Journal of Taiwan Agricultural Research

Author:Bo-Jiun Yang, Hsin-Ju Yang, Tsang-Sen Liu, Yi-Ting Zhang, and Chien-Hui Syu*

Abstract:

Soil carbon sink is the second-largest natural carbon sink globally, surpassed only by the ocean. Soil carbon sequestration is recognized as playing a crucial role in climate change adaptation and mitigation. Simultaneously, soil organic carbon (SOC) has positive effects on the physical, chemical, and biological properties of soil. Therefore, the development of accurate mapping techniques is essential for estimating SOC stocks and quantifying soil functions at a regional scale. The objective of this study is to apply digital soil mapping to estimate the SOC stocks in the surface soil (0–30 cm) and subsoil (30–50 cm) of the Zhuoshui River basin. This involves creating spatial distribution prediction maps and conducting uncertainty analysis. Additionally, the study also aims to compare the differences in soil carbon stocks under different topography and land cover. The results show that Regression Kriging (combined with Cubist) has the best predictive performance (surface soil: R² = 0.46; subsurface soil: R² = 0.48), with soil order, elevation, and mean annual temperature (MAT) identified as crucial environmental parameters for predicting SOC stocks in both layers. Uncertainty analysis indicates a higher prediction range in forested areas due to fewer soil survey points. In terms of different land cover types (forest, paddy, upland, orchard, other), the study reveals that the surface soil organic carbon stock is highest in mountainous forested areas (11.2 kg m^-2), while no significant differences are observed in subsoil among land cover types. According to the prediction results, the estimated organic carbon stocks in the surface and subsurface soils of the Zhuoshui River basin are approximately 28.22 and 15.14 million tons (Tg), respectively. The findings of this study can serve as a reference for soil carbon sink estimation, ecosystem services value assessment, and carbon-farming planning in the Zhuoshui River basin.

Key words:Soil organic carbon stocks, Digital soil mapping, Zhuoshui River basin, Machine learning

Download: PDF Links