Soil organic matter (SOM) is a significant carbon pool on a global scale. Accurately mapping the spatial distribution of the SOM is crucial for achieving the “double carbon target” and promoting sustainable agricultural development. However, the impact of using diverse remote sensing data sources on high-precision SOM mapping in hilly and mountainous areas remains unclear. In this study, the Jiangyou City, located in Sichuan Province, China, was chosen as a typical example of hilly and mountainous regions. We devised 15 distinct feature combinations by utilizing three remote sensing variables (Sentinel-1, Sentinel-2, and Landsat-8) along with DEM data. Next, the Boruta algorithm was employed for feature selection. Finally, the RF, SVR, Cubist, and INLA-SPDE models were adopted to create spatially detailed distribution maps of SOM for the region, and an uncertainty analysis was performed on the SOM mapping results. The results indicate that: (1) the INLA-SPDE model, which integrates both data information and spatial structure, achieves the highest ac-curacy and the less uncertainty in SOM mapping, with an R² of 0.647 and an RMSE of 4.227 g/kg; (2) optical images are more important than SAR images, but their combination enhances model accuracy. Specifically, Sentinel-2 data significantly influenced SOM prediction in hilly and mountainous areas, followed by Landsat-8 data; (3) the predicted spatial distribution patterns of SOM by the four models are similar, indicating lower SOM content in the southwest, higher SOM content in the central and northeast. This study serves as an important reference for future large-scale and high-spatial SOM prediction and verifies the importance of the spatial resolution to the SOM prediction accuracy in hilly and mountainous regions.
 

Soil organic matter; Multi-source remote sensing; Hilly and mountainous areas; IN-LA-SPDE