 制图的贡献。结果表明，ST对SOC预测精度的贡献最大，其次是MI、VI和BI，R2分别提高了29.27、26.83、19.51和14.43%，这归因于ST包含更丰富的遥感光谱信息。最优SOC预测模型综合了土壤性质、地形因素、区位因素、景观格局指数以及遥感变量，RMSE和MAE分别为15.05和11.42 g kg-1，R2和CCC分别为0.57和0.72。Shapley加性解释分别解释了土壤水分、植被状态、土壤亮度和SOC之间存在的非线性和阈值效应。与传统线性回归模型相比，可解释机器学习在预测精度和揭示反映景观特征的变量对SOC的影响方面具有优势。总体而言，我们的研究不仅揭示了遥感衍生变量如何有助于了解破碎农业景观中的 SOC 分布，而且还阐明了它们的功效。通过可解释的机器学习，我们进一步阐明了 SOC 变化的原因，这对于可持续土壤管理和农业实践具有重要意义。

Abstract:

Accurately mapping the spatial distribution of soil organic carbon (SOC) is crucial for guiding agricultural management and improving soil carbon sequestration, especially in fragmented agricultural landscapes.  Although remote sensing provides spatially continuous environmental information about heterogeneous agricultural landscapes, its relationship with SOC remains unclear.  In this study, we hypothesized that multi-category remote sensing-derived variables can enhance our understanding of SOC variation within complex landscape conditions.  Taking the Qilu Lake watershed in Yunnan, China, as a case study area and based on 216 topsoil samples collected from irrigation areas, we applied the extreme gradient boosting (XGBoost) model to investigate the contributions of vegetation indices (VI), brightness indices (BI), moisture indices (MI), and spectral transformations (ST, principal component analysis and tasseled cap transformation) to SOC mapping.  The results showed that ST contributed the most to SOC prediction accuracy, followed by MI, VI, and BI, with improvements in R2 of 29.27, 26.83, 19.51, and 14.43%, respectively.  The dominance of ST can be attributed to the fact that it contains richer remote sensing spectral information.  The optimal SOC prediction model integrated soil properties, topographic factors, location factors, and landscape metrics, as well as remote sensing-derived variables, and achieved RMSE and MAE of 15.05 and 11.42 g kg–1, and R2 and CCC of 0.57 and 0.72, respectively.  The Shapley additive explanations deciphered the nonlinear and threshold effects that exist between soil moisture, vegetation status, soil brightness and SOC.  Compared with traditional linear regression models, interpretable machine learning has advantages in prediction accuracy and revealing the influences of variables that reflect landscape characteristics on SOC.  Overall, this study not only reveals how remote sensing-derived variables contribute to our understanding of SOC distribution in fragmented agricultural landscapes but also clarifies their efficacy.  Through interpretable machine learning, we can further elucidate the causes of SOC variation, which is important for sustainable soil management and agricultural practices.

Key words: soil organic carbon , remote sensing-derived variables ,  Shapley additive explanations ,  efficacy and interpretability ,  fragmented agricultural landscapes